These notes, memos, and hints report, change, and govern my behavior.
Conclude from contemplation on templates:
Journal your thinking and feeling.
Predicates are
Use a thermometer.
Good ones are exceptional people who happen to be psychologists.
The conditional connective ‘only if’ is a grammatical partical and ‘implies’ is a predicate.
A predicate is NOT a singular term which puports to designate its extension as the class (or relation) of each item (or list of items) which it denotes.
(working draft 202608161553)
Grammatical constructions of logical languages denote each of their compounds with its components. Compounds are also components. Lexicals are only so. Constructions of categoricals match. Only categoricals with matching like components match.
Each interchange of each interchangee for each interchanger at each locator in each interchangeand matches its interchangeand and its locator is lexical; or matches its interchanger, its interchangee matches its interchangeand, and each component of its locator is lexical; or is categorical with its locator and its interchangeand, and each of its components matches the same at the like component of its locator in the like component of its interchangeand.
Each interchange matches its interchangeand and its interchangee matches its interchanger at each locator of its interchangee in its interchangeand. Substitutions are interchanges at largest such locators. Lexsubs are substitutions of lexicals for perhaps extra lexicals.
Lexsubs in validities/consistencies denote everywhere/somewhere, and in conclusions of implications they denote where the same in its premises do. Equivalence is mutual implication. Equichanges are interchanges for equivalents.
Equichange keeps validity, consistency, implication, and equivalence. Equichanges in equivalents are equivalent since each lexsub is lexical or matches the categorical of the same in each component.
Subsitution for lexicals keeps validity, implication, and equivalence since lexsubs of substitutions for lexicals are lexsubs– take the component substituted for as an extra lexical.
When there are lexicals, validities are consistencies. Components are self equivalent; and imply themselves, validities, and what their conclusions do. Only validities are implied by and equivalent to validities. Equivalents are mutually equivalent.
Other than where grammatical constructions denote, they misdenote. Lexsubs in invalidities/inconsistencies misdenote somewhere/everywhere, and in conclusions of nonimplications they misdenote where the same in their premises denote. Nonequivalents are mutually nonimplicative.
Equichange keeps invalidity, inconsistency, nonimplication, and nonequivalence. Substitution for lexicals keeps inconsistency. When there are lexicals, inconsistencies are invalidities. Components are implied by inconsistencies. Only inconsistencies imply and are equivalent to inconsistencies.
[More on nonimplication and nonequivalence is needed here!]
Constructions are
Compounds are
Matching compounds of denotative functional constructions jointly denote or exclusively misdenote, where and only where (waow) each of their like components jointly denotes or exclusively misdenotes. For example,
Denotative functional constructions denote eternal bivalent compounds with eternal bivalent components.
Alternations of eternal bivalents with their negation are valid: they denote waow their eternal bivalent denotes or misdenotes, which is everywhere, and hence each of their lexsubs denotes everywhere.
Conjunctions of eternal contradictions with their negations are inconsistent: they misdenote waow their eternal noncontradiction denotes and misdenotes, which is everywhere, and hence each of their lexsubs misdenotes everywhere.
Resolution and Alternational analysis #2026-0816-1239
Existential closures denote waow there is somewhere denoted by their component.
Existential croppings denote part of waow their component does.
Notes on CLassical Mechancis Lecture 2 by Susskind
https://youtu.be/mYDrufxpW9E?si=SMUveQRzJ8VApIrM
Force was proportional to velocity for Aristotle. This is true when friction is pervasive. Susskind uses the “stroboscopic” outlook to derive the equations of motion with respect to discrete changes.
With Aristotle’s law of motion, what is the motion of a particle where the force is also proportional to is position along some x axis? Irreversible exponential slowdown towards the origin: can not distinguish starting points without “infinite resolution”.
What is the motion of an object subject to a force proportional to its displacement from the origin along the principle axis? By Aristotle’s law, the velocity is proporitional to the force which is proportional to the displacement. Since velocity is the time derivative of displacement, the solutions are functions proportional to their derivative i.e. natural exponentials. Without loss of generality, assume the constant of proportionality is an additive inverse, i.e. a negation. Then, the position with respect to time is the product of its initial position with the exponential of the product of the proportion with elapsed time.
Thus, over a long enough interval, every initial condition on a particle puts it arbitrarily near the origin. This very strongly suggests that such a system is irreversible. The conclusion that it is irreversible is supported by our tools for measuring: they have a lower bound on their precision below which all particles near the origin are indistinguishably near it no matter what their initial conditions were.
Susskind is not clear on where the math begins and ends, and where the physics begins and ends. He is also no clear on where theoretical and experimental analyses begin and end. Though there may be no clear or distinct boundary between any of these methods, there is at least some region which can be established through an appropriately vague definition, however arbitrary it may be.
A canonical example of such a distinction is that between logic and math where identity can be brought under the methods of logic when only languages with an itemization of predicates are contemplated, but where anything more, such as a set theory or category theory, are clearly out of the bounds of pure logic. Another canonical example is with an effective definition of implication: one sentence implies another when the conjunction of the negation of the one with the other is inconsistent and inconsistency is established by Quine’s main method. Whatever criticisms come for these definitions must be prepared to deal with the dismantling of powerfully effective methods i.e. they must supplant that power with one much greater on firmer grounds.
My vague boundary between theoretical and experimental practices is that between libraries and laboratories. One is more literary than the other, but both are maintained by strong verbal behavior, presumably the strongest of the human practices.
A key question is now this: what are the origins of Aristotle’s theory and what are the variations and consequences which selected its evolution through Newton and Einstein? Contributing factors to Aristotle’s theory were identified in the introduction to the lecture by Susskind. The example is something like this.
Put a book on a table and you must maintain contact with it while pushing it to move it across the table. Stop pushing it and the book stops moving. Having a history with geometry where proportions rule the day, you are likely to notice that the harder you push the book the faster it moves across the table. This proportion between strength of push and change in position occurs with a change in time.
While writing the previous note on the characterization problem for first-order theories, I had to look something up on my phone and it opened to where I had left it: Twitter. This is one of many defects in my local programmable environment. Right next to the beautiful benefits of modern technology are the worse burdens on human practices in history. No doubt it could be worse.
I happened to read a tweet, in my ‘for you’ feed, from Paul Graham. He continues to be a loci of control over a few of the more powerful self-aggrandizing social controllers in the modern world. He has had profound influence over their verbal behavior, and, consequently, over what they think and otherwise do. He is not some sort of puppet master. No one is. Such push and pull theories of control are, at best, ancient holdovers from the early days of the experimental analysis of respondent behavior, or, as it is pejoritively called and rightfully so, stimulus-response behavior.
What he does is to construct a verbal environment which selects and maintains the kinds of behaviors characteristic of fledgling social controllers. Every human controls the behavior of an other human. This is part of our genetic endowment. It is part of the behavior of organisms that their behavior is shaped by the part other organisms play in their evolutionary and individual environments. Other organisms, of the same or different species, are a stable part of the evolutionary environemnt of a given species. The seperation of ontogenic and phylogenic contingencies contributions to present behavior is made difficult by the weakness of most experimental analyses of onotogenic contingencies. The contribution of an ontogenic contingency to the strength of a present behavior is to be contrasted with the strength of that behavior from the contribution of phylogenic contingencies. Since we have direct access to the arrangement of ontogenic contingencies, and have no way of similarly doing so for phylogenic contingencies (with some remarkable exceptions), we are left to control ontogenic behavior as strictly as we can so as to factor out the complementary contributions from a much more distant past.
The remarkable exceptions to the analysis of phylogenic behavior, are the attempts made by work on genetics to simulate, through a present intervention, the effects of contingencies of natural selection. These simulations of natural selection are only as good as our analysis of natural selection is accurate. We are bound by the chemical consequences and variations which permit the biological changes seen as natural selection. Eventually, though I do not know how long it might take, the analysis of natural selection will be strong enough to accurately simulate natural selection from a proximate change in the biochemistry of an organism. I do hope that we need not wait until such changes frequently occur before we have built up an effective study of ontogenic contingencies of behavioristic selection. It would appear that such biochemical changes are unlikely to occur but from work on factoring out the contribution of ontogenic contingencies to the effects of phylogenic contingencies.
But, looking back at all I have written here, it is not clear to me how I have come to write my way into a corner on the topic of behavior when I began by reporting on my misuse of Twitter.
…
Upon seeing a Tweet from Paul Graham which said something like “I saw a trend in the founders I met for office hours. Many were using open large language models. What I have retweeted here independently verifies what I noticed.” I was reminded immediately of Justine Tunney’s, of https://justine.lol/, work on llamafile: https://justine.lol/matmul/.
I am a fan of Justine’s work on LISP. They demonstrated how tiny a fully functional LISP can be, and this challenged FORTH people to demonstrate the same. Between the two, the problem of hardware is emphasized, as it already was in the history of both languages, e.g. LISP machines and the GA144.
I am not yet familiar with any attempt to show that one langauge requires fewer transistors than the other overtly. But, there are projects, like My4thXS http://mynor.org/my4th_xs.htm which uses only 13.75 ICs, which demonstrate something of the straightforward circuit design for FORTH which I have not yet seen done for LISP. Though, now that I write that, I am going to search for one before I conclude this paragraph.
A quick search reveals that the closest candidate for a LISP machine out of 7400 Logic is one that works for brainfuck: http://grapsus.net/74/. LISP and FORTH have both been implemented through brainfuck and anyone can happily test the Church-Turing hypothesis.
There is more to be done with this that brings the geometry of the world into the reach of any proposed programmable solutions.
…
I no longer recall what prompted this note in the first place, and hence I shall stop while I am ahead.
Picking up where I left things at the end of the last note, the premises of N, or “non-logical axioms” as Shoenfield presents them, are used in an example application of the consistency theorem as a contrast between its finite proof and the proof of Henkin’s theorem which is not. In order to better understand the role of free variables in Shoenfield’s work, something which has puzzled me after having read Quine’s clean methods and explanations in “Methods of Logic Fourth Edition”, I shall perhaps follow through on the construction that Shoenfield leaves behind and which demonstrates the clear difference between the consistency theorem and Henkin’s completeness theorem.
First, finitary proofs are contrasted with abstract proofs by Shoenfield, e.g. “a finitary proof often gives more information than an abstract proof, simply because the restriction to finitary methods often requires us to prove more than is stated in order to obtain the desired result.” pg.51. In the paragraph before,
“finitary proofs can throw light on the nature of the concrete and the abstract. For example, the finitary proof of the consistency of N shows that Hilbert’s view of abstract mathematics discussed in section 1.2 is tenable for N. We can think of the variable-free formulas of N as expressing concrete results, and of the quantifiers as being abstract notions introduced to prove concrete results. Our consistency proof then shows that any such concrete results is correct, i.e., that any variable-free formula A provable in N is valid in (curly N). For otherwise, (not N) would be valid in (curly N); and the above proof woudl then show that N[(not A)] is consistent, which contradicts the corollary to the reduction theorem for consistency.” pg. 51
Here is a key insight into the role of variables, and also perhaps an access point to understanding the obsession with constants. Shoenfield claims that the variable free formula of N are concrete. Such formula are like formula with free variables but where the free variables have been substituted throughout for constants, functions of constants, functions of functions of constants, and so on. While Shoenfield uses ‘term’ to denote variables or functions of terms, this definition leads to much confusion which is unhelpful to logical method. Better to define ‘singular terms’ as functions of singular terms, and, as must already be done with Shoenfield’s definition, emphasize that zero place functions, or what Shoenfield calls 0-ary functions, are also called constants and assure us that our inductive definition is wellfounded.
Why ‘singular terms’? Because, ‘general term’ is another phrase for ‘predicate’. Then general terms can be divded into absolute or relative. Absolute terms are also known as one place terms, they denote one and only one thing at a time, and relative terms are also known as many place terms, the denote many things at a time (presumably the same many of things each time, so that a two place general term denotes two things at a time, a three place general term denotes three things at a time, and so on).
The distinction between singular and general terms leaves the word ‘term’ open for further classifications while happily including singular and general terms among its denotata. Such a helpful cross classification is that between concrete and abstract terms:
“Concrete terms are those which purport to refer to indiivduals, physical objects, events; abstract terms are those which purport to refere to abstract objects, e.e., to numbers, classes, attributes. Thus some singular terms, e.g., ‘Socrates’, ‘Cerberus’, ‘earth’, ‘the author of Waverly’, are concrete, while other singular terms, e.g., ‘7’, ‘3+4’, ‘piety’, are abstract. Again some general terms, e.g., ‘man’, ‘house’, ‘red house’, are concrete (since each man or house is a concrete individual), while others, e.g., ‘prime number’, ‘zoological species’, ‘virtue’, are abstract (since each number is itself an abstract object, if anything, and similarly for each species and each virtue).” pg. 259 Quine “Methods of Logic Fourth Edition”.
Note that, finally, terms come to be our means of referencing the rest of the world. Where do the variables fall as terms? To answer that question, more on the history of ‘term’ is required.
Concrete terms carry abstract ideas. This much is admitted by almost anyone, though perhaps they are insensitive to the distinction between concrete and abstract. The one place, absolute, general term ‘idea’ is said to denote each idea. Ideas are some of our prime examples of abstract things, if things they even be, and this bestows upon the term ‘idea’ its abstractness.
Following the principle that “Concrete terms carry abstract ideas.” Quine goes over some history in the first chapter “Ddays of Yore” of his book “From Stimulus to Science”:
“The purpose of langauge is said to be the communication of ideas. We learn a word from our elders by associating it with the same idea, and we use it in the communication of ideas. How do we know that the words we use to express our ideas are conjuring up the same ideas in the minds of our listeners? Words and observable behavior are all we have to go on, and the idea idea provides only the illusion of an explanation.
John Horne tooke urged the point in 1786, ten years before Hume’s death.
“I only desire you to read [Locke’s] essay over again with attention to see whether all that its immortal author has justly concluded will not hold equally true and clear, if you sbustitute the composition, &c, of terms wherever he has supposed a composition of ideas … The greatest part of Mr. Locke’s essay, that is, all which relates to what he calls the abstraction, complexity, generalization, etc., of ideas, does indeed merely concern language. PP.37-38”
Took was a kindred spirit of his remote predecessor William of Ockham and other medieval nominalists, who had dismissed abstract objects as flatus vocis, vocal breeze. Tooke’s was a majory step toward what Richarde Rorty has called the linguistic turn. If there is sense to be made of the compounding of ideas, clearer sense can be made of the compounding of language. Words, unlike ideas, are out where we can see what we are doing.
Much further freedom in the constructing of ideas— or, now, terms— was achieved by Jeremy Bentham a few years later in his theory of fictions. he observed that to explain a term we do not need to compose a synonymous phrase. We need only explain all sentences in which we propose to use the term. It is what is now called contextual definition. Bentham’s motive was ontological: he wanted to be able to introduce one or another useful term without being charged with assuming some controversial object for it to designate, or objects for it to denote. The seeming object or objects could in this way be dismissed as innocent fictions.” pg. 6-7 Quine “From Stimulus to Science” 1995
Those who are apt to take ideas as abstract are just as apt to take words, or their technical kin ‘terms’, as concrete. Yet, upon further scrutiny, words or, now, terms come to be
…
Got distracted by Twitter. Some day I may be rid of it. Not today. The next note says more.
The previous entry lists the definitions needed to approach the first lemma of the section on the consistency theorem in the chapter on the characterization theorem in Shoenfield’s “Mathematical Logic”. Following that list is a short investigation of the role that special constants, otherwise called witnesses in Ebbinghaus et al., play relative to Quine’s method of taking free variables as dummy singular terms, along side predicate letters as dummy predicates and sentence letters as dummy sentences.
Though Quine’s methods are schematic and Shoenfield’s are formulaic, there is complete correspondance: the formula are seen as instances of the schemata. Shoenfield uses schematic methods in English and distinguishes them from the formal methods of theories and structures. Quine maintains the entirety of his methods in English, leaving formalities as more a matter of mathematics than logic. This is the same outlook which he had to further explain in an appendix to later editions of his “Mathematical Logic”. There, he began with descriptions of a formal system much more like Shoenfield’s, but such methods are now entirely ancient even though they continue to prevail within matheamtical practices. It is the perverse prevailance that primes me to write here on logic as I have so as to perhaps reinforce a prevailing change in method.
The consistency theorem is a theorem on open theories. For Schoenfield, open theories are are those whose nonlogical axioms are open. A formula is open if it has no quantifiers. Presumably an open theory may contain axioms which have free variables. Later in the section on the consistency theorem (which can be summarized as a finite proof of the characterization theorem for open theories) an example application to the open theory N of a system for natural numbers is mentioned. That theory is given by the following nonlogical axioms:
“The nonlogical symbols of N are the constant 0, the unary function symbol S (which designates the successor function), the binary function symbols + and *, and the binary predicate symbol < .” pg. 22
The use of the word ‘symbol’ is misleading here. In mathematical logic, there is already an interpretation planed for a formalism. The formalisms of interest are theories, and theories are divided into two parts: their logical part and their nonlogical part. The logical part is given by specifying theorems as special formula which are either initial, in which case they are called axioms, or peripheral, in which case they are derived from the axiomatic theorems as theorems generated by the rules of inference. Rules are formalized as generation rules e.g. “Expansion Rule. Infer B or A from A.” pg. 21, which are poorly explained predicates of the logic of theories with proofs under consideration.
For example, the predicate ‘is generated by the expansion rule from’ can be introduced in its relevant premise with the mathematical pidgin English:
Each item is x, y, with z such that z is the alternation of x with y only if z is generated by the expansion rule from y.
A reason, perhaps the main one, to write out such a premise and predicate for specifying a theory of first-order theories, is to make excruciatingly clear what are the objects of the theory and what are its predicates. It is the burden of notation that it escews the logic. Another reason, perhaps the main one again, is that such pidgin English can be easily paraphrased further into English idiom to enhance the strength of rhetorical effects without distorting the aims of the project. The aims are not to give a theory of theories which has the same kind of precision as the theories it is about. To take such aims is to further obscure logical method.
This also makes it clear that if we are given a theory of the grammar of a language with the predicate ‘is the alternation of’ then we can define the predicate ‘is generated by the expansion rule from’ as
x with y such that some item is z such that x is the alternation of y with z
This must be done if the line of investigation taken by Shoenfield is to be followed because it eliminates overtly, rather than covertly as the case is, the mysticism of ‘inference’ as some intervening mental operation. It is overtly clear that these are ‘purely formal’ instruments as much as it is clear that the grammatical construction of alternation is formal.
Ultimately, I have no clue what is formal and what is not formal, without some supplementary definition of the term. At the end of “Mathematical Logic” Quine makes the case for never settling with a claim about the nature of formality without making clear what the definition of ‘formal’ in the claim’s context is.
Yet another attempt to write my way through Herbrand’s theorem as proved in Shoenfield’s “Mathematical Logic”. I reviewed all that Quine had to say in “Methods of Logic Fourth Edition” on his free variable methods. In the final part, free variables finally come to play the role of dummy singular terms i.e. singular term letters. In the terminology of Shoenfield, and most other ‘first order logic’ methods, the free variables do the duty done by constants and function symbols.
I avoid all mention of ‘symbols’ or ‘meaning’ not just because Quine explains them away, or at least submits good reason for why they should not be relied upon as the roots of reference, in his philosophy of logic, but also because Skinner provides a coherent theory of verbal behavior which clearly does away with any purported other worldly entites that would stand as the meanings of any terms whatsoever. This is still a controversial conclusion, but it is a technical one which serves a purpose above and beyond the convictions those like Russell had for items known only through the events of their naming. As dummy singular terms, free variables can be eliminated by singular descriptions and the problem of nondesignating singular terms can be dealt with using nothing more than the same methods of logic that appeared in the earlier parts.
Those following Shoenfield’s pattern, the one which prevails across texts since Hilbert’s work on logic, allow for free variables but admit them as a kind of holdover from the propositional functions of Frege and Russell. Schemata with free variables are then said not to have a single meaning, but rather many meanings dependong on which singular terms come to replace the free variables in a given instance. Shoenfield goes further and permits the introduction of names as special kinds of constants when giving an elaborate definition of validity with respect to structures.
A structure has a universe of discourse which is a set, and each item in that set is given a name which is included among the singular terms of a theory when speaking of the logical language of the strcture under investigation. Then, structural instances of schemata are abstract objects of truth or falsity, that is once the names of items of the universe of discourse of the structure replace the free variables in the schemata and the predicate letters are replaced by the appropriate set of tuples, and the singular terms by the appropriate functions (where constants are zero place function letters). This roundabout way assigns a truth value to each schemata relative to a structure in which it is said to be valid (this being Shoenfield’s terminology). A stronger kind of validity, logical validity, is ascribed to those schemata that are valid in each and every structure.
Structures match schemata to sets, interpretations, contrary to Quine’s use of the term, match schemata to formula. Technically, Shoenfield does not use schemata. Only formula are contemplated. Formula are what are otherwise called phrases or grammatical constructions of other languages. I am unable to maintain talk of formula because they are sometimes schematic and sometimes note depending on the langauge contemplated. This adds a kind of confusion to the relevance of certain methods.
Either way, Quine’s vague definition of interpretation, which may either associate with each predicate letter an actual predicate or any other item which indicate what items it is true of and which it is not, be it a set as in the structures and models of mathematical logicians, or something more exotic such as an algebra or topology.
This all makes it quite hard to confront, from Shoenfield’s work, the problem of nondesignative terms. The very definition of structure imposes the same simplification as Quine’s methods but with no overt note as to the nature of it all. Shoenfield gives the definition of structure by requiring that “for each constant e of L, eA is an individual of A” (pg. 18). Then, as if to fill things out for good measure, the names are introduced for each item of the universe designated by ‘A’. The names are to be taken as distinct from the constants already introduced by the language the structure is contemplated with respect to, even though any constants of the langauge already designate an item of the universe.
All of this confuses the problem of meaning with the methods of logic e.g. as if variable free sentences have one meaning or as if sentences with free variables have many meanings.
Here are some tricky definitions that must be kept in view when reading Shoenfield because of the problem of not being able to use free variables as dummy singular terms (or, what Shoenfield would have to use as actual singular terms from a given language, which, of course, does explain why all these strange notational distinctions).
| “Suppose that the first-order language L’ is an extension of the first-order langauge L, and let A’ be a structure for L’. By omitting certain of the functions and predicates of A’, we obtain a structure A for L. We call A the restriction of A’ to L, and designate it by A’ | L. We also say that A’ is an expansion of A to L’.” pg. 43 |
| “A theory T is a Henkin theory if for every closed instantiation (exists x)A of T, there is a constant e such that | -T ((exists x)A only if A_x[e]).” pg. 45 |
And this is where the consistency theorem begins, with all those definitions in hand and all their minor consequences too. Further technical definitions issue forth as the theorems, lemmas, and proofs proceed. I must start with the consistency theorem to get at Herbrand’s theorem because it is relied upon in Shoenfield’s presentation.
Lemma one is as follows
“If T proves A, and A’ is a closed instance of A in L(Tc), then A’ is a tautological consequence of formulas in D(T).” pg. 49
This lemma does not translate directly into the terminology of Quine’s methods of logic because he does not introduce function letters at all and hence does not introduce the zero place function letters Shoenfield and others otherwise call constants. Quine goes back to Russell’s work and with him rejects the irrelevant distinction between function letters and what he calls functional descriptions or singular descriptions.
In a theory with function letters, predicate abstracts reveal the relevant transformations between functions and predicates as follows:
The context dealt with there was that of identity. In general a sentence in Shoenfield’s text such as “pf..x..y” becomes “p(the u such that Fu..x)..y” which, through notational conventions of partial concretion of predicate abstracts (which can be traced back to Russell’s “collective relate”) can be recast simply as “p(the (F..x)..y” which is more than familiar to those who program in a functional programming langauge like LISP (but now, to be excruciatingly clear, this is pure logic).
I have gone to such lengths to describe the different outlook because it is key to bringing all of these methods to pure logic where function letters are merely a distraction from the use of free variables as dummy singular terms. A remark for anyone reading this who already knows a thing or two: of course all of this is standard, e.g. the elimination of function letters, but I am returning to the conveniences of logic and not those of mathematical logic which have largely been corrupted by the miscarriage of the metaphor of meaning, as if singular terms must designate, rather than simply purpoting to.
Now to integrate the method of singular descriptions with Shoenfield’s method of special constants. A special constant, also called a witness in Ebbinghause et al., is presumably, a disguised definite/singular description. Technically, special constants are more like how English speakers use the indefinite articles ‘a’ or ‘an’ to introduce something for the first time. Among sentences there are those which begin with ‘Something is’ as in ‘There is an (x such that Fx)’. With each one place predicate, e.g. ‘F’, of a theory there is an associated special constant, e.g. ‘r’, such that ‘There is an F, only if Fr’ which is called its special axiom, or special sentence. We say “‘r’ is THE special constant for ‘F’”.
But, Shoenfield would rather say “r is the special constant for (exists x)F.” My inclination is to Quine’s “The objects of a theory are that which are denoted by a one place predicat of it.”
Since I have no interest in special constants, I shall just keep to free variables as dummy singular terms in the same way that predicate letters are dummy predicates and sentence letters are dummt sentences. I know that Shoenfield is not using schematic methods (he only uses them as part of English).
Compound one place predicates are not easy for Shoenfield to depict and that is why the special constants and witnesses are so hard to pin down. With predicate abstracts it is easy to see that one place predicates can all be brought into a standard form ‘(x such that Fx)’.
…
Sadly, I can not finish my thought.
My definition of the logic of a theory as that which it implies does not permit the stronger claim that a theory is that which is implied by some of its implications and not others, a particular case of which is a theory whose implications are those of an itemization of some of them called, now, premises. This particular case is the naive view of Barwise in the first chapter of “Model-Theoretic Logics” where predicate logicians, such as Quine, are likened to mathematical logicians with their axioms and rules of inference which they prove generate the implications of a theory as theorems. In this case the itemization of premises becomes the itemization of theorems given the speical name ‘axioms’. Theorems are, under such interpretations, items of a calculus as part of a theory of concatenation or classification or composition.
The mathematical analysis of mathematics is not a logical analysis of mathematics. The difference was well appreciated by Frege and Russell who were concerned by the erroneous arguments tolerated by mathematicians who, for one reason or another, were insensitive to the methods of logic and their scientific consequences. Somehow modern methods lost touch with their origins and lost something of their better reasons.
However Barwise came to feel restricted by the methods of “first-order logic” or what might better be called, pejoratively, “deductive logic”, he need only release himself from the strictures of mathematical methods and escape to those of logical methods. No doubt, he would claim that he somehow had, but his talk of mathematical entities which he somehow knows from direct contact betrays his convictions. Oddly, he felt it necessary to say in the introductory chapter that he had initiated the study of the first order logics of set theories used in the metamathematical analysis of the first order logic of set theories. As if to hedge a bet, or profess some kind of apologetic outlook. I’m uncertain what his ultimate aim was, but it seems to have been used to excuse himself from dealing with the grave problems and solutions presented by, e.g., Quine’s work on logic. Thus we have found the explanation for the obtuse results reported as model-theoretic logic.
Somehow, Barwise found solice in Godel’s incompleteness proofs, as if they demonstrated the insufficiency of the methods of logic to address the problems of mathematics. The heterlogical paradox is perhaps a better lesson for those like Barwise, those of us who are determined to deal with the pains of accuracy rather than rempage through the labyrinthian halls of inarticulate precision, accuracy be damned. What we learn is that the prior convictions of past scientists that there is a single predicate ‘denotes’, which can be contemplated as a shortening of ‘is true of’, is false. This is the same kind of shift in practice proposed by Russell’s paradox when he wrote of it to Frege. But, whereas Russell could work around his trouble by restricting grammatical distinctions, the trouble with denotation is found wherever language occurs, at least those languages which are rich enough to supply us with immanent validities.
As much as I would only like to write on logic and the peculiarities of my particular programming methods, I can not. The science of behavior looms ever larger over my shoulder each time I look at the latest news about countries and companies at war over the control of the world. I do not porpose that the problem is control and that the solution is an escape from it. The experimental sciences uncover that we trade one kind of control for another and never escape controlling circumstances, however unkown they are to us. Even the most sacred of our societies’s sentiments, that gem of freedom, is a kind of control and if we are to strengthen it then we must deal with it accurately as a kind of control.
The problem posed by the modern world is the power of new positively reinforcing controls. In the past, when punishment was threatened to condition escape and avoidance which was in the favor of the punisher, those so punished were primed to dismantle the troublesome contingencies which maintained their deplorable circumstances. Such countercontrolling practices remain strong among those who participate in cultures which prize individual autonomy and the autonomous actions of groups in isolation from each other.
I have little doubt, though if I returned to reading more particulars then my doubts would very likely grow, that the best of the modern world is a consequence of fighting for freedom as an escape from punitive measures of control, and I would never wish upon any future generation anything like the punitive measures of the past or present. In a sense, these are the obvious problems to solve when it comes to working upon our shared social environments so as to establish stronger practices among us. The problems which threaten us most are not those of punitive control. They are those of positive control, the kind that we have no history of countercontrolling effectively.
The effects of positively reinforcing control are felt now, and it is their deferred consequences which threaten our survival later. While there are some people who spend much more time contemplating how to control others, it is not as if they are special in that they control others. Humans evolved, as every other organism on Earth has, from having controlled other organisms. The blunt instruments of war are present without the tools of destruction. It is metaphor to ascribe war to groups of organisms which do not have social behavior e.g. the swarms of microorganisms which daily feed upon each other as the spoils of a conflict which has raged on the molecular level for eons. War is human as much as love is in that the social environments which build and maintain them are a consequence of our individual behaviors and not those of other organisms.
This may not be the case if we select definitions for social behaviors like those which compose what we call love and war wildly enough. This does not mean that humans are cosmically special in some way. It means that whatever definitions we may choose, humans are conditioned by the world in which they live which are distinct from how every other organism on Earth is conditioned by that same world. The origins of these differences are to be found in the selecting consequences from past variations on physical, chemical, biological, behavioral, and social scales of change (maintanence is a special kind of change).
Newly verbal and social machines, owned by a few self-aggrandizing social controllers, take advantage of prevailing social practices which neglect the science of behavior and the technologies of behavior which it has introduced, e.g. as with modern gambling systems, and promises to introduce, e.g. as with robotic systems. I write here to say what those scientists of behavior who are only with us in their, mostly, written works can no longer say aloud. But, I do not claim to speak for those who spoke for the trees. My personal history is different enough from theirs to make my responses, broadly, distinct in their controlling conditions. Metaphorically, a phase change in the resolution of contingencies arranged by programmable machines has permitted newly powerful kinds of positively reinforcing control.
What was once only accessible to a few laboratories and set up by intervening pully systems, as in the averaging of cumulative recorders across dozens of individually arranged experimental environemnts involving rats and levers, is now accessible “on the wire” to almost anyone. Little use has been made of what is available, and the more educated the public is about the science of behavior, the greater use is likely to be made of it.
It is for science to tell us how to use it. For, the behavior of humans when they are making moral decisions or building newly powerful technologies, is behavior, and is the subject of the science of behavior with all its theoretical and experimental practices.
In the past, the philosophers where called upon to whip up the public into ethical conundrums that tested their readiness to respond well to the changing world in which others lived and worked with them. Any appeal to such past practices is as weak as any other science making a similar appeal. Even the greatest of the sciences, logic, has undergone profound changes since the two thousand year reign of Aristotle’s syllogistic. A return to earlier advice born without anything like an experimental analysis, is not just pre-Baconian, it is bankrupt.
Yes, we are to give up most of what we have been burdened with from past generations. In a sense, this is obvious. How else are we to change? Are we not lucky that the young are not born with the social conditioning of the old? But, what promise is there to this seemingly desperate alternative? Have I done anything more than promise a brave new world? No, and the devil is in the details.
Prevailing cultural practices are born of a presceintific world. Until little over a century ago, a science of behavior was an oxymoron. Even now, the public is made to think that the uncertainty in any such psychological science is beyond what can be tamed by technology, even in a laboratory environment. Field work is as good as it gets and what we get are scraps of self-reported observations. No kind of conspicuous control over behavior is even wanted, much less demonstratable in the field. The control demonstrated in the lab is capricious. It is too frail to survive the tumult of the realms beyond laboratories and libraries. For those committed to such convictions, I have little to say for now. My powers of explanation are limited, and I am unable to justify the methods of science and logic beyond what can be seen from the controlling technologies which have resulted from them.
This is little more than an experimental exposition, a way of discovering what I have to say. More is likely to come if only because I am prompted by every report I read from the frontlines of modern technology. Thankfully, I am not so isolated as to be committed to what I have written here. I have made no final conclusions on such matters as are well beyond me and presumably well beyond even our most prized practices of science. Such weaknesses do not excuse a sort of indifference to further investigation. The future has not happened, and thus has no control over the present, no matter how vivid are the stories of it from times past and present.
We are not free in that we can NOT escape the controlling consequence which have selected present biologies, behaviors, and practices. Yet, we have happily evolved practices of self-management and self-control to the extent that one part of our behaviorial repertoire has come to control others and vice versa. No doubt our brains are a big part of mediating these complex countercontrolling practcies throughout our prized social environments. But to fixate on brains, like famished zombies, is to miss the behavior whose mediation they are to explain. It is also to miss what is immediately accessible to our countercontrolling practices of self-management and self-control.
Very little of the selecting consequences mediated by human behavior are covert. There is a world beneath our skin which we are in contact with, thorugh our peripheral nervous system, that others are not, yet. Our perspective on the inner world is born from prescientific practices e.g. those of philosophers and other storytellers. That our descriptions of our inner world shall begin anew from what the science of behavior has to tell us about our contact with it and the social behavior conditioned upon that contact is a bright light at the end of a very long tunnel.
Francis Bacon found himself at odds with the establishment much as scientists today find themselves. What was once ruled by kings and queens has been replaced by lest punitive kinds of social control. We may have exchanged the wage for the whip, but we have not escaped the punitive control. Worse, we have not even begun to prepare ourselves for the deferred and destructive consequences of past and present positively reinfrocing kinds of control. My hope is that at least those of us who care for the survival of our strongest scientific and logical practices, and the human species necessary to mediate them, shall not contribute to a future worse than the present.
Notes on Standford’s Susskind’s “Classical Mechanics: Lecture One”
https://youtu.be/ApUFtLCrU90?si=oHLPj6Cf4qnTs7Zm
What are the laws of the laws of physics? After having collected a few theories of the physical world, such as those in classical or quantum or electromagnetic mechanics, what similarities are there between them that might point us in a better direction as we assemble stronger theories?
A space of states, e.g. the heads or tails of a coin on a table, and the changes from one state to another by the abstraction of motion. A coin which starts as heads or tails and stays in the same state is described by the identity function (assuming a “stroboscopic” world where motion is discrete as a function of a discrete dimension of time). A coin which flips from heads to tails or vice versa without end is described by a function which is the negation of its prior state where heads is 1 and tails is -1.
A state diagram is a directed graph each of whose points is labeled by a state of the system and each of whose lines is a motion from one state to the next. Cycles are conserved quantities. We can choose representative states of conserved quantities e.g. we are on the cycle that includes a given state.
These require a few provisions on the construction of the graphs: only one arrow goes out of any point. In classical mechanics
“the state of the system” or “configuration” is what you need to know to perfectly predict where it will go next.
The arrows of closed systems are never changed.
There can be an infinite number of states (so far only coins and dice have been been contemplated). The states/configurations are the integers, and a law of motion says that you add one: this creates a cycle (even though it doesn’t come back around itself). Another law is jump two ahead: this gives a conserved quantity we call the oddness or evenness of the state.
Laws of physics not allowable, by example. Heads goes to tails, tails goes to edges, and edges goes to tails. You get stuck in an endless edges tails sequence. Why is this odd as a law? It is complete predictive into the future, but it is not predictive into the past! If you know you are at tails then you may have come from heads or from edges. While you can predict the future you can’t retrodict the past.
The system is not reversible. How to check for reversibility? Reverse each arrow and you must have a completely predictive situation. The general law is that one arrow in and one arrow out.
Classical physics is reversible.
What are the limits on predictability? If we have a perfectly predictive system of equations it requires that we know the initial conditions. We need to know what the rules are and where we start. In statistical mechanics we simply do not know exactly where we start, and consequently we do not know what happens next since small changes in some initial conditions may or may not give large differences in the next state.
The laws of classical mechanics are deterministic into the past and the future.
…
Moving onto the world of particles. Point particles moving around in space.
A point of space will be labeled by three coordinates.
What are the configurations of a point particle?
The first lecture of Schuller’s “Geometrical Anatomy of Theoretical Physics” is oddly imprecise. There is such a distaste for the methods of logic that even in a course aimed at establishing an effective language for a clear and exact theory of physics little commensurate attention is played to the logical language in which it is all ultimately cast.
This is cosmically unfortunate.
…
Yesterday and today I have been reading Barwise and Feferman’s 1985 “Model-Theoretic Logics”. This follows from my reading of Lindstrom’s theorems at the end of Ebbinghaus et al. That reading was itself ultimately prompted by my attempt to paraphrase Shoenfield’s reduction theorem into my methods of logic. There it occurred to me that two distinct kinds of implication could be contemplated:
I do not yet have a good name for the two, though the phrase ‘immanent implication’ and ‘transcendent implication’ occurred to me. These seem in appropriate because when speaking of grammar and denotation there is no escaping the problem of definitions aimed at one language or at many, and the distinction between such distinct aims is well suited to the terms ‘immanent’ and ‘transcendent’ because they are a canonical and concrete example of an ancient problem.
An alternative terminology which does not quite satisfy me is that the familiar implication is still implication, but that the other is a ‘multiple implication’ or ‘universal implication’ or, if I am being deviant, ‘endless implication’. My reason for prefering ‘endless implication’ is to emphasize taht it is, as it is, unworkable.
There are at least two reasons that it is unworkable. The first is that supplementary predicates which are not included in the theory of grammar and denotation which admits the staring definition of validity as truth of each lexsub, or, better, universal denotation of each lexsub, are not obviously relevant. These predicates are in the theory of axiomatics where, e.g., there is a predicate for ‘is an axiom’ and the definition of the endless implication is no longer schematic in that it becomes:
The similarity between this and some deductive method of inferential proof which masks a calculus is wildly misleading.
…
Going to do dishes rather than get lost in thought.
I read the first section of the first chapter by J. Barwise in “Model-Theoretic Logic” and was astounded by the ease with which he dismissed Quine’s philosophy of logic, as if Quine was nothing more than a mistaken ideologue.
…
I can not worry about that now because simultaneously I just watched one of the final lectures of Frederic Schuller’s series on the “Geometric Anatomy of Theoretical Physics” and did not realize that the familiar method of introducing Quantum Mechanics does not work the moment you chnage coordinate systems (you can show that one of the key operators does not commute as hoped). Consequently I have begun to take proper notes on it.
…
Without further ado, here are notes on the first lecture.
Institute for theoretical physics III
Introduction/Logic of propositions and predicates 01 Frederic Schuller
Theoretical physics uses mathematics to fully explore the implications of our theories of the world. If we do not have a precise mathematical description of the physics of the world then we probably don’t know something about it.
The aim of the course is to provide proper mathematical language for classical mechanics, electromagnetism, quantum mechanics, and statistical physics.
This course combines analysis, algebra, geometry into differential geometry to deal with physics. In analysis and geometry are classical mechanics, electromagnetics, and gravity. In geometry and algebra is statistical physics. In analysis and algebra is quantum mechanics.
The course is structured building block style. At the bottom is logic (propositional and predicate) and upon that is a theory of sets. The need for continuity demands topology. For physics we narrow on topological manifolds. For derivatives, we then build differentiable manifolds. Then bundles, most of modern physics can be cast in this language. In bundles you learn that there is no such thing as a wave function. Ontop of bundles is geometry e.g. symplectic and metric geometry. Then from the geometries the physical theories can be developed.
…
Chapter 1. Axiomatic Set Theory
1.1 Propositional logic
(He presents logic in the old way, that is as an algebra of proposition and then one of propositional functions, though he does speak of predicates rather than propositional functions.)
Def. A proposition p is a variable that can take the values ‘true’ or ‘false’. No others.
Can build new propositions from given ones.
Logical operators
Theorem. (p only if q) if and only if ((not q) only if (not p)). Proof: truth table.
Corollary. We can prove assertions by way of contradiction.
…
This is not how logic works at all, but I will go with it because I have dealt with much of logic elsewhere and will be able to paraphrase to that.
…
1.2 Predicate logic
Def. A predicate is a proposition-valued function of some variable(s).
ex. P(x) true or false, dependent on x; Q(x,y) dependent on x and y.
We can construct new predicates from given ones:
Remark: quantification for predicates of more than one variable.
I am only going to write down stuff that seems unexpected from my past.
1.3 Axiomatic Systems and Theory of Proofs
Def. An axiomatic system is a finite sequence of propositions or propositional schemata a1, a2, .., aN which are called the axioms.
Def. A proof of a proposition p within an axiomatic system a1,..,aN is a finite sequence of propositions q1,..,qM,p such that for any j inclusively between 1 and M either (A) qj is a proposition from the list of axioms, (T) qj is a tautology, or (M) “modus ponens” some m and n less than j such that “qm and qn, only if qj”.
I am uncertain why he has phrased modus ponens this way.
Remark. This definition allows to easily recognize a proof.
Def. An axiomatic system is consistent if there exists a proposition q which can not be proven from the axioms.
This first lecture is so very unsatisfying. The world suffers from a lack of effective logical method.
In #2026-0828-1446 I began proving the main theorems of Shoenfield’s “Mathematical Logic” to show how mathematical logic has obscured the methods of logic. The aim is to strengthen logical methods throughout the sciences, including the mathematical ones.
What Shoenfield calls “the reduction theorem” is a weak way of saying that implication is validity of the conditional. The statement of the theorem is couched in his technical terminology of theories and burdens a given theory with a particular method of proof. This is an unhappy circumstance.
My statement and proof of the reduction theorem remains relevant to both Shoenfield’s aims and my own more logical ones. In the last note I discovered that it is better to begin by speaking of the logic of a theory than of a theory generally since there is more to a theory than its logic. The logic of a theory is the many implications of its premises. Instead of saying ‘the logic of a theory’ I shall just say ‘the theory’ as nothing more than the logic of theories is dealt with here.
Rather than say that implications of a theory belong to it— and raise concerns of a vague theory of classing classes, e.g. a set theory— I shall say, somewhat abruptly, the appropriate conjugation of ‘is of’ as in “This conclusion is of that theory.” It can sound biblical, but it is merely a hint that some unspecified predicate is at work when clarifying what is or is not of a theory.
A theory whose implications are of an other is extended by it. Theories which extend each other are equivalent. The implications of a theory with premises are of the theory or implied by the premises. Since implication is reflexive, the premises of a theory are among its implications and, thus, are of it.
That is all that need be said of (the logic of) theories in order to paraphrase talk of them in Shoenfield’s statement of the reduction theorem on page 42.
In order to more clearly reveal that the reduction thoerem obscures implication as validity of the conditional, some work on cedents, my take on Gentzen’s sequents, helps.
A cedent is a conditional whose antecedent is a conjunction and whose consequent is an alternation. The components of the antecedent conjunction are called its antecedents and the similar for its consequents. Since a conditional is equivalent to the alternation of the negation of its antecedent with its consequent and the negation of a conjunction is equivalent to the alternation of the negation of its components, then a cedent is equivalent to the alternation of the negation of its antec3dents with its consequents.
An iterated consequent conditional is a conditional with a nonconditional consequent or one which is an iterated consequent conditional. The equivalent alternation of a cedent is equivalent to the iterated consequent conditional of the antecedents with the alternation of the consequents.
Multiple premises imply a conclusion when each lexsub in the conclusion denotes where the same in each premise does. In terms of truth rather than denotation, multiple premises imply a conclusion where each lexsub in the conclusion is true when the same in each premise is true.
The problem now posed is somewhat different than that of Shoenfield’s reduction thoerem: when multiple permises imply a conclusion is that conclusion implied by a conjunction of them? Technically, to phrase it like Shoenfield, the question is asked “When multiple premises imply a conclusion is it the consequent of a valid iterated consequent conditional of them?” Or, conveniently, is it the consequent of a valid cedent of them?
At first, this seems to me to be a dividing line between what is called on one side constructive methods and on the other nonconstructive. Examples should help. The simplest example I can think up is a theory each of whose premises are all the variants of some one. Since each variant implies each other, then nothing implied by all of them isn’t implied by one of them.
Next that occurs to me is a theory like Peano’s arithmetic, where each instances of the schema of induction is a premise.
It has just occurred to me that this is the problem which is now known as Lindstrom’s theorems. Specifically the second. This has profoudn implications for my own methods of logic.
The problem can be given by asking if the following is true: each lexsub in the conclusion denotes where the same in each premise does, if and only if some compound is such that it is compounded only and solely of premises and implies the conclusion.
In this case I am relying on the assumption that compounds do not have an endless sequence of subcomponents (subcomponents are also known as ‘proper subcompounds’).
…
Now for the long and laborious task of downloading each chapter of “Model Theoretic Logic” by Barwise and Feferman from Project Euclid
(this is one of those things that AI people would tell me is trivial now and to which I would ask “at what cost?”)
…
In the mean time, my logical methods go unhindered by factoring out the definition of implication which corresponds to validity of the conditional, and that which purports to no longer so correspond.
I’m tempted to say “immanent implication” and “transcendant implication”.
I was told to watch the movie “I Swear”. It is a 2026 British biodrama a friend saw on Netflix.
Recommendations shall now be recorded here more often than they ever were.
The logic of a theory is its implications. Instead of saying ‘the logic of a theory’ I shall just say ‘the theory’. There is more to a theory than its logic, but not here.
Since implication is reflexive, the premises of a theory belong to it. Better to say that they are of it so as not to raise problems posed by a theory of classing classes e.g. set theories. The phrase ‘are of the theory’ yields to paraphrase into ‘x such that some y..z such that y is a premise, .., and z is a premise and the conjunction of y,.., with z implies x’.
A schematic approach is needed to deal with theories indifferently, and the peculiarities of a given theory follow from the particular instance of such a schema. Though, as mentioned in the earlier note on Quine’s extension of his schematic methods to theories of identity, Russell’s comments at the beginning of “An Introduction to Mathematical Philosophy”, to the effect that the predicates and constants of a purported theory of arithmetic do not specify an interpretation (e.g. ‘zero’ can be played by 100 or perhaps even a googol as long as the others are interpreted correspondingly), remains in force. Hence my complete lack of talk of schema in my memo on logic.
Said another way, there is actually no particular benefit from working with schema as opposed to the grammatical constructions of a given theory. This is far less surprising than it sounds. The key is that substiution for lexicals permits the kinds of work once done by schema. This is all the more simple in a pure predicate logic as one without variables.
…
This fragment contributed to my grasp of the logic of theories. While I accept, quite tentatively, that there is more to a theory than its logic, I can not reach whatever this more is other than through its logic.
I wonder if I am ready to write publicly about “AI”. My outlook is unique. It is that of a radical behaviorist. There are not many of us. There are even fewer of us that have any significant background in logic, math, and physics. I mention them because they play a profound part in dealing well with a new science in a turbulent world.
…
I was not ready.
I shall drown the failure of the last note with my reading of “A Contemporary Rhetoric” by Robert W. Daniel of 1967. It is a delightful thing, and very well written as far as I have already read through it. It is a profoundly Aristotilian document. The substance of a document is its meaning and it is good when it functions well with respect to its essence. Though the essence is now quite modern in its “form” as the content, idea, or meaning veiled by the words spoken aloud or written on the page.
“This book therefore begins with the substance of writing.” pg. v
“methods of attaining concreteness, and, second, the importance of a theme or thesis as th at element of the substance which controls the choice of details to be included for the development of the meaning.” pg. v
There is some backlash in the modern age over the emphasis of structure over and above substance. This is violently demonstrated as an attack on the five paragraph essay. The five paragraph essay is nothing more than a text with a beginning, a middle, and an end. The middle is concrete and the rest is abstract. Consequently, the concrete core is three paragraphs and the beginning and end are one. The problem as people see it is that such writings are structures without substance.
If it wasn’t clear that we were in Aristotle’s territory then perhaps now it certainly is. What is strange is that nothing of substance is very often taken from Aristotle when it comes to how we are to discover the nature of substancies i.e. the forms immanent in them and their ends. Never mind, it is no matter, for modern rhetoric is religated to ‘style’ as something for artists or taste makers to worry about.
Thankfully, modern noise had not intruded upon the sixties, and Daniel’s book is one man’s earnest attempt to help the reader write well. Unlike modern books on matters of style, he gives a philosophy of writing. This is much better than much of what goes as rhetoric these days. It is not a bad thing to have reasons laid down for why you write one way and not another. If that is all that rhetoric is, the amassing of such reasons and perhaps even the straightening out of any kinks, then I am all for it. Were I not for it then I would surely end thought all together.
For me, logic and rhetoric are, not always distinct, sorts of control over behavior. They are to be classified as verbal controls over behavior in that they are a special kind of social control over behavior which has been conditioned precisely because of the reinforcing effects of similarly conditioned repertoires. Social behavior which has this feature is surely special, and most likely worthy of being called verbal. I wish I knew more of the unknown history of the evolution of cultural practices. There is no way to examine the past that proceeds modern verbal practices. There is a gap between basic social behaviors like those observed in any experimental analysis which involves the behaviors of more than one organism in such a way that the response of one is contingent on the response of one other, and the verbal behaviors observed right here as reasponses of reading what is written on this page.
…
But, here I have strayed far from Daniel’s words, sentences, paragraphs, pages, chapters, and book. Before I go off and lose myself like I did in the last note, I must drop this wayward writing and return to reading the next page of his text. Presumably that is what occurred before the following three dots appeared below.
…
Family visited.
A more faithful paraphrase of Shoenfield’s “Reduction Theorem” in his “Mathematical Logic” than the one given in a late note, is given in terms of iterated consequent conditionals. An iterated consequent conditional is a conditional or one whose consequent is an iterated consequent conditional. These are grammatical constructions appropriate to one generalization of implication to multiple premises.
Each such compound is equivalent to a cedent as a conditional whose antecedent is a conjunction and whose consequent is an alternation. The components of the antecedent conjunction are usually just called its antecedents and the similar for its consequents.
Iterated consequent conditionals are equivalent to the cedent whose antecedents are those of each of its nested consequent conditionals and whose consequent is that of its deepest consequent conditional.
Where the reduction theorem becomes most interesting is in giving a particularly thorny definition of implication with respect to many premises (and with respect to Quine’s analysis of validity into either truth and grammar, or denotation and grammar):
One component is implied by many others when each lexsub in it denotes where the same in each other does.
Here I have given implication in terms of denotation, but there is the classical method with respect to truth:
One component is implied by many others when each lexsub in it is true where the same in each other is too.
This is quite different than a definition along the lines of:
One schema is implied by many others when each elxsub in it is true where the conjunction of some of the others is too.
As I am writing about it now, there is no reason to avoid ‘if’ for ‘where’, though I shall continue to do so because it is in line with Quine’s method of vagueness which is necessary for this kind of presentation lest I have to first do damage to the reader that must later be undone (however impossible that is).
In a text like Shoenfield’s on “Mathematical Logic” the mathematicians have direct contact, or so they say, with the sets they talk about e.g. sets of sentences, sets of sets of sentences, and so on. Then, instead of saying that many sentences imply another, they simply say that a set of sentences implies another. They use almost the same definition as I did with ‘implied by many’ but now restrict the many by their membership to a set:
One schema is implied by a set of others where each lexsub in it is true when each lexsub of each sentenece in the set is too.
For the mathematician who is already in touch with sets, this encourages the further use of methods involving collections when setting up arguments for or against proposed methods of mathematical logic.
Since my access to existence, or my contact with the world, is limited in Quine’s natural way, I find no comfort in the precision of a definition couched in the sentences of a set theory. It is said that there is nothing to fear from these methods, because they are carefully considered in the end, and, ultimately, there are only countably many sentences anyway so no grand abundance of sets is ultimately involved in the arguments presented through set theory.
Along this vein, some have taken the arithmetization of set theory quite far in a direction very much unexpected by those who began arithmetizing languages. Set membership is analyzed as a predicate of natural arithmetic and limited by the less than obvious consequences of its premises on such an interpretation.
For my part, I am only given the vague theory of grammar and truth from which validity is analyzed as truth of each substitution of lexicals for perhaps extra lexicals. This gives me, vaguely, the items of the grammatical constructions included in such a theory. Though, as much as possible, I have found it helpful to put off the peculiarities of particular grammatical constructions, like conjunction or negation or alternation, and deal with their classifications relative to the inheritance of, e.g., validitiy of their compounds from their components or vice versa.
Thus, my appeal to conjunctions of premises when giving a generalized definition of implication with respect to multiple premises. The temptation with Shoenfield’s methods, and with all those that follow him in form such as Ebbinghaus et al., is to fixate on what they call the semantics of the language as what it means when declaring the essence of implication. The mistake in this method is supposing there is something for sentences to mean.
But, there is more to it, and I hope thinking here about the reduction theorem makes it easier for me to point up just what is the more there is to it.
…
An important step which Quine takes in the presentation of the methods of logic in his “methods of logic” is to speak of interpretations of schemata vaguely. He points to interpretation of predicate, sentence, and singular term letters as giving, in some unspecified way, a paired predicate, sentence, or singular term. Then this is contrasted by pairing each letter with some perhaps nongrammatical entities which settle whether the sentence letter is to be true or false, what the predicate letter is to be true or false of, and what the singular term letter is to designate. Finally he says that one of the reasons we do not care yet about what exactly we mean by this vague definition of interpretation is because when we are given a predicate to go with a predicate letter such a predicate is true or false of some things and not others, the pairing sentence is true or false, and the singular term designates this or that singularly, however unknown all of this is to us as logicians.
Then, when he finally confronts the problem of eliminating the vague definition of interpretation, he appeals to the hilbert-bernays-lowenheim completeness thoerem which shows that there can be no consequential difference between the purportedly different interpretations of interpretation in a world, purportedly liek this one, with the arithmetic of the natural numbers.
Whether this is ultimately satisfactory to the reader is by the way. Interpretation can be left vague and the definition of validity can be broken into pieces as each person administers their own analysis of interpreation into substitution of sentences, predicates, and singular terms for sentence, predicate, and singular term letters, or into associated truth values, sets, and items, or whatever else your heart desires, be it even propositions, propositional functions, and particulars.
…
This is not all as obscure as it may seem. For example, proofs in Shoenfield’s book are taken as concatenations of formula where formula are concatenations of symbols. Most people assume that concatenation is an innocuous thing e.g. there are no endless concatenations. Tarski’s editor’s premise says otherwise. My grammatical methods are equally suspect.
…
I have failed to simplify the statement and proof of the reduction theorem here.
I had to make an unfortunate edit to a previous entry on the premises of a theory of the arithmetic of natural numbers. The note was copied from pen and paper and I changed lettering as part of the transcription. What was a ‘w’ became a ‘v’. This mistake is at least somewhat excusable due to the similarity of their shapes.
I did not go back and correct another sorta error in the proof of my paraphrase of Shoenfield’s rendering of the reduction theorem in his ‘mathematical logic’. I mention that iterated consequent conditionals (that is, either a conditional or a conditional whose consequent is an iterated consequent conditional) as equivalent to a single conditional whose antecedent is the conjunction of the antecedents of each relevant conditional in the iterated consequent conditional, but then proceeded to use the equivalence in the wrong direction (it should have been from the single conditional to the iterated one, but I did it backwards).
Another mistake: I mentioned the principle of universal generalization in the second half of the proof, but it had already been used in the first half and thus could have been mentioned there.
…
I left to go play the piano for about half an hour.
The ‘reduction theorem’ on page 42 of Shoenfield’s “mathematical logic” seems to be a special case of implication as validity of the conditional. Its relation to this fact is obscured by the word ‘theorem’ and the method of proof there adopted (a Hilbert style deduction system i.e. a concatenative theory of schema). Is it?
First, the theorem as it is:
“Reduction Theorem. Let G be a set of formulas in the theory T, and let A be a formula of T. Then A is a theorem of T[G] iff there is a theorem of T of the form B_1 only if … only if B_n only if A, where each B_i is the closure of a formula in G.” pg. 42
As the theorem is presented in Shoenfield, I can not make sense of it. Bold letters are used (I have not inserted them here because they serve no purpose not already served by the bold insertion of capital letters into regular discourse), no quotation marks are used, and even though I am told repeatedly that there is no ambiguity to the autonomous use of a concatenation as its own name, I can not grasp the grammatical structure of the theorem as is. This does not even begin to scratch the surface of the notational abbreviations at work.
Here is my first attempt at paraphrasing the theorem:
Reduction Theorem. Let G be a set of formulas in the theory T, and let A be a formula of T. Then A is a theorem of T with G if and only if there is a theorem of T that is a conditional whose antecedent is a conjunction of (universal) closures of formula in G and whose consequent is A.
Specifying a theory in Shoenfield’s text, and in the methods of mathematical logic generally, is a burden. I shall begin by attempting to give a consistent account of how I deal with theories.
The sentences of a theory are implied by its premises. A theory whose sentences are those of another is extended by it. Theories which extend each other are equivalent. The sentences of a theory with additional premises are sentences of the theory or are implied by the additional premises. Therefore, a theory with additional premises extends it.
Now, a second attempt at paraphrasing:
Reduction theorem. A sentence of a theory with additional premises is the consequent of a conditional of that theory (without additional premises) whose antecedent is a conjunction of the closures of some of the additional premises and vice versa.
So, it may be easier to say the other way around as follows:
Reduction theorem. The consequent of a conditional of a theory whose antecedent is a conjunction of the closure of some sentences is a sentence of that theory with those sentences and vice versa.
If notation is still needed, because crossreferencing is hard, then here is a further paraphrase:
Reduction theorem. The consequent of a conditional of a theory T whose antecedent is a conjunction of the closure of some sentences S is a sentence of T with S and vice versa.
But, either of the last two paraphrases is then a special case of implication as validity of the conditional. The argument for the reduction theorem in this form reveals it clearly enough:
Note, this argument is simplified by appealing to my definition of ‘cedents’ which are analogous to the sequents of Gentzen. A cendent with a consequent cedent is equivalent to one without it. Cedents are equivalent to iterated consequent conditionals.
The reduction theorem might be better called something like ‘the logical elimination of theories’, or, if we must maintain some two word phrase like ‘logical validitiy’, we can talk of ‘theoretical validity’ as validity of the appropriate conditional. Expedience of method is my only concern, so if talk of theories is dispensible, as was just shown, then I shall dispense with it.
Also, talk of theories makes validity out to be some degenerate theory e.g. one without premises. What a surprise it must be that there are sentences of a theory without premises.
Hopefully the argument just given shows how the reduction theorem obscures implication as validity of the conditional. From such a profound principle, talk of theories is as trivial as I have tried to make it in my definition given only for the sake of paraphrasing the reduction theorem in Shoenfield’s book.
Done bad at keeping up with notes.
Prenexed alternational form of a truth functional and quantificational schema can be given an inductive definition as follows.
The prenexed alternational form of a(n)
Full distribution of a conjunction can be thought of as the purification of conjunction with respect to alternation, though ‘with respect to alternation’ is implicit in the purification since there is no distribution of conjunction over negation which is appropriately called a purification since the putting of a conjunction under a component negation introduces not a conjunction but rather an alternation as in the equivalence of ‘p and not q’ with ‘not (not p or q)’.
Since each truth functional combination is equivalent to a compound of negations and alternations of its nontruthfunctional subcompounds and existential quantifications are equivalent to the negation of the universal quantification with respect to the same variable of the negation of their component, each schema can be put into a prenexed alternational form.
The prenexed alternational form of a schema is not unique under the routine given above. This is clear enough from the careless restraint on disjoint variants. Alphabetizing the lexicon and extending it lexicographically to compounds wtih respect to their length imposes an order on variants which obtains a standardization, but does not enforce uniqueness.
Uniqueness can be sought in a deviant way by demanding the postfix be the lexicographically earliest and shortest alternational form, but there remains the problem of parallel quantifiers prior to disjoint variant conversion.
It may so happen that the first quantifier of each prefix of the prenexed alternational form of each component of an alternation are identical existentials in which case a preliminary step can be inserted that takes them out and prepends them to the prenex alternational normal form of the alternation of components with one less leading existential quntifier.
Does this further standaradization confer uniqueness on the prenex alternational form of a schema?
…
A word of warning that occurred to me while trying to answer the question I just asked. Although a schema’s validity is indifferent to universal quantification, this does not say that a schema is equivalent to its universal closure. The validities ‘Fx only if some item is y such that Fy’ and ‘each item is (x such that Fx), only if Fy’ are certainly not biconditionals.
The root of the problem seems to be that the appropriate biconditionals are not vacuously valid. In a universe without any items whatsoever, there is no thing designated by the free variables, and we are to interpret the predication with an unbound variable as we would its elimination as a dummy singular description via the use of an appropriate predicate abstract to emphasize the abundance of predicates e.g. ‘Fx’ is equivalent to ‘some item is z such that Fz and x=z’ which is equivalent to ‘some item is z such that Fz and z is (u such that x=u)’ whereupon a one place predicate ‘X’ can be introduced as coextensive with ‘u such that x=u’ so that ‘Fx’ is equivalent to ‘some item is z such that Fz and Xz’ where the existence and uniqueness of such an item as purportedly designated by ‘x’ in the open sentence ‘Fx’ can be given as a relevant premise on the one place predicate ‘X’ i.e. ‘soem item is u such that Xu’ and ‘each item is u with v such that Xu and Xv, only if u=v’.
Alternatively, we can go straight to singular descriptions as ‘Fx’ is equivalent to ‘F(the u such that Xu)’ which the method of contextual elimination by Russell’s method gives ‘some item is u such that Fu and each item is v such that Xv if and only if u=v’.
Bought used books. Many in the complete works of Mark Twain. Completed my set of the Great Works. Got highschool USA civics and history texts from 1930. “Mansions of Philosophy” by Durant. Odd one: “the art of thinking”.
Must index library.
While I am thinking about it I have to get down the latest edition of my premises for the arithmetic of natural numbers in terms of the two three place predicates “adds” and “multiplies” as in “Five adds three with two.” and “Six multiplies three with two.”
The premises for ‘adds’ are the closures— closures are always universal— of
The premises for ‘multiplies’ are the closures of
The premises for ‘adds’ and ‘multiplies’ are the closures of
The schematic premise for finite induction is the closure of
Notes on the 14th lecture of “A History of Philosophy” by Arthur Holmes on Aristotle’s Ethics
Aristotle’s philosophy is teleological i.e. end oriented. His ethics comes from the functions of the soul. The good is proper function, “a complete life in accordence with reason, with virtue”. Virtue is proper function.
Virtue is good and proper function; functioning with respect to the inner telos. The greek term for virtue is ‘arete’. It is looser than virtue e.g. it means ‘excellence’, ‘quality’, etc. A human quality, a whole life lived in accordance with what it is to be distinctively human.
Virtue has to do with the whole life: the outward behavior and the inward disposition. Virtues as moral dispositions. Disposition is what disposes you to certain kinds of behaviors. Virtuous people are inner directed, they do not just respond to external stimuli.
How do we acquire proper functions? How do we actulize this capacity for functioning as humans?
Aristotle emphasizes habit and habit formation. “A habit of the heart” is a settled moral disposition. What we talk about today as moral development.
Character is a set of virtues or vices unified in some way.
How does habit formation work?
Disposition in accordance with virtue is a habit formed with respect to reason. There must be deliberation and choice/decision. When there are two or more ways where you should go you must deliberate about ends, the telos, and the means to those ends. Then choose in line with it. Doing it once doesn’t make a habit, you must do it repeatedly: repeat, repeat, repeat.
“One sparrow does not a summer make.”
Among those who can not rule themselves by reason must be rulled by those who can e.g. children and slaves. Discipline sets up proper function.
Now, how does the deliberation work? What kind of process is it? How does it guide? Govern? It seeks a mean between extremes.
Good functioning in vegitative life:
Good functioning in Animal soul:
Good functioning in the rational soul:
The virtues of intellect
On the sensitive life. Feelings can be felt too much or too little. We must feel at the right times, the right objects, the right people, the right motive, in the right way. Such a balance is what is to be found in deliberation. Moral virtues are a matter of controlling the moral life.
Example Moral virtues:
For the presocratics, what is just ordered or good is balance.
What is the place of pleasure in the moral life to Aristotle? Pleasure is not the same as happiness, where happiness is well-being of proper functioning. Pleasure is more of an emotion, it is subject to that deliberation on the mean. Pleasures are intermittent, contingent, beyond our control, and come in a wide variety each of which is of dubious moral worth. Pleasure is a side effect or byproduct of actualization.
What is the function of government and the place of the arts in moral development? Plato said they contributed to the betterment of the soul. Aristotle agrees. Government brings out the good.
Aristotle sees government through the human telos. Humans are both rational and social. Everything by nature has its final cause, by virtue of its nature, its essence, its form.
The ideal state is one that is properly functioning in the light of the nature of the state and its proper end. A just society is rationally ordered for the common good. The common good as the good for the individual at large.
Proper functioning of
Art as immitation. Art is immitation of life to Aristotle. Art is representation of life, of characters, their emtions.
Art helps us to see the forms through the particulars. The universal through the particular. Thus, poetry is more scientific than history. History narrates particulars, but poetry arranges particulars to uncover universals.
A good tragedy must produce catharsis, of the emotions of fear and pity because they easily rob us of our courage. Catharsis purges emotion.
Epistomology as a corrolary to metaphysics, then everything else follows.
Raymarching lets a blind bull safely charge in a china shop.
Imagine someone blindfolds you and takes you into a room filled to the brim with delicate glass sculptures that shatter into millions of pieces the moment they are touched. They spin you around until you’re dizzy. Your goal is to walk as far in the direction you are pointing as you possibly can without bumping and breaking any of the glass.
Once you regain your balance, you are told that you are only three steps away from the nearest sculpture. This means that if you take a little less than three steps you are assured you will not run into anything. You take a little less than three steps. Now you are told that you are still a little less than three steps away from the nearest glasspiece. A little less than three steps later you are told you are about six steps away from the nearest. You take about as many steps but not more.
Now, they say you are two steps away from the nearest. You take a little less than two steps and now you are told you are very near a piece. You’ve gone a little less than 14 steps from where you started. If you go any further you will certainly shower the floor with glass.
Raymarching is the same process. All you need to make it work is someone, or some thing, that can tell you how far away you are, or the tip of your ray is, from the nearest surface. To each point of the world we assign a number that says how far away it is from the nearest surface. Such an assignment is a mathematical function called a field. If we give negative numbers for when we are inside a surface, the field is called a signed distance field, or sdf for short.
The value of the sdf at any given point tells us how far we can safely go in any direction without running into a surface. If the value is negative it tells us how much we have to backtrack to surface.
We can construct complex sdfs from basic functions and build a virtual world that raymarching lets us visualize as if we were standing in it with only a rangefinder for .
Here is all the javascript code you need to set it up and view a virtual sphere from scratch:
// theory
let mm=0.001,dam=10,bf=255/dam,l=43*mm,w=36*mm/2,h=24*mm/2
,W=256,H=Math.floor(W*h/w),u=2*w/W,v=2*h/H
,head=[0,0,0],eye=[1,0,0],hat=[0,1,0],ear=[0,0,-1]
,L=(x,a,y,i=3)=>{while(i--)x[i]+=a*y[i];return x;}
,march=(f,p,u,N=64,s=1,d=mm)=>{while(s<dam&&mm<=d&&N--)s+=d=f(L(p,d,u));return s;}
,dot=(x,y,i=3,a=0)=>{while(i--)a+=x[i]*y[i];return a;}
,norm=x=>Math.sqrt(dot(x,x))
,unit=x=>L([0,0,0],1/norm(x),x)
,fog=(f,r)=>march(f,[...head],unit(r))
,ray=(i,j)=>L(L(L([...head],l,eye),h-v*i,ear),w-u*j,hat)
,pel=(f,i,j)=>bf*fog(f,ray(i,j))
,set=(px,f,i,j)=>px[3+4*(W*i+j)]=pel(f,i,j)
,rnd=n=>Math.floor(Math.random()*n)
,draw=(px,f,n)=>{while(n--)set(px,f,rnd(H),rnd(W));}
,sphere=(center,radius)=>p=>norm(L([...p],-1,center))-radius
// experiment
,px=new Uint8ClampedArray(4*W*H),id=new ImageData(px,W,H)
,C=document.createElement('canvas'),ct=C.getContext('2d')
,put=(ct,id)=>ct.putImageData(id,0,0)
,show=(px,f,n,ct,id)=>{draw(px,f,n);put(ct,id);}
,scale=Math.min((window.innerWidth-10)/W,(window.innerHeight-10)/H);
C.style.width=scale*(C.width=W);C.style.height=scale*(C.height=H);
C.style.background='white';C.style.imageRendering='pixelated';
document.body.append(C);show(px,sphere([2,0,0],0.5),200000,ct,id);
The units are meters e.g. ‘dam’ is short for decameter.
Another collection of books that needs to be properly indexed in order to better integrate my sources into my work e.g. a part of the complete works of Mark Twain, some books with primary sources in physics, and a few histories of cultures.
Unhappily I have been watching a playthrough of the recent James Bond game during the time when I would otherwise watch Dr. Holme’s lectures on physics. Soon I shall get back to regularly scheduled notes on various relevant lectures.
I have completed the first part on my work on the vitual world of my programming environment. Though I would like to write about it now, I have to eat. It is a complete simplification of the logic for rendering static scenes. It is the foundation for the rest of the more sophistocated graphical methods.
Simplicity starts anew every day.
On aspect ratios: https://www.braddailey.com/aspect-ratio-display-image
No good way of organizing the virtual world has occurred to me. The past few notes seemed to work at first, but then when I sat down to really program my way through them it was hard to match the math to what mattered.
What I missed from the beginning is that giving the position of each coordinate frame relative to the previous frame introduced ended up adding complexity. The problem to solve is to find the world coordinates of the pinhole of the camera and the representative point of a picture element, ‘pel’ for short.
Rather than keeping my eye on the world, I got caught up on the relative frames. The only way such relativity works out is with an eye to the mathematics of it all. When programming, math is secondary to effective calculation. The projection of the theory down to the program is less important than the calculations generated by “physical intuition”. I am not committed to this, but rather trying it on as a new approach to solving the problem of simply simulating the programming environment.
Moving the camera through the world can be achieved by moving the world around the camera. I already know that many simulations implement the latter because the further away the camera is from the origin of the world the more problems there are with accumulating errors in calculations.
I’m not going to do that yet though because it breaks the concrete approach I am going for e.g. we tend not to talk about the world moving around us when we walk through the world.
I am going for a concrete approach not a physical one. Physical rendering is far more complicated than what I have in mind. I am more interested in a program with the fewest moving parts that achieves the greatest articulation needed for simulating the virtual world that is the programming environment.
Maybe I should just take the more mathematical approach. Transform to and from one coordinate frame to another, and then compose as needed e.g.
wcp=[0,0,0]// origin of the world coordinate frame
w0=[1,0,0]// first basis vector of wcp
w1=[0,1,0]// second
w2=[0,0,1]// third
wv=(x,y,z)=>L(L(L([...wcp],w0,x),w1,y),w2,z) // world vector with coordinate (x,y,z)
Ugh, no this goes to far into the math already.
Maybe if I start telling a story.
…
You are led into a dark room with a pin in your hand. The door is closed behind you. Your eyes adjust. Cardboard is taped over the windows on the front wall. Dim light leaks in around the edges.
The front wall is full of windows covered with cardboard. The tape holding it up lets a dim light through around the edges.
A pin pricks a hole in the center of the front wall of a dark room. Light through the pinhole is cast upon the back wall.
…
No, this isn’t working either. It is just slowing me down. What even am I aiming at here? A simple explanation that almost immediately yields the appropraite implementation.
Maybe if I try jumping forward by talking about how it should all come together. This may simplify some of the design decisions that I am faced with, or give me something to get me over this hump.
…
This will have to be for tomorrow me. I am apparently too tired to keep writing.
The last note gave the building blocks of the pinhole camera method of imaging virtual worlds. It set up world, camera, image, and receptor coordinate frames. Raymarching from a point of a pixel of the receptor through the pinhole of the camera and into the rest of the world gives a way of displaying the virtual world onto the screen of a computer. It solves the problem of predicting what parts of the world get projected through the pinhole of the camera onto its image receptor.
Pixels are rectangular regions of the image receptor of the virtual camera. There are also pixels which emit or augment light which is seen by the computer programmer looking at their computer screen. Since most people talk about screens pixels when they say “pixel”, it is better to call the elements of the image receptor “picture elements” or “pels” for short. Had this occurred to me prior to writing the last note, then I would have used that term throughout it. As the language changes, the problems and their solutions change. I am no longer solving the same problem I started with, nor am I necessarily solving a simpler or more complex version of some old problem. All I can say with any certainty is that I am solving a new problem, or at least a fresh one.
Raymarching occurs in the world coordinate frame, but the tail of the ray marched is first identified as a point of a rectangular picture element of the image receptor of the camera. The tip of the ray is teh pinhole of the camera, and that is also the origin of the camera coordinate frame. The position of (the origin) of the camera coordinate frame is already given in world coordinates, so there is no more work that needs to be done to prepare it for raymarching.
Pels are arranged as rectangular regions of the receptor in the image plane.
The position of (the origin) of the receptor coordinate frame is given as an ordered pair of image plane coordinates.
By convention the origin of the receptor frame is the upper left hand corner, ‘ulc’ for short, of the pel in the zeroth row and zeroth column, i.e. the (0,0)-pel.
Thus, the image coordinate of the ulc of the (0,0)-pel is [-rw*0.5,-rh*0.5] in image plane coordinates where rw is the width of the receptor and rh is its height.
In the past I forgot to fix the units of everything and was plagued by errors that I could only report as “it doesn’t look right.”
In the past I didn’t even include the dimensions of the image sensor because I thought it didn’t matter and that I could just go directly from the pixels on the sceen to the picture elements of the image sensor. Some combination of adhoc “field of view” calculations along with other black box methods were my answer. That no longer helps me because I now understand just enough to see that there is greater clarity to be gained from a little more precision i.e. the black box only obscures the potential simplicty to come from someone like me looking for a unifying logical method to it all.
I will stick with, as horrible as it may sound, the standard 35mm “full frame” sensor i.e. 36mm wide and 24mm tall. Apparently the focal length of the 35mm camera is usually 43.3mm. The aspect ratio of 35mm film is 3:2. I just did an experiment where I put all these values into the adhoc version of the virtual world that I have running while I write this and it works wonders. It is a wonder to me why more people who write up explanations of computer graphics do not make everything as concrete as possible e.g. 35mm film has been around for a long while and there are a lot of really good explanations about how the geometry of it all works. Sure, modern technology introduces all kinds of new technicalities, but the geometry of space has not changed considerably since the introduction of standard film size.
The programming environment is a virtual world projected through a virtual pinhole camera onto a virtual image receptor hooked up to the pixels of the screen of your computer.
// extrinsic parameters of the camera
,wcf=[0,0,0];// origin of the world (coordinate frame)
,up=[0,1,0];// opposite direction of gravity in wcf
,ccf=[0,0,0];// position of pinhole of the camera (coordinate frame) in wcf
,back=[0,0,1];// direction of the 1st axis of ccf to the back of the camera in wcf
,top=[0,1,0];// direction of the 3rd axis axis of ccf to the top of the camera in wcf
,side=cross(back,top);// direction of the 2nd axis of ccf to the right hand side, by counterclockwise convention, of the camera in wcf
// intrinsic parameters of the camera
,fl=0.1;// effective focal length of camera, i.e. distance between image plane and ccf
,icf=[0,0,fl];// position of image plane (coordinate frame), i.e. principle point, in ccf
,i0=[-1,0,0];// direction of 1st axis of icf in ccf by opposite side convention
,i1=[0,-1,0];// direction of 2nd axis of icf in ccf by opposite top convention
,rw=;// image receptor width
,rh=;// image receptor height
,rcf=[-rw*0.5,-rh*0.5];// position of upper left hand corner of image receptor (coordinate frame) in icf
,hpc=;// horizontal pixel count per row of rcf
,vpc=;// vertical pixel count per column of rcf
,hpd=rw/hpc;// horizontal pixel density
,vpd=rh/vpc;// vertical piexel density
,u=[hpd,0];// direction and magnitude of 1st axis of rcf in icf
,v=[0,vpd];// direction and magnitude of 2nd axis of rcf in icf
The units of the image receptor are ‘pixels’ and the letters ‘i’ and ‘j’ are given to specify the i-th row and j-th column of the (i,j)-th pixel of the receptor. Since each pixel is a rectangular region of the image receptor there is not a unique point in each pixel associated with its region. The tacit convention in the construction thus far is that the upper left hand point of each region is identified with its pixel. This aligns the origin of the image receptor coordinate frame with the upper left hand corner of the (0,0)-th pixel.
The hue hit in the world from a point on a pixel through the pinhole of the camera may or may not change from a change of point on that pixel. Some methods march multiple rays from multiple points of a single pixel and combine them to simulate the contribution of potentially different hues. This is a complex process that I shall skip entirely.
Most people go for the center of the pixel when looking for a representative ray to march. This has been my habit.
The pixels on the screen of your computer are not the same as the pixels of the virtual image receptor of the virtual camera in the virtual world. For now, I defer working on the pecularities of a given computer screen.
Raymarching solves the problem of projecting a point from a pixel of the image receptor through the pinhole of the camera onto a surface of the world. Raymarching works first by seperating points of the world into three classes: exterior, boundary, and interior. Each point gets a sign: positive if exterior, zero if boundary, and negative if interior. Each point also gets a distance: the distance to the nearest boundary point. The function which assigns to each point in the world a signed distance to the nearest boundary point is called a “signed distance function”. Functions which assign values to points are also called “fields” so it is just as well to say “signed distance field”.
Raymarching starts with the ray whose tail is on a pixel and whose tip is the pinhole of the camera. The signed distance function, ‘sdf’ for short, is consulted at the pinhole. If the value is negative then the camera is inside some part of the world. When the camera is inside some part of the world it is safe to assume, for now, that it is broken, e.g. if the camera becomes part of the interior of a solid wall then it no longer works. Later, when I introduce the physical simulation of the world, and not just static pictures of it, provisions can be made to prevent the camera from entering the interior of any part of the world we would rather it not. Yet, we may find ourselves entering the interior of things, be it water or jello, but for now we defer such contemplations.
Upon consulting the sdf at the pinhole of the camera we are told how far away the nearest boundary of the world is from it. Thus, if we move that much further along the ray we are assured that either we will hit such a boundary or we will remain at an exterior point of the world.
I have made yet another tacit assumption: to get to an interior point from an exterior point you must go through a boundary point and vice versa. Sneaking past a boundary is yet another problem for another day (what would that be like and how would that work?).
The programming environment combines the kind of thinking I learned from physics with the kind I learned from math. On the mathematical side is calculus. On the physical side is something like the eye of the artist.
When solving physics problems you see the situation to yourself. In the beginning there are irrelevant details that distract from solving the problem. The first experiments in physics are always blocks on an inclined plane. It brings in the breadth and depth of the geometric problems. More importanly, it brings the physical situation under the conspicuous control of the experimenter.
While most books begin with motion in one dimension as with hovercrafts on tracks, it is hard to get a hold of such things outside the lab. Dropping bodies from heights does little to aid in navigating the physics at work.
Perhaps starting even earlier, the introduction of a scale, or a lever arm, gives the most practical instruction, that which is most likely to release the potential of the work available to any given organism.
The importantance of the physics when implementing the programming environment is first in setting up the graphics. How it helps is in diagnosing where the errors in the algorithms that draw the scene occur.
Great progress has been made on this latest version of the programming environment. First, here are various sources which I have open but which I have done a horrible job of reporting as I used them:
This material should be part of every calculus course. It brings the work to a beautiful application. I’d prefer fixed point methods, but javascript is a sad thing, and forcing fixed point methods into it would just slow everything down.
Right now I am reviewing how Lambertian Lighting works. The urge to get lighting working always takes over because the moment you get a glimpse of a world that you have programmed with the power of math, you get excited like a child playing pretend.
The logical presentation of everything will be much simpler than this hodgepodge assembly of other people’s advice. Vector calculus is best appreciated through physics. It is my hope that the implementation I eventually set up will be in standard units. Without standard units it is surprisingly hard to check whether the simulated physics is accurate to life. There are a lot of shortcuts that past programmers were forced to take because the hardware was too slow to go straight to numerical approximations. Now, the algorithms are faster, tinier, and the hardware remains bulky slow and needlessly complex. There are only a few kinds of calculations that must be done to solve almost all problems that can be given a mathematical form, and yet they are not primitive from hardware all the way up to software. This is supremely unfortunate.
For Penumbra Shadows this is what I am watching right now
I am very confused by how I am going to get colors to work with my method. I do not want colors right now, I just want an appropriate grayscale for everything.
Phew, the allergies really got to me today. Who knows if any of these notes even make sense. It is like writing while severly drunk.
Unfortunately, whatever great thinking went into simplifying my sdf+raymarching+xpbd simulation yesterday was lost to history. Where is the story in it? There was one, but I don’t know how to tell it yet.
Just showing code, like I did in the last note, is not helpful for anyone but me, and even then after a little time passes I won’t remember the real story behind the code.
I do as I did in much earlier notes, and look for a way to evolve my writing as conspicuously as possible.
This is the first time I am writing here about what I have come to call “THE terminal”. A while back, now perhaps years, I began experimenting with computer graphics and found it all to be so wildly complex that there was little to recommend it as a serious interest. The complexity of software and hardware is not to be reinforced through its continued use either by study or through new products. Upon finding the radicaly simplicty of raymarching signed distance functions, and then discovering the wildly unexpected simplicity of extended position based dynamics, I changed my tune entirely.
My earlier attempts at programming THE terminal were mixed up with my study of the J programming language. I still find the methods of J are, more likely than not, closer to being timeless than I had originally expected. It is not as if the notation is of any grand importance, it is rather the logic of its interlocking operations. The rank function is the crux of it all.
Sadly, the logic of the J programming language is overkill for the methods that run the terminal. “The term” for short.
Here are all the basic functions that go into it:
let dup=x=>[...x]
,add=(x,y,i=x.length)=>{while(i--)x[i]+=y[i];return x;}
,mns=(x,y,i=x.length)=>{while(i--)x[i]-=y[i];return x;}
,sub=(x,y,i=x.length)=>{while(i--)x[i]=y[i]-x[i];return x;}
,mul=(x,a,i=x.length)=>{while(i-i)x[i]*=a;return x;}
,lin=(x,y,a,i=x.length)=>{while(i--)x[i]+=a*y[i];return x;}
,dot=(x,y,a=0,i=x.length)=>{while(i--)a+=x[i]*y[i];return a;}
,cross=(x,y)=>[x[1]*y[2]-x[2]*y[1],x[2]*y[0]-x[0]*y[2],x[0]*y[1]-x[1]*y[0]]
,sqrt=x=>Math.sqrt(x),norm=x=>sqrt(dot(x,x)),unit=x=>mul(x,1/norm(x))
,dis2=(x,y,a=0,i=x.length)=>{while(i--)a+=(x[i]-y[i])*(x[i]-y[i]);return a;}
,dis=(x,y)=>sqrt(dis2(x,y)),sin=rad=>Math.sin(rad),cos=rad=>Math.cos(rad)
,rot=(x,u,s,c)=>lin(lin(mul(cross(u,x),s),u,dot(x,u)*(1-c)),x,c)
,e=[[1,-1,-1],[-1,-1,1],[-1,1,-1],[1,1,1]],de=[[1,-1,-1],[-2,0,2],[0,2,-2],[-2,0,2]]
,grad=(f,p,d=0.001,g=[0,0,0],i=4)=>{while(i--)lin(g,e[i],f(lin(p,de[i],d)));return g;}
,march=(o,r,sdf,M=10,m=0.01,N=128,s=norm(r),ds=m)=>{mul(r,1/s);
while(N--&&s<M&&m<=ds)s+=ds=sdf(lin(o,r,ds));return s;}
,step=(x,V,dt,C,c,G,d,p=dup(x))=>{lin(x,lin(V,gv,dt),dt);dt=1/dt;
if(c=C(x)){G=grad(C,x);d=dot(G,G);lin(x,G,-c/(sqrt(d)*(d+0.001*dt*dt)));}
V=mul(sub(p,x),dt);}
Perhaps the best thing is that the tetrahedron technique for approximating the gradient for reflections and shadows is as appropriate to approximating the gradient of the constraint function at each point in the xpbd simulation. All the basic techniques follow directly from the methods of finite differences (yet another subject I have to write up as a result of primitive recursive analysis).
The first place where I actually ran into the tetrahedron method was Inigo Quilez’s wonderful website, where I am also likely to find many other things of wonder and fun that bring the theoretical principles that make up most of my personal history.
I got taken away on a different task!
The code for the most recent version of the machine is
var d=x=>x?.[0],b=x=>x?.[1],Q=[]
,go=x=>{if(!x.q&&(x.q=true)&&Q.push(x)==1)while(Q.length)Q.shift().step();}
,cell=(...arcs)=>{let tx=false,stack
,f=()=>stack=d(stack)
,s=()=>stack=[[d(d(stack)),b(d(stack))],b(stack)]
,v=()=>stack=[[d(d(stack)),d(b(stack))],b(b(stack))]
,n=()=>undefined===b(stack)
,p=x=>{x.list=[b(stack),x.list];f();}
,q=x=>{stack=[stack,d(x.list)];x.list=b(x.list);}
,it={arcs,rx:true,list,step(){
q(it);if(n()){p();return;}
v();s();p(it); v();f();v();f();// next micromand, decode,
v();f();if(n()){f();f();return go(it);} // if drop, drop flag & drop.
v();f();if(n()){f();stack=[stack,b(stack)];return go(it);} // and so on
v();f();if(n()){f();s();return go(it);}
v();f();if(n()){f();q(it);s();p(it);p(it);return go(it);}
v();f();if(n()){f();q(it);q(it);s();p(it);return go(it);}
v();f();if(n()){f();stack=[d(d(stack)),[b(d(stack)),b(stack)]];return go(it);}
v();f();if(n()){f();v();return go(it);}
v();f();if(n()){f();stack=[stack,undefined];return go(it);}
v();f();if(n()){f();stack=[[d(stack)],n()?b(d(stack)):b(d(d(stack)))];return go(it);}
v();f();if(n()){it.rx=true;for(let i=arcs.length;i--;go(arcs[i]));
if(it.rx)return;return go(it);}
v();f();if(n()){tx=true;for(let i=arcs.length;i--;)if(it.arcs[i].rx){
tx=it.arcs[i].rx=false;
it.arcs[i].list=[b(stack),b(it.arcs[i].list)];
go(it.arcs[i]);}}
f();}}
return it;
}
and the code for the most recent version of the operating system is
onkeydown=({key})=>{i();for(k=0;k<40&&key!=" .,'0123456789abcdefghijklmnopqrstuvwxyz"[k++];i(),x());
if(u()){v();f();x();return show('rune');}// if rune, decrement, compile, return.
f();if(u()){x();i();return show('word');} //Drop, if no empty word, compile, new one.
f();v();s();p();x();p(); // Drop. Load program.
while(--l){ q();if(n()){q();i();break;} // Execute: If no next, return.
v();m();q();s();x();p();s();p();v();// increment program. get word. Part last.
if(n()){i();x();i();x();x(); // if green, make red,
m();q();m();q();m();p();s();p(); // dup word and program.
while(--l){ if(n()){f();f();f();break;} // find: If no prior, not found, break fetch.
v();s();p();x();m();p();v();s();f(); i();p();// decrement program & get its word, nil list tree parts,
while(--l){if(u())s(); // same: if no, then swap.
if(u()){v();q();x();p();s();v();q();x();p();continue;} //if no nil, right parts on list, test left parts.
s();if(u()){q();x();x();break;} // is nil. swap. if not nil, list parts, break.
f();f();q();if(n())break;v();v();p();} // Nils, drop. If no parts, break. Test next parts.
if(n()){f();f();break;} // If no parts, drop flag & word, found, break.
f();m();q();q();};continue;} // Else, drop flag, dup word. If no more, drop & fetch.
v();f();if(n()){f();p(); // if gray, drop flag, load mand.
while(--l){q();if(n()){f();break;}// mand: If empty, drop & fetch.
v();s();p(); v();f();v();f();// next micromand, decode,
v();f();if(n()){f();f();continue;} // if drop, drop flag & drop.
v();f();if(n()){f();m();continue;} // and so on
v();f();if(n()){f();s();continue;} v();f();if(n()){f();q();s();p();p();continue;}
v();f();if(n()){f();q();q();s();p();continue;} v();f();if(n()){f();x();continue;}
v();f();if(n()){f();v();continue;} v();f();if(n()){f();i();continue;}
v();f();if(n()){f();q();f();q();q();f();f();break;} f();} continue;}
f();f();} show('exd'+bins(y));} // inert hue, drop, fetch. done.
with error investigation tools
// probing & debugging tools
var abc=" .,'0123456789abcdefghijklmnopqrstuvwxyz";
function ln(x){return o==x?0:1+ln(d(x));} // length (or height) of stack
function a(n){return abc.length<=n?'?':abc.at(n+1);} // nth letter of abc
function letter(x){return a(ln(x));} // letter of stack height
function letters(x){return o==x?'#':letters(d(x))+letter(b(x));} // letters of stack
function lwords(x){return o==x?'#':lwords(d(x))+letters(b(x));} // words stacked last top
function rwords(x){return o==x?'#':letters(b(x))+rwords(d(x));} // words stack first top
function prog(){return '{'+lwords(b(d(d((y)))))
+'['+letters(b(d(y)))+']'+lwords(b(d(d(d(y)))))+'}';}// edited program
document.body.style.font='12pt monospace';document.body.style.whiteSpace='pre';
var l=1<<12; // loop limit
function show(s='',...r){document.body.textContent+=`${l} ${prog()} ${s} ${r.join(' ')}\n`;}
function bin(x){return o==x?'0':bin(d(x))+bin(b(x))+'1';} // bitree -> bistring
function par(x){return o==x?'()':'('+par(d(x))+' '+par(b(x))+')';} // bitree -> parens
function lbin(x){return o==x?'':bin(b(x))+'\n'+lbin(d(x));} // bistring stack
function rbin(x){return o==x?'':rbin(b(x))+'\n'+bin(d(x));} // bistring list
function bins(x,s=''){return rbin(b(x))+'\n-'+s+'\n'+lbin(d(x));} // bistring tree
Finally introducing the first memo to this document. It is a working draft. That is the only way I was able to get it out as a memo.
My original plan had been only to present memos in their final form, but it became necessary to place their latest working draft in the memos because otherwise I would have trouble finding the draft material among all these other notes.
I have yet to incorperate the grammatical constructions from the hints which give a breadth of basic predicate functors for the introduction of narrower kinds of validity like that in a recent note on validities of identity which follows the pattern of validities of denotative functionality, validities of closure, validities of cropping, validities of recombination, and so on.
For now I shall add, in a very preliminary and tentative way, before denotative functional grammatical constructions are introduced, the negationalistic, alternationalistic, conjunctionalistic, conditionalistic, and mutally conditionalistic classificatoins in terms of validity and inconsistency.
Negationalistic grammatical constructions are
Alternationalistic grammatical constructions are
Conjunctionalistic grammatical constructions are
Conditionalistic grammatical constructions are
Mutually conditionalistic grammatical constructions are
The terminology is atrocious, but I prefer it over nothing. It is not yet clear if these distinctions are even worth it. Tehy take advantage of Quine’s definition of validity in terms of grammar and truth which does not mention any particular grammatical constructions.
The definitions as given may need refinement. They permit demonstrations that, e.g., negation is negationalistic because it denotes waow its component misdenotes and vice versa.
It occurs to me that I can define vacuous validity after validity by simply saying
…
Without further ado, here is the memo on Logic (working draft 202608161553)
Grammatical constructions of logical languages denote each of their compounds with its components. Compounds are also components. Lexicals are only so. Constructions of categoricals match. Only categoricals with matching like components match.
Each interchange of each interchangee for each interchanger at each locator in each interchangeand matches its interchangeand and its locator is lexical; or matches its interchanger, its interchangee matches its interchangeand, and each component of its locator is lexical; or is categorical with its locator and its interchangeand, and each of its components matches the same at the like component of its locator in the like component of its interchangeand.
Each interchange matches its interchangeand and its interchangee matches its interchanger at each locator of its interchangee in its interchangeand. Substitutions are interchanges at largest such locators. Lexsubs are substitutions of lexicals for perhaps extra lexicals.
Lexsubs in validities/consistencies denote everywhere/somewhere, and in conclusions of implications they denote where the same in its premises do. Equivalence is mutual implication. Equichanges are interchanges for equivalents.
Equichange keeps validity, consistency, implication, and equivalence. Equichanges in equivalents are equivalent since each lexsub is lexical or matches the categorical of the same in each component.
Subsitution for lexicals keeps validity, implication, and equivalence since lexsubs of substitutions for lexicals are lexsubs– take the component substituted for as an extra lexical.
When there are lexicals, validities are consistencies. Components are self equivalent; and imply themselves, validities, and what their conclusions do. Only validities are implied by and equivalent to validities. Equivalents are mutually equivalent.
Other than where grammatical constructions denote, they misdenote. Lexsubs in invalidities/inconsistencies misdenote somewhere/everywhere, and in conclusions of nonimplications they misdenote where the same in their premises denote. Nonequivalents are mutually nonimplicative.
Equichange keeps invalidity, inconsistency, nonimplication, and nonequivalence. Substitution for lexicals keeps inconsistency. When there are lexicals, inconsistencies are invalidities. Components are implied by inconsistencies. Only inconsistencies imply and are equivalent to inconsistencies.
[More on nonimplication and nonequivalence is needed here!]
Constructions are
Compounds are
Matching compounds of denotative functional constructions jointly denote or exclusively misdenote, where and only where (waow) each of their like components jointly denotes or exclusively misdenotes. For example,
Denotative functional constructions denote eternal bivalent compounds with eternal bivalent components.
Alternations of eternal bivalents with their negation are valid: they denote waow their eternal bivalent denotes or misdenotes, which is everywhere, and hence each of their lexsubs denotes everywhere.
Conjunctions of eternal contradictions with their negations are inconsistent: they misdenote waow their eternal noncontradiction denotes and misdenotes, which is everywhere, and hence each of their lexsubs misdenotes everywhere.
Resolution and Alternational analysis #2026-0816-1239
Existential closures denote waow there is somewhere denoted by their component.
Existential croppings denote part of waow their component does.
What has been called truth value analysis is better called truth functional analysis. It is the mistake of Frege to introduce what he called ‘truth values’. Truth values obscure the role of validity and inconsistency in the methods of logic. They also undermine the serious contributions of Aristotle to logical methods e.g. the principle of noncontradiction.
There is much to Aristotle that can be revived with fresh interpretations of his work through modern methods. An unexpected occurrence is that his analysis into subject and predicate is as strong as it ever was. This is even in the light of Frege’s extension of arithemtic functions metaphorically to propositional functions. The propositional functions of Frege, Russell, and Whitehead are best taken as commensurrate with Aristotle’s analysis into subject and predicate via the methods of the relative clause as enshrined in Quine’s predicate abstraction and concretion.
Predicate abstraction abstracts a relative clause from a sentence by prepending the mathematical pidgin “such that” clauses with the purported singular terms of the sentence now firmly identified as free variables which are then bound in the appropriate predicate abstract e.g. “Tom with Dick such that Tom loves Dick” abstracts ‘Tom’ and ‘Dick’ from the sentence ‘Tom loves Dick’ such that when predicate of Tom with Dick, e.g. “Tom with Dick is (Tom with Dick such that Tom loves Dick)”, says of Tom and Dick what the sentence does but now only by predicating a predicate of Tom with Dick. The principle of variants makes this more vivid where, e.g., “Tom with Dick such that Tom loves Dick” is equivalent to “x with y such that x loves y” which, for most people, is more obviously a relative clause as a predicate abstract, equivalent to the two place predicate ‘loves’ outright, so that the predication “Tom with Dick is (x with y such that x loves y)” is equivalent to both ‘Tom with Dick is loves’ and ‘Tom loves Dick’.
All of this mathematically pidgin English is just a way of smoothing the paraphrase directly into a logical langauge. It is of the same kind as B. F. Skinner’s programs of instruction which build newly effective verbal behavior from the verbal repertoire available to the organism.
…
The rules of resolution govern truth functional analysis, better ‘denotative functional analysis’. Traditionally, they are given in terms of truth values or denotative values. This obscures the part that validities and inconsistencies paly in the methods of logic. It also introduces avoidable complexity early in any organiation of logical methods.
The methods of truth tables further the confusion, and truth trees follow suit. The confusions born of the division of logical methods into the syntactic and the semantic continue the mistakes made with truth values and obscure the strengths of decision methods as primary and proof methods as secondary. That every method of proof is half of a method of decision, the other half being a method of disproof, remains remote to many mathematicians and logicians. This is gravely unfortunate.
A first step in releasing the potential of the latent methods of logic is to give the principles of truth functional, or denotative functional, resolution in terms of validity and inconsistency rather than in terms of truth values, truth and falsity, or denotative values, denotational and misdenotational. For now I shall give the principles of resolution with respect to the classical grammatical constructions of negation, alternation, conjunction, and conditionalization. In a recent note I identified that grammatical constructions can be classified as negational, conjunctive, alternational, and conditionalizational, without mentioning the classical grammatical constructions specifically and then the classical constructions can be shown to belong to the appropriate classification. I submit that this aids in emphasizing the major consequences of extensionality earlier rather than later as some sort of secondary or abstract concern.
The principles of resolutiona are a part of the method of truth/denotative functional analysis. The presentation I shall give disentangles the metaphorical extension of truth or denotativity to validity, and falsity or misdenotativity to inconsistency as follows:
There are others which go further into detail or provide dodges of various problems cause by the vague difference between outright validities or inconsistencies and the others which are practically reached by truth functional analysis fascilitated by the principles of resolution.
It is a helpful consequence of switching from truth functions of sentences to denotative functions of predicates that the old truth values are dropped for what matters relative to logical method: validity and its dual inconsistency.
Predicates denote some whereabouts and misdenote some whereabouts. In either case, there may be no such whereabouts, i.e. no such occassions. Denotative functional analysis establishes each possible contribution of a lexical predicate to the validity or inconsistency of the compound in which they occur. It makes no good sense in the case of denotating predicates to presume an analysis of whether they denote or not on each of a potentially endless list of occasions because it is not at all obvious how this might ever be done to the satisfaction of logical method. That this can be placed as a burden upon validity itself, in Quine’s way where each substitution of lexicals for perhaps extra lexicals in validites, or inconsistencies, denote, or misdenote, everywhere, is a boon. It is an explanation for the survival of logical methods.
As a matter of economy, in a denotative functional analysis, the challenge is termination i.e. when does the denotative functional analysis stop? It stops when an outright validity or inconsistency appears from the operation of resolution. The budding logician is given the minimum of starting equipment with Aristotle’s principle of noncontradiction in precisely the way in which he describes it.
From the classical grammatical constructions of negation, conjunction, and alternation, the principle of noncontradiction lets us substitute at each occurrence of any nondenotative functional component of a grammatical compound with, on the left, its alternational validity (the alternation of itself with its negation) and, on the right, its conjunctional inconsistency (the conjunction of itself with its negation). Resoultion proceeds until an outright validity or inconsistency is reached, or until denotative function analysis begins again.
Careless analysis leads further from uncovering an outright validity or inconsistency (no matter how many we may have accumulated from past analysis with Aristotle’s manfest validities and inconsistencies from his principle of noncontradiction).
As an aside, there is a strengthening of noncontradiction that may yield to a wider audience who tends to forsake such humdrum matters as “excluded middle”. It can be said that no grammatically constructed component is both valid and inconsistent. But, every grammatically constructed component is both vacuously valid and vacuously inconsistent, where trivial validity is validity in a vacuous world (universal quantifications are vacuously valid and existentials are vacuously inconsistent).
One way of avoiding a careless analysis is to enforce a method of substitution for outright validities or inconsistencies with respect to some other nondenotative functional component. Schematically, e.g., ‘P and Q, only if P’ becomes ‘P and, P or not P; only if P’ then ‘P and, P and not P; only if P’. Thus reducing the method of denotative functional analysis to that of the denotative functional analysis of denotative function compounds of a single nondenotative functional subcompound.
As another aside, I finally figured out how to appropriately depict the method I use when grouping, as I so often must, the horribly ineffecient infix notation of pidgin logic. It is already implicit in the punctuation of standard English, where the semicolon is stronger than the comma. The progression of stronger punctuation issues in by repeated appending so that, from weakest to strongest, they are:
Where this becomes most important is in the generalization of connectives to their iterated kin, e.g. ‘P, Q, and R’ is short for ‘P and Q and R’ which is short for ‘P Q R and and’, i.e. with parenthesis ‘P and (Q and R)’ or with commas ‘P, and Q and R’.
There are endless diversions to these methods that are all eliminated by postfix notation which corresponds to the stack machine interpretation of predicate functor logic.
Now, the reason people embark upon a truth functional analysis or a denotative functional analysis is to establish the validity or inconsistency of the compound undergoing the analysis.
This is all sidestepped the moment it is revealed that alternational normalization is nonbranching and that a comopound is inconsistent if its alternational normalization is empty (alternatively we can give a method of alternational normalizatoin which holds onto each of its nondenotative functional subcompounds by presenting them in a canonical conjucntion of appropriate inconsistencies). The principles which permit a method of alternational normalization are broken into principles of
The generalization of biconditionalization is mutual conditionalization of which biconditionalization is the special case of two component mutual conditionalization. This distinction is important because iterated biconditionalization is not in general equivalent to the corresponding mutual conditionalization. But, mutual conditionalization is equivalent to the conjunction of the biconditionalizations of the first component with each of the rest since equivalents to mutual equivalents are equivalent and each compound is equivalent to itself. This is not a problem much covered in classical logic texts.
The principles of conditional elimination are
Thus by the appropriate interchange of equivalents, i.e. equichanges, a compound is equivalent to one without biconditionals and then one without conditionals all together.
The principles of negational purification are one half of the principles of passage for negation:
The result of negational purification is a compound whose only negated subcomponents are predications in a quantificational logic and predicates in a predicate functor logic (though in the predicate functor logic you have additional principles of passage for the recombic functors). For a truth/denotative functional analysis it is enough to leave any negations of a nontruth/nondenotative functional analysis untouched. This is a very very very very very (yes FIVE ‘very’s) important application of Quine’s principle of paraphrasing inwards. Whenever logical methods are applied, one must not intrude upon interior elements unless otherwise forced to. It is best to encapsulate them as lexicals.
It just occurred to me that the principle of inward paraphrasis is a specific case of the maxim of minimum mutilation.
The principle of distribution of conjunction over alternation is just one half of the appropriate equivalence of distribution:
Fixed point iteration of the one half of the principle of distribution of conjunction over alternation ends with an alternation of conjunctions of nontruth/nondenotative functional compounds or their negation. Each such conjunction is called a clause and each component of the clause is called a literal. An engineer might call a clause a bus and literals lines. An electrical engineer might call a clause a series of relays and the whole alternation parallel series of relays. Whether or not the relays are “normally open” (N/O) or “normal closed” (N/C) then corresponds to whether they are not or are negated. This is the import of Shannon’s paper on the logic of switching circuits.
The principle of redundant elimination can be seen as a combination of two steps or of one generalized step. The two steps are first alphabetization and then matching adjacent elimination. The one step is matching component elimination.
Alphabetization takes advantage of the principles of associativity and commutativity of conjunction:
Alphabetization begins with an order on the lexicon and the grammatical constructoins which is inhereted by the grammatical compounds in the so called lexicographic order. Then a method of sorting in lexicographic order by operations of associativity and commutativity is given. There are many options, but the one which first presents itself is that you commute components of a conjunction of nontruth/nondenotational compounds whose first component is lexicographically after its second component. Associativity is used excruciatingly to isolate the left most pair of nondenotational subcompounds (in the prefix order of their grammatical analysis as a labeled binary tree each label of which is the grammatical patricle corresponding to the grammatical construction of which it is a compound) and proceeds from the left most pair onward until an application of commutativity occurs as prescribed after which everything begins again at the left most pair. This continues until the rightmost pair is reached and no application of commutative step occurs.
Matching adjacent elimination combines the left to right associativity routine with the principle of self conjunctional elimination as half of the principle of self conjunction: components are equivalent to their self. So, conjunction of identical components are equichnaged for their component and the associative scan begins again from left to right as in the alphabetization case.
Inconsistent clausal elimination takes advantage of the principle of equivalence to alternation with inconsistencies: compounds are equivalent to their alternations with inconsistencies. Then a conjunction one of whose components is a negation of the other is inconsistent. With an appropriate lexicographic ordering we can guarentee that a component is adjacent to its negation, if at all, and then easily check, again by associative scanning from left to right, pair by pair for patent inconsistency (instances of “P and not P”) and if one is found then the entire clause is inconsistent and the whole alternation replaced by the alternation of all but that inconsistent clause.
Finally, if each clause is checked and found to be inconsistent then we conclude that the initial compound, being equivalent to each intermediate compound generated by the combining each step in this multistep routine, is inconsistent because compounds equivalent to inconsistencies are inconsistencies.
Compounds are valid if their negations are inconsistent.
It occurs to me that I made a mistake when I said that there is no branching in the alternational method: there is one branch at the beginning where a compound and its negation are placed at the top and both undergo alternational analysis in paralell or one after the other (only if the first does not end in inconsistency). If the left analysis (the non negated side) ends without vanishing then the compound is consistent, and if the right analysis (the negated side) ends without vanishing then the compound is invalid.
Thus, there is no more reason to mention truth tables or trees.
The procedures outlined here can be augmented to provide profound speedup towards their inevitable termination.
Interestingly, no use is made of the vague terms “outright validity” and “outright inconsistency”. Only and solely the inconsistency corresponding to Aristotle’s principle of noncontradiction occurs. In this way, we can see that his conviction as to its supremacy remains wellfounded.
There is a technicality with respect to the conclusion of consistency or invalidity that is worth noting. It hinges on the the substitution of lexicals for perhaps extra lexicals, and emphasizes the “perhaps extra” part in an interesting way.
For consistency there is a clause of the terminal alternation that has not vanished. For each component of that conjunctive clause, it or its negation appears, but not both otherwise the clause would have disappeared. If the unnegated component appears then we introduce a lexical which we take as equivalent to the alternation of it with its negation. If the negated component appears then we introduce a lexical which we take as equivalent to the conjunction of it with its negation. Substituting each lexical for corresponding supplemental lexical as it is or is not negated in the clause under consideration produces an outright nonvacuous validity which demonstrates, hence, consistency of the unsubstituted compound.
The dual argument deals with invalidity.
As a result of a confusion between universal quantification and conjunction, and existential quantification and alternation, Wittgenstein introduced the term ‘tautology’ alongisde his truth tables. Furthering the confusion, truth functional validity has been called ‘tautology’. Along with truth functional validity, truth functional implication has been called ‘tautological consequentiality’ or ‘tautological implication’ similarly for equivalence.
Further confusion over the the purported supremacy of model theoretic or proof theoretic methods of establishing validity have introduced misguided distinctions between, e.g.,
These are just some of the many confusions that must be eliminated if we are to overcome the problems posed by modern methods which are, in surprising cases, far weaker than is supported by past work.
I do not claim that there is a final way to distinguish each from the others, but only that they can all be put to an exact use which benefits all who use them otherwise or not at all.
Aristotle’s categories become grammatical categories in Quine’s predicate functor logic.
What was once spoken of as a change in theory can be approached as a change in grammatical category e.g. between truth functional validity and monadic validity, or quantificational validity, or, from my recent analysis of Quine’s schematic methods with respect to identity, identificational validity.
…
A denotative functional analysis is an outright validity, an outright inconsistency, or a tree whose children are denotative functional analyses of the reductions of the substitutions of a predicate in their parent for an outright validity then an outright inconsistency.
The distinction betwen an outright validity and one which is not is required in this definition of a denotative functional analysis, otherwise the analysis of complex compounds begets an infinite regress, the leaves are themselves potential validities or inconsistencies of dubious origin. This happens to also be the way in which, I submit, most people become familiar with what is unhappily called ‘truth values’. What are called truth values are a weak way of mentioning the obviousness of it all, a sort of head smacking obviousness.
…
since the completeness and incompleteness and undecidability theorems are about a century old, the practices of mathematicians and logicians have not yet come under their full control e.g. proof has been around much longer than decision and tends to steal the show even in the presence of a decision method.
…
The items of a theory are denoted by the one place predicate of the theroy. They exist in as much as some one place predicate denotes them. THus, predicates of existence of a theory correspond to its predicates of denotation i.e. ‘exists-n’ for ‘x such that some item is y such that y denotes-n x’. DOes this correspond to ‘x such that some item is y such that y is identical to x’? Go through denotation, truth, existence, and identity.
…
Predicate abstracts give Aristotle’s analysis into subject and predicate new life: they are relative clauses and are predicates which can be constructed from any sentence by predicate abstraction to denote the complex of things mentioned in the sentences according as to what the sentence says of them.
…
Logical analysis is paraphrase into logical languages.
Bob Dotson’s textbook “Make it Memorable” on journalism is coming on Monday. I am unfamiliar with the specific methods of journalism.
https://en.wikipedia.org/wiki/Bob_Dotson
Lots of little things popping up to me as I do something else.
It occurred to me that there may be a way to classify grammatical constructions purely in relation to the validity, inconsistency, invalidity, or consistency of their compounds and components.
Negational grammatical constructions are valid if and only if their component is inconsistent, inconsistent iff their component is valid, invalid iff their component is consistent, and consistent iff their component is invalid; alternationals are inconsistent iff each of their components is; conjunctionals are valid iff each of their comopnents is; and conditionalistics are valid iff their first component implies their second.
This occurred to me as a result of taking Quine’s definition of validity in terms of grammar and truth as the starting point rather than any particular grammatical construction as is tradition.
This also lets me discuss the key results on the traditional truth functional constructions of negation, alternation, conjunction, and conditionlization (and mutual conditionalization) without having to introduce them or any of their details. I submit that it shall embrace the nontraditional methods of logic by point directly to the strengths of one method over and above others.
This is in the spirit of Lindenbaum’s theorems. The difference is that no underlying set theory is present, it is merely a theory of grammar and denotation (of which truth is a degenerate case).
This kind of note is written with the expectation that there is some obvoius flaw or missing template which I have not yet contemplated adn which will appear as obvious to someone else (of which I am to myself). Stuff like this tends to happen when you are sitting around a table at a cafe after a series of talks, at least this was my experience at, e.g., the joint math meetings and the offshoots that happen thereabouts.
What good is reading if it isn’t intertwined with writing? A self satisfying silence is nothing more than ruin in progress. What a waste of upfront potential.
To imagine that I have not yet written about my work with a poet friend. A new person in my small social environment. There is so much more to say that may never come, but at least I can look back and see how I have helped others say what they must.
Most of the time, the only thing a poet, performer, or writer needs is a good audience. An audience of one can be the world to one who lives on their public acceptance, or at least its prospects. Who doesn’t want to be known well?
The cosmology of self can be all consuming.
Saw “Guys and Dolls” today. Reminded me that I played the lieutenant in a middle school production and played in the pit band of a production during high school.
Book recommendations which appeared to me on my Twitter:
For some time I have made it a task to weaponize my library as much as I am able. The first step is to index it, to gather together the books as they are arranged by topics from my everyday use of them. There remains a wall of books that I have yet to set up, but which will come along soon. Another wall is already full of books that I have read, and which form, from the center outwards, the core and periphery of my outlook on the world.
Triangulating with thinking from days past is perhaps my own sense of connection to the world. I do not feel in step with the world around me. In the past I may have said that I feel like an alien, but that is just a weak way of saying that the words of the small world immediately around me do not move me like those of the world at large.
It is not clear to me why I struggle to write about my impressions from reading ficiton. There are so many things I am compelled to write which settle themselves mostly in the realm of logic. There is something poetic about the depiction of a proof by Quine’s main method. Especially when it is cast in its functional normal form. But for a lucky selection of instances, there is only the potentially endless wave pool of not quite inconsistent instances. It certainly explains why proof tends to steal the show from decision.
There is almost a similarity between prayer and proof, but there is nothing quite so firm in prayer like there is in proof. The main method makes the procedure of proof as inevitable as it can ever be. There is something like a spiritual relief in it, but nothing that reaches the heights of revelation.
I do not myself have any religious convictions though I was raised in one which loosely held together what used to be farm land. The tradition of it all was far more important than the scripture. The organ pleased me, as it does so many of the mathematicians I once knew, but it never raised my spirits as much as it engaged my ear.
Like the difference between a chance attraction and a true love, the mystery of it all remains lost on me. This has not stopped me from trying to understand, from taking an active interest in the role religion plays in the lives of others, and, hence, the role it plays in my own as a member of a largely spiritual species.
Of those who are under spiritual contorl I distinguish between the supersticious and the contemplative. The supersticious collateral is perhaps too much for me to encourage, but the contemplative are not mindlessly enduring rituals for their own sake.
There is so much more that I don’t know than that I shall ever know, and the more I learn of logic the the less certain I am about anything else, that is anything else that comes to select so much of prevailing practice. Sure, I wish it was more science than supersticion. I submit that almost everyone accepts this in the end.
There is no reservation in my outlook. Like Aristotle, I push out the rest of the world from that little part that I have any conspicuous self-control over. While it was the boundary beyond the stars for Aristotle, it is the parts beyond the horizon of my ignorance that I respect like a tiger in the wild.
Though I have written mystically here, I would rather that than obscure my lack of commitment to much more than what the moment brings me. Even with all the logic in the world I do not see my behavior being under any kind of strict control, social or otherwise. At the same time, I do not see myself as a skeptic, a perpetual doubter who may never glimpse a scrap of knowledge. These are all natural things, ones which, upon being appropriately paraphrased, are more articulate and less certain than they seem. I find that all the more wonderful than sham certainty.
It can be hard for others to accept the presumed contradiction between my logical practice and my self proclaimed lack of commitments. As an other to myself I feel reticent at times. Is this an admission of some sort of guilt? No support either way comes.
What isn’t a blunder under this estimation? Goodness knows.
I am very unhappy with the work that I have done over the past two days. Though there has been much less time to write here than in the days before, that does not excuse me from drawing on whatever behavior I have left to at least write something rather than nothing.
The semi-significant saying from Dr. Larson is “A small fish is better than an empty dish.” My family and friends hear me say this all the time because of how powerfully consequential it has been when what I am likely to do after having said it is different from what I was likely to do before having said it.
For example, a friend was making a sale and found himself unable to budge on a price point that he had picked for his own satisfaction. I told him that “A small fish is better than an empty dish” and that he would be so much happier tomorrow having made the sale than having held to a pernicious price point. In matter of fact, he did feel bad, not the next day, but at the end of that very day.
Another letter to a friend.
“I have a book of idioms, actually I have two. When I was younger I read through many of them, but I have forgotten them all, just like the algorithms for the rubik’s cube. “Toot your own horn” must sound strange. it is already a strange thing to say in English. I enjoy such strange things because they make it clear just how foolish it is to try and communicate with language. As if there is some idea in my head that I am now transmitting through my words, like a soul stone imbued with the spirits of my thoughts, and which you are to place your hand upon to absorb into your skull’s inner gray matter, like the caramel center of a candy bar, but haunted.
I’ve been trying to figure out a bit on the history of economic theories. I have a few books on macroeconomics and some on the history of capitalism, but none which really go over the history of theories themselves. This occurs to me because I was just made aware of Thorstein Veblen’s social economic theories. He came up with the term “conspicuous consumption” as a way of “signalling” that you have social status. Buying things for more money just because you want people to see that you have more money. It is not a purely prescientific view, but Veblen insisted on keeping the individual’s behavior seperate from the group’s practices. This is a mistake in as much as individual behavior is all that can be referenced when explaining what we call group pracitces. There is no social entity above and beyond the sum of individual behavior.
B. F. Skinner wrote on economic behavior and his main criticism of past theories was that they may very well make it easy to calculate certain values according to some group statistic, but until that stitistic is established as a result of individual behavior, there is no way of ultimately explaining the role of such calculations in controlling the economic behavior of individuals. Pointing to some group above adn beyond the individual just makes it that much harder to see where the science has yet to be done.
OH PERON! I remember you telling me about him, he is an example of a piece of history that I wish more Americans got to know.
I would guess that the reason we are not taught much about it is that if more people know how it works then they will make it that much harder to extract deals. I haven’t thought too much about it. The social part of my program is so far away, and yet it ckeeps confronting me at every turn. If only there was a way to speed up my work on logic. The only way would be if someone else had already done the hard work, but no one wants to simplify science and logic for the same reason they would prefer education only work well enough to make more workers who follow orders without thought.”
https://en.wikipedia.org/wiki/Juan_Per%C3%B3n
Some story about paper boys from a book a friend read led to me learning about Thornstein Veblen. It would be grand if I could find a series of lectures on the history of economic theories. The view I have now is so distorted even though I have more than a passing familiarity with economic history.
https://en.wikipedia.org/wiki/Thorstein_Veblen
Nelson was unable to seperate logic from mathematical logic and this contributed to his redacted claim that the theory of arithmetic of Peano was inconsistent.
The risks posed by the inconsistency of the theory of arithmetic of Peano is such that it is more important for us to have the computers of the world working on it rather than something like searching for signals from organisms across the universe.
There are many more almost as important problems of that kind, and that they do not consume more of our energy is alarming. It is evidence of the broad lack of effective technologies of teaching logic.
Before leaping back to complete the note on Nelson’s proof of Hilbert and Ackermann’s consistency theorem, I wanted to explore the application of the main method to the completeness of identity theory.
It seems to me that this is a prime example of what I am about to call “lexical methods”. The methods are lexical in that there is a predicate letter which is reserved for substitution by a lexical. In the case of a theory of identity, the lexical predicate which we have in mind is the two place predicate of identity as in “This is identical to that.” Sometimes we say ‘equality’ for ‘identity’. These two terms are so overloaded by history that there is practicaly no escape from a difference in opinion over whether these are two names for the same thing, or different things, or that we have got the whole deal wrong.
How does Quine approach identity theory in “Methods of Logic Fourth Edition”? Part four of the book is titled “Glimpses Beyond”. It is impossible not to snicker at the similarities between it and “Bed, Bath, and Beyond”, i.e. “truth fuctional logic, quantificational logic, and beyond”.
The first chapter of the fourth part, chapter 41, is on singular terms. After that is the chapter on identity. Identity follows singular terms because the cases where identity shows up as being of interest is when there are different names for the same thing as opposed to either just different names of different things.
Aristotle declared, as many did before him, that every thing is identical to itself and nothing is identical to something else. Better to say it as Quine does: “Everything is identical to itself and with nothing else.” Why? Because, the phrase “nothing is identical to something else” asks us to paraphrase ‘something else’ which goes rocky e.g. the predicate ‘x with y such that x is other than y’ is presumable paraphrased further as ‘x with y such that not x is identical to y’ so that ‘nothing is identical to something else’ comes out ‘not each item is x such that some item is y such that y=x and not y=x’ which is equivalent to ‘some item is x such that each item is y such that not y=x or y=x’ which, by the main method or otherwise, is an instance of the patent validity ‘some each F or not F’.
A more forceful way of saying what Quine has said is “Everything is identical to itself, otherwise it is not.” Or, more oddly, ‘Identity is true of everything with itself, otherwise it is false.” This leaves the mystery in ‘otherwise’.
The mystery of identity is in its elmination in a language with only finitely many lexical predicates. Indiscernibility with respect to lexicon is coextensive with identity in a logical langauge with finitely many lexical predicates.
Though talk of identity tends to enter into the vernacular of a new speaker through its uninterrupted proximity to names, which come out as singular terms in a quantificational logic (also known as ‘function letters’ of which ‘constants’ are a degenerate case), it is the crossreferntial work of variables as the logical import of pronominal crossreference.
“Statements of identity consisting of ‘=’ flanked by singular terms are needed because two singular terms may name one thing. But the need of ‘=’ flanked by variables arises from a different peculiarity of langauge; viz., from its use of multiple variables of quantification (or their pronominal analogues in ordinary language). Two variables are allowed to refer to the same object, and they are also allowed to refer to different objects; and thus the sign of identity comes to be neede when, as in the above four examples, there arises the question of sameness or difference of reference on the part of the variables.” pg. 270
Already I feel as if I am seeing to myself a mistake that Quine makes in the following paragraph that also helps to explain why he may have missed the overt application of his main method to a completeness proof of identity theories:
“The logic of identity is a branch not reducible to the logic of quantification. Its notation may be thought of as comprising that of the logic of quantification plus the one additional sign ‘=’. Thus the schemata of the logic of identity are the same as the quantificational schemata except that they may contain, along with the clauses ‘p’, ‘q’, ‘Fx’, ‘Gxy’, etc, additional clauses of the form of identities: ‘x=y’, ‘x=z’, etc. Validity may be defined for schemata of the logic of identity precisely as it was defined for quantificational schema (Chapter 27)”. pg.270-271
I submit that there is nothing which distinguishes ‘=’ from the predicate letters ‘F’, ‘G’, ‘H’, and so on. This Quine misses just as I missed it when I thought up calling these ‘lexical methods’.
A more awkward way of making my point is to say that we pick a predicate letter, e.g. ‘E’, and permit interpretations which only and solely interpret ‘E’ as the predicate of identity. This avoids the introduction of a secondary collection of special predicate letters. Technically, we can go backwards and define this restricted kind of secondary predicate letter by wrangling up the restrictions on subsequent interpretations, but this merely makes a selection from what we have already been dealing with, i.e. nothing new is required of the main method which otherwise would be revised to accomodate a predicate like ‘=’ rather than the predicate letters familiar to it.
Now, under my restriction we must reinterpret what Quine says of those schema which end up valid because of something purportedly special about the predicate of identity:
“there are further valid schemata whose validity depends specifically on teh meaning of identity; one such is: (I) ‘Fx and x=y, only if Fy’. For, consider any choice of universe, any interpretation of ‘F’ therin, and any assignment of objects to the free variables ‘x’ and ‘y’. If the object assigned to ‘x’ is the same as that assigned to ‘y’, and is an object of which ‘F’ is interpreted as true, then (I) comes out true through truth of its consequent ‘Fy’; and in any other case (I) comes out true through falisity of its antecedent.”
Here we simply extend the part of the argument which says “and is an object of which ‘F’ is interpreted as true” by beginning with ‘Fx and Gxy, only if Fy’ and then ‘If the objects assigned to ‘x’ with ‘y’ are true of the interpretation of ‘G’ as identity and ‘x’ is an object of which ‘F’ is interpreted as true, then …”
This naturally puts a restraint not on the schema, but on the interpretations. Validity so constrained is no longer worthy of the name ‘validity’. This much even Quine must acknowledge as he does later down the same page:
“Any schemata obtained by putting identities ‘x=x’, ‘x=y’, etc. in place of say ‘Gxx’, ‘Gxy’, etc., in valid quantificational schemata are, it has been explained, valid schemata of the logic of identity. But such putting o identies for schematic clauses can, when thorough, yiled a sentence instead of a schema, e.g.: ‘x=x only if some item is y such that x =y’. The concept of validity may be applied to such sentences along with the schemata, following the precedent of Chapter 28.
This example is even quantificationally valid, in the sense of that chapter. Of those whose validity depends specifically on the meaning of identity, on the other hand, the simplest is: (II) ‘x=x’.” pg. 271.
Here the hint for a good name is already at hand: ‘quantificationlly valid’. It is just as good to say ‘identificationally Valid’ as an even narrower kind than ‘truth functional validity’, ‘monadic validity’ and ‘quantificational validity’.
This is all well and good, we have a new name for a new kind of validity, but what of it? It is merely another way of speaking of a restriction on interpretations. The main method is a seperate matter all together.
My first response is to now take something like Russell’s early outlook in “The Principles of Mathematics”. Identificational conditionals are a grammatical category of schemata. Their antecedents are a conjunction which is best specified by a new predicate functor. A reminder for those unfamiliar with predicate functors: the negation of a schema, e.g. ‘not Fx’, is a compound in whcih the grammatical partical ‘not’ occurs, and a generalization occurrs so that ‘(not F)x’ is take as short for ‘not Fx’ and ‘(not G)xy’ is taken as short for ‘not Gxy’ and ‘(not H)xyz’ for ‘not Hxyz’, and so on so that ‘(not F)’, ‘(not G)’, and ‘(not H)’ are each taken as compound predicate schema which when predicated of a list of variables is equivalent to the negation of the predication of those variables with the component predicate letter, i.e. ‘F’, ‘G’, and ‘H’ in these examples.
Better to just call them grammatical constructions, since Quine ends up using them to combine predicate letters with variables and sentence letters and predicate abstracts.
The gramamtical construction which I now introduce to aid in the schematic description of the category of indentificational conditionals is as follows:
Then, each instance of
is an identificational conditional. The identificational validities are then the valid identificational conditionals. But, each is an instance of the first case only since
so it suffices to say the identificational conditionals are all instances of ‘totally reflexive F and F substitutive G, only if H’.
It remains to be seen if it is better to call the consequence of an identificational conditional out as an identificaiontal validity or the whole conditional. I submit that since it is the whole conditional which is valid then it is the whole conditional which is especially indentificationally valid. This fits well with the traditional phrases “truth functional validity”, “monadic validity”, and “quantificational validity” in mathematical logic where the role of implication is overemphasized by axiomatic methods and “logical validity”, a very unhappy term, is taken as the validity of degenerate conditional with empty antecedent.
That’s all for now, more to come, as always (why I say this at the end of so many notes is mysterious, except, upon immediate reflection, it may just be a trace of my everlasting hope).
I saw the movie “Stand By Me” for the first time a few days ago. I knew it was based on a Stephen King story– I did not know that its title was “The Body”– but I did not know it was Rob Reiner film.
This continues the work done earlier on Shoenfield’s proof of Herbrand’s Method in the chapter on consistency from his book “Mathematical Logic”.
These are specifically notes on Nelson’s lecture in Fall 2004 on the consistency theorems from Shoenfield’s book: https://web.math.princeton.edu/~nelson/503/cons.pdf.
As with many mathematical logicians, Nelson’s skills were perfected through his work on Shoenfield’s text. The text is a loci of control over many a modern practice in mathematical logic, even though there are now far more wide ranging results to be found in the fertile grounds of category theoretic works and their descendant consequences through type theory and vice versa.
Nelson’s immanent set theory relies heavily on the technical distinctions given by Shoenfield. His theory is the closest attempt I have seen to realizing Quine’s proposal in his paper “Immanence and Truth”. I submit that sets are not the way to do it and neither are categories. It is the appropriate arithmetic, as in Conway’s Surreal numbers, that is where we are most likely to find effective action.
I agree with Nelson’s initial distinction “an axiom system in which one is required to understand the meaning of some of the axioms is not an axiomatization at all”. This is better put as a distinction between the Euclidean method of axiomatization where an axiom is a self-evident truth and the modern form of axiomatization where the axiom is an item in a theory of syntax, such as a theory of concatenation or trees.
I disagree with Nelson’s claim that by distinguishing between syntax and sematics we have made some sort of leap that was inaccessible to those who did not make that distinction. It is something like the mathematical version of Descrate’s division between material and mental things. Quine’s solution is the one I follow: if you must speak of a classification in terms of the material and the mental, then let it be of the predicates i.e. material predicates and mental predicates. Sort out their implications as you must depending on the scientific theory that is a going concern.
Nelson suggests that dividing formalism into the syntactic and the semantic allows for us to narrow reason down to computation. I can not support this analysis because I am familiar with the methods of logical programming where the practice of programming is shown to be a specialized practice of logic e.g. under the main method.
I am inclined to agree with his judgement that “Teh major technical achievement of the formalist program is the Hilbert-Ackermann consistency theorem”, but can not commit myself to this inclination because I do not yet fully grasp the theorem itself (hence why I am even reading this or going into it). It is relevant to me because I must see clearly how the main method can be adapted to a completeness proof for the corresponding method befitting predicate functor logic.
Consistency theormes are distinct from completeness theorems, but are closely related. Godel’s completeness theorems (Each model theoretic interpretation of a schema is true if and only if it is a theorem) does not give a method for demonstrating the theoremhood from having established the truth of every model. This, the consistency theorem does. But, also, this is what Quine’s main method does.
The main method can be described in Godel’s way: every valid schema has a proof where a proof is an inconsistent conjunction of truth-functional instances of the schema. The main method then actually tells us how to find this inconsistent conjunction if it exists.
I submit that the main method makes whatever it is the consistency theorems say relative to the completeness theorems much clearer than it has been made hitherto.
It also seems to me that the main method clearly demonstrates the superexponentiality of the potential search.
How the main method does this is by being aimed at inconsistency and by entirely avoiding axiomatizations as Nelson has championed in the introduction (and as all those following the formalist program, however strong or weak it is now taken). The main method is a clear indication, both in its exposition and proof, that the division into syntax and semantics weakens clarity and exactness in method and theory. This much should already be known from the decision procedures of truth functional logic, one place predicate logic, and those of fragments of full blown truth functional and quantificational logic.
Thankfully, Nelson cuts to the chace, as he always does: “We shall pursue cut elimination as the basic result. A cut is an inference of B or C from A or B and not A or C”. I am unable to render the sentence as he gave it, but thanks to Quine’s schematic methods (which come through von Neumann’s introduction of them) I can rewrite that sentence as follows.
‘B or C’ cuts ‘A or B’ with ‘not A, or C’.
Now, ‘cut’ is clearly a three place predicate in a formal theory, and not some mysterious inference. The sentence letter ‘A’ is the cut schema. Gentzen demonstrated that cuts need not occur in proofs from his sequent formalism.).
I submit that Nelson wrote up this paper for almost the same reason that I have been compelled to write this recent series of notes on the constistency theorems: they are so important that they must be understood in isolation from the overwhelming number of technical distinctions which Shoenfield makes and which are special to his presentation.
I was going to continue taking notes on the relevant definitions that contribute to the statement of Herbrand’s theorem, but, thankfully, Nelson has once again saved me the trouble (the other place where he has saved me a great deal of trouble is in his Radically Elementary Probability Theory about which I do plan on writing more later once I have the methods of logic sketched out to apply predicate abstracts in talk of nonstandard natural numbers).
Rather than taking proofs as expressions, e.g. as the conjunction of each schema in a sequence, Nelson regards “a proof as an expression, the concatenation fo the formulas in the proof. (An expression in a first-order langauge can be a concatenation of formulas in only one way.)” pg. 3. This is essentially the method that Gentzen promotes when he introduced sequents, though it is a sorta generalization to any appropriate grammatical construction in the formal language.
“As we establish cut elimination we also produce bounds on the expression lengths of proofs with cuts eliminated.” pg.3
Note, “a theory is open in case each of the nonlogical axioms is open”.
To each open theory Nelson introduces an omega theory whose language is the same as the open theory, but whose axioms are all the instances of the open axioms of the open theory, adn whose rulse are
to be completed
Surprised that upon receiving my copy of “Butlers & Household Managers” the acknowledgements mention Hubbard. There is nothing more to make of this now other than it is an interesting factoid.
…
Don’t forget to write about the chemical saftey board (or whatever CSB stands for) and its relation to “AI Safety”. The perspective makes it out to be the joke it is.
…
I shall soon come to write more on what makes Y Combinator work as a way of potentially designing an effective counter Y Combinator culture. Preliminary conclusions have already been made relative to the remote contact that I have to the oblique reports through Paul Graham and Jessica Livingston. I submit that they see themselves as butlers. Specifically, butlers to founders where investors are like their guests. This is only a result of vauge contemplation on wayward principles.
…
The best writing must be done with an eye to ease of translation. Even though this is frowned upon in the short term, and there is good reason for such frowns, it is an aid to clarity and exactness and effective action to do so.
Notes on the 13th lecture of “A History of Philosophy” by Arthur Holmes on Aristotle’s Epistemology and the Human Soul
https://youtu.be/AISyO6hR_fI?si=JHNErSQdZc6fFeMm
Posterior analytics is about our knowledge of first principles and the development of a science of first principles. First principles have to do with the essences, the forms, and, of course, other causes involved.
From the first principles of science follows a chain of deductive inferences so that the logical structure of any body of theoretical knowledge is that of a deductive system. I.e. science is deductive, not emperical as we presently take it. From universal first principles you can infer principles about particulars.
Deductions about particulars are demonstrative, they can be logically deduced, but the first principles are not demonstrated or deduced from other principles (because if they were then either there is an infinite regress or a circular dependence). Therefore, any chain of deductive inferences must go back to indemonstrable premises.
…
Holmes, very happily (!), makes the distinction, that so few are insensitve to today, between axioms and postulates. Axioms, in Euclid’s and Aristotle’s sense, are self-evidently true. For us, and since Hilbert’s formalism of Euclid’s geometry (it surprises many that Hilbert’s was the first actually logical development of Euclidean geometry up to truth-functional and quantificational standards), axioms are items in a theory of formalism e.g. a theory of conatenation or, better, binary strings, or a theory of ordered pairs, or binary trees.
I go on to dismiss postulates all together, making only the distinction between the antecedents and consequences of a conditional sentence and the premises and conclusions of a conditional argument. This is the wise course to take after having demonstrated, over this past century, that there are many equivalent sets of premises that each offer various tradeoffs with respect to a given deductive strategy and with respect to their contribution to emperical checkpoints.
For example, independence is sought along Hilbert’s lines so that no one premise is implied by the others, but it is often the case that when teaching the formalism of a theory (which MUST be distinguisehd firmly from teaching its logic!) it often helps to have redundant premises. Much like Russell and Whiteheads starting points in Principia Mathematica, it is only later on that we uncover an independent set of equivalent premises.
Perhaps unsurprisingly, the prime implicants of an alternational normal form are one way of getting at the fewest effective premises for a given theory. A clause of a schema in alternational normal form is redundant and can be dropped if it implies the alternation of the rest of the the clauses. A literal of a clause of such a schema is redundant and can be dropped if the rest of its clause implies the whole schema itself. (Dropping all redundant literals is a way to check for validity because it reduces the entire such schema to an outright validity e.g. an alternation of complementary components.)
Dropping all redundant literals leaves prime implicants as clauses none of can lose a literal. The consensus of clauses having complementary literals (a predication and its negation) is the conjunction of the other literals in each clause (without duplicates, of course). Including every consesus into the alternation of prime implicants and working out systematically the shortest equivalent (through various alternative sequences of dropping redundant clauses) yields a shortest logically independent set of premises as that of each prime implicant from a shortest alternational normal form (there need not be one shortest such schema, there can be many of equal length but of distinct clauses).
…
Coon’s revealed to modern philosophers that they might want to look at history of science when they make evaluations of whether its methods are deductive or not. Hypotheses as emperical generalizations i.e. Hypothetico-deductive approach.
Aristotle doesn’t accept hypotheses as first principles because emperical generalizations are falsified by one counterexample and hence are uncertain.
The problem with this outlook is revealed by Quine’s naturalism. Falsification involves theory. There is no empirical falsification independent of theoretical proof.
Posterior analytics is where he points up the problem with getting indemonstrable first principles.
How can we know indemonstrable first principles? Plato does this easily: we already know the forms, we just uncover what we already know.
Aristotle gives an induction procedure (not the kind of inductive reasoning of modern science). Sir Francis Bacon’s Novum Organum repudiates Aristotle’s induction and replaces it with our present experimental outlook.
Aristotle Induction:
These all support this “process” outlook.
Experience is the product of sensations, and contrasted with empiricism which says experience is sensation.
Aristotle’s inductive method is the intuitive recognition of the abstracted form of classified experiences as organized sensations.
“Syllogism assumes an audience that accepts its premises, induction shows us the universal implicite in the particular.”
“To know a thing’s nature is to know the reason why it is. This is equally true in so far as they are said without qualification to be as opposed of being possed of some attribute in so far they are said to be possed of some attribute such as equal of two right angles.”
“It is strange if we possess them from birth for it means that we possess apprehensions mroe accurate than demonstration and fial to notice them.” How can you know something if you don’t know that you know it?
“We must posses a capacity of some sort but not such as to rank higher in accuracy than developed states.” We have a capacity/potential for knowing these natures/forms i.e. potential intellect. It is a passive potential i.e. passive intellect.
What does it mean tos ay that a young child who hasn’t learned a thing is a rational being? It has a potency that is not yet actualized.
How does passive intellect get to be active/actualized intellect?
“an obvious characterist of all animals, they posses a congenital discrimintive capacity of sense perception. But though sense perception is innate/natural in all animals, in some the sense perception persists and in others it doesn’t. Animals in which it persists have either no knowledge at all outside the act of perceiving or no knowledge of objects of which no impression persists in which this persistaence does not come they have no knowledge. ON the other hand, animals in which it does come into being, they retain this sense impression, and when that persistence is frequently repeated a further distinction arises between those which out of persistent sense impression develop a power of systematizing.”
“Out of sense perception comes to be memory, out of frequently repeated memories of the same thing, experience, from experience again, taht is from the universal now stabalized in its entirety within its soul, the one beside the many is a single identity from this experience stablized in the soul origninate the skill of the craft. We conclude that these states of knowledge are neither innate nor developed from higher states of knowledge but from sense perception. It is like a route in battle stopped by first one man making a stand and then another until the original formation has been restored, teh soul is so constituted as to be capable of that process.”
…
It goes without saying, that I shall be taking notes on the entirety of Aristotle’s written works as well as Plato’s. Though I have done some of this elsewhere, it was only in handwritten notes or in documents that are now long lost.
The main reaosn to really go deep into Aristotle is that he is the person who I know of that has done the sort of encyclopedic rewriting around logical principles like that I must do around human behavior. Though Liebniz skethed out much of the same in an evolved form, I am not as familair with his works. They are only now being translated at https://leibnizpapers.org/ by Lloyd Strickland https://leibnizpapers.org/about.htm.
…
A number of memories constitute a single experience.
“All men by nature desire to know.”
“Experience is knowledge of individuals, art of universals.”
…
Theories of Knowledge we have met:
…
Criticism of foundationalism: the quest for certainty is inconsistent with humans. Coherentism vs. foundationalism: knowledge as justified belief vs. certainty.
…
Moving onto Aristotle’s Ethics and the Human Soul
‘Soul’ and ‘life’ are the same term. Life is distinct from matter because it has certain functions that matter doesn’t have.
Hierarchy of Souls/Lifes:
…
Aristotle’s “capacities” seem to evlove eventually into Quine’s “dispositions”.
…
Growth of a living being actulizes the potential of the soul. The rational soul has capacity for immortality.
The good is defined on the basis of the nature of the soul because Aristotle’s metphysic is teleological: everything is end oriented to final causes.
Supreme good or Highest Good or Summum Bonum: by virtue of its teleology it must be intrinsically good and not instrumentally good by nature of the rational soul. It must be inclusive: it doesn’t exclude any lesser goods, it embraces them all. Applicable to the human species: the species can participate in the good.
What is happiness i.e. “Eudaimonia”: the diamon’s are the gods or spirits that inspired socrates on his mission. Best translation “Well Being”. Is it “fulfillment”? They mean “actualizing the potential”. This doesn’t work because it is associated in modern times with “individualized fulfilment” i.e. “self actualization”.
Actualizing the rational capacity of the soul is the good.
The good life is “a complete life lived in accordance with reason” and virtue.
Complete life? A human is not a soul trapped in a body, it is a rational animal!
Among the videos to take notes on:
My reading of fiction is a study of vernacular. It finally occurred to me that technical terms are too difficult to introduce to the public without padding their introduction with a swarm of vernacular comforts.
To speak technically to a nontechnical audience is to remove them from the language of their household. Like a child whisked away from their parents by their first bus ride to school, it is traumatic no matter how much the results of it are celebrated.
While this is a defect in prevailing cultural practices, it is the state of affairs. It must be dealt with or we are likely doomed to live only the vernacular life.
Vernacular is not in itself defective. It is the absence of effective technologies of teaching technical terms that demonstrates the defect. When people learn technical things they are an accidental product of our times. This happens to be why we prize them when they do what no one expected or can explain.
Now that we live in the age of midling averages thanks to the newly minted verbal machines, the world is made to think that there is no need for anything more than the vernacular.
…
leaving this incomplete because I was whisked away and lost the thought.
Notes on the 12th lecture of “A History of Philosophy” by Arthur Holmes on Aristotle’s God
Every kind of thing that is, process of change or becoming, has to be explaiend in terms of four causes: the material cause as the subject of change, the efficient cause that exerts the force, the formal cause the esential nature of what is going on, and the final cause as the purpose.
This is true of earth, planets, stars, etc. i.e. the cosmos. The circular spatial motion must have the four fold causual motion. The material nature of the thing, basic elements, the ether, and the planets, these are the material causes. The efficient cause is the movement of the fixed stars that produce changes in the ether and maintains the rotation in the planets, that maintanes changes in the atmosphere of the earth, and the processes of the earth. Formal cause, the nature of the fixed stars, it is their nature to spin, an actualized potential. Beyond the perimeter of the universe there is an unmoved mover that does not move but moves things which are moved. It is unchanging and is the ultimate cause of all kinds of change by virtue of its influence on the fixed stars.
The unmoved mover is pure actuality. (potentiality and actuality) Every change is the actualizing of some potential. “I’m potentially over here and now I am actualizing that potential.” Every process of change is the actualization of some potential it has. Pure actuality means there is no unactualized potential.
Processes of locomotion: linear, rectalinear, circular. Circular is an unchanging locomotion. The god of Aristotle is not a creator, it does not bring things into existence from nothing. It is not an efficient cause, it does not exercise power, force, doesn’t push things around, force them to exist etc.
The materials of the cosmos always existed so there was nothing to create or destroy. The cosmos is an everlasting gobstopper.
Motion has to be produced and maintained, it is not inertial to us modern physics.
God is final cause and not efficient cause. Things are moved by wonder and love at it. Thus, the stars have some sort of soul that can love this unmoved mover and wonder at it.
The actuality of this final cause is that it thinks on its own thinking because if there was something that it perceived then it would be changed and hence no longer be an unmoved mover.
Since there is no possibility of it being unactualized then it is perfectly good. Pure actuality is pure goodness.
This is an example of what we now call “natural theology”. A theology based on inferences from what we know about nature.
Not all final causes involve consciousness.
Henry Veatch “Aristotle is telling us that man is not the measure of all things, god is.”
…
Looking at Aristotle’s logic and epistemology.
Brought together by students as “The Organon” i.e. “The Method”. Works are:
The categories deal with the way terms function in our thinking. On interpretation deals with the logic of propositions, where a proposition is somethign which asserts or denies. Prior analytics deals with the syllogism, the composition of propositions e.g. major and minor premises lead to conclusions. Posterior analytics is on scientific reasoning as “how do we get to our first premises” i.e. how do we know our basic premises are true. Topics is about dialectic, before he discovered the syllogism, used in rhetoric. Sophistcal refutations is about logical fallacies.
More on the categories:
More on the posterior analytics: how can we know unchanging truths about classes of things for certain? How do you know a general truth about a species? You have to know its differntia, its essence, its form etc.? So, how can you know the forms? You want your premises to be about the forms, the essential nature of things.
We can not know the forms by sensation: sense perception gives us particular and is relative. Innate knowledge: no, people do not know everything. So how can we know them… good time for the video to end.
I am unsatisfied with the progress that I have made over the past week or so. No great good seems to have come from my work. There have been mistakes.
The biggest one is that the javascript code at the end of the note on the proof procedure for truth functional validity (which follows from the existence of an alternational normal form) does not implement the exact proof procedure specified.
Rather than beginning with a prenexed schema and performing the proof procedure on its postfix (aka matrix), it constructs, from each connective, a compound already in prenex alternational normal form.
For example, a sentence letter is taken as a schemata in prenex alternational normal form whose prefix is empty and whose postfix is the altneration of a single conjunction of a single predicate of zero places. Predications fair similarly in that their prefix is empty and they are the degenerate alternation of a degenerate conjunction of an n place predicate predicated of n variables. Each such variable is free in the resulting schema.
The conjunction of two schemata in prenex alternational normal form is equivalent to one in prenex alternational normal form. It is equivalent to many since variants of any one are equivalent to each other. The prefixes of each component are pulled out after the second component’s quantified variables are replace by the alphabetical continuation of the first’s. The postfix of the one is distributed over the other such that inconsistent clauses are dropped in the process. Inconsistent clauses are identified as conjunctions containing a predication (sentences as zero place predicates are treated as zero place predications) and its negation. Duplicated predications in a clause are dropped during the check for outright inconsistencies.
All of the predicate letters and variables are put into an adhoc order as they are placed, as strings, into a dictionary if they are not already part of the dictionary and their index in it is what is actually manipulated.
Strictly, an n place predicate letter predicated of n variables is kept as a pair whose left part is the index of the predicate letter in the dictionary and whose right part is a list of the same for its variables. When the index of the predicate letter of a predication is negative it is interpreted as the negation of that predication in its surrounding clause. Variables which are free are positive, and those that are bound are not. To discover whether a bound variable is existentially or universally quantified, index the absolute value of the bound variable into the prefix and if the result is positive then it is existential, otherwise it is universal.
It makes better sense to make bound variables nonnegative and free ones negative since then there is no intermediate operation of sign change between a bound variable and indexing into the prefix to discover its polarity. There are a lot of things that would make better sense, but that is for the next version.
So far I have only really described how the conjunction of two schemata in prenex alternational normal form is constructed in the javascript program. Now I move on to negation, and then after that existential quantification, and then the rest are defined in their familiar way from there (this is to save space, but there are reasons to program them out individually to take advantage of skipping small steps).
Negation is, at first, profoundly simple: negate each value of the prefix so that universals turn to existentials and vice versa. Then, take the dual of everything in the postfix (matrix) and simplify. In reality, this means taking every conjunctive clause of the schema and turning it into an alternation of one place conjunctions of each literal of the clause. All together this produces a large conjunction each component of which is an alternation of (degenerate conjunctions of) the negation of the corresponding literal from the corresponding conjunctive clause. Happily, this turns out to just be iteration of the familiar conjunctive construction already given (restricted to the postfix and no longer involving the prefix). As long as outright inconsistencies and duplicates within clauses are dropped, this rarely baloons nearly as large as it otherwise would if done naively. Since we already need to drop outright inconsistencies and duplicates anyway, this comes as a boon.
Existential quantification appends a one to the prefix and replaces each occurrence of the index of the quantified varaible in the dictionary with the negated length of the prior prefix.
Now, if the postfix comes out empty, you have uncovered an inconsistency. Validity is an empty postfix from negation. In otherwords, negating a schemata also checks its component for validity if it comes out with an empty postfix.
Here again is the whole dirty procedure written in Javascript.
function A(...x){return x.every(Array.isArray);} // is Array
function C(x,y){return D(x).concat(D(y));} // concat arrays
function D(x){return A(x)?F(x,D):x;} // deep dup arrays
function E(x,y){return A(x,y)?L(x)==L(y)&&x.every((e,i)=>E(e,y[i])):x===y;} // identically equal
function F(x,f){return x.map(f);} // map
function IJ(x,y,f){x.forEach(i=>y.forEach(j=>f(i,j)));} // 2d iterate
function L(x){return x.length;}
function M(x,y){return y.some(e=>E(x,e));} // member?
function S(a){return `[${F(a,x=>!A(x)?x:S(x)).join('')}]`}
function wh(x,y,i=L(y)){while(i--)if(E(x,y[i]))return 1+i;return y.push(x);}
let s=[];function sn(x){return wh(x,s);}
function pr(p,...v){return [[],[[[sn(p),F(v,sn)]]]];}
function n(e){return [-e[0],D(e[1])];}
function d(i,j,w){return j.every(e=>!M(n(e),i)&&(M(e,i)||w.push(D(e))));}
function nd(x,y,z=[],w){IJ(x,y,(i,j)=>d(i,j,w=D(i))&&!M(w.sort(),z)&&z.push(w));return z;}
function m(x,f){return F(x[1],i=>F(i,([p,v])=>[p,F(v,f)]));}
function and(x,y){return [C(x[0],y[0]),nd(x[1],m(y,e=>0<e?e:e-L(x[0])))];}
function ot(x,y=[[]]){x.forEach(i=>y=nd(F(i,e=>[n(e)]),y));return y;}
function not(x){return[F(x[0],e=>-e),ot(x[1])];}
function exi(x,y){return [C(x[0],1),m(x,e=>e==y?-L(x[0]):e)];}
How I came to mistake this for the other proof procedure given in the previous note is not clear. An earlier version of this procedure was implemented in that way, but it is so long and cumbersome that I abandoned it early.
[It just occurred to me that if the dogs didn’t run away the neighbor’s cat then the cat would eat the birds that come to my yard for seed. The reason this just occurred to me is that there are two doves that appear frozen in time. They did so after all the other birds flew away and all the chirping stopped. They only just started moving again when three or so other birds flew down within their view. Now a squirrel is making its approach. I know there is a hawk that lives near here, perhaps that is what set this little episode in motion.]
Where to go from here?
…
An example of the application of my logical methods to arithmetic (they are only mine in that they are supposedly well known methods peculiarly applied) begins with two three place predicates: ‘adds’ and ‘multiplies’ as in ‘Five adds two with three’ and ‘six multiplies two with three.’
Then,
or, without finishing that list and leaving intermediate definitions behind:
or, with ‘A’ for ‘adds’ and ‘M’ for ‘multiplies’:
or just
or just
Now what?
…
I found a symmetric way of describing the operations of my machine which is simpler than before in one way and more complicated in another.
There are now two kinds of drops: a left drop and a right drop. They do what you would expect in that the left drop drops the top of the left stack and the right drop drops the top of the right stack.
Now, the novelty is that all other complex operations are shoved into this third one, THE op, which first puts a duplicate of the top of the right stack on the left one and vice versa, and then puts whatever else is ever needed by any other operation ontop of either stack in any order.
So, if m new items are put on the left stack and n on the right stack then m left drops and n right drops leaves you with what you started with. If m+1 and n+1 new items are put similarly, as they must because of the initial provision on duplication, then m left drops and n right drops leaves you with the swapped duplicates atop their respective stacks.
Now, how to get our good old operations of dup, drop, push, pop and swap? Drop is the easiest, you just let ‘drop’ be short for ‘left drop’. Push and pop are also easy enough: a push is the op followed by m drops and n+2 right drops, and a pop is the op followed by m+2 drops and n right drops. Now, dup is the op, m+1 drops, n right drops, and a push. Finally, a swap is a pop, the op, m drops, n right drops, a pop, a drop, three pushes, and a drop.
It is probably more comfortable when thinking about this to factor out the recombic operations from the others. Thus the three primitive recombic operations are then the left drop, the right drop, and the symmetric op which puts a duplicate of the top of the right stack on the left one and vice versa. Then it is easier to see that a classical drop is just the same as a left drop; a push is the (symmetric) op followed by a right drop; a pop is the op followed by drop; a dup is the op, a drop, and a push; and swap is a pop, the op, a pop, a drop, three pushes, and a drop.
…
Now, this simplifies the presentation of the recombic predicate functors as follows:
One of the main reasons that I started reading more fiction was to learn to write about what does not yet exist. The future is unknowable in that it has no effect on the present or the past. What we talk about as the future is a projection, often enough through science and logic, from what has happened on to the things that we theoretically calculate shall happen. What shall happen shall happen whether we calculate it or not, and there is no thing that it is that shall or shall not happen. This is confusing, but it is also vaguely nonsensical.
Let me avoid these cosmic conundrums and begin here and now. The work to be done is minute. The complexity of our machines and of the programs that schedule the contingencies they mediate is a barrior to effective action. Logical simplicty begets practical actions. It makes it easier to see what can be taken advantage of just by speaking validly.
Validity may seem other worldy until it is clearly defined as the contribution of two components: grammar and truth. It is truth that pins the language to the rest of the world, and it is truth that we chase when we speak grammatically in a logical language. Validity is truth from each substitution of a lexical for a perhaps extra lexical in a sentence of a logical language.
But what does this really have to do with practical action? Practical action is shaped by the contingencies we are exposed to during our life time and those to which our species were exposed. Among those that we are exposed to during our life time, there are the social contingencies. The social environment, and the contingencies arranged within it, can undergo profound change during the lifetime of an individual, but can also have a history which stretches across generations. Each consequence arranged by these natural, operant, and social contingencies selects, and maintains, the behaviors that ultimately appear from the individual. Whether that behavior is subsequently reinforced after having been selected asks the question of practicality.
In general, behavior which contributes to the survival of one or the other social environment can be seen as an ultimate value in that those who do not care for such survival are, presumably, already contributing to its extinction. Having not been taught to care for the survival of their culture, the prospects for their cultures survival would seem to have diminished.
We start by simplifying our own lives and push out that simplicity from there.
…
I don’t know if I believe any of this or if any of what I just wrote is true. It was something of an exercise.
Notes on the 11th lecture of “A History of Philosophy” by Arthur Holmes on Aristotle’s Metphysics Part Two
Metaphysics is the science of all sciences i.e. the most general principles. He contrasts his theory of forms (Immanent Teleology) with Plato’s and gives four classifications of causes: efficient, material, formal, and final.
Metaphysics is the science of being (in general i.e. “being qua being”) of which there are ten different categories e.g. Substances, qualities, places, relations, etc. There are laws of being. The categories and laws of being and also the laws and categories of thought (as far as we know).
There are three laws of being/thought:
The contrary to white is nonwhite: black and green are contraries to white because they are both nonwhite.
Three senses of ‘substance’:
Particulars are also called “primary substances” and forms are also called “secondary substances”. Particulars are becomings for Plato but they are beings for Aristotle.
Forms combine in material as particulars. Paticulars are distinct entities. Forms are not, except for god. Forms are drawn out of the potential of matter.
To some, Plato is to Aristotle as idealism is to realism.
Forms give particulars its nature and potential e.g. the material infant grows and the form gives it the potential to be a human adult.
The telos is deduced from the potential of the form.
Essence is distinct from accident: “Things are said to be first in an accidental sense and second by their nature.”
Rationality is essential to the nature of a human being, and having blue eyes is not, it is accidental.
…
Moving onto book 12 on God.
“There are three kinds of substance. One that is sensible of which one subdivision is eternal and another is perishable.”
Four kinds of change:
Two kinds of particulars: perishable and everlasting.
Then a bunch of stuff about the Earth being at the center of the cosmos, the planets going round and round the Earth, stars spinning like tops, and the good way out there keeping them spinning without there being anything else to keep it doing what it is doing (because it is THE good after all). The ether is in there too, between everything that he doesn’t explain otherwise.
Over all it is a grand presentation when Holem’s speaks of it, but it is a stretch for me to embrace it beyond folklore. Thankfully, that is the end of this lecture.
Self-aggrandizing social controllers ask “Why make what you can take?”
A friend whose opinion I respect, and whose opinion of me I am sensitive to, has asked me, on more than one occassion, to seriously examien why I avoid better understanding where he is coming from with respect to science and logic.
I really don’t know where to start because I do not quite grasp the problem to solve. Part of the problem to solve may be that when I write I do not write with the happiness and comfrot of the reader in mind. This is especially the case when I write about things that I do not think the reader is likely to enjoy hearing. Rather than avoid what it is that I have to say by making it feathery soft, I sharpen and harden it so that it sticks out like the leg of a table you stub your toe on.
I show disrespect towards the reader when I do not take their feelings into account. I must say what I can how I can so that it makes the reader feel good while informing them of some uncomfortable truth. I am averse to this because I have no personal experinece with such love from writers or speakers. Sometimes I chalk this up to being different. That is, I do not ever quite feel like I am the audience that speakers are speaking to. Perhaps other people do not feel this way most of the time, but it is something that I feel almost all the time.
Most of my hunts for good books are the result of having eventually found a writer who speaks to me in that peculiar way that mixes science and logic with clarity and exactness. I used to find some sort of satisfaction from reading Russell, but now I do not because even he is less than clear and exact about that which Skinner and Quine have found better ways of speaking.
…
Ok, this doesn’t seem to be getting me anywhere. Perhaps it is just because I am writing to myself and not to myself as someone else.
The consequences of science and logic are inconspicuous and remote to every day experience, but that does not mean that they are absent nor that they are weak.
The amount of time that I spend justifying the strengths of science and logic to people is troubling.
It occurs to me that the word problem for recombic predicate functors is the core of the undecidability of predicate logic.
Is love more than the residue of survival?
Quine’s method of taking free variables as dummy singular terms and then eliminating singular terms as singular descriptions paves the way for a new proof, and I would say logically coherent proof, about Skolem Normal Forms and similarly arguments involving in Shoenfield’s case ‘special constants’ and in Ebbinghaus, Flum, and Thomas’s (EFT) case “witnesses”.
The supplementary axioms of Shoenfield’s special constants are then just the appropriate descriptional premise and substitutional premise. The substitutional premise is substantially what Peano once used for his descriptional analysis of singular terms.
Thses ‘substitutional premises’ are what would otherwise be called a definitional premise in EFT’s case, and either the same or similar in Shoenfield’s case.
In EFT’s case, the definitional extension theorem is a consequence of results on syntactic interpretations. Unfortunately the methods of EFT are model theoretic.
If I did not have the three books– Shoenfield’s, EFT’s, and Avigad’s– then I would not so easily see that they are all circling around, what appears to me now, as passage between singular terms and singular descriptions, and, most importantly, between singular terms and their elimination.
What is interseting, is that rather than speak of the elimination of singular terms, I am interested in their introduction! A pure predicate logic has no singular terms, and has no variables to boot!
The descriptional premises are the universal closures of the existential cropping of the collective relate of identity with a singular description (open or closed).
The definitional premises are the universal closure of the biconditional of the collective relate of identity with the singular description and its descriptional predicate (the relative clause, predicate abstract, which the descriptinoal premise otherwise states is functional).
Aside, Herbrandization is the dual of Skolemization. No, this is not just an aside. After reading https://en.wikipedia.org/wiki/Herbrandization it is clear that Herbrandization is the method by which singular terms are introduced from a pure predicate logic.
It appears that Herbrandization is the method Dreben used when Quine references his method in the chapter on functional normal forms.
The appropriate analogy is that Herbrandization is to Skolemization as validity is to inconsistency. It is erroneously reported that they are as validity to consistency, but these are not duals of each other.
Descriptional premises are of the form “Each item is ..x such that some item is y such that y is identical to the z such that z is F to ..x.” where ‘F’ is in the position where a descriptional predicate goes. A descriptional predicate is nothing more than a total functional predicate and this is the origin of the elimination of descriptions via equivalence of the descriptional premise to “Each item is ..x such that some item is y such that each item is z such that z is F to ..x if and only if z is identical to y.” Which can be broken into premises of existence and uniqueness: 1) Each item is ..x such that some item is y such that y is F to ..x 2) Each item is x, y, ..z such that x is F to ..z and y is F to ..z, only if x is identical to y.
Modern methods stress the existence and uniqueness without mentioning descriptive predicates or total functional predicates.
without
Alternational normal form of truth functional schemata gives a proof procedure for truth functional logic (unhappily called ‘propositional logic’ by most modern mathematical logicians) which includes no axioms.
I am uncertain if this is already known or not. It is certainly not a method of natural deduction, about which cuts and such are fawned over (which only makes sense in a world without Quine’s main method). Natural deduction is characterized by the rules of inference keeping implication. That is, natural deduction gives a proof procedure for implication, whereas Herbrand’s method, in his axiomatic form, gives a proof procedure for validity, and Quine’s main method gives a proof procedure for inconsistency.
Validity remains of primary interest to logicians only and solely as an economical starting point for the rest of their science. With the advent of Quine’s main method we now have good reason to reinterpret Aristotle’s logical methods with respect to the high praise he placed on his principle of noncontradiction or, what I would call, the law of inconsistency i.e. validity is inconsistency of negation. That these are similar principles is not exactly clear to see at first glance, but there is enough of a hint that such a regimintation is possible that I shall not linger. Ultimately, there is not much to be said about vague phrases in the vernacular other than to present an appropriate paraphrase in a logical language.
The proof procedure for truth functional logic which I present is one which does not branch and, under appropriate constraints, only one step of which can be taken at any stage.
The basic operations are:
They are given in order of their priority, or if they are taken as their multistep generalizations (that is, their fixed point forms, e.g. where purification of negation is applied iteratively until its input is identical to its output) in the order of their application.
This method can be used on the postfix (matrix) of the prenex form of a quantificational schemata to prove it truth functionally inconsistent.
Why a proof procedure for inconsistency instead of validity (the validity method is just the dual of the one given here, that is it follows from conjunctional normal form)? It is a matter of space and examples. When the procedure ends with an empty schemata it is concluded that each earlier truth functional schema is inconsistent. That is, each of the operations keeps inconsistency (back and forth if we allow the empty schemata as a degenerate inconsistency). This follows from the rules of equivalence that permit each of one way halves of the equichanges behind each basic operation.
Proofs of inconsistency first balloon and then, as outright inconsistencies are eliminated, they shrink. If space is at a premium, then the basic operations can be tried in prefix or postfix order rather than on the prescribed infix order which I prefer.
Various speedups can be obtained by early identification of appropriate validities and inconsistencies which fall out from the appropriate equivalences and corresponding equichanges.
Here is an inscrutable implementation I cooked up in JavaScript which does not clearly factor out the basic operations as I have above (and which includes the appropriate methods for producing the prenex form).
function A(...x){return x.every(Array.isArray);} // is Array
function C(x,y){return D(x).concat(D(y));} // concat arrays
function D(x){return A(x)?F(x,D):x;} // deep dup arrays
function E(x,y){return A(x,y)?L(x)==L(y)&&x.every((e,i)=>E(e,y[i])):x===y;} // identically equal
function F(x,f){return x.map(f);} // map
function IJ(x,y,f){x.forEach(i=>y.forEach(j=>f(i,j)));} // 2d iterate
function L(x){return x.length;}
function M(x,y){return y.some(e=>E(x,e));} // member?
function S(a){return `[${F(a,x=>!A(x)?x:S(x)).join('')}]`}
function wh(x,y,i=L(y)){while(i--)if(E(x,y[i]))return 1+i;return y.push(x);}
let s=[];function sn(x){return wh(x,s);}
function pr(p,...v){return [[],[[[sn(p),F(v,sn)]]]];}
function n(e){return [-e[0],D(e[1])];}
function d(i,j,w){return j.every(e=>!M(n(e),i)&&(M(e,i)||w.push(D(e))));}
function nd(x,y,z=[],w){IJ(x,y,(i,j)=>d(i,j,w=D(i))&&!M(w.sort(),z)&&z.push(w));return z;}
function m(x,f){return F(x[1],i=>F(i,([p,v])=>[p,F(v,f)]));}
function and(x,y){return [C(x[0],y[0]),nd(x[1],m(y,e=>0<e?e:e-L(x[0])))];}
function ot(x,y=[[]]){x.forEach(i=>y=nd(F(i,e=>[n(e)]),y));return y;}
function not(x){return[F(x[0],e=>-e),ot(x[1])];}
function exi(x,y){return [C(x[0],1),m(x,e=>e==y?-L(x[0]):e)];}
I opt to include as much of personal letters to others as I can publicly share without causing them harm or without their direct permission. Since much of what I say to others is what I would say to anyone, and often do, in a public arena, there is little that must be left unsaid here, yet.
I must say ‘yet’ because there is no reason for what I say to ever go unsaid. This is absurd to most people. It is “playing life on hard mode” for me.
Without further ado, here is a good example of how I write to people individually. It is clear that I could learn a lot from myself. You, after all fare reader, are yet another individual I, perhaps, have yet to meet or know. Perhaps, and this is ego, I am no longer here to respond as I would if you could speak with me at this exact moment. These are the only artefacts of my behavior which remain from this wonderful moment.
“That we live in a world made to be beyond our understanding is clearly a grand weakness and one which shall end us if we do not end it first. Public opinion has been warped by economic practices to make it appear as if paying for things to be done is somehow more noble than knowing how to do them and having the time to do them yourself. It is not as if I am saying we should all do everything for ourselves, as if on some isolated island estate. I am saying that we must have effective countercontrol over our immediate social environments which does not automatically involve an unnecessary third part inserting themselves into every moment of our life.
If the people who built the phones in my pocket or the computers on my desk or the television in front of my couch were as close to me as their products were then it would be obvious that they need to get out of my house and that what they are doing is generally creepy and needs to stop now.
…
I have to go to a different computer in order to update my website now. This is probably the main thing that has stopped me from entirely leaving twitter, that and I am kinda excited to watch all the new ways that it breaks so that I can get a feel for the future that everyone will be living in if they do not make the changes I mentioned above to the world.
My technology is mostly ten years old, and the internet just doesn’t work. What is funny is that you really understand what I am talking about. You are living through the same thing with your music stuff. Americans are so isolated from teh state of the world outside their dumb borders.
What is it with borders and boundary lines that makes people go wacko. The wind blows over them without a care, and earthquakes do not say “oh, no, forgot my passport”. Imagine a virus needing to apply for a green card before entering the U.S.A.
I find that none of what I have to say about teh world is profoundly new outside of what I might say about it with respect to the science of behavior. Otherwise Plato’s Republic is as up to date as any other book you might pick up from a living expert on politics and government.
…
Having to put my cumulative record onto a single usb drive and run it over to another computer to update it publicly has made me lazy. I have now written over ten notes (or so, I tend to exaggerate a little, as you may have noticed :) ) that are just waiting to be sent to the public. This used to only happen between my handwritten notes and my typing them up. Now I hand write a bit less and update about as often as I used to type up important notes.
…
Reunions are exhausting, but so very important in this modern world with littel social Chernobyles everywhere. Family are a line of defense against “the dark arts” unlike any other. Friends are sometimes better, but a good friend is hard to find. There is no substitute for time shared together, and like planting a tree, you would always have done better to do it twenty years ago.
…
I am still learning how to write better from all the fiction I have been reading. I recently got my first Ian Fleming book “casino royale”. I was told by more than one person that Roald Dahl (children’s author of “Willy Wonka”, “James and the giant peach”, etc.) is the template for James Bond. Ian Fleming reported to the head of the british navy and ended up digesting all the messages sent through to him. This gave him access to reports on Dahl’s missions. Dahl’s adult writing, I am told, is also highly erotic and very often involves seducing ladies much like Bond, James Bond, or shoudl I say Dahl, Roald, Dahl?
…
Now I get to toot my own horn. I had a friend call me who works with all this AI stuff that of all the people he has met who talk about how to deal with AI he keeps thinking about teh things that I say. I actually got told “Your values are the only ones I have heard that actually make sense”. I was over the moon.
I am going to try to do more to put together these nonmathematical things in my notes, but so much of what I am working on is mathematical, logical, or programmatic that I just haven’t gotten around to writing on social and ethical things.
But, I have taken some good notes on a beautiful sequence of lectures on the history of philosophy by a Dr. Arthur Holmes that you can find on youtube, adn that I woudl highly recommend. He is a grand speaker and knows a lot about what he is saying because you can tell he has taught the class for deacdes (he mentions a story from his world war II days that gives you a sense of his purpose, which is clearly not to do another world war two if we can avoid it, if).
Taking notes on his lectures has showed me that the message boards of the old dead internet were a recapitulation of the entire history of philosophic discourse. This is not a good thing, it is a bad thing: it means that we just keep saying the same shit over and over again and expect something differnt to happen. I already mentioned this when I said that Plato’s Republic is as up to date as the m ost recent textbook on government and politics. This is exaggeration but all the same…”
“The Long Road” by Steven King is “The Wizard of Oz” if only Dorothy made it to the Emerald City alive.
Two notes ago I concluded that the reference in Shoenfield’s “Mathematical Logic” to Hilbert and Ackermann in his theorem “The Consistency Theorem (Hilbert and Ackermann)” was certainly to the proof in their “Principles of Mathematical Logic”. Since then, I went ahead and started reading their book from the very beginning. Last time I read it I must not have read the preface, but now I have and this theorem gets curious. Not only did Quine have to insert a supplemental paragraph into their book to correct a non sequitor (section 8 pg. 87 used to end having only established that a formula and its Skolem normal form are jointly provable or not provable as the case may be but then go on to rely on having established that they are jointly valid or not valid, Quine noted that the same argument with respect to provability works with respect to validity but that though a proof exists which is almost exactly the same as the one for provability, it is a distinct endeavor), but Church had to fix the exact part of the proof of completeness that I am convinced is the one being referenced by Shoenfield.
Now, it occurred to me that I might have only searched for “Hilbert and Ackermann” and not gone all the way to “Hilbert and Ackerman Consistency Proof”. Having searched for the latter, I am almost certain that I did not do this exact search earlier because the result which pops up is one that I would remember: Edward Nelson’s notes from Fall 2004 Lecture at Princeton for Mat 503.
https://web.math.princeton.edu/~nelson/503/cons.pdf
Upon reading some of the paper, I am now convinced, as much as I was previous in what I see as an error, that the “Hilbert-Ackermann” consistency proof referenced by Shoenfield is from a paper of Ackermann’s prior to the publication of their “Principles of Mathematical Logic”.
If someone wanted to know more about it I would point them to Nelson’s paper.
As a result of reading Nelson’s paper, I turned my attention to the first few chapters of Jeremy Avigad’s “Mathematical Logic and Computation” which gives a delightful presentation on free-variable equational theories. These were used to great effect by R. L. Goodstein in his “Constructive Formalism”, “Recursive Number Theory”, and “Recursive Analysis”. They were also presented in his “Development of Mathematical Logic”.
I was once under Goodstein’s spell because I had not yet met Quine’s methods of logic and seen profound consequences of mistaking use for mention. Almost all matheamtical logicians prefer to speak of the “autonmous use” of their formula (the word ‘formula’ is strange to me, ‘sentence’ does so well, similarly for ‘abstract syntax tree’ versus ‘grammatical analysis’ or just ‘grammar tree’). The autonomous use of a formula is defended by emphasizing that the writer is speaking in their native language when they are talking about mathematical logic and, hence, there can be no confusion that they are mentioning the formula when they use it in their natural language.
There is nothing wrong with this, except that if the distinction between use and mention is not taught, then it goes unnoticed and we have generations of mathematical logicians who take a truth-functional and quantificational schema as if each sentence letter, predicate letter, and free variable was more than a dummy, and on par with its bound variables, as if you could form the generalization of a sentence with respect to one or more of its sentences or predicates.
Alas I am pissing in the wind.
The benefit of Jeremy Avigad’s text is that he does a much better job of emphasizing the part well-foundedness plays in the exposition/explanation of logical methods.
This is something that Quine failed to do in his “Methods of Logic Fourht Edition”. You can see the fragments of his vision from his early books “Mathematical Logic” and “Set Theory and its Logic” (both revised or second editions, the earlier ones contained interesting errors that spawned a lot of great papers, and you can read more about them in the introductions to the later editions of each book respectively).
At the end of “Mathematical Logic” Quine has a grand chapter on “Protosyntax” and in it he slides from the logical language for his theory of sets to the logical language for the theory of the logical langauge of his theory of sets. In the beginning of “Set Theory and its Logic” he introduces the virtual theory of classes which goes on to turn into a full blown logic of predicate abstraction and concretion in his Methods. The origins of the relative clause as the logical import of the predicate abstract can be found in his “Word and Object” where much more time is taken to cover the ground between the reader’s vernacular and the foreign logical languages and the technical terms with which they are talked about in the vernacular.
Thus, the end of Quine’s methods could have included, rather than just a fragment of set theory, the theory of concatenation which formed, substantially, his theory of protosyntax and thereby addressed undecidability and incompleteness which he only mentions in a chapter for omission soon after the chpater on the main method, as he otherwise did at the end of his “Mathematical Logic” (though there, if I recall accurately, he only demonstrated incompleteness).
Also, he might have gone out of his way to include the arguments which form the crux of his roundrobin analysis of validity in equivocal terms of proof, models, valid schema, and, finally, grammar and truth.
I shall have much more to say on well-founded predicates as they are, like schemata of equality, not purely schematic: they have strictly lexical predicate letters in addition to predicate letters. It has just occurred to me that “lexical predicate letters” are probably the proper phrase, and that he methods are lexical methods i.e. they are schematic relative to a given lexicon of predicates. Here I am speaking, as I must, of (pure) predicate (functor) logic and not of truth-functional and quantificational logic (though they are, as is well known to anyone who knows me, easily paraphrased between each other by mechanical methods).
Continuing where I left off in the last note.
I shall only allow myself the luxury of diverting to Quine’s methods. Otherwise, I shall aim to stick firmly to grasping Shoenfield’s proof of
“Herbrand’s Theorem. Let T be a theory with no nonlogical axioms, and let A be a closed formula in prenex form in T. Then A is a theorem of T iff there is a quasi-tautology which is a disjunction of instances of the matrix of A_H” pg. 54
First, to trace down what Shoenfield means by ‘theory’:
“A first-order theory, or simply a theory, is a formal system T such that
- the language of T is a first-order language;
- the axioms of T are the logical axioms of L(T) and certain further axioms called nonlogical axioms;
- the rules of T are the expansion rule, the contraction rule, the associative rule, the cut rule,a dn the E-introduction rule.” pg. 22
Now to get after ‘formal system’:
“the first part of a formal system is its language … this should be chosen so that, as far as possible, the structure of the sentences reflects their meaning” pg. 3
“The next part of a formal system consists of its axioms.” pg. 4
“We need a third part of a formal system which will enable us to conclude theorems from the axioms. This is provided by the rules of inference, which we often call simply rules.” pg. 4
Where ‘language’ is
“We consider a language to be completely specified when its symbols and formulas are specified … We shall designate the language of a formal system F by L(F).” pg. 4
and ‘symbol’ is
“In English, the symbols would be the letters, the digits, and the punctuation marks. Most of our artificial languages will have infinitely many symbols.” pg. 3
and ‘formula’ is
“certain expressions of the language are designated as formulas of teh language; it is inteded that these be the expressions which assert some fact.” pg. 4
with ‘expression’ as
“Any finite sequence of symbols of a language is called an expression of that language … a symbol may appear several times in an expression; each such appearance is called an occurrence of that symbol in the expression. The number of occurreences of symbols in an expression is called the length of that expression … allow the empty sequence of symbols as an expression; it is the only expression of length 0. … one expression to appear within another expression … called an occurrence of the first expression in the second expression” pg. 3
Axioms are
“Our only requirement on these is that each axiom shall be a formula of the language of the formal system.” pg. 4
rules are
“Each rule of inference states that under certain conditions, one formula, called the conclusion of the rule, can be inferred from certain other formulas, called the hypotheses of the rule.” pg. 4
and theorems
“satisfy the two laws: i) the axioms of a formal system are theorems; ii) if all of the hypotheses of a rule are theorems, then the conclusion of the rule is a theorem. … Define a theorem to be a formula which is a theorem on teh basis of laws (i) and (ii).
.. explicit description of the theorems. Let S0 be the set of axioms; .. Let S1 be teh set of formulas which are conclusiosn of rules whose hypothese are all in S0; .. Let S2 be the set of formulas which are conclusions of rules whose hypotheses are all in S0 and S1; .. let Sw be teh set of formulas which are conclusions of rules whose hypotheses are all in at least one of S0, S1, …; … continue in this way until no new theorems can be obtained by (ii);” pg. 4
Thus a theory is a special sort of formal system to Shoenfield, but there is more to formal systems than has been said thus far. It is unfortunate that the method of vagueness which is so important to Quine in both his theory and practice is not also invoked by Shoenfield, alas, there is only one life to live and not everyone accumulates wisdom.
As an aside, and I know this breaks my rule of only going aside to Quine, R. L. Goodstein’s “Constructive Formalism” can be synthesized with Solomon Feferman’s “Finitary Inductively Presented Logics” to give the logic of a theory of ordered pairs (contemplated as binary trees) which, more or less, captures, in one instance or another, what many matheamtical logicians have mentioned as formal systems. There is much more to say about this all at a much later date (though I hope not too late) with respect to logic programming.
Symbols are to expressions as expressions are to arguments, but, for some unknown reason, most do not call a sequence of expressions an argument and then pick out, among arguments, the formal ones as those each expression of which is a formula of a formal theory, and then pick out among the formal arguments the logical ones as those each formula of which is either an axioms or a theorem of its priors in the theory.
I submit that there is even more to be got from paraphrasing a logical argument as the conjunction of its formula (negate the last and apply the main method to prove that the rest imply it and you can repeat this dropping each last one until they are all gone to learn that the earliest ones imply the rest).
The word which is preferred to ‘argument’ in mathematical logic is ‘proof’, but before getting ot proof
“A rule in a formal system is finite if it has only finitely many hypotheses.” pg. 5
then
“Let F be a formal system in which all the rules are finite. By a proof in F, we mean a finite sequence of formulas, each of which either is an axiom or is the conclusion fo a rule whose hypothese precede that formula in the proof. If A is the last formula in a proof P, we say that P is a proof of A.” pg. 5
from which yet another characterization of thereomhood is given
“a formula A of F is a theorem if and only if there is a proof of A” pg. 5
Having exhausted the definition of ‘formal system’ it is time to narrow in to
“A first-order language has as symbols the following: a) the variables, b) for each n, the n-ary function symbols and the n-ary predicate symbols, c) the symbols ‘not’, ‘or’, and ‘some’.
For each n, the number of n-ary function symbols may be zero or nonzero, finite or infinite. The same holds for predicate symbols, except that among teh binary predicate symbols must be the equality symbol ‘=’.
A 0-ary function symbol is called a constant. A function symbol or a predicate symbol other than ‘=’ is called a nonlogical symbol; other symbols are called logical symbols.
It will sometimes be convenient to have a fixed ordering of the variables to refer to. We call the order in which they are listed above alphabetical order.” pg. 14
where “the order listed above” is
“x, y, z, w, x’, y’, z’ w’, x’’, …” pg. 14.
It is unfortunate, though not unexpected, that the phrase “first-order language” is used by Shoenfield and by most mathematical logicians. There is, e.g. through Carnap, a history of contemplating “higher-order logics” like those that Frege dealt with when he appeared to be quantifying with respect to the predicates (what were then called propositional functions) in his schema. Quine, more than anyone else, has made clear the methodological error in such higher-order logics, often emphasizing that they are set-theories in disguise.
They need not be just set-theories in disguise. They may be any other theory in disguise which does the work presumably done by mistaking an n place predicate true of each list of n items for the name of some item which when predicated of some unspecified predicate (in the set theoretic case this is “membership”) with respect to n other items is true in those cases which, presumably, do not generate a paradox (there is nothing stopping such a theory from embracing the paradox as an inconsistency from which all else is implied, but what of such a theory?)
I am fine with ‘first-order language’ applying to languages with a purportedly infinite number of predicate or function symbols because such languages have already conceeded the fight for the infinite which launched so much that is grand in logic e.g. all the way back to Aristotle’s continuum.
I for one have not seen it all settled and, as Shoenfield has already conceeded when restricting his formal systems to finite rules, I shall continue to separate the languages of predicate logic from the first-order languages. Sometimes by predicate logic I mean the logic of truth-functions and quantification, but most of the time I mean predicate functor logic. More importantly, the logical languages are those whose lexicon is finite and whose grammatical constructions are finite place predicates.
The grammatical generation of an endless category of variables is achieved with a single lexical variable, e.g. ‘x’, and the grammtical construction of accentuation which is a two place predicate subject to the obvious principles of distinction.
Speaking of distinction, it can not be introduced into a first order language without becoming a seperate and complexifying concern. This is sadly the path that Shoenfield has picked. The predicate of identity is mentioned exceptionally among the predicates of the first order languages and, later, when axioms are given for the theories in which first order langauges play a part, there are separate ones in which the predicate of identity occurs (the ones Shoenfield uses correspond to those intorduced by Godel in his completeness proofs).
This is very unfortunate because in a predicate logic indiscernibility is grammatical compound built up from lexicals and coextensive with any lexical which purports to independently be identity. For those wedded to second order languages, the definition of indiscernibility in a theory is the conjunction with respect to the lexicon of Liebniz’s definition (taking the universal quantification into the “predicate variable” as ranging only over the finite lexicon and thus being equivalent to the appropriate conjucntion). Then it is shown that each compound predicate satisifies the appropriate instance of Liebniz’s laws and, this being all of the grammatical compounds, you get that which is equivalent to Liebniz’s definition with respect to all the compound predicates generated from the lexicon.
This method can fail tragically when there are an endless list of lexical predicates for it can be supposed there are an endless list of one place predicates each of which is false of some item which the priors are true of. This is not far fetched because by way of a construction liek that of Cantor’s, and what, I guess, we would now call a diagonlization argument (though that is not quite accurate of Quine’s construction), Quine demonstrates that in a theory of which its predicates are items, there is a predicate which irreducibly supplements the theory (this is given towards the end of “Word and Object”).
I have now gone far afield against my advice at the beginning of this note. So, I will stop here to stop myself from going further afield.
(Life is more than the residue of work.)
The note before the last one began by mentioning a theorem and proof from Schoenfield’s “Mathematical Logic” that I was to report on reading. The theorem is as follows:
“Herbrand’s Theorem. Let T be a theory with no nonlogical axioms, and let A be a closed formula in prenex form in T. Then A is a theorem of T iff there is a quasi-tautology which is a disjunction of instances of the matrix of A_H” pg. 54
I did not present it in that note becuase I digressed into Quine’s chapter on Herbrand’s method of proof in “Methods of Logic Fourth Edition”. There I failed to report that I was reading out Quine’s chapter to prepare myself for translating Schoenfield’s methods into Quine’s. Quine’s methods are supperior in that they make distinctions through definitions which contribute to an understanding of the whole of the methods of logic.
The same can not be said of Schoenfield’s work, even from a generous reading of it. This is also true of Kleene’s “Introduction to Metamathematics” and the projection of that work into his smaller book “Mathematical Logic”. While “Mathematical Logic” is easier to read than “Introduction to Metamathematics” it is characteristic of text on mathematical logic that they claim a kind of bankrupcy with respect to cogency. As if the methods of logic are not precisely the result of arriving at an excruciating kind of clarity and exactness.
Without having been exposed to Quine’s “Methods of Logic Fourth Edition” I do not know if I would have made as much sense of the world as I have (or at least purport to, as it is my aim to demonstrate what I can here in this massive document).
Another digression occurs to me, and again I am indebted to Mark Twain’s work on his Autobiography for knowing to indulge it.
Schoenfield’s depiction of Herbrand’s theorem and its proof are given in the fourth chapter “The Characterization Problem”. It is given in the fourth section of that chapter. The prior section is titled “The Consistency Theorem” and it refers to
“Consistency Theorem (Hilbert-Ackermann). An open theory T is inconsistent iff there is a quasi-tautology which is a disjunction of negations of instances of nonlogical axioms of T.” pg. 49
I happen to have Hilbert and Ackerman’s book “Principles of Mathematical Logic” and do not yet know if it is from this book that Schoenfield is referencing “Hilbert-Ackerman”. Let me look up if there is some other obvious paper from them titled something like “Consistency of Predicate Logic” or perhaps some other book on mathematical logic that they collaborated on that may be the source of Schoenfield’s reference. I wish Schoenfield had written up a bibliography: there is only an index at the end of the book preceeded by an appendix on the word problem.
A quick search shows that this text is likely the one that Schoenfield is referencing since it is the text which Ackermann helped to write from Hilbert’s 1917-1922 lectures on mathematical logic and which were the first presentation of an axiomatic system for predicate logic in the formal sense which is still easily recognizable with respect to today’s standards. Frege’s Begriffsschrift is the earliest, but it is alost entirely unrecognizable to the eyes of a modern mathematical logician.
Reading Hilbert and Ackerman, I can not tell if Schoenfield is referencing the proof of what is there called one of strict completeness, and which, to correspond with Schoenfield’s interpretation, is merely a fragment of the proof of such strict completeness. It occurs from pages 95 to 99. If so, then Hilbert and Ackerman reference Godel as their governing expositor.
I am presently convinced that this is the theorem and proof from Hilbert and Ackermann that Schoenfield calls the consistency theorem. My conviction comes from seeing that Schoenfield’s definition of “a special sequence” and its subsequent use corresponds with the proof of one alternative in the proof of the strict completeness theorem from Hilbert and Ackermann on page 96. There, the result is given for a formlua that is in what Hilbert and Ackermann call “Skolem Normal Form” which is a special kind of prenex form where all existentials are out in front of all universals. I shall need their proof, or parts of it, somewhere else, but for now, nad not to get drawn into Hilbert and Ackermann’s wonderful text again, I shall return to Schoenfield and my translation of his work into Quine’s methods.
My first response upon looking back on page 54 of Schoenfield is that I do not appreciate his mathematization of logical methods e.g. the insertion of identity and equality axioms in his logical axioms and the frequent use of names for each item in the class he callst he universe of discourse. Most importantly, I do not appreciate why the modern world is wedded to schematic function letters rather than just sticking with functional predicates. Even though one of the problems at the end of chapter four of Schoenfield prompts the reader to prove that a formula with function letters is equivalent to one without function letters under supplementation of the predicate letters in the theory, it does not make it clear how this radically simplifies the arguments of the chapter, leaving most of the results as little more than applications of logic methods to those with singular descriptions.
This is something which may be a holdover from Frege and Russell transmitted through Church and Kleene where naming has a sacred value for maintaining reference to meanings.
Quine’s predicate functors provide a completely new outlook on Aristotle’s categories and Liebniz’s Scientia Generalis, Ars Combinatoria, Calculus Rationcinator, and Characteristica Universalis.
Aristotle’s categories become the grammatical categories. It is easier to carry over Aristotle’s outlook in a predicate functor logic because there is only one type of thing in its lexicon: predicates. When appropriately constructed, the predicate functor logic of an immanent theory of the world clearly and exactly determins the categorical distinctions made in Aristotle’s sense i.e. vaguely, what can be spoken of meaningfully.
Liebniz’s Ars Combinatoria is both a recognition that settling the shape of the predicates in the lexicon of a predicate functor language with respect to the world and then including the appropriate recombic predicate functors in the logical language of which the lexical predicates participate is not yet an obvious affair. It is governed by the logical language and the rest of the world, and constrained by these complementary parts which together exhaust the whole (though it is not clear if they are strictly complementary or if there is some sort of penumbra cast by Quine’s indeterminacy of reference). Thus Liebniz’s Ars Combinatoria splits the world into the complementary methods of the Scientia Generalis and the Characteristica Universalis. The Calculus Rationcinator bridges the two as the contribution of the Characteristica Universalis to Scientia Generalis in effective action.
Now for a much needed note that has been a long time coming. This is an example of how a human reads a theorem and its proof. Specifically, a theorem and a proof in a book and not in a paper. Papers are free floating in ways that most books are not. They rely on an assembly of sources that often times report on work that was substantially done other than through the words presented in the final paper. Papers tend to speak only of the tip of the iceberg whereas books tend to give the whole iceberg, or substantially more than what is floating above the surface.
These are my notes on the proof of Herbrand’s Theorem from Schoenfield’s “Mathematical Logic”. They are taken to not only force myself to learn his terminology (which is also largely the terminology that is most familiar to mathematical logicians), but also to prepare myself for the full blown proof of the completeness of my proof procedure for Quine’s predicate logic.
First a word on Quine’s “Methods of Logic Fourth Edition”. Chapter 36 is titled “Herbrand’s Method”. It is first and foremost a demonstration of how one method of proof, Quine’s main method, gives way to others, Herbrand’s method in this chapter, Dreben’s method in an early chapter on functional normal forms, and the more familiar deductive methods in a later chapter.
Quine proves that Herbrand’s method is sound in two ways. The first is directly by duality with respect to the main method. This duality only appears to work, as presented by Quine, for the soundness of Herbrand’s method. That is because the soundness of the main method (“that it generates a truth functional inconsistency only when the premises are inconsistent” pg. 192) is assembled from methods of logic which are blessed with decision procedures.
Specifically, Quine invokes the method of truth value analysis to decide truth functional validity and inconsistency (which is called by computer scientists “binary search tree on the lexicographic order of binary strings corresponding to a partial interpretation of sentence letters as truth values whose values at leaf nodes are the truth value of the truth functional schema analyzed and whose values at the non leaf nodes are the resolution under the truth value interpretation corresponding to that node’s key), and what he calls the method of existential conditionals to decide monadic validity.
Only validities resulting from these decision procedures are used in the proof of the soundness of the main method. It is a very short proof, but it is given, as is Schoenfield’s of Herbrand’s method or Kleene’s proof of Hilbert-Bernay’s completeness in “Introduction to Metamathematics”, with respect to a concrete example rather than in full generality because the notational technology available remains unfit for presentation to humans.
Most importantly, the briefness of the proof of the main method is not betrayed by its naturalness: the main method of proof is the same as that in Plato’s dialogues with Socrates.
The completeness of the main method “says of any combination of (one or more) quantificational schemata that either they come out jointly true under some interpretation in a nonempty universe or some truth-functionally inconsistent combination of unquantified schemata can be derived from them by the familiar operations: convert to prenex form, reletter bound variables that match free ones, and then apply UI and EI, using only new instantial variables in EI.” pg. 205
This is the first way in which it is stated, there are two alternatives which turn up as his proof progresses:
The completeness proof of the main method is a two part procedure. The first part establishes a “rigid routine” for exectuing the main method i.e. it gives a strict program for generating instances of the premises, appropriately relettered and in prenex form, such that “each existential line gets instantiated once, and that each universal line gets instantiated with each free variable that ever turns up” pg. 205. Sometimes this program halts and sometimes it does not. In either case completeness is established by Quine’s “Infinite Conjunction Theorem” which is just that if each conjunction of truth functional schema from a set are consistent then there is an interpretation that makes each in the set true i.e. we call such a set consistent. (Remember, a conjunction of schema is a construction of which only finitely many of the schema in a set are components of).
Quine makes no note of what you get when the rigid procedure for the main method halts. He merely notes that when it halts it may with or without having produced a truth functional inconsistency. Those abreast to the modern jargon of formal methods would recognize that when the rigid routine halts without an inconsistency you have generated a finite model of the premises and the negated conclusion. I hesitate to mention how much money goes into ever more obscure hunts for such finite models in the formal verification of computer programs.
He also makes no extension of the main method to the method of logic programming, even though it is astoundingly simple to see that the rigid routine for the main method when it produces one inconsistency may be allowed to go on and produce further inconsistencies each of which can be interpreted as the result of a database query. Again, I hesitate to mention money.
It is not clear that Quine’s proof of completeness for his main method does not give a dual argument for Herbrand’s method. I must distinguish between Herbrand’s method as it appears frmo the main method, and Herbrand’s method as it appeared from Herbrand. Herbrand’s version of his main method is introduced by Quine as the first example of an axiomatic method for quantificational theories: the axioms are all truth functional validities and the rules of inference are clausal universal generalization, clausal existential generalization, and deletion of duplicates.
The missing ingredient to dualizing Quine’s proof of completeness for his main method to its dualization as Herbrand’s method is that Quine gives nothing like the rigid routine of the main method. Such a routine would lead back, presumably, to the dual of the law of infinite conjunction. Presumably, the law of infinite alternation would trade “consistency” for its dual “invalidity” and say
An infinite class of truth-functional schemata is invalid if each of its finite subclasses is
where invalidity of a class of truth-functional schemata is to say there is an interpretation which makes each schema in the class false.
The proof would go through as follows, dualizing each of Quine’s steps:
Slowing down the proof of lemma 0:
Hark, there is no way for one and the same K-alternation to be in conflict with both interpretations of ‘p0’ without immediately contradicting the assumption that each K-alternation is invalid i.e. such a K-alternation would be an outright validity. The dual of this remark is missing in Quine’s text.
(I wanted to use ‘hark’ in a sentence, though I also contemplated ‘Lo’.)
Next is a definition and lemma 1:
and so on. Taking any schema from K it has an alphabetically largest letter as subcomponent so that the corresponding lemma shows that the corresponding assignment of innocuous truth values is contradicted by the supposition that the assignment resolves the schema to truth, so, since it resolves to either truth or falsity under an assignment of truth values for each of its letters, it must resolve to falsity. Therefore K is invalid.
Had I known how much fun I was going to have writing here on just Quine’s work I would not have mentioned Schoenfield yet, but alas, here we are.
It seems clear enough that a mechanical procedure can be given that is dual to that of Quine’s for Herbrand’s. It also seems clear that this mechanical procedure may be the only way that I firmly grasp how to carry the finitary proof of completeness of Herbrand’s method from Schoenfield back to Quine’s main method.
Got Ian Flemming’s “Casino Royale” today. It is a 2023 edition which greets me unhappily with the following.
“This book was written at a time when terms and attitudes which might be considered offensive by modern readers were common place. A number of updates have been made in this edition, while keeping as close as possible to the original text and the period in which it is set.”
Such edits are worse than whatever was edited away.
It is a very fast read. I sat down to read a bit and met the fifth chapter with surprise. It feels like an ellucidation more than a story.
Notes on the 10th lecture of “A History of Philosophy” by Arthur Holmes on Aristotle’s Metphysics Part One
https://youtu.be/kLKHpSPoNKI?si=q8LUUaHI-eFQjWFN
Aristotle’s causes are far more diverse than our mechanistic force-based definition.
Aristotle’s causes:
For example, wood or stone carver:
Holmes is speaking very quickly here, and I think it is because in the previous lecture he had planned on covering the causes but spoke longer on Plato and ended off schedule. If I did not have the power to pause and rewind, then I would not get all that he was saying. His verbal behavior is shaped in a way which tends to be the mark of rhetorical excellence, and I would guess that he has spoken large parts of these lectures to many audiences. He clusters the key terms right next to each other as emblems of the philosopher, or at least of modern talk about them.
It occurs to me that the causes may be related as follows:
this is only a loose interpretation.
Ultimately, the way of translating is through contingency analysis.
The example of sexual reproduction:
“The woman is just the material cause. We have come a long way, I would hope.” Dr. Holmes
Aristotle says early Presocrates only spoke of material causes: What is the basic stuff? Thayles said water, anaximenes air, heraclitus fire, empedocles earth air fire and water. Empedocles also talks about love and hate, moving beyond what Aristotle would call material clause.
Heraclitus logos, Anaxagoras nous, etc. show that there is more. Aristotle makes it clear that each cause is part of each process as a full blown generalization.
The final cause is what Aristotle sees as his distinct contribution.
The final causes are what I would otherwise call the selecting consequences.
Chapter 3 of metaphysics: “causes are spoken of in four senses:
(or thereabouts)
Aristotle asks how particulars participate in forms (questioning Plato’s doctrine).
Switching from the metphysics to the physics: the same distinction between causes. “We must now consider why nature is to be ranked among causes that are final or purposeful.”
There is a natural teleology, purposes, end achieving nature, we must speak of necessity.
“What is it that inclines physical things towards ends that are good, ordered, beautiful?”
Innate endowment, natural potency, natural potential, which is actualized in all processes of chanage.
“Nature never makes anything without a purpose.”
“God and nature create nothing that has not its use.”
Natural things are “end oriented”.
Art is a matter of discovery. The musician does not create music, he discovers the potential in the physics of sound. The sculpter does not invent designs or shapes, he uncovers the natural potency of the texture of the stone and actualizes them.
Nothing occurs in nature uncaused i.e. randomly. But, due to complexity of cuases, there are accidentaly or incidental causes from which we get chance. For example, it is by chance that accidents occur. A chance event is not an uncause event, it is a complex of extraneous causes and are not the good at which the process originally aimed.
Nature is always in process, in change, and time is the essence of nature. Time is the measure of motion. Time is how we measure the processes of change. Aristotle answers Zeno’s paradoxes using time as a continuum. Zeno left time out of the motion and got contradictions.
Aristotle’s view of nature is called “Immanent Teleology”.
Occum, Luther, and Calvin worked to overturn Aristotle’s control over the middle ages so that God could enter as an intervening operator on what would otherwise be natural processes.
Intrinsic necessary causes vs. extraneous contingent causes.
Aristotle’s “cause” is “first principle” or “explanatory factors” as in “what are the first principles” is “what are the first causes”.
Aristotle was a “biological vitalist” in that life is essentially something different from material processes.
Science to Aristotle: if you can formulate these principles, then you can deduce conclusions about particular cases. Formulate premises and deduce conclusions. Dominates until the 14th century.
Moving onto Being and its Categories.
Metaphysics is the science of being. What do we me when we say that something is? Different ways in which we speak and think of being and in which they actual are are Categories.
Encyclopedic scientific interests:
They are both categories of thought and categories of being.
There are laws of thought that correspond to laws of being. The basic law of thought is the law of noncontradiction: a thing can not both be and not be at the same time and in the same respect. Aristotle’s “most certain principle”.
Can a principle like this be demonstrated? No, because you have to assume the law of logic in order to prove the law of logic. A negative proof: “The starting point of such an argument is not that our opponent say that something either is or is not but that he say something that is significant. If he means nothing then he is not capable of reasoning. If one grants this then and so on…” Try to say something significant and you are stuck with the law of noncontradiction.
For better or worse, I am looking back over past notes, and reading #2025-1103-1456 which gives a monolithic version of the interpreter for my machine. There I planed on introducing conditional execution into the interpreter. This was achieved, I think only provisionally in that note, by introducing compound gray words each letter of which is a microinstruction.
Thankfully I didn’t get lost in thought on this and simply got up and walked away.
Notes on the 9th lecture of “A History of Philosophy” by Arthur Holmes on Summing Up Plato and Aristotle’s Metaphysics.
In conclusion, Plato’s metaphor of the divided line governs the rest of his philosophy. On the epistomological side there is knowledge above and opinion below, and on the objective side there are forms above and particulars below. From this follows his theories of ethics, soul, education, god, ethics, politics, etc.
Plato’s view of history as in The Republic is cyclical. It is contrasted with the modern notion of linear or cumulative change.
The cycles extend to the soul i.e. the many bodies occupied by the same soul. Remember, you are your soul to Plato.
The first linear view of history appears with St. Augustine in “The City of God”. God was moving historty to its destiny. May explain the origins of historical progress e.g. as in the scientific revolution.
Why cyclical? The divided line. There is no stability, no unchaning achievement, without full participation of forms. To participate in time is to be caught in change and decay. Hence the fixation on the disembodied afterlife was part of bringing the soul nearer to the forms, and the form of the good especially.
The christian resurrection was something of an odditity: who wants a body back? The story of creation as a purposeful good created by a god was part of changing public opinion on the relative importance of the resurrected body.
Cycle in goverment means again and again, recycling:
“Until philosophers are kings or kings are philosophers we shall never have peace or justice on this Earth”, he means the lover of wisdom, the good, with dialectical understand sees what is good and does it.
For Plato, god is the form of the good, i.e. the form of beauty. Aesthetically, beauty is the highest good, where beauty is unified ordered rational harmony. Truth to knowledge, beauty to aesthetic, and goodness to ethics are the same to each other as the good. The soul is related to the good by the virtue of love. “Love is a moving force. Love for the good makes you want to be like it.”
Our vernacular is saturated with the philosophy of Plato and Aristotle through the etymology of our words. The weight of this history makes it hard to change the course of vernacular.
“I hope you go back to Plato in your thinking and reading. After all, you’re not educated unless you’ve read The Repulbic.” That last part is put in a cheeky tone not a condesending one.
…
Now, to Aristotle, the student of Plato.
What is the difference between the appearence of good and the essence of good. The differene between appearence and reality follows in Plato’s divided line argument.
The good is our goal or purpose. The good is god. There are unchanging essences which in some way represent the ideal of the things.
His primary criticism of Plato is that the forms are nonsense because particulars are said to participate in them and yet there is no evidence in any particular of this formal participation. Furthermore, there is evidence that mathematical things, which are in Plato’s philosophy forms, are part of the essence/reality of the things we come to know.
Aristotle does not write dialogues, he writes papers, or, at least, papers are what have survived from his work through his students. In general, Aristotle is far more analytic and systematic in his writings than Plato is.
“Essence” or “esse” is “what it is” or “to be”. He asks “What is the what-is-it of this?” “What is a substance? A being. What is an essence? What the being is.”
Aristotle’s Metaphysics disagrees with Plato’s divided line. He begins by defining the scope of metaphysics:
Aristotle is making the distinction that I make as the difference between theoretical and experimental. Experimental practices include the methods of engineering which manufacture the technology.
Metaphysics is the science of being, the first causes of all causes, processes of becoming for every kind of being (substance).
“being qua being” means “being in the sense simply in being” rather than “being in the sense of being an animal, a chair, or a warm day.”
In 1800 we got “ontology” as “the study of being qua being”, and the division of philosophy of “cosmology” apart from physics or astronomy, of “psychology” apart from specific sciences, and “theology” apart from a specific religion etc.
The best way of making the distinction Aristotle made is between the methods of logic which are not about anything in particular and its applications in theories.
Aristotle on Forms and Causes:
In Book II chapter II of the physics: Forms are transcendent, materials are immanent. Mathematical objects, forms to Plato, can not be seperated from material objects (physical objects) then forms can not be seperate from matter. Plato should have known physical things have mathematical order to them.
Plato’s form of the good is an ideal, only to be contemplated. It has no power itself over the particulars of the material world. Particulars have no form in then only matter. So, what inclines the particular towards orderedness, beauty, etc.? It’s good. What gives material things the impetus towards the good if forms are transcedent and not immanent in matter? How do you explain change in the physical universe, naturally producing order, beauty, goodness, when there is no form in the material things and the transcdent forms have no power. To which Plato would say “that’s what the world soul is for”. Aristotle is open to “the good as god” but “a world power” sounds too much like story. If material objects have mathematical properties then there must be the form of mathematical objects in material objects, similarly for plants, animals, and natural processes spontaneously produce order and beauty there must be potency of form within the material things and processes. Why talk about transcendent forms at all? Why not all immanent?
Particulars are composite of form and matter. Hylomorphism: material forms.
If forms inhere in particulars, then particulars have inherent capacities or inherent potential or inherent end/telos in its very essence (remember, essence is what a thing is). The form makes a thing the kind of thing that it is, it defines its essence, defines what it is, then it has the potential to be that kind of thing. The potential to be actualized in the acorn is the oak tree. Every kind of thing has a potential that can be actualized. Sometimes “potency” is used to capture the sense of power e.g. “A acorn has the power to become an oak tree.”
The essence of a thing is such that it has capacity, potency, potential, to be realized, the end/talos is the good of that thing. “What’s the good of those acorns if they don’t grow into oaks?”
The ethical question is “What is the human telos, capacity, essence by virtue of which we have a corrosponding potential, the actualization of which is like a flourishing.” Qualitative achievement out of inherent potential.
Note to tomorrow self: collect sources for my “one true terminal”.
Start here:
Try to actually explain the program if you can, but it seems likely that you will end up returning to the machine and its language so that you can more harmoniously integrate the ice (interpreter, compiler, executer) into the play (grab, move, show).
One new design to implement is to make resolution a function of the scroll wheel in complex scenes.
I have not yet implemented the sdf work on rendering text: there are opportunities to make it much more responsive.
If you happen to get stuck in the weeds of, e.g., a javascript implementation (again), then try to go back to the entry here on the logic of rendering.
Start here on the logic on rendering: #2025-0923-1940
First some notes on a closer reading of Schoenfield’s “Mathematical Logic”.
Section 2.5 is on structures. Structures are classes generated from the items of the universe of discourse. The universe of discourse is the class of all things we are talking about with the formula of the formal “Let A be a structure for L. For each individual a of A, we choose a new constant, called the name of a.” pg.18
The structures of which Schoenfield speaks are the same for most mathematicians: they are sets of mathematical objects, objects which are often themselves sets, which we come to talk about with language. The language we use to talk about the mathematical items is a logical one in that it
…
I am going to try something different here. Usually when I write about something I do not stop to say it in any better way than how it came out as it occurred to me. This is not the best way of writing publicly because members of the public are not sensitive to the nonverbal response that come along with my verbal responses. It is hard to tell from my limited contact with my behaving body through my peripheral nervous system whether a given nonverbal response is collateral to a verbal response or not. Some may be prior to the verbal response, others may be subsequent to it. Responses such as are described by the phrases “seeing of, to myself” or “feeling of, to myself” are descriptions of what I can only claim are collateral responses to the verbal responses that have left these marks upon the digital page.
…
I am using elipses to indicate that my verbal behavior goes weak between the prior and present paragraphs. These are only records of verbal responses and they do not accurately report the contingencies in which these forms of response play a part. It would be strange for a modern human to read a cumulative record of the consequence, response, and occasions which compose the verbal contingencies that occur as I write this here and now.
The only reasy it would be strange is because most modern humans are obssesed with the human brain as the originator, intiator, or creator of the behavior that ultimately appears from the body with the brain. The mind is what the brain does and we can not inspect our brains as we can the overt consequences of our verbal behavior. Taking verbal behavior as a mere consequence of some verbal process in the brain is the error of cognitive psychologists and scientists. It is promoted because the computer is the metaphor of our time. Everything is a computer. Everything is a computation. Everything is a nail to a hammer.
…
Writing while sleepy is hard in that my entire verbal repertoire is weak. How do I know this? Because I have conveniently marked each pause in verbal behavior with the elipses. If there was a cumulative recorder that I could present here… well I can’t present one here yet because that is something I am not prepared to do. But, if I had one that I could present here, you would see it flatline before the response which made the elipses.
…
People who write fiction are much better at dealing with these bursts of nonsense without being entirely turned off by the disorganization of it all. They tend to get that not all verbal responses are under the strict control of a repertoire of accurate reporting. We are not calculating our way through the contingencies of life. Sometimes we respond without self-control, or, specifically, without autoclitic behavior. There is no way to read of the kind of behavior from a record of the responses alone. That is all that these words, sentences, and paragraphs are: records of response forms.
The forms of response which are occurring now are fingers tapping keys. These are called discrete responses as opposed to continuous ones because the consequence which selects and maintains the form of response appears only and solely with respect to the compact support of some material body. It is bounded by a neighborhood of finite radius. Crossing out from one neighborhood and into an other includes an intermediate ride through the nonempty exterior of both. Up to approximation, these are Euclidean neighborhoods and there are finitely many of them and they are disjoint, that is the intersection of any two is empty.
What a strange way to say what was already said with the technical phrase “discrete response” familiar to any novice scientists of behavior. Yet, these technicalities are what people cherish when they talk about cognitive processes. Instead of saying a discrete response, they demand something like “a finite family of disjoint open neighborhoods in a topological vectors space that is locally Euclidean with three space like dimensions and one time like dimension”. What a waste of breath. Or, of finger fidgiting.
…
Yeah, this is going to be a messy note.
I had planned on copying down some hand written notes on Schoenfield’s “Mathematical Logic”. Earlier today I finally sat down and did more than respond to myself while reading his book. I actually wrote out some of the covert responses that occurred to me. They were almost all about the absurdity of Schoenfield’s methods, and hence about those of most mathematical logicians. An example of an absurdity is Schoenfield allowing for a unique name to each item of the universe of the structure of a theory.
I find it absurd to speak of the meaning of a theory. Or, directly, to speak of what the grammatical items of the theory mean. It was only upon reading Quine that I found someone else who not only noticed there was a problem, but had the sentences that seemed to cut through the mysticism of it all. Perhaps Bertrand Russell actually knew the propositions he kept talking about in some way that is beyond me. I accept that as much as I accept that there are those who speak of their contact with their god(s). I accept that as much as I accept that there are those who sepak of their contact with their minds, or the mental world, or some such Cartesian clarvoyance.
I can feel a lull coming on.
…
Oh, here is some stuff I had already written but which somehow got shoved down here rather than left above.
For Schoenfield, the mathematician is confronted by a universe of abstract things about which he speaks when talking with other mathematicians. Such things as points, lines, figures, sets, relations, funtions, groups, rings, fields, graphs, trees, automata, procedures, and so on. Somehow, these things are known to the mathematician, at least, the mathematician speaks of them as if they are known in some direct way through some abstract action of mind called reason.
Since we can not reach into the mind of an other, we are left with what mathematicians say when they talk about their direct contact with these abstract entities. The phrases of mathematics appear to have some relation to the abstract entities, or so many mathematicians insist. All we have to go on in conversation is what the mathematician says about their contact with a (the?) mathematical world. Most mathematicians accept that when someone talks about triangles they are talking about some thing and not just saying “triangles”.
…
It is hard for most people to accept that I have all the thoughts and feelings that they do, but that I do not think thoughts or feel feelings. I think of, as in “bring to myself”, and feel of, as in “feelings of the keys of my keyboard” in the absense of the keys of my keyboard. That is, I can “type to myself” without this keyboard being present and without overtly moving a finger. It is like air guitar but without the air.
Why this is so hard for people to deal with is hard for me to deal with. It is hard because our vernacular is not conditioned on the experimental analysis of behavior. It is largely the result of Plato and Aristotle… still.
Such are the frightening things that keep me typing here even though I should be relaxing with a nice book or show.
I wish more people where sensitive to their behavior and its consequences and notice how little there is to explain. All around me I see people talking about how surprised they are that machines have been made to mediate fragmentary and weak verbal contingencies set up from calculations on the forms of responses recorded on the internet.
Happily, I seem to have found a way to present these newer notes publicly alongside everything else. Privately, this document is a singular one. It is a pure text document written with the syntactic style of Gruber’s markdown. The only difference between the document as displayed online and as it is displayed in plain text is that there are hyperlinks there which are not here. They add little to the text, or at least they should. A basic textual search is all that may ever be needed, but only to navigate the jungle of notes with their immanent imperfections.
Having finally made these much needed updates, I can go back and read, in detail, what I wrote in these latest notes. I have worked more on my writing than I have on anything else over the past two or three months. Consequently, what I write tends to reach beyond what I can immediately recall, and I am only just exploring the benefits of this verbal blessing. In particular, the freedom with which I move between topics is born from Mark Twain’s autobiography and his advice to Grant in writing his. This is not so much an autobiography as it is a cumulative record, but there are sufficient similarities that much is to be gained from treating the one as the other.
It appears more important than ever to have a cumulative record of my verbal behavior because of the newly strengthed verbal and social machines that have all but taken over our digital worlds. What once was a back ally has been overshadowed as little more than the slums beneath an opulant highway on which the wealthy zip and zoom. Much like Bradbury, I can only imagine the green blurs they come to call grass. In my case, it is not grass, but math that has been most recently distored by the short sighted need for speed. All the way back in 1931 Bertrand Russell pointed up many of the problems which come from the addiction to optimization promoted by a naive outlook on science (in his wonderful book “The Scientific Outlook”). That there is much today which reads well from his words is a sadness. It is almost a century later and not much has been done to avoid such cataclysmic outcomes as those so easily forseen and forestalled. Alas we are “living during interesting times”.
The introduction to Goldfarb’s “Deductive Logic” begins by comending the earlier editions of Quine’s “Methods of Logic” and condemning the later editions. I submit that Goldfarb’s condemnation is the result of misapprehension. My plan to reveal this is to bring out Quine’s later methods through a personalized translation of results from Shoenfield, and perhaps also in tandem from Ebbinghaus, Flum, and Thomas. Vaguely, my plan is to extend Quine’s methods so as to cover the traditionally obfuscated paths through mathematical logic.
Perhaps the most significant difference between Quine’s methods and those of Shoenfield or EFT is that Quine inserts the intermediate methods of what he calls Boolean term schema. These comprise, towards the end of the second part of “Methods of Logic Fourth Edition”, the monadic, or one place predicate, schema. They issue forth from the previous part on truth functional logic as accomodating what was first Aristotle’s syllogistic and later, vaguely, Boole’s logic of terms. What Quine calls “Boole’s logic of terms” is the logic of one place predicates. In a previous note I link to a more recent paper which explaisn in greater detail that the abstract algebra preseneted by Boole (or what we come to translate from his writings as an abstract algebra) is not what is commonly called “Boolean Algebra”.
A fantastic book on Boolean algebras a they are traditionally conceived is R. L. Goodstein’s thin text “Boolean Algebra” which is available as an inexpensive Dover book. It encompases only indirectly the logical methods of one place predicates in as much as there is a correspondence between Boolean algebras of sets and validities of monadic logic. Problems arise, as we well know from the paradoxes, when each one place predicate letter is substituted not with by a one place predicate, but by the purported name of a class as the extension of each item true of some apporpriate one place predicate. In the latter case, the grammatical construction of predication is naively interpreted as a two place predicate of set membership. Elsehwere I have mentioned that it is better to speak of “classing classes” than “beloning to sets” or “set membership” because the vernacular of “classifcation” is closest to modern theories of sets which are special cases of theories of classes. Classes obtain the kind of ultimate status attributed to classifications in the Aristotilean tradition when they are themselves classed in no class. Sets are then a special case of classes which are classed in some, not necessarily distinct, class.
From afar, that is prior to embarking on this verbal discovery expedition, there should be a dual version of Shoenfield’s proof of Hebrand’s theorem that provides a finitary proof of the completeness of Quine’s main method.
I can not be even remotely sure if my guess is right because there is a great deal of confusion in texts like Shoenfield’s and EFT’s over the boundary between logic and mathematics. The result is that Shoenfield and EFT make distinctions, as almost all mathematical logicians do, between syntactic methods and semantic methods which do not arise in Quine’s work because, unlike them, he finds no firmer basis for speaking of such objects beyond what is provided by, e.g., the main method itself.
Where I come in is in finding the harmony between Quine’s definition of validity in terms of grammar and truth, which I generalize to validity in terms of grammar and denotation (true of), and his main method. More than that, I have narrowed down the home of this harmony in a pure predicate logic whose grammatical constructions are those of Quine’s predicate functors (in Carnap’s sense and not in the category theoretic sense).
It is not clear to me that Quine’s proof of completeness of the main method is directly translatable into the broken worlds of syntax and semantics. There is something critical in Shoenfield’s use of “special constants” in making Henkin theories and subsequent use of them in teh Hilbert-Ackermann consistency theory which is a finitary solution to the characterization problem for the consistency of open theories, and then the finitary solution fo the charcterization problem for theories generally.
In EFT, Zorn’s lemma is invoked to cover what they call “Satisfiability of consistent sets of formuas (the general case)”, and rather than use the terminology of ‘special constants’ from Schoenfield, EFT uses the terminology of “Witnesses” (and negational completeness).
It is these disperate kinds of terminology which have caused me the most trouble, that and the obscurity of the methods of logic amongst the mathematical methods. Specifically, the methods of logic have been pasted onto mathematical methods in both EFT and Shoenfield and the confusion is itself justification for Quine’s strict distinction between logic and mathematics.
The method of ‘witnessse’ or ‘special constants’ is suspiciously like Quine’s elimination of singular terms by singular descriptions with their accompanying descriptional premises.
How does Quine avoid the problems caused by breaking the world into two where one is made of syntax and the other made of semantics? He follows Carnap and Russell in speaking only of interpretations. Vaguely, Quine hints at what you might call syntactic interpretations and semantic interpretations, where semantic interpretations are the closest thing to set theoretic models you can get. He avoids all of this ultimately from his work on the definition of validity in terms of grammar and truth, and from the methedological import of what he calls “Lowenheim-Hilbert-Bernay’s completeness”.
In general, I find it better to speak of the grammar and interpretation of schematic theories of logic. Grammar cover vaguely what is narrowly construed as syntax, and interpretations cover vaguely what is narrowly construed as semantics. The emphasis is on ‘schematic’ since interpretations of actual logical languages is as nonsensical as talk of meanings, propositions, information, etc. This is where Quine’s methods are starkly distinct from those of most mathematical logicians. Quine’s elimination of singular terms by singular descriptions is at peace with his more general rejection of talk of items beyond what can be true of a one place predicate in a language of predicate logic.
This troubles the mathematical logician who already knows, in some way which is beyond me and was beyond Quine, the mathematical objects of which he speaks without any appeal whatsoever to the ultimately logical langauges by which others come to tell me of their contact with them. It is important that I mention the difference between nongrammatical verbal behavior and grammatical verbal behavior. Quine insists, because he must, that talk goes on smoothly with short sentence whose, often, single word later comes to play the part of a general term without invoking reified entities. Russell accepts this somewhat as well when he mentions the surprising role which substances still appear to play in the primitive vernacular of early speakers.
Such insistence on Quine’s part is less absurd sounding when we contemplate an early native of language more smoothly translatable into a pure predicate (functor) logic, a langauge, that is, without singular terms or what might be called “proper names” in English and sufficiently similar nonlogical languages. The free wheeling use of function letters in presentations of mathematical logic like Shoenfield and EFT mask these alarming issues under the guise of ‘constants’ where a constant is a degenerate zero place function letter. Mathematician’s talk of functions and sets is beset with trouble most familiar to scientists who eliminate the metaphorical control of vernacular term by introducing technical terms which do not harbor collateral miscarriages of such prescientific metaphors e.g. as from ‘rat’ to ‘rodent’. Though, it is hard to get across the point that mathematician’s talk of sets and functions is prescientific when such terms are already treated as technical to the vernacular of the public. Here we find how it came that Frege is compared with Aristotle.
Whisked away again.
Quine’s main method of proof is the simplest and most important method of proof that I know of. It can be put into a single short sentence if you know the jargon (which you already must to even begin talking about proofs): make a truth functional inconsistency from instances with new variables for existentials of the prenex form of a sentence to prove it inconsistent.
Each key phrase brings up an important part of the methods of logic. Truth functions are grammatical constructions such that any two identical constructions are jointly true or exclusively false when and only when their like components are jointly true or exclusively false. Compounds and components that are jointly true or jointly false are said to have the same truth value. This excludes the case where compounds or compounents are simultaneously true and false, or neither. These cases are excluded by excising them from a strict definition of logic. As far as we are free to make those distinctions which contribute the most to the survival of our species and its practices, we are free to accept the consequences of acting “with vigor without absolute certainty”. Those who would argue that there are meaningful or significant constructions which are not restricted to the truth functional as defined here are not without their reasons and there reasons are not without merit. For now, and, I submit, for a very long time, it is best to set aside such deviations as deviations from a strictly defined logic. If a more persuasive argument is needed then I need only point to Lindstrom’s theorem that truth functional and quantification logic is, vaguely, characterized among alternatives by its completeness and compactness. For more see “Jörg Flum, Characterizing Logics , chap. III of Barwise, Feferman (eds.), Model-theoretic Logics , Springer Heidelberg 1985” here: https://projecteuclid.org/ebook/Download?urlId=pl%2F1235417268&isFullBook=False&isResultClick=False. How far Lindstrom’s theorem(s) go in reinforcing the distinction between logic strictly and logic loosely is not without its vague boundaries.
In truth functional and quantificational logic, inconsistency of schema is validity of negation and validity is commonly given as truth under all interpretations of the sentence letters for sentences, predicate letters for predicates, and free variables as designators of one and only one item in the universe of discourse which is assumed to be nonempty. Often, the utility of free variables is cut short by beginning with schematic function letters whose degenerate case is counted as so called ‘constants’. Under such, often mathematical, methods, constants are nothing more than reserved variables, those reserved for use in positions where singular terms or, better, singular descriptions might fit. The nonemptiness of the universe of discourse is a minor convenince, but not a trivial one. Contemplation of the empty universe has led some to introduce so called ‘free logics’. Whatever they may be, they go in the deviation bin.
The contextual elimination of singular descriptions as the replacement for any purportedly singular term (by, minamally, the use of relative clauses ‘the x such that x is identical to y’ where ‘y’ is the free variable of interest gives that ‘x such that x is identical to y’ is a one place predicate and we replace it everywehre by a newly minted one place predicate, say ‘Y’, subject to the premises ‘some item is Y’ and ‘each item is u with v such that u is Y and v is Y, only if u is identical to v’) renders the schema in which free variables occur as one devoid of free variables and subject to the simple test of vacuous validity which interprets existentials as false and universals as true. I appreciate that this expedient can be troublesome to those who wish to entertain a strict distinction between general and singular terms, but such distincts are a hold over from pre predicate functor methods which do away with variables all together by way of the so called recombic constructions which usher in compound predicates from component ones and which do the work once done by variables.
It can still be said that something is lost by starting with logical methods whose grammatical components do not include anything deemed a singular term, but this becomes a problem of translation which is already pervasive and it is only from a consideration of the history of translational issues that we even come to speak of grammars and other such precise instruments of effective analysis. A native of a pure predicate logic, one devoid of variables free or bound, is not one devoid of designation as reference by a singular term which purports to name one and only one item. Designation is defined in such a pure predicate logic culture as denotation by a one place singular predicate where a singular predicate is just another name for a one place functional predicate i.e. where, for the quantificational native, ‘F is singular’ is short for ‘some item is F and each item is x with y such that Fx and Fy, only if x is identical to y’. The mystery of designation as it is traditionally explained through, e.g., theories of knowledge is left as a problem of paraphrasing into appropriate predicates and premises.
This is all to say, by way of introduction, that truth functional inconsistency is a special case of inconsistency as validity of negation where validity is first taken specifically as truth of each interpretation of a truth functional and quantificational schema.
Schema are like escrow accounts in that they guarentee that validity projects through them to concrete sentences as instances of the schema when in fact such schema are valid.
Schema are truth functionally valid if they are grammatical substitutions in valid sentential schema where sentence letters and truth functional compounds of sentential schema are sentential schema. What makes a substitution grammatical is that they are substitutions of sentence letters or schematic predications for sentence letters and n place predicate schema for predicates in n place predications subject to appropriate restrictions on component variables that are best dealt with using the principles of predicate abstraction and concretion.
That truth functional inconsistency enjoys a decision procedure is not clear from the definition of truth functional inconsistency from quantificational validity.
Here I note that if you are a casual reader, then I am sorry for the headache this must all give you if you ever happen to read through this note to this point. There is so much that I have yet to say that I must say and sadly this is how I am saying it.
My general method of establishing the truth functional inconsistency of a quantificational schema is derived from a generalization of the alternational normal form method of deciding truth functional inconsistency of sentential compounds. My presentation is very much unlike that of Quine’s which begins with the truth functional and carries on step by step as new problems are confronted but not prior to their confrontation. This is a consequence of his definition of vagueness from “Word and Object”: “A singular term naming a physical object can be vague in point of the boundaries of that object in space-time while a general term can be vague in point of the marginal hangers-on of its extension.” pg. 126. And, “Commonly a general term true of physical objects will be vague in two ways: as to the several boundaries of all its ojbects and as to the inclusion or exclusion of marginal objects.” pg. 126. And, finally, “Also, vagueness is an aid to coping with the linearity of discourse. An expositor finds that an understanding of some matter A is necessary preparation for an understanding of B, and yet that A cannot itself be expounded in correct detail without, conversely, noting certain exceptions and distinctions which require prior understanding of B. Vagueness, then to the rescue. The expositor states A vaguely, proceeds to B, and afterward touches up A, without ever having to call upon his reader to learn adn unlearn any outright falsehood in the preliminary statement of A.” pg. 127
Whether there is always an appropriate path of vagueness for that of linear discourse under the adverse conditions just quoted is a question whose answer is as elusive to me as that of an equally appropriate paraphrase into a logical language.
For my own sake, I am going to go a bit further into Quine’s theory of vagueness. The setences I have quoted just two paragraphs ago make it seems as if only physical terms are contemplated as vague. This is not true, but I have been vague in my introduction of vagueness. Quine begins section 26 on vagueness with the somewhat cryptic literary description of what goes on that brings us to even speak of vagueness:
“Vagueness is a natural consequence of the basic mechanisms of word learning (cf. section 18). The penumbral objects of a vague term are the objects whose similarity to ones for which the verbal response has been rewarded is realtively slight. Or, the learning process being an implicit induction on the subject’s part regarding society’s usage, the penumbral cases are the cases for which that induction is most inconclusive for want of evidence. The evidence is not there to be gathered, society’s members having themselves had to accept similarly fuzzy edges when they were learning. Such is the inevitability of vagueness on the part of terms learned in the primitive way; and it tends to carry over to other terms defined on the basis of these.” pg. 125-126
Quine’s literary dint shows in his phrase ‘penumbral objects’. When I first read that phrase I saw to myself some of the examples of ray marching rendering algorithms for the penumbra of shadows cast by objects and their edges in a given arrangement of lights about a room. This seeing was helpful at first, but far too strong for me to respond appropriately to the role of the phrase ‘penumbral objects’ in the vague introduction to vagueness. A key consequence of what Quine has said which escaped me at first is that vagueness is to be contrasted with the indeterminacy of reference as demonstrated by his puzzles of radical translation. Such indeterminacy is felt from the task of translation, but remains unsettled even after having been felt and acted upon where as vagueness is a matter already settled by the practices of a given verbal community. I say settled, but perhaps it is better to say that vagueness remains under the control of the purported indeterminacy of reference and never the other way around.
Whether the practices of verbal communites are actually subject to the effects of indeterminacy of reference or independent of their fancied existence is a question to which I have no good answer yet. I’m inclined to say that Skinner’s analysis of verbal behavior reveals that only upon introducing autoclitic verbal behavior is the possibility of such indeterminacy raised, and that it is perhaps only properly inflicted upon grammatical behavior as a special kind of autoclitic behavior. From which such indeterminacy of reference is a pathological consequence of linguistic precision which looms larger than it is when set out as a spurious consequence of this or that formal description of a response in isolation from its occasion or consequence.
I find myself traveling down these paths which appear to diverge from that set out by my first steps from my starting point without worry that I am actually lost. Those first steps suggest a straight line where I see none. Rather than make a big deal as to where these first steps seem to lead I prefer to take further steps and look back from there.
Thus I look back and see that I left my generalization of the alternational normal form decision procedure for sentential schema (what Quine calls merely ‘truth functional schema’ given his method of presentation in the first three parts of “Methods of Logic Fourth Edition”) unexplained.
My generalization is a tiny one which takes advantage of the rules of passage allowing quantifiers to pass up and down the truth functional subcompounds of a schema without disrupting equivalence. I must note that equivalence here is “nonvacuous” in that we remain warned to check our original schema in the vacous case as a special or trivial concern. Some of the rules of passage do not survive in a vacuum.
My generalization takes the classical grammatical constructions and paraphrases them into appropriate compounds of a strict form: prenex alternational normal forms. Just as we can take the grammatical constructions of joint denial or alternative denial as basic and give the other truth functions by appropriate paraphrase, so too can we paraphrase each truth functional construction into ones belonging to a strict category that eases mechanical calculation.
Free variables are marked so that they do not get mangled by the relettering which newly alphabetizes the prenexed quantifiers of the compounds to be constructed. Alternatively, the method of taking all free variables as dummy singular terms eliminable by a newly minted one place predicate subject to the appropriate existence and functionality premises could be instituted without imposing the full elimination of variables by predicate functor methods (though it is the elaboration of such methods which govern almost all of my work here).
The prenex-alternational-normal-form-conjunction of two components in prenex alternational normal form is constructed by first substituting each quantified variable in the second component with those which continue on the alphabetized order of the quantified variables from the first component. The quantificational part of the constructed compound is then just that of the first part concatenated with the reletterd second part. A predetermined method of distributing the alternations of the matrices of the first component over that of the relettering of the second as with the classical equivalence of alternation over conjunction is applied. Then inconsistent clauses (a clause of a schema in alternational normal form is the subcompounds of conjunction, though it occurs to me that I have generalized to alternations and conjunctions of other than just two components by way of associative equivalence) are crossed out. Inconsistent clauses can be identified by the occurrence of a predication and its negation among its components. If all clauses are crossed out then the compound may be taken as some canonical inconsistency which displays whatever clauses seem to help most in the degenerate case. Otherwise the remaining clauses are put in some sort of lexicographic order and appended to the block of quantifiers in an alternation.
One of the benefits of dialogue, in Plato’s sense, is that it produces intermediate products which the public can somewhat interpret. This partial passage between the vernacular and the technical is helpful to those who are not already familiar with the technical aspects of dialogues from times past.
This is also unforunate because it reveales how weak are our methods of teaching that we are having dialogues which reoccur throughout millenia without anything like an eternal conclusion. Eternal conclusions are the kinds of sentences that make science a going concern as part of prevailing social practices. Had the sentences of scientists not been marked out by their almost endless standing then there would no longer be talk of science outside those peculiar communites that stay on like a vestigial limb to the rest of our evolving social worlds.
Hence the significance of Liebniz’s many projects which called for some universal methods meant to unify, regiment, and marshal the disperate methods of science and logic so that fewer pieces might slip through the cracks in history. How much capital has been spent as of late on scanning old books in order to extract something from them that individual scientists may leave unnoticed because they are committed to some institutionalized timesink? Over and over again we have probably scanned the same book into relatively similar sequences of ones and zeros that have been replicated at what may seem like a low economic cost, but which hits basic productivity right in its boldest bits.
The theory that capital controls will sort out these inefficiencies is supported by little more than wishful thinking taped together with slam pow prose and pointed graphs with highlighted midsections for greater emphasis. The experimental analysis of behavior reveals that the probability of a response occurring on a given occasion is a theoretical conclusion from experimental records of changes in rates of conspicuous consequences from categorical responses on carefully controlled occasions. Anything less than such vivid reports from the lab and through it into the field are scientifically bankrupt. Anything else is less than a ‘fact in the bag’. The slightest bit of predigested data is suspect to the experimental scientist. The strong experimentalist is always in the ‘show me’ state.
I shall surprise myself if anything I have written in these past two notes comes together coherently upon further reflection.
Unsurprisingly I have been unable to update my notes publicly where my earlier notes are on display. This is largely because the internet as it was known no longer works as expected. It is also because I tend to have a stricter control over my contact with machine mediated social and verbal contingencies than other people. Such nonsense is easy to excuse by pointing to my weaknesses, e.g. general anxiety. This would not be inappropriate, but it is more than likely to be followed by a general kind of contemplte, e.g. for those things that make me different from the majority of people who live in our shared, yet divided, digital world.
I will find a solution to this problem of posting, but wish it was not just one thing after another when it comes to dealing well with computers and the people who vaguely control them.
The term ‘vague’ was once one that I avoided because it had no definite role to play in my writing. Upon rereading Quine’s “Word and Object” over the past few days I have finally grasped what he was getting at in the chapter on ‘vagueness’. I do not have complete control over the contingencies in which ‘vague’ now plays a part, but responses in which it occurs are stronger than they have ever been. The similar appears to be tentatively true of ‘ambiguous’.
Quine’s method of vagueness is one that I had already noticed from my extremely close readings of “Methods of Logic Fourth Edition”. Vagueness is the only passage from the vernacular to the technical that seems to be open to me given my sensitivity to truth e.g. I can not stand sentences in the vernacular which purport to be true or false when all I can see of such predicates is their application to the items of a theory of the grammar of logical languages. Vagueness permits me to speak in the vernacular without purporting anything like truth, denotation (of each item true of a general term as opposed to designation of one and only one item by a purportedly singular term), or, worse, propositions. I say worse because the science of behavior points clearly beyond such ancient ruminations whose origins seem to pass through ‘substances’ or its modern update as ‘stuff’.
While such terms may hold up nicely when only the vernacular is at play, the moment technical terms pass through such literary boundaries, we come to the benefits of experimental control in laboratory environments. These benefits are largely that of effective action.
Now that I have entirely forgotten what it was that set me down to write this note in the first place, I am lost. Such verbal discovery expeditions as these are not all downers. Heading out from basecamp is always something of a risk. The jungle is thick. The blades are dull. There is much that we do not command or control. It is wild out here in the world of words and things.
Ah, something is coming to me now. I had sat down originally to say a few things about Mark Twain. His writing inspires me to write as I wish as much as I want as long as I am saying what I would if I was trying to make a friend laugh or at least share in the joys of talking. For all that we do not control about our complex and interconnected social environments, we have profound control over our immediate verbal environment. We have less control that we used to, but more in other ways. There are more microphones than ever, and these are all access points to crippling kinds of control over verbal behavior. As verbal machines grown in strength they come to mediate complex contingencies which were once rare or nonexistant and which mediate the speaking and listening of others we otherwise would have never met or been in any contact with whatsoever.
I am not so much concerned by prying eyes or ears, there is nothing I have to say that is of any kind of threat to anyone who might ever be interested in such snooping, rather I am worried about the new kinds of verbal behavior that can be freshly extinguished or reinforced and which was once beyond the reach of even the most fastidious teacher of a language. As long as the public continues to speak of languages in the ways set out by linguists we are doomed to come under the control of half baked verbal machines.
The science of linguistics is as important a science as any other, but it does not reach the limits of language as the reinforcing practices of verbal communities. This is unfortunate because grammatical behavior is far beyond where the science of verbal behavior begins and those like Russell and Quine have already gone out of their way to show what are the logical and epistomological (and ontological) grounds for doubting the supremacy of grammar in daily life. Quine vividly describes the problem of radical translation which has its roots in unavoidable errors of reference when terms come to stand for more than globs of events as part of the passing show which bombards the ends of our sensory nerves on the boundary between the exterior and interior of our bodies. Russell vividly describes his inability to justify the premises of any workable theory of the world which must make due with our limit contact with the external world and our knowledge of it through this pinprick of one off experiences, beliefs, and other such ultimately mystical and mythical things.
Back to Mark Twain. He encourages me to be the imp that I am. I would say ‘gadfly’ if others were not so quick to pin Socratese on me. He was certainly a memorable character, but what I see of him is more myth than man. His purported mastery of rhetoric and logic, though he would appeal to some innate uncertainty as to either claim, is beyond anything that I have seen evidence to acknowledge beyond the spritely part in the dialogues of Plato. They are synthetic dialogues, and though I have seen and heard great debates thanks to the modern methods of audiovisual recorders, I have seen nor heard anything like the repeated triumph of Plato’s Socrates, more like ‘Socra-tease’.
What a waste of words this has been, or so it seems to me. Unlike when I compose memos, these sentences and phrases will not get a sharpening. They will be left like fat and grissle. Perhaps an aid to flavor but not to the eating.
Shall I move on to a new note?
The previous note is very important as it contains much of the updated material to be found in the latest draft of the memo on logic. I am not going to finish here what I started there. It is late and I must write something or feel the small death which comes from leaving the page unmarked.
Since I began the note I started yesterday I have reread much of Russell’s “An Inquiry into Meaning and Truth” and “Human Knowledge” as well as Quine’s “Word and Object. I return to these texts again and again each time I come to a problem which I do not yet have the words to solve.
Russell’s method of philosophy links vernacular phrases to technical ones. He acts innocent when he writes his way towards the kind of conclusions which have already occurred to him. This feigned ignorence is purportedly to the benefit of the reader who does not yet know all that Russell has previously contemplated, much less what exactly he has written in the book in their hands. This is not a critique of Russell. It is a fact which all writers must wrestle with, and it is the problem of teaching beyond the vernacular of the reader.
The reader is part of an audience and the audience is part of a language, though we allow that audiences be composed of perhaps members of different languages. The problem of translation is hardly dealt with well in a book composed by a speaker who belongs to the same langauge as the reader. Such belonging is always a matter of degree, and this is made clear when we take languages as the reinforcing practices of verbal communities and ask to what extent an individual is under the control of a langauge.
Such contemplations are beyond those like Russell who wed themselves to the vernacular where speaking is nothing more than a byproduct of thinking in that what is said expresses what is felt, thought, believed, or some such other operation of the soul as the mental self or process of the brain or whatever modernization of the Platonic is popular these days.
Why have I spent hours reading Russell and Quine yesterday and today? When I stopped the note from yesterday it was because I did not know what names were best for dealing with the distinction between items categorized by the following:
Finally, in Quine, I found supplementary terms to ‘bivalent’ and ‘contradiction’ that are tentatively appropriate as follows.
These are only of importance to those who are more like Russell or Carnap and less like Quine in that they want to contemplate entities that, in the degenerate case of denotation, are not true or false somewhere or everywhere, or that are true and false somewhere or everywhere. What is to be gained from my accomodation of these digressions is simply to satisfy those who bring them up but are unable to carry through the consequences of their convictions.
I tire and yet I can not sleep.
My work on my memo on logic has been done here in much earlier notes and, lately, has been done with pen and paper over the past few months. The return to these public notes is prompted by repeated shifts in public opinion away from the methods of science and logic. The strengths of science and logic are the greatest of human practices, and if they are lost or broken apart then so may be the survival of the species and its practices.
Without further ado I shall type up the most recent revisions into the initial fragment of the most recent draft of the memo on logic.
WARNING! This is a draft, and I do not have all the details of duality (between ‘denotes’ and ‘misdenotes’ as in “Contradictions denote and misdenote everywhere.”) worked out e.g. there is more I must say about nonimplication and nonequivalence, and the denotative functional constructions are sometimes only half defined. Quine fails to address the consequences of his theory of stimulus meaning, from “Word and Object” in terms of assent and dissent, and thus fails to meet Russell’s legitimate concerns over ‘believes’ and ‘disbelieves’ as related to his definitons of ‘fact’, ‘belief’, ‘truth’ (and ‘knowledge’) in “An Inquiry into Meaning and Truth” and “Human Knowledge: Its Scope and Limits”. Surprisingly, B. F. Skinner’s theory of qualifying autoclitic verbal behavior in “Verbal Behavior” addresses Quine and Russell (and Carnap and many others) prior to and in anticipation of his analysis of scientific and logical behavior. Skinner shows that both Quine and Russell are accurately describing parts of a bigger picture uncovered from the experimental analysis of human behavior.
SUPER WARNING! The words ‘denote’ and ‘designate’ were divided into distinct methods of reference (or naming) by J. S. Mill, and Quine’s predicate functor methods of logic hinge on this critical distinction. Singular terms purport to designate one and only one item. One place general terms denote each item they are true of. N place general terms denote each N-tuple of items they are true of, and so on through ever more complex shapes of general terms. Since my logic descends from Quine’s predicate functor logic (‘predicate’ is just another way of saying ‘general term’), I have avoided Quine’s dodge of “denotes” with “is true of” to emphasize that truth is, at best, “degenerate denotation” by a zero place predicate. This must be done to properly address many problems related to propositions, meanings, facts, and information as, somehow, the things with respect to which zero place predicates are true (or false).
202607241505 Grammatical constructions of logical languages denote each of their compounds with its components. Compounds are also components. Lexicals are only so. Constructions of categoricals match. Only categoricals with matching like components match.
Each interchange of each interchangee for each interchanger at each locator in each interchangeand matches its interchangeand and its locator is lexical; or matches its interchanger, its interchangee matches its interchangeand, and each component of its locator is lexical; or is categorical with its locator and its interchangeand, and each of its components matches the same at the like component of its locator in the like component of its interchangeand.
Each interchange matches its interchangeand and its interchangee matches its interchanger at each locator of its interchangee in its interchangeand. Substitutions are interchanges at largest such locators. Lexsubs are substitutions of lexicals for perhaps extra lexicals.
Lexsubs in validities/consistencies denote everywhere/somewhere, and in conclusions of implications they denote where the same in its premises do. Equivalence is mutual implication. Equichanges are interchanges for equivalents.
Equichange keeps validity, consistency, implication, and equivalence. Equichanges in equivalents are equivalent since each lexsub is lexical or matches the categorical of the same in each component.
Subsitution for lexicals keeps validity, implication, and equivalence since lexsubs of substitutions for lexicals are lexsubs– take the component substituted for as an extra lexical.
When there are lexicals, validities are consistencies. Components are self equivalent; and imply themselves, validities, and what their conclusions do. Only validities are implied by and equivalent to validities. Equivalents are mutually equivalent.
Other than where grammatical constructions denote, they misdenote. Lexsubs in invalidities/inconsistencies misdenote somewhere/everywhere, and in conclusions of nonimplications they misdenote where the same in their premises denote. Nonequivalents are mutually nonimplicative.
Equichange keeps invalidity, inconsistency, nonimplication, and nonequivalence. Substitution for lexicals keeps inconsistency. When there are lexicals, inconsistencies are invalidities. Components are implied by inconsistencies. Only inconsistencies imply and are equivalent to inconsistencies. More on nonimplication and nonequivalence is needed here!
Matching compounds of denotative functional constructions jointly denote or exclusively misdenote, where and only where (waow) each of their like components jointly denotes or exclusively misdenotes. For example,
Validity is inconsistency of negation and vice versa. The following need further contemplation relative to contradictions, bivalents, and etc.:
Here I am unable to go forth with the edits outlined in pen on the paper I have copied here. The problem is that denotation is not obviously misdenotation of negation as I had felt. This may occur first from a deviant lexical which happens to neither denote nor misdenote somewhere. Secondly it comes from deviant constructions which are similarly constrained.
I had hoped to have this resolved before typing up my notes here, but it is a happy accident to find that I had not.
Validity, and inconsistency, have been defined as in Quine’s “Philosophy of Logic Second Edition”. While Quine admits that there may be occassions where …
I leave this note undone as it was when I was wisked away from writing here. The rest of this is fragmentary responses which may not even be complete sentences, but I very much wish to leave them as they were so that there is no mystery as to what goes on when you make notes.
Contradictions denote and misdenote somewhere, in bivalents denote or misdenote everywhere, Contradictions denote and misdenote somewhere. Bivalents denote or misdenote everywhere. Noncontradictions denote and misdenote nowhere. Validities, inconsistencies, and denotative functionals are bivalent.
Each alternation of each bivalent with its negation is valid: it denotes waow its bivalent denotes or misdenotes, hence each lexsub in it denotes everywhere. Each conjunction of each nocontradiction with its negation is inconsistent: it denotes waow its noncontradiction denotes and misdenotes, hence each lexsub in it denotes nowhere.
Existential closures denote waow there is somewhere denoted by their component.
Existential croppings denote part of waow their component does.
Notes on the 8th lecture of “A History of Philosophy” by Arthur Holmes on the ethics of Plato.
Out of nothing more than habit, I tend to take notes when I watch lectures, but have found myself doing so when watching something as casually as a documentary by Ken Burns.
The tripartite theory of souls by Plato, better the three elements of the theory of the soul by Plato:
By spirited he means something like vavoom or ambition or vim or pep or vigor or vitality or initiative or etc.
Each element is identified with a corresponding part of the body:
Souls are gods of individual behavior in that they are the initiators, originators, or creators of the behaviors said to be mediated by the appropriate parts of the body. Plato would not have said as much as I have just, but what I have said ties his work into the long and persistent history of prescientific explanations of human behavior.
Holmes mentions the similarities between the three elements of the theory of souls by Plato and the three elements of the theory of psyches or minds by Freud. Even the word ‘psyche’ was manufactured by the Greeks and Romans as the personifcation of the soul: Psyche was the wife of Cupid, otherwise known by “Eros” an etymological ancestor of “erotic”.
Specifically, the theories of behavior from the 1800 divides the origins of behavior into
Loosely these map onto the elements of the theory of Freud as
and these are handed down into modern cognitive-behavioral psychologists by the metaphors of
The proliferation of purported explanations of behavior thorugh various alchemical interventions are one of the largest threats to an effective science of behavior because they make it hard for the new scientist to escape the various controls imposed by their vernacular upon their experimental analyses.
Corresponding to each of elements in the theory of Plato are its activities or goals or “telos” or the good of it:
These emerge parallel to the classic Greek cardinal virtues (prudence, justice, fortitude, and temperance) conditioned on cosmic justice like that on display in the Homeric epics.
Plato’s ethics appears then as the claim that the just person is one who obtains these telos harmonously so that the intellect controls the spirit and the spirit controls the appetite.
It is unfortunate that this philosophy still governs prevailing social practices. The science of behavior is the appropriate alternative, even though it may be young and not have the same pedigree as the popularized theories in our earliest and latest literature.
The myth of the charioteer pits two horses against the driver in a dynamic metaphor of the life of the human soul where the charioteer is the intellect, the lead hours is the spirit, and the other hourse is the appetite.
Teleological is “goal oriented”: “telos” means “end”, “goal”, or “purpose”. Natural, operant, and social selection by consequences from variation replaces the theoretical role of a telos in explaining the origins of species, behavior, and practices, but it has been hard to overcome the millenia of prescientific stories which continue to overwhelm prevailing practies.
Plato uses “eros” like “desire” as in “What do you want?” from which the question follows “What are you going to do to get it?” or “What is the highest life?”.
The soul’s love moves the individual. “The good life is one inspired by love and guided by reason.” Bertrand Russell.
How do you get people to know and love the good? What are values and how do they work? Values are ideals we are said to pursue.
The life of the soul is abundantly practical in its wonder for knowledge. The highest kind of thought, invocation of the soul, is contemplation, as that which is beyond logic and rhetoric and identifying the good, to delighting in, pondering, reflecting on, absorbing, looking at, wondering at, admire, etc. it.
Wisdom results from the contemplative delight in the acquisition of knowledge. It becomes a practiced capacity for making sound judgements in particular cases with moral or aesthetic convictions. Naturally guides the spirit and controls the appetite.
“Rhetoric without dialetic is a tool of the appetites.” We need knowledge of ideals from dialectic.
The Pythagoras coined “philosophy” as “the love of wisdom.”
Eros vs. Phileo: two types of love, where phileo is a love for the things own sake. The four classic loves are agape, storge, eros, and phileo.
Blind fate gave way to cosmic justice in the epics, and it is cosmic justice which primed Plato.
Intrinsic v. instrumental goods of Aristotle: good in itself v. good as means.
The improvement of the soul is valued because it is eternal and the body is not, thus it must be liberated continuously to be like the form of the good i.e. to be like god where god is the form of the good. Good souls immitate god.
What is a virtue, “arete”, quality, excellence, full realization of potential, moral vitue? To have virtue is to be the good. Good acts are by the way. To be good is what the soul must reach through improvement as pursuit of the telos.
Justice is the ordered unity of the virtues.
Pleasure: the Gorgious and Philebus dialogues. Pleasure is not the highest good, but it is a good. Satisfying the appetites, hedonism, is not the highest good because the soul is composed of the higher elements of intellect and spirit. But pleasures are part of, by products of, higher activities such as contemplation. Pleasures crown the good life.
It is easy to get caught up in all this wishful thinking because it has become our vernacular. The history of philosophy, like any other history, is gathered as an empirical checkpoint, however odd the term may seem when applied to such unreliable reports from field work of such ancient provenance, on theories of behavior and their specialization to human behavior.
Though my memo on logic has been a very long time coming, it must continue to await completion as the longer I contemplate it the stronger it becomes. Just yesterday I solved a big problem with the dual words ‘denote’ and ‘misdenote’ which are a sorta neutral version of Quine’s ‘assent’ and ‘dissent’ and Russell’s ‘belief’ and ‘disbelief’. Mine are technical terms, ‘misdenote’ isn’t even a generally recognized English word. As always, Russell’s ‘belief’ and ‘disbelief’ are the most comforting access point to anyone not already familiar with teh problems posed and solved by their duality.
The problem is this: how to define the negation of a sentence without the definition including the negation of a sentence. The traditional definition is this: The negation of a sentence is true if and only if the sentence is false and vice versa. This appears, at first to avoid including the negation of a sentence as a subcomponent, but we fall into trouble when we ask what the definition of ‘false’ is. The traditional definition is this: A sentence is false if and only if it is not true and vice versa.
You can try to slip out of the error by playing around with negation and falsity as much as you want but will always find yourself in a similar pickle. Russell’s way out is to appeal to abstract entities such as facts, propositions, and beliefs. His line is that of the vernacular. Most people accept without doubt that there are facts and beliefs. They can easily be made to accept that there are propositions by talking about the meaning of sentences in different languages being the same. Propositions come out then as what sentences mean when they are true or false. He leaves the signification of sentences wider than might be reached by truth or falsehood following Frege and his ilk.
A proper synthesis of the work of Quine and Skinner eliminates the problems posed by meaningfulness. Sentences fall away as mere records of certain kinds of verbal responses, grammatical ones at that. Meaningfulness is found to be a restriction, one which we are freed of from an experimental analysis of behavior. As much as scientific practice is taken as vernacular to the speculative philosopher, there is no doubt that Quine and Skinner are meandering in a helpful direction. Whether we arrive anywhere like where they point is still to be seen, though if I must bet I would bet on them more than any others.
This has been such a delightful few paragraphs that I shall include it in the newest note of my cumulative record.
Most of my work on the memo on logic has continued with pen and paper and without any computer mediated word processing. This is the first note where I plan to write out the latest evolution of the memo. Work done prior to the long interval between these latest notes and earlier notes has brought greater clarity and exactness to the methods I previously championed in the hints which are, at this moment, enshrined in the few hints on logic given above.
The memo which has evolved is one which covers similar grounds with paths whose routes are well charted and well worn. The entierty of my logical methods derive almost exclusively from Quine’s definition of validity from grammar and truth in his 1986 “Philosophy of Logic Second Edition”. When dealing with logical languages generally it is best to follow Quine’s grammatical definition since it does not depend on the peculiarities of the grammatical constructions of a given logical language.
This mixes well with von Neumann’s schematic methods. Quine more than any other that I have read does the most to articulate the full strength of schema, but, there remains a further use of schema which von Neumann pointed up and Quine more or less avoided except when its passing necessity occurred. It is a practice which mixes von Neumman’s “axiom schema” each instance of which is to be counted as a theorem in the axiomatization presented with respect to it, and Quine’s predicate functors which fully generalize the syntactic practices of Carnap in his “Logical Syntax of Language” and even more so in his “Introduction to Symbolic Logic”.
A concrete example of the practice I have only hinted at is at the end of the present edition of my memo on logic:
“F is wellfounded” is short for each substitution for ‘G’ in “Each item is F if each item is x such that x is G if each item is y such that y is G if y is F to x.”
I do not yet submit that I have articulated the practice with sufficient clarity and exacteness to give it the same status as that of Quine’s predicate functors. Predicate functors are clearly grammatical constructions each of whose compounds and components are only and solely predicates.
The introduction of a sorta generalized predicate functor of the sort shown in “F is wellfounded” is that its occurrence in a dialogue carries the effects of not one schemata but rather a purportedly endless accumulation of instances each of which is itself not a predicate but a predicate schema.
Just now, two paragraphs ago, I mentioned predicates as the only and sole compounds and components of grammatical constructions in the logical languages of predicate functors, but what was said there is not limited to predicates. It includes predicate schema and hence predicate letters.
Methods which smear the contributions of predicates and predicate schema are called schematic methods by me. The schematic theories of indiscernibility are my go to examples. Logical langauges with finite lexicon include their predicate of indiscernibility and it is coextensive with what is otherwise identified as a predicate of identity in theories of identities like those of Godel or Wang.
My choice of wellfoundedness as an example of the schematic methods I am advocating for is picked because it encapsulatesd the entirety of the inductive methods characteristic of formal theory and its logic.
I think of Quine’s proof of his law of infinite conjunction. It follows from a purportedly endless list of lemma each of whose proofs follows in part from those prior to it. It is problematic in that it depends on an endless schedule of consequences which can not be exhausted decisively.
Since this note has become its own thing rather than merely turning out as an introduction to the latest draft of my memo on logic, I shall begin a new note for it. Before that I shall meditate briefly.
The calculus of variations is the second most important missing piece to an effective high school education in mathematics. The first such missing piece is actually a missing method. It is the method of the uniform calculus, from which the point calculus is confronted only upon meeting Maclaurin and Taylor series.
An online text for the caluculus of variations can be found here: https://liberzon.csl.illinois.edu/teaching/cvoc/node59.html.
A canononical text which does not reach into the complexities of functional analysis is Gelfand and Fomin’s 1963 “Calculus of Variations”: https://archive.org/details/gelfand-fomin-calculus-of-variations.
Again, the uniform calculus of variations is likely a simpler starting point than the traditional methods.
The last chapter of Gelfand and Fomin is on “Direct Methods in the Calculus of Variations”. These are to be found as a result of primitive recursive analysis.
For now Gelfand and Fomin is the most elementary yet rigorous presentation of the calculus of variations that I have found. More work must be done to compare and contrast it to other texts.
While I have studied philosophy since the summer of my sophomore year in high school, I have not had a formal course on its history. I finally corrected this error by watching Artuhur Holmes’ History of Philosophy: https://www.youtube.com/playlist?list=PL9GwT4_YRZdBf9nIUHs0zjrnUVl-KBNSM.
The text used in that course is Walter Kaufmann’s “Philosophic Classics, in two volumes: I. Thales to Ockham II. Bacon to Kant”. I have not yet found a link to the text, but there may be something from the wiki: https://en.wikipedia.org/wiki/Walter_Kaufmann_(philosopher).
The memo on logic is coming along quickly. It has divided into two alternatives which are difficult to contemplate concurrently. The one is traditional in that it follows the evolved methods of logic admited under Quine’s strict analysis which eliminates meaning. The other is equally traditional in that it follows the methods admited by Russell which embrace meaning. The difference is starkly contrasted by their respective terms for dealing with the origins of their methods: Quine’s ‘ascent’ is to Russell’s ‘belief’ as Quine’s descent’ is to to Russell’s ‘disbelief’. Quine ascents to public science and Russell believes in private sensations. Neither denies they other outright, but both contend their one is the better starting and stoping point i.e. Quine begins and ends with science where Russell, earlier more than later, starts and stops with sensation.
Skinner covers the convictions of both Russell and Quine without relinquishing any of their key articulations. The contingencies characteristic of Skinner’s science unify the methods of Quine and Russell uniformly. Ordered pairs exemplify this unity in their social efficacy, their concrete expedience, and their abstract economy.
It is better that I record my thinking here or with pen and paper than to leave it unrecorded. However ridiculous my thinking may get, it is not as ridiculus as it might otherwise be had I not at least been taught by so many talented teachers. Perhaps I am not in a proper position to judge the effectiveness or appropriateness of their teaching as their student. Though I might say “I am more effective from having been taught by them than I might otherwise be” this may not contribute to a firm conclusion like “I am more effective because they taught me.” Such self doubts suggest themselves as a consequence of my teachers’s teachings.
…
During the large gap between these latest notes and those much earlier than that, I have made great strides in my work on logic. The memo on logic has appeared in a far more complete form than it has hitherto. This comes in part from the realization that if I were to spend the rest of my life looking for stronger logical methods I am unlikely to find better than what I have already found. The extent to which these methods were found or created is irrelevant since I can point to the books the strengthed them and my explanations of them.
It reamains to be seen if I shall return to writing here at semiregular intervals. It also remains to be seen if I shall ever share this updated version of my notes, memos, and hints with the public. As it stands, I see myself sharing more again, though what may come of it remains unknown.
For a brief time I had succeeded in reading my way ferociously through some of the best books from the best writers I could get my hands on. This petered off a week or so ago as the rate at which unexpected interruptions increased for reasons which remain largely beyond my control. When the conditions that maintain unfortunate events are largely beyond my control, I tend to change what I do so as to minimize the deleterious effects of that which is beyond my control by changing that which is within my control.
Such obvious methods of dealing with the less than happy parts of life and work are much more difficult to execute in practice than they are in principle. This appears as a generality in that practices may be governed by principles. A similar projection of a similar principle is that advice may govern behavior. To change the behavior of someone by giving them advice requires that they already have a history which has taught them the kind of self control that activates advice as a governer of behavior. Who among us really has such a wonderfully effective personal history?
…
The internet suggests that “Calculus of Variations” of 1963 by Gelfand and Fomin gives the proofs of the main consequences which compose the theory of the calculus of variations as it extends the most familiar methods of the derivative and integral calculus popularly taught in a scientific education. A version can be read online here https://archive.org/details/gelfand-fomin-calculus-of-variations.
It is an English translation of a Russian text. Liberties in the loose constraints on any translation are exercised by the translator to include problem sets which were not part of the original text as well as explanations which are tempered by the evolved practices in the field.
It is strange that there is no logically tight presentation of what Newton and Liebniz handed down to us so many years ago. Real analysis, and its big brother Functional Analysis, do not achieve the generality of method which is otherwise introduced to those more focused on the calculations required to set up strong laboratory environments. I find no more comfort in the set theoretic methods of Cantor than I do in the category theoretic methods of Louis D. Nel of Carleton University described briefly here https://arxiv.org/pdf/math/0504565.
I am having great difficulty reading David Deutsch’s “The Beginning of Infinity”. Why? Because it presents just enough of what David knows about the history of science and logic to support his outlook. It takes advantage of what the reader does not know, e.g., about those who objected to the methods he promotes.
Who cares if I am having difficulty? Perhaps I am as self-aggrandizing in my selective history as I claim Deutsch to be. Yet, here it is that I diverge from Deutsch and so many others. I have begun by questioning my authority as a writer. Who am I? Who am I to tell you the reader what is true and what is not? Do effective methods of science and logic work upon readers by surrepticiously taking advantage of prevailing social practices and presumed authority?
Forgive me reader for wasting your past life with my words. I was a fool to think that you might care for what I said. Fortuitously, you are not me and need not be burdened by the trivia and tedium that make my past writings intelligable. I now know that it is best to avoid making an ass of you.
It is not good enough to tell you what I have read that has changed me as a writer over the past three months. Who gives a damn what someone says when you can see what they do? Thanks James Baldwin.
No one cares about science and logic because no one is taught to care. No one writes well enough about science and logic to be read by children and adults alike. Everyone likes a good story, and, purportedly, science and logic are the best stories in history, but they are not told like the little engine that could or sung like the itsy bitsy spider. Science and logic are not a luliby to the world yet. Who is going to make them so? Why are so many mystics hell bent on making science and logic into an imaginarium rather than into what it is that intelligent people use to deal well with the world?
What is it about pseudo science that makes it indistinguishable from science for most people throughout history?
Why do we hate smarty pantses who prop themselves up on a stage and say “look at me!” when it is really the world that science and logic look at with an all consuming intensity that burns brighter than the most brilliant nuclear blast?
What am I going to do about it? Nothing? Am I going to just wait for the world to unwind while I delude myself into thinking I can do something to stop it? Or, am I going to try to stop it and truly discover where my delusions begin and end? The delusion of ambition frightens me when I see it in others. Does it do the same when I see it in myself? Whose problem is that?
So what if I want to write about science and logic. If I do not do it well then nothing I write matters. It might feel good to me to see my words on the page, like a bit of candy to a diabetic.
Self indulgent writers die of a self administered overdose.
…
Forgive me reader for robbing you of life with my wasteful words in the past. I was a fool. There is no other reason for me to write than to be read by you. Anything less is soil on the grave of a slow death.
These words are all that you have to judge me by and if you are to judge me well then it is only because I have recently read those who are better writers than I shall ever be.
I am still a fool, but not the same one I once was. Only my words can save me from being the fool I once was.
I stopped writing here last November because it kept reinforcing the kind of writing that is only helpful for those who already know something about what I have to say. It is a poor thing to leave the reader so alone without any other help than saying “as you know” or “recall” or “continuing from where I left off here” where ‘here’ begins itself with the same message leading back, message through message, to the beginning of these notes until all interest has been fully expunged from the wayward reader.
That I have started writing here again can not be taken as indicating some sort of literary accomplishment on my part. I have yet to learn how to write well for the reader which does not yet know something about what I have to say, but I am certainly better equipped to do so now than I was months ago.
I want reading my writing to feel like sipping a smooth cup of warm coffee. If you do not like coffee, then think of tea instead. If you don’t like coffee or tea, then you are helpless.
In the interim I have read far more than ever before with the explicit aim of assembling sentences that read like a smooth cup of coffee. of picking up the tricks of the writing trade that make for smooth and satisfying reading, even if it is of a stupifyingly technical sort. There is no reason that technical things must be tiring. They are not tiring for me and I like to think that this is more than just that I have had such wonderful teachers in the past. If that ends up being the only reason I am fond of the technical things that delight me, then I must at least be something like a teacher to you, my reader.
I know how to teach you in person but not how to teach you in writing. It hit me that I can write as if we are here together. Our shared language links us remotely accross time and space. The worse I write the further apat we feel. If you trip over this carpet of words it is because I either wove it wrong or let it bunch up when I rolled it out.
More fiction has been cast upon my reading eye over the past three months than over any other interval of time in my life. Young me read everything he saw, but even so it never added up to all that I have consumed as of late. I could list all that I have read but that would be boring and nothing is worse than boring writing.
There are a number of results involving Conway’s surreal numbers that I would like to collect here, but each time I sit down to collect them I am reminded that I have not yet collected together in a single place the methods of logic as I find them.
The hints that I have collecte thus far are aimed at predicate logic generally and begin there with the definition of validity with respect to supplemental lexical substitutions. The denotative functional validity is not limited to the supplemental lexical substitution. It admits substitutions of non denotative functional components.
Here then the definition of denotative functionally valid is needed. To say that a predicate is denotative functionally valid is to say that each substitution of like predicates for like non-denotative functional components.
The main difference between the denotative functional compounds and the quantificational compounds is that the quantificational compounds denote or not with respect to occasions other than the present one. That is, the component of an existential closure may not denote on this occasion but its existential closure might if there is an occasion where it denotes.
I was reading through Patrick Suppes’s “Axiomatic Set Theory” (a delightful dover book) for the second or third time and it finally occurred to me that I’d never really written down the Tarski-Peirce definition of ‘finite set’.
One of each subset of a family from a finite set is not a proper superset of the others.
I also have not tried to produce the corresponding schema appropriate to the predicate functor ‘finite’.
The ‘family of subsets’ part is just a two place predicate, e.g. take it as F. To say “there is one subset of the family that is not a proper superset of the others” is to say “some item is (x such that each item is (y such that not y=x, only if some item is (u such that Fyu only if not Fxu)))”.
Then a one place predicate F is finite when each instance of the schema - each item is (u such that Fu if and only if some item is (v such that Gvu)) and some item is (x such that each item is (y such that not y=x, only if some item is (z such that Gyz only if not Gxz)))
is true (where ‘G’ is a (schematic) predicate letter). It may be better to say ‘F itemizes’. It is interesting to note that since ‘F itemizes’ is not short for a schema or an instance of a schema, then it can not be used as a component in a grammatical construction. For this reason, it is probably inappropriate to speak of ‘itemizes’ as a predicate functor.
But, this happens often and the solution is often enough (always?) the introduction of an appropriate predicate functor e.g. ‘F itemizes with respect to G’ is a schematic predicate functor (here ‘F’ no longer designates a one place predicate and is now a predicate letter on par with ‘G’) each instance of which is a complex predicate and hence may be a component of further complex predicates or schema.
I was unable to finish the note I started yesterday. The aim was to show that each of the denotative functional abbreviations of logic were denotative functional according to the provided definition.
I’m reminded of the benefits of the methods of logic over the methods pursued in
Lehman, Leighton, Meyer 2018 “Mathematics for Computer Science”
So much of what is presented as if it was set theory, number theory, or type theory is purely logical and the benefits to presenting it as such is uniformaty of methodological contemplation e.g. the application of the methods of logic to mathematics and the other sciences obscured by neglecting logic.
There is an arithmetic theory which presents itself as epistomologically prior to the theory of natural arithmetic. It is the theory of binary trees. Among its premises are
Write ‘empty’ for ‘x such that each item is (y and z such that it is not the case that x pairs y with z’ so that the premise of emptiness is equivalent to ‘There are empty pairs’. Exstentionality implies indistinguishability of empty pairs. Therefore, extensionality and emptiness imply the singular description ‘the empty pair’ designates. Write ‘0’ for ‘the empty pair’.
Similarly, extensionality and formation imply the singular description ‘the pair of x with y’ designates. Write ‘(x,y)’ for ‘the pair of x with y’.
Then perhaps the most important premises are those which are instances of the schema of induction:
The methods of logic are purportedly reported by conclusions of this theory of the arithmetic of ordered pairs. Such a theory is not categorical in that it has multiple modesl. If you do not like that then provide additional premises which are relevant to your interests. Alternatively, you can take advantage of the noncategoricity Those who submit that the theory is ‘not categorical’ in that it has ‘multiple models’ need only supply the premises they submit are missing or take advantage of such noncategoricity.
The induction schema given is not yet enough for all of mathematics. All of mathematics is presumably realized by introducing precisely those principles of induction that permit the construction of Conway’s surreal numbers.
I’m collecting references to assemble the logic of such an arithmetic outlook on mathematics.
It has come to my attention that the order that Conway gives in the construction of the surreal numbers may be the order upon which the stronger form of induction may be based.
This appears to correspond to the order established via the operation of transitive closure described by Feferman in 1989 “Finitary inductively presented logics” and, depending on how the operation of transitive closure is defined, this appears to correspond to the Kleene-Brouwer order.
First, the definition of Kleene-Brouwer order.
The pair of sequences, indexed by ‘<’, x with y of items linearly ordered by ‘<.’ are denoted by the postorder ‘<:’ derived from ‘<.’ when there is an index i such that j<i only if x(j)=y(j) and, either 1. x(i) exists and y(i) does not 2. or x(i) <. y(i) (i.e. they both exist).
Post order is left subtree, right subtree, tree.
Wait, there’s a lot more going on here than these stupid wikipedia sources suggest.
The first thing is a prefix order. A two place predicate is a prefix order if it is antisymmetric, transitive, reflexive, and downward total.
A two place predicate F is downward total when - Fac and Fbc, only if Fab or Fba.
A history of an item x in the field of a prefix order F is the predicate {y: Fyx}, and its future is {y: Fxy}.
https://en.wikipedia.org/wiki/Prefix_order
Oh, okay, so “trees” in set theory are defined as follows.
A two place predicate F is a tree if each item is (x such that x in the field of F only if (y such that Fyx) is well ordered).
A two place predicate < is well ordered if each instance of the following schema is true: - F is included in the field of < and some item is (x such that Fx), only if some item is (y such that y is indistinguishable from the (z such that Fz and each item is (u such that u is indistinguishable from z or, Fu only if <zu)))
OH a generally helpfull book on lattices and order: - Davey, Priestley 2002 Introduction to Lattices and Order 2nd Edition
How denotative functional logic works.
Grammatical constructions are (denotative) functional when each like compound exclusively denotes or not where and only where (waow) each like component exclusively denotes or not. Examples:
Introducing joint denials by saying “joint denials denote waow each of their components don’t” may be ineffective. The alternative way is with “joint denials denote waow (neither) their left (component) nor right (component) does”, or, better, “joint denials denote waow (neither) their left does nor their right does”. Then alternations denote waow their left or right does and conjunctions waow their left and right does.
The bigger reason to do things with ‘nor’ rather than ‘each not’ is to avoid the problem of ‘infinite conjunction’ as some endless conjunction. So, beginning anew.
Oh, another efficiency is to show that these examples all reduce to combinations of alternations, conjunctions, and negations so that the decision procedure (the compound predicate version of truth trees) comes out quickly.
Grammatical constructions are (denotative) functional when, like compounds denote or exclusively don’t, where and only where (waow) their like components denote or exclusively don’t. Examples:
complementary conditionals are swapped complementary converse conditionals, conditionals are negations of complementary conditionals, alternative denials are conditionals with negated consequents, conjunctions are negations of alternative denials, exclusive alternations are conjunctions of alternative denials with matching alternations, biconditionals are negations of exclusive alternations, and cedents are conditionals of conjunctions with alternations.
Functional constructions on lexicon generate chains (a grammatical category).
I have still yet to take notes on methods of unification e.g. from the youtube series on “artificial intelligence: knowledge representation and reasoning”:
My question is what does unification look like in predicate logic (i.e. “standard” predicate logic with a finite lexicon of predicates) and how are the more familiar methods inherited from there?
At best, substitutions give singular descriptions for each variable i.e. ‘x = the {y: ..stuvw..}’.
Or, in general, the general problem is an itemization of indiscernabilities like ‘the {x: ..abcd..} = the {y: ..stuvw..}’ to some substitution.
The general problem is where the predicates are functional in their top place and an equation like ‘f(x,y,z)=g(s,t,u)’ is “Func F, Func G, Fwxyz and Gwstu”.
Well now I’m really confused. It seems that Herbrand already dealt with this but in some different way.
Now I’m trying to figure out the solution to a simpler problem: is there any way to know for sure that the prefix normal form of two schema are variants?
New diction of the history of ideas https://ia601604.us.archive.org/35/items/NewDictionaryOfTheHistoryOfIdeas/New%20Dictionary%20of%20the%20History%20of%20Ideas%202005.pdf
Such a dictionary provides a testing ground for any effecitve theory of (social) behavior.
These are notes and quotes on theories and experiments of the science of behavior. One of the aims is to get out publicly the definition of verbal behavior as a special kind of social behavior in a way that solves the problems of Skinner’s idiosyncratic (and ultimately necessary) definitions in “Verbal Behavior”.
Just like a sewing machine selects teh practices of sewing so do cumulative recorders, or their digital analog click-tickers, select the practices of science.
A clickticker is a gray box with two buttons, one white and the one black, filled with a spool of gray tape such that when the buttons are pushed correspondingly colored dots are stamped onto the gray tape that ejects from the box.
A dot on some tape comes out of a box from a tap on a key of matching hue.
Colored dots on tape appear from taps on keys of matching hue.
Trace a sequence of responses with a sequence of dots.
Behavior appears on occasion. What is the probability that this behavior appears on this occasion?
Calculate probabilities from rates and collect rates from records of repeated occurrences.
Rates are a ratio of clicks over ticks.
“We tend to say, often rashly, that if one thing follows another, it was probably caused by it– following the ancient principle of post hoc, ergo propter hoc (after this, therefore because of this)” About Behaviorism (AB) pg. 10 by B.F. Skinner
“many of the things we observe just before we behave occur within our body, and it is easy to take them as the causes of behavior” AB 10-11
“We assume that other people feel as we feel when they behave as we behave.” AB 11
“what is felt or introspectively observerd is not some nonphysical world of consciousness, mind, or mental life, but the observer’s own body” AB 18-19
“It does not insist upon truth by agreement and therefore considers events taking place in the private world within the skin.” AB 18
How does the verbal community solve the problem of privacy? Science.
“A natural datum in a science of behavior is the probability that a given bit of be3havior will occur at a given time.” Contingencies of Reinforcement (COR) pg. 75 by B. F. Skinner
“Social behavior may be defined as the behavior of two or more people with respect to one another or in concert with respect to a common environment.” Science and Human Behavior (SAHB) pg. 297 B. F. Skinner
“Behavior reinforced through the mediation of other persons” Verbal Behavior (VB) pg. 2 by B. F. Skinner
“behavior reinforced through the mediation of other persons [who] must be responding in ways which have been conditioned precisely in order to reinforce the behavior of the speaker” VB 225
“Verbal behavior is behavior that is reinforced throught eh mediation of other people, but only when the other people are behaving in ways that have been shaped and maintained by an evolved verbal environment, or language.” Upon Further Reflection pg. 90 by B. F. Skinner
“Verbal behavior is operant behavior whose properties are selected by the reinforcing action of a mediator on teh basis of their correspondence to the conventions of a community.” Maria De Lourdes R. Da F. Passos “Skinner’s definition of verbal behavior’
“In the history of logic and science we can trace the development of a verbal community especially concerned with verbal behavior which contributes to successful action.” VB 418 ch.18 logical and scientific behavior
“The scientific community encourages the precise stimulus control under which an object or property of an object is identified or characterized in such a way that practical action will be most effective.” VB 419
“The scientific and logical community sharpens the discriminative control of verbal responses with classificatory schemes.” VB 419
“Generic extensions are tolerated in scientific practice, but metaphorical, metonymical, and solecist extensions are usually extinguished or punished.” VB. 419
“The contingencies of reinforcement which create a special scientific repertoire and sharpen its stimulus control provide for a kind of behavior which serves the listerner as (1) an optimally effective discriminative stimulus in evoking any behavior he may possess with respect to a situation and (2) a fruitful source of instruction in altering his behavior with respect to new situations.” VB 420
“The logical and scientific verbal community has slowly accumulated a set of techniques for the construction of effective verbal behavior. The speaker moves from one set of responses to a possibly more useful set.” VB 422
“Logical and scientific verbal behavior, as well as the practices of the community which shape and maintain it, have been analyzed in logical and scientific methodology. .. Three steps appear to lead to this sort of methodological inquiry: (1) some kind of verbal behavior, including appropriate relation and quantifying autoclitics, prove to have important practical consequences for both speaker and listner, (2) the community discovers and adopts explicit practices which encourach such behavior” VB 430
The voluptuary (sybartic) solution: “Reinforcement is maximaized while the unfortunate consequences are either disregarded– on the principle of eat, drink and be merry for tomorrow we die– or prevented.” COR 51
“It is not difficult to promote the sybaritic solution. Men readily subscribe to a way of life in which primary reinforcers are abundant, for the simple reason tha tsubscribing is a form of behavior susceptable to reinforcement.” COR 52
Here are some paragraphs I wrote in an attempt to disentangle verbal behavior from non-verbal social behavior.
Verbal behavior is a special kind of social behavior. Social behavior may include the behavior of an other organism vacuously. Verbal behavior involves the other organism essentially. There is non-verbal social behavior that involves an other organism essentially e.g. breast feeding. The other organism behaves socially perhaps without having been conditioned by a social environment e.g. breast feeding.
An organism may behave in ways it otherwise wouldn’t but for the social behavior of others. These mild forms of constructive collaboration are touched with that feature which distinguishes verbal behavior from non-verbal social behavior. The other organism may be conditioned by the social behavior of others so that they are likely to reinforce the social behavior of others e.g. as when teaching teachers. Behavior selected by such reinforcing practices is verbal.
It may be that there is more to verbal behavior than what has already been mentioned e.g. there may be an intermediate categorification of collaborative social behavior which is appropriately placed between social behavior widely and verbal behavior narrowly.
The distinction between socializers and socializees in a social episode. Verbal episodes as social episodes.
water, food, sex, and violence
Abbreviation of itemized binary pairs.
A change to the infix notation for ordered pairs, e.g. ‘(x,y)’, to postfix notation (or prefix notation in a mirrored presentation), e.g. ‘xy,’, yields a method of introducing itemizations of finite trees as abbreviations for complex predicates.
The two place predicate of indistinguisability is designated by ‘=’ (note, I only contemplate languages with finite lexicon, and here, indistinguishability is coextensive with identity).
The ‘{(x,y)}’ for ‘{uvw: x=u, y=v, and w=w}’ by reading ‘(x,y)’ as ‘xy,’.
Then ‘{(x,(y,z))}’, i.e. ‘xyz,,’, for ‘{stuvw: x=s, y=t, z=u, v=v, and w=w}’ and ‘{((x,y),z)}’, i.e. ‘xy,z,’, for ‘{stuvw: x=s, y=t, u=u, z=v, and w=w}’.
Then, writing ‘is’ for the copula of predication, it appears as if ‘(x,(y,z)) is F’ for ‘{stuvw: x=s, y=t, u=u, z=v, w=w, and stuvw is F}’.
Furthermore, given ‘F’ is of shape ‘0’ and ‘G’ is of shape ‘(0,0)’, i.e. ‘00,’, then ‘(F,G)’, i.e. ‘FG,’, for ‘{stuvw: s is F and (t,u) is G}’ so that ‘(x,(y,z)) is (F,G) if and only if x is F and (y,z) is G’.
Obviously, other than, e.g., ‘w=w’ such a component of a conjunction can be dropped entirely within the appropriate predicate abstract.
This would then require the introduction of the kind of notation that Quine used with his virtual theory of classes and, as such, there is a cleaner way of dealing with this.
This is related to the fundamental problem of unification algorithms investigated by Herbrand. Why did he have to die in that ski accident.
I’ve been away from the computer a lot over the past two weeks. More than in the past month. Less than in general.
Looking up papers and their references has been a headache. It used to be that you could click through things and, bam, there’s the source.
Now there’s more broken links than ever. This frustration was mentioned in a past note. It’s mentioned here because it has once again blocked me from digging into an important topic: (finite) trees and their logic.
Maher’s 1988 paper on axiomtizations of algebras of finite trees (and others) is not accessible anywhere and yet it is referenced in a large enough number of papers to make its inaccessibility a frustration.
I also have a question about the transitive closure of each predicate of a logical theory (with finite lexicon). The transtiive closure theory of a given theory includes the schematic premises which amount to saying that the new predicate playing the part of the transitive closure of some predicate in the initial theory is transitive (not a schema) and the schema that any predicate of the theory that includes the initial predicate and which is transitive is included in the predicate playing the part of the transitive closure.
Well founded predicates are part of what I’m reaching for.
Kleene-Brouwer order https://en.wikipedia.org/wiki/Kleene%E2%80%93Brouwer_order is relevant. It gives the logic of post order traversal in a way that generalizes beyond the obvious. It also yields the following result: “For trees over a well-ordered set, the Kleene–Brouwer order is itself a well-ordering if and only if the tree has no infinite branch.”
There is no other general way to deal with this all than to continue work on the abbreviations of predicate logic e.g. the tree abbreviations as part of the graph abbreviations?
Some notes on trees. I’m trying to get a good feel for the standard terminology and the various ways that predicates purportedly about trees are not easy (impossible?) to introduce in predicate logic.
Perhaps a good starting point is:
Some of its references are
These continue notes from #2025-1108-1258 on Kleene’s proof of Hilbert-Bernay’s completeness in “Introduction to Metamathematics” (IM).
The last note ended with a list of what Kleene called “the postulates of the formal system” and, in the context of the proof of Hilbert-Bernay’s completeness, they are the system through which the nonconstructive part of the completeness theorem is to be formalized. The formalism is given only in outline. It is not carried out. Kleene and almost all others refer to Hilbert and Bernays rather than present something like their proof again. This strange avoidance is an alarm bell to me that there is something painful about it and that there is something that can be learned from it.
Like most text on mathematical logic, the methods of logic are erroneously divided into propositional and predicate. At best they can be divided between the truth functional and the quantificational. Furthermore, there is no distinction made between the methods of logic and the methods of mathematics through which the methods of logic are to be analyzed. The so called ‘propositional calculus’ and ‘predicate calculus’ are calculi, they are not methods of logic. They are theories of calculi/protosyntax/concatenation which are beholden to the methods of (predicate) logic. If this distinction is not maintained then the methods of logic are often lost and it is thought that theories of sets, categories, arithmetics can be taken as going concerns when examining the methods of logic.
So, what of the relationship between theories of concatenation and the methods of logic? A theory of concatenation is governed by the methods of logic. That it just so happens to be applied in the analysis of logical method is of little consequence to the logician except in that metaphorical methods are often involved in discovery. How ever strong the metaphor may be, that is, however strong the corner of the theory of concatenation may be, to the logician there is only the interveining theory to be eliminated. A theory of concatenation may have pointed up a metaphorical conclusion about some part of logical method, but the logical method must be dealt with directly rather than some metaphor which might otherwise be a phantom.
Quine made all of this far more aparent by metaphorically describing logic as the vector sum of grammar and truth, better grammar and denotation. Denotation is where the limits of theories of logic in the investigation of logical methods are often found wanting. The response of matheamtical logicians has been to introduce more elaborate calculi as if they are strong logical methods. It goes unnoticed by most that they have only made a more elaborate theory and have not repudiated the logical methods they find wanting.
This problem of logic and theory is made all the more difficult by the lack of general knowledge about the science of verbal behavior. The methods of grammar and the part they play in the science of linguistics is well enough apprecaited by anyone working on “natural language methods” in logic, but the linguist’s prescription to distinguish between performance and competence breaks apart the cumulative analysis of logical behavior as specialized verbal behavior and verbal behavior as specialized social behavior and social behavior as specialized operant behavior. The autoclitics of Skinner’s science of verbal behavior provide a graduated analysis of the verbal behavior that ultimately supports and explains the logical methods written about, e.g., in Quine’s “Methods of Logic 4th edition” (MOL).
All of this is made clearer by properly and uniformly regimenting logical methods relative to, e.g., MOL. While theories of concatenations are popular because they most conspicuously match the analysis of records of marks made by those behaving verbally, it is now widely accepted that the method of binary trees is far more effective at dealing with grammatical analyses of concatenations. And, since concatenations can be simulated in a theory of binary trees as lists of symbols (both lists and symbols are binary trees in this analysis), it is better to dispose of concatenations outright.
What Kleene calls ‘postulates’ shall be here mentioned for what they are: components in the definition of the complex predicate that kleene calls ‘immediate consequence’ and that is more often simply called ‘consequence’. In what theory is this compound predicate of consequence defined? It is a theory with a single three place lexical predicate ‘pairs’ as in “This pairs it with that.” The premises (premises are not axioms, but axioms are frequently governed by premises) of this theory of pairs are assembled as needed.
The most basic premise is the extensional one i.e. the one which says when two order pairs are indistinguishable:
better, the universal closure of 1. The next assures us that whatever there is, there is an ordered pair of such things:
This is sometimes called the law of formation since it says that given this and that we can form an ordered pair from them. Which reminds me that the ‘order’ in ‘order pair’ is got from the following conclusion:
Though I have not said it explicitly, the methods of logic require that the lexicon of the language of the theory be finite (beter “itemized”) and can be limited entirely to many place predicates (so that constants and function letters are nothing more than abbreviations for appropriate predicates as in Russell’s method of contextual elimination of singular description). Consequently, the two place predicate of indistinguishability is taken as relative to the lexicon unless otherwise noted (note that methods of inner models are an example of where ‘indistinguishability’ is not to be taken only and solely as relative to the lexicon but rather either with respect to the lexicon or with respect to the compound predicates of the language of the inner model). I’ve written extensively on identity and indistinguishability in other notes and shall not say more about it here and now.
The premises of extensionality and formation link up cooperatively with the premise of the empty tree:
Which says that there is an empty tree, i.e. one that does not have parts. I have not given exact definitions of ‘part of’, ‘left part of’, and ‘right part of’, but shall do so now:
‘is (a) part of’ is short for ‘x and y such that some item is (z such that y pairs x with z or y pairs z with x)’
‘is (the) left part of’ is short for ‘x and y such that some item is (z such that y pairs x with z)’
‘is (the) right part of’ is short for ‘x and y such that some item is (z such that y pairs z with x)’
The importance of the predicate ‘is part of’ can not be under emphasized by me. The later chapters of Russell’s “An Inquiry into Meaning and Truth” go much deeper into the the relevance of such a predicate in any effective theory of logic.
The premise of extensionality and the premise of empty tree give that any two items purporting to be an empty tree are indistinguishable:
where ‘is empty’ is short for ‘x such that each item is (y and z such that not x pairs y with z)’.
Thus, the description ‘the empty tree’ is singular and there exists an item named by it.
Remaining premises are usually introduced under a schema of induction.
Some neat notes on the model theory of arithmetic by Tin Look Wong from 2014/2015 at the Kurt Gödel Research Center
In the third lecture Wong mentions that arithmetic and finite set theory are biinterpretable (the same as Quine in a theory of concatenation and arithmetic).
This is all relevant because the weak konigs lemma over the base theory of RCA (recursive comprehension axiom) is equivalent to the arithmetized completeness theorem (the standardized version of Hilbert and Bernay’s completeness). This is discussed in the fourth lecture:
I need to write down all the sources I’ve been stumbling upon while trying to figure out the details of a simplified and accessible version of the arithmetic completeness theorem.
Metamathematics of first order arithmetic https://projecteuclid.org/ebooks/perspectives-in-logic/metamathematics-of-first-order-arithmetic/toc/pl/1235421926
These are notes on chapter 14 of Kleene’s “Introduction to Metamatheamtics” (IM). Specifically, they are on section 72 “Godel’s completeness theorem” and are aimed at uncovering all the moving parts involved in the proof of Hilbert and Bernay’s completeness.
With the help of Quine’s “Interpretations of Sets of Conditions” (ISC) I have been able to read Kleene with a new kind of clarity. It has revealed that Quine’s proof of completeness does not reach the result now known as “arithmetized completeness”.
A direct and simple proof of Arithemtized Completeness has not yet appeared in the literature that I have access to (and I must mention this because there is only so much that I can get at from my self imposed isolation). Quine’s proof of the law of infinite conjunction in “Methods of Logic 4th ed.” (MOL) carries out the proof from his ISC. In MOL he emphasizes that the method of proof is “not constructive”. He does not do the same in ISC.
The phrase “constructivity” is like “formalism” in that it does not yet have a canonical theory with a canonical predicate logic. I support something like Fefferman’s “Finitary Inductively Presented Logics” (FIPL). In particular, I support the introduction of a theory of formalism that is indistinguishable from an appropriate theory of arithmetic.
This is something like Kleene’s “metamathematics as a generalized arithmetic” (the title of section 50 of chapter 10 of IM). Hilbert and Bernay’s enhanced Godel’s numbering into a general method of numeric interpretation. Kleene’s “generalized arithmetic” abstracts from strictly numeric methods. It is what we might now call a multiple successor arithmetic with multiple zeros. In that way it is as Kleene says “Herme’s 1938 considers an arithmetic with the empty expression as the zero, and the operations of suffixing one of the formal symbols as the successor operations” (IM pg. 246).
Goodstein mentions multiple successor arithmetics in his “Introduction to Mathematical Logic” (IML) in his chapter on primitive recursive arithmetic. Whether we look to Hereme, Kleene, Goodstein, or more directly to Skolem, it is through Rózsa Péter’s “Rekursive Funktionen” that the life blood of Hilbert and, mostly, Bernay’s “Grundlagen der Mathematik” (GM) and Hilbert and Ackermann’s “Grundzüge der theoretischen Logik”, Herbrand and Godel, and, dare I say, Frege’s “Begriffsschrift” made its way into our world full of programmable machines. The role of Russell (and Whitehead) in all of this is the reminder that whatever is done in logic through mathematics is subject to the theory of mathematics we do it in and, most importantly, its logic.
While MOL makes it methodologically clear that predicate logic is the logic (in that cooperation and coordination of action is unavoidably intertwined with the methods of logic), he does a much better job of dealing with the general problem of logic in “Philosophy of Logic 2nd ed.” (POL). For those who know of Quine only as that snob who said all was set theory and the rest of the world of mathematics be damned, I strongly recommend they read the later parts of POL and the later part of the introductory chapter to MOL. Quine was stubburn, but where he disagreed with Carnap he was onto something fundamental. That fundamental thing was in front of him and he sometimes noticed it, but never quite got it. It is the role of the science of verbal behavior in logic and science (or, concretely, theory and experiment, or, even more concretely, libraries and laboratories).
Kleene’s work is governed by the swirl of history that surrounded him, and he worked hard to slather himself in honey and roll around the floor of the jungle that is mathematical logic.
Kleene’s generalized arithmetic is essentially that of Feferman’s FIPL. The difference is that Kleene didn’t have the history of conspicuous consequences uncovered by LISP programmers to lean upon. All he had were Rózsa Péter’s results (he had Hilbert and Bernay’s too, and Godel and Herbrand’s). The difference between them is largely the inclusion of what is substantially Godel-Bernay’s set theory. The introduction of set theoretic simulations in FIPL is seen by me as a misstep. Well founded trees are made of sturdier stuff.
My ultimate reasons for diving deep into the thickets of these often complex theorems is to unearth the origins of their complexity and to eliminate it as much as possible. I see the complexity as a pain which is alerting the world to a learning opportunity. What I submit we shall learn is that our canonical principles of arithmetic are to be conditioned on completeness and not on our first touch with, e.g., counting. I submit that counting is obviously included in such an arithmetic, and point to the strengths of a programming language like LISP as a concrete example. The strength of LISP comes from its arithmetic of binary trees. We have already seen how much stronger is a world where bit strings prevail. I would like to see a world where binary trees supercede bit strings. It is no secret that I submit that the arithmetic which comes from simplifying completeness is that like the primitive recursive arithmetic of binary trees.
The theorems, lemmas, and corollaries of Kleene’s IMT are marked by notation which distinguish whether the result holds in an intuitionistic logic or “classically”:
“In terms of the derived rules of Theorem 2, this means replacing negation elimination by weak negation elimination. When we wish to consider this system also, we call the original system with Postulate 8 the classical system, and teh system with Postulate 8[superscript I] instead the (corresponding) intuitionistic system. Our results are marked with teh symbol “[supersctipt circle]” in every case when the demonstration we give is not valid for both systems, but only for the classical (and no demonstration which the reader is expected to discover for himself is available for the intuitionistic system).” IMT 101
“For later reference, we summarize our present conclusions in a theorem and corollary, which we mark with the letter “[superscript C]” to show that they do not belong to our sequence of metamathematical theorems, but are only established here by use of non-finitary classical methods. Although we are using ‘set-theoretic’ for the present kind fo predicate logic, some of teh results are only on the level of classical number-theory;” IMT 176
The distinctions made are critical. In light of Quine’s later work, it is important to make these distinctions more sharply and simply. When we speak of predicate logic we do so in the same way that the radical behaviorist speaks of verbal behavior. The practices of predicate logic are special kinds of verbal practices. When we speak of “intuitionistic (theoretic) predicate logic” or “number theoretic predicate logic” or “set theoretic predicate logic” we are speaking through the the science of verbal behavior of the science of verbal behavior.
It is a mistake to speak of ‘meta’ as ‘self-referential’. That is the mistake of the modern world. The original Greek is “beyond” or “after”. Aristotle spoke of ‘metaphysics’ (if he ever spoke of it as such at all) as in “after physics” i.e. after introducing physics this is the next thing.
Said another way, rather than think of ‘speaking thru the science of verbal behavior of the science of verbal behavior’ as ‘speaking of the science of verbal self-referentially’ (whatever that may come out to for you) think of it as “speaking of an even more narrow kind of verbal behavior that is even more narrow than scientific verbal behavior and is entirely limited to only that science which has as its subject verbal behavior”. The way in which we come to the science of verbal behavior is not the same as the way we come to the science of the science of verbal behavior. The latter has a history which the former entirely lacks.
The key, in any case, is “beyond” or “after” and NOT “above” or “within”. I must admit that the English language is so heavy laden with phrases that make mention of ‘feedback’ and ‘self-reference’ that I can not often avoid making metaphorical mistakes myself.
I have found it helpful to point up the methods by which we are taught predicate logic as a way of getting at the difference between speaking under the control of predicate logic and speaking otherwise.
We can speak of a theory without speaking of its predicate logic, and this is most often what we do on our initial approach to predicate logic (because we are unable to speak other than what our history has already selected and predicate logic is not an overt part of that history in this contemplation). What Kleene calls “metamathematics” is a theory of predicate logic which is sometimes supplied with its predicate logic and sometimes not. When specifying “number theoretic predicate logic” we know that the predicates of this theory of logic include “adds” and “multiplies” as in “This adds it to that” or “It multiplies this with that.” When specifying “set theoretic predicate logic” we know that the predicates of this theory includes “classes” as in “This classes that.”, or, more familiarly but more obliquely, “in” or “belongs to” or “is a member of” as in “This is in that.”, “This belongs to that.”, and “It is a member of this.”
I hope these things are made clearer to me from this work.
The theorem of Hilbert and Bernay’s is given on page 394:
“Theorem 36[superscript circle]. The addition to the postulate list for the predicate calculus of an unprovable predicate letter formula G for use as an axiom schema would cause the number-theoretic system as based on the predicate calculus and Postulate Group B( section 19) to become omega-inconsistent (section 42). (In fact, a certain formula would become refutable which expresses a true proposition of the form (y)D(y) where D(y) is an effectively decidable predicate.) (Hilbert-Bernay’s Completeness theorem, 1939).” IMT 395.
Omega-inconsistency is something which Quine was suspicious of e.g. in “Omega-Inconsistency and the Infinite”. The inspiration in that paper was due to a lack of proof that the empty set was not a member of the Frege numerals in Quine’s Natural Foundation. It was later proved, but the comments on the terminology of “omega-consistency” and “omega-inconsistency” remain relevant. The term Quine substitutes for “omega-inconsistency” is “numerical insegregativity”:
“The property of numerical insegregativity may be restated intuitively thus: every sentence phi which is demonstrably true of all natural numbers (i.e. numerically general) is also demonstrably true of something else (i.e. ‘some item is (x such that x is phi and x is not psi)’ is demonstrable for some numerically general psi). In such a system there is no proper translation of ‘is a natural number’, but only an infinite series of better and better approximations.” Quine’s “Selected Logic Papers Enlarged Edition” pg. 118-119
The term Quine picked is terrible. Even reading it on the page is a chore. It’s also unclear if there is anything to his claim about the predicate ‘is a natural number’. His invocation of it is an echo of Russell’s “Introduction to Mathematical Philosophy”. There is at least some tiny thing to his talk: ‘non-standard model’ is not to be accepted as a kind of final judgement.
Back to articulating Kleene’s statement of Hilbert-Bernay’s completeness with greater self-sufficiency. First, what is “the number-theoretic system as based on the predicate calculus and POstulate group B (section 19)”?
Section 19 is title “Transformation Rules”. These are common to anyone familiar with calculi and so called “deductive” systems. Their origin is in Hilbert’s work on Frege’s logic. While Kleene goes out of his way to make some distinction between the methods of logic and the particular methods presented in this chapter titled “A Formal System”, there remains great confusion among modern matheamtical logicians (and especially those who call themselves “computer scientists” or “computational theorists”) between math and logic.
The items contemplated, the ones subject to transformation rules, are concatenations. They are better taken as in Quine’s later editions of “Mathematical Logic” or in his paper “Concatenation as a Basis for Arithmetic”. What is here called ‘a formal system’ is a theory of what Quine called ‘protosyntax’, and what we can now simply call “a theory of concatenation”.
The lexicon of a theory of a concatenation puports to have a single predicate ‘concatenates’ as in “This concatonates it with that.” As Quine shows most clearly in “Concatenation as a Basis for Arithmetic” predicates which are true of finitely many concatenations can be quoted by a single concatenation (what is familiar to programmers as the method of “comma-seperated values” or “CSVs” and is a well worn plain-text way of dealing with tables of values) in a theory of concatenation with at least two non-empty items (a concatenation is empty if it concatenates itself with itself). Such theories of concatenation are rightfullly called bit string theories after the zeros and ones of an information theoretic outlook on computation with registers as concatenations of fixed length.
The transformation rules of such a formal system are then predicates which compose what some modern mathematical logicians call “the consequence relation”. It is not a relation, it is, most often, a complex predicate of a theory of concatenation i.e. it is a truth functional and quantificational combination of the three place predicate ‘concatenates’. The consequence predicate is usually written using notation descendent from Frege’s turnstile and Russell’s adaptation of the same. For this reason it is to be deplored. Both Russell and Frege were unable to adequatly distinguish between use and mention in their commentary. Both Russell and Frege are redeemed for this mistake in that dropping their commentary (Frege’s in Begriftshrift and Russell’s in Principia) leaves a presentation which respects the methods of semantic ascent (Quine’s term for the generalization of the methods of quotation and disquotation i.e. semantic ascent and descent).
But, there is nothing we can do about it now except to always take time to warn future readers of the pitfalls of the turnstile. The double turnstile is then introduced as a kind of semantic parallel to the syntactic commitments of consequence. That there is even a consequential distinction between the term ‘syntax’ and ‘semantics’ is brought into question by the very theorem under investigation here. Hilbert and Bernay’s completeness purports to dismantle “constructively” any such distinction: they are interconvertable and logically indistinguishable methodologies. Walter Dean’s 2017 “Bernays and the Completeness Theorem” goes into this more deeply. MOL takes it to heart and is presented with respect to such a disollution. POL goes into much greater detail on the most general consequences.
So to get to it.
“Postulates for the formal system.
Dramatis Personae. For Postulates 1-8, A, B, and C are formulas. For Postulates 9-13, x is a variable, A(x) is a formula, C is a formula which does not contain x free, and t is a term which is free for x in A(x).
Group A. Postulates for the predicate calculus.
Group A1. Postulates for the propositional calculus.
1a) A only if (B only if A)
1b) (A only if B) only if ((A only if (B only if C)) only if (A only if C))
2) B is an immediate consequence of A with (A only if B)
3) A only if (B only if (A and B))
4a) (A and B) only if A
4b) (A and B) only if B
5a) A only if (A or B)
5b) B only if (A or B)
6) (A only if C) only if ((B only if C) only if ((A or B) only if C))
7) (A only if B) only if ((A only if not B) only if not A)
8[superscript circle]) (not not A) only if A
Group A2. (Additional) Postulates for the predicate calculus.
9) (C only if each x is such that A(x)) is an immediate consequence of (C only if A(x))
10) (each x is such that A(x)) only if A(t)
11) A(t) only if some x is such that A(x)
12) ((some x is such that A(x)) only if C) is an immediate consequence of (A(x) only if C)
Group B. (Additional) Postulates for number theory.
13) A(0) and ((each x is such that (A(x) only if A(x’))) only if A(x))
14) a’=b’ only if a=b
15) not a’=0
16) a=b only if (a=c only if b=c)
17) a=b only if a’=b’
18) a+0=a
19) a+b’=(a+b)’
20) a.0=0
21) a.b’=(a.b)+a
I like to see how it’s all going to end. More than that, I want to see how it gets there. What happens between now and later that makes all the pieces fit together? What makes it click?
I start taking notes and end up reading and reading and reading. Maybe I’ll come back later and take some more notes. The moment I do, the moment I flip to that page, I make a few remarks and then I’m off to reading and reading and reading.
Reading for me is a lot like thinking through soemone else. When you read math, you’re in touch with the same things the thinker touched. Its right there on the page. You’re on the adventure with them even though they may be half a word away or half a millenia dead. It is the best of fiction and the best of matters of fact. The laboratory of the mind is before you.
Religions are no better than governments or economic systems. We the hints we get from science more than anything. Logic and science are both the most well known access points to truth. Revelation, law, and utility are mists. They becloud the interconnected network of theory and experiment that leverage and release human potential.
Leaders and followers are both locked into a culture of control which looms beyond them both. That we can take control of the controlling environment is the closest thing to a blessing bestowed upon us. We may thank natural selection though it is indifferent to our praise in precisely those ways which made us us.
Upon flipping back and forth between Quine’s “Interpretation of Sets of Conditions” and the chapter on the completeness of his main method from his “Methods of Logic 4th Edition”, and the later chapter of Kleene’s “Introduction to Metamathematics” which contains an outline of the proof of Hilbert and Bernay’s completeness, I am no longer convinced that Quine’s presentation establishes anything like Hilbert and Bernay’s completeness as I had hoped.
R. L. Goodstein’s Theory of Constructive Formalism has not yet been given a proper logical analysis. Although one of his students wrote their thesis on the metamathematical properties of multiple sucessor arithmetics as they were called then, there has not been much of a modern approach to it all.
Goodstein’s Constructive Formalism is presented as what he calls an equation calculus. It is a calculus in that if it were formlized without further reflection it would be, as with any other calculus, a subtheory of Quine’s protosyntax. That is, appropriate predicates in a theory of concatenation with at least two distinct atoms is sufficient and necessary for dealing with the “constructions”, i.e. the equations, that he presents.
For a long time I was stuck with his methods as a calculus and saw the methods of logic as something derivative from that. This is the outlook promoted by modern type theorists. They are unable to see the difference between logic and mathematical logic, and thus can substitute mathematical logic for logic wherever logic itself doesn’t fit their intuition. This is a dangerous road to walk down. It misled me far longer than I would like to admit.
First let me present Goodstein’s methods as he gave them. The entirety of his program is layed out in 1951 “Constructive Formalism: Essays on the Foundation of Mathematics”. As a publication it is a collection of his conclusions from earlier papers. I am not in the mood to search each of his papers out and list them here. It should be easy to do so with today’s technology, but it isn’t. Like every other bit of science that was written down and scanned into the computers of the internet, there are an endless number of hoops, traps, and detours that one must take before reaching the most basic data. I hope future generations look up on these times as the dark ages because of how foolishly destructive are our self imposed social practices.
Though he revised the exact ‘inference rules’ which generate the equations of his calculus over time, they are substantially what he presented in 1957 “Recursive Number Theory”. Chapter 5 of that book is titled “Formalisms of Primitive Recursive Arithmetic” and there the materials of his calculus are assembled.
Before quoting the set up, some context may help to explain the significance of Goodstein’s work. Goodstein was Wittgenstein’s student. He was greatly influence by Wittgenstein’s philosophy and his later philosophy in particular. The later philosophy had more to do with verbal behavior than the earlier work of the Tractatus. Wittgenstein’s earlier works are to linguistics as his later works are to the science of verbal behavior.
This confusion, between between language as the reinforcing practices of a verbal community and language as syntax and semantics, is largely responsible for the confusion between logic and mathematical logic. The confusion can be made overt by examining the arguments that are meant to support the mathematical logician. If they are not logical then they lack the appropriate properties of extensionality to support effective coordination and collaboration (that is, by their own standards).
Goodstein’s work follows immediately from Skolem’s and very likely would have been strengthend by Herbrand’s work had he not died from that ski accident. Without further ado, Goodstein’s introduction to his equation calculus on page 104 of “Recursive Number Theory”:
“In this chapter we consider a codification of primitive recursive arithmetic in which the only axioms are explicit and recursive function definitions, and the only inference rules are the substitution schemata
F(x)=G(x)
--------- Sb1
F(A)=G(A)
A=B
--------- Sb2
F(A)=F(B)
A=B
A=C
--- T
B=C
where F(x), G(x) are recursive functions and A, B, C are recursive terms, and the primitive recursive uniqueness rule
F(Sx)=H(x,F(x))
--------------- U
F(x)=H^x(F(0))
where the iterative function H^x(t) is defined by the primitive recursion H^0(t)=t, H^(Sx)(t)=H(x, H^x(t)); in the schema U, F may contain additional parameters but H is a function of not more than two variables. In Sb1, the function G(x) may be replaced by a term G independent of x, provided that G(A) is also replaced by G.” Goodstein, Recursive Number Theory pg. 104
All of this appears to be largely independent of the methods of logic e.g. without mentioning truth functional connectives like ‘and’ or ‘not’ or the quantifier ‘something is x such that’ or a lexicon of basic predicates like ‘sums’ as in ‘It sums this with that’, it appears that Goodstein has set up something mathematical. In fact, he goes on to define the truth functional connectives within his calculus, and to even prove the deduction theorem (which states that if one schema can be proved from another with the axioms then their conditional can be proved from the axioms alone).
This is all a mistake of use versus mention e.g. the grammatical atoms ‘and’ and ‘not’ and the grammatical compounds constructed from them are not identical to an item like a truth function or its values (which when applied to tuples of truth values are identical to a unique truth value). Whatever notational expedience there is to calculating with tables of truth values (or, better, truth trees) is only as effective as it ultimately is shown to correspond to a method of logic e.g. in a consistent theory of truth predicates which also contains truth functions, following Russell and Tarski (and Quine), it is expected that the truth function said to correspond to the grammatical construction of conjunction assigns the value of truth to the truth values of components of a true conjunction. Without an argument that such is the case, there is no effective exchange between arguments in a theory of truth functions and arguments in a theory of truth predicates.
Note, it is possible to construct “truth sets” which do the work of truth values, truth value assignments, and truth functions (and more!) which are made out of items from grammatical categories, but there you’ll find a similar problem linking up use with mention. Now, there is a significant method which confronts this problem head on and it is outlined in Quine’s paper “Immanence and Validity” where he contrasts the transcendent metatheoretic methods with immanent metatheoretic methods. While it is not entirely inappropriate to enter upon a further exposition on these differing methods, it is also not something I’m prepared to undertake at this moment.
By way of concrete example, let me take the formalism as Goodstein has presented it and make overt its covert logic. Quine is the most vicious when it comes to repremanding those who confuse “is inferred by rules from’ with ‘is implied by’ e.g. in the somewhat cryptic chapter 13 of “Methods of Logic 4th edition” titled “Axioms”. To quote
“Lukasiewicz swore by axiomatic truth-function logic as a training ground for axiomatic method in more demanding domains. I swear rather by the sufficinecy, unto the day, of the evil thereof. Axiomatic logic, with its schemata and its specific rules of inference, is very unlike postfoundational axoim systems. Trainin in the latter , if wanted, is best sought as such. Its main component is training in the recognition or proof o implication, since implication is what relates postfoundational axioms to theorems; and such training is a pervading purpose of this book. There will be some specific notice of postfoundational axioms in Chapter 42 and 48. Meaningwhile I have included this brief account of axiomatic truth-function logic partly for contrast an dpartly because of the prominence of the subject in earlier literature.” (pg. 88-89)
where
“Axiomatic logic is one thing; the application of full-blown logic to extralogical axioms is another. The contrast between them is the contrast between what Sheffer called foundational and postfoundational systems. Foundational systems have their autonomous rules of inference; postfoundational systems just infer their theorems from their axioms by logical implication, deferring to logic for the analysis and the technology of that relation.” (pg. 87).
It is precisely on this point that I was lost and confused when I first confronted Goodstein’s constructive formalism without having yet read Quine’s “Methods of Logic 4th edition”. Now I expressly speak of “premises” as sorta short for “postfoundational axioms”. It is not appropriate, in my eyes, to speak of “postfoundational axioms”. Let axiomatists keep their word. A word which Quine uses repeatedly, and which suggets the nature of axiomatic rules of inference, is “generation”. It is a bit wider than “construction” and emphasizes the difference between “generating validities” establishing by logical methods if a schema (better, a sentence) is valid.
So, to get concrete. What is a rule of inference? I can speak of a rule of inference within a theory of concatenation (such as Quine’s theory of protosyntax) or, more specifically, a theory of bit strings. Alternatively, theories of ordered pairing or specifically theories of binary trees are where rules of inference are revealed to be special predicates which, when added to a consistent theory of truth of the theory said to be axiomatized, interlocks instructively with the reachable predicates of implication.
Theories of axiomatization are ways of setting up the logic of formalism i.e. saying just exactly what it is that we’re dealing with when we speak of formalisms. As long as the formalist avoids speaking in a predicate logic, there is little that can be said for or against what they claim except for the overt issues of extensionality e.g. we have no way of agreeing when two things are identical, we may substitute a true sentence for a true sentence in a compound and the latter may turn true when the former was false, or coextensive predicates may not denote each of the same items! It can be difficult to teach those unfamiliar with the methods of logic that extensionality is helpful and that it is lost when predicate logic is avoided or dropped. For those mathematical logicians who should know better, it should be enough to point to Lowenheim’s theorems, but this does not often have the effect expected for the very reasons that are given by it!
Don’t forget the Boolean Prime Ideal theorem: it appears to be the most effective access point to compactness results. https://en.wikipedia.org/wiki/Boolean_prime_ideal_theorem
Added hint on how consistency, implication, and equivalence work. All it does is to give their definitions from a restatement of the definition of validity as each supplemented lexical substitution denoting everywhere. This was done in response to writing my notes on Quine’s 1954 “Interpretation of Sets of Conditions” because there is, in general, a great deal of confusion over terminology throughout logic and mathematical logic. The problem is twofold: 1) people do not know about the grammatical definition of validity, 2) people do not know the consequences of Hilbert and Bernays completeness theorem.
In particular, modern texts in mathematical logic are obsessed with proofs and models as if they are the only methods by which logic proceeds. The assumption of a surrounding set theory is also so prevelant that it precludes itself almost overtly, but in the end it sadly does not. While some are able to deftly distinguish between internal and external concepts (or whatever they are calling them today), this is not the same as a distinction between logic and mathematical logic which is where the error has occurred in their methods.
These continue notes on Quine’s 1954 “Interpretations of Sets of Conditions” which shows how to construct an interpretation of a consistent set of quantificational schema in the theory of natural numbers as in Hilbert and Bernay’s Grundlagen and later as in Kleene’s Metamathematics and elsewhere.
The reason this is significant to me is because consistency is nonvalidity of negation and while Lowenheim and Skolem showed that there is some arithmetic interpretation of a consistent schema this does not imply or even suggest that such an interpretation can be constructed!
To say that an arithmetic interpretation can be constructed is to say that for each predicate letter in a consistent schema there is a predicate in the theory of natural numbers (whose canonical lexicon contains just “sums” and “multiplies” as in “It sums this with that.” and “It multiplies this with that.”) which when each substituted for their corresponding predicate letter yields a compound predicate that denotes everywhere.
I must emphasize that I say ‘denotes everywhere’ rather than “denotes each appropriate tuple of natural numbers” because it is a logical mistake to suppose there are items called ‘natural numbers’ above and beyond what is denoted by a given theory of arithmetic.
This warning must be made overtly because its neglect has wasted the work of many lives. Whatever contempt some may have of Quine’s simple definition “To be is to be denoted by a one place predicate” or his much more well known “to be is to be the value of a variable” (variables are where the problem started and their elimination eliminates all these problems outright), there is nothing to be gained from flatly rejecting his advice without providing a full supplement for what is subsequently lost from its denial. Much more is lost than most think, and I advise those who would deny it to act with vigor.
Technically, Hilbert and Bernays constructed an arithmetic interpretation of a consistent schema and yet Lowenheim and Skolem say more: there is an arithmetic interpretation for any set of consistent schema. To be even more technical, Lowenheim showed that a consistent schema has an arithmetic interpretation and Skolem showed that any set of such had such. Note, Lowenheim implies that finite sets of consistent schema (what should more properly be called perhaps “itemized” sets) have an arithmetic interpretation since a conjunction is consistent if and only if each of its components are.
Now the result Quine explains is a later generalization of Hilbert and Bernay’s construction of an arithmetical interpretation of a consistent schema to the construction of an arithmetical interpretation of a set of consistent schema by Kleene (just as Skolem generalized Lowenheim from a consistent schema to a set of the same). It is not entirely like Kleene’s (which is presented in his 1952 “Introduction to Metamathematics” i.e. Quine’s paper was published two years later) in that it includes some of the methods of Godel’s completeness theorem.
More on Godel’s completeness theorem can be found here https://en.wikipedia.org/wiki/Original_proof_of_G%C3%B6del's_completeness_theorem. The relevant events are these:
Now a note on compactness and completeness. Under Quine’s main method, compactness and completeness are made as close as their equivalence appears to allow (thus far). In the 4th edition of “Methods of Logic” Quine presents compactness as ‘the law of infinite conjunction’: “an infinite class of truth functional schemata is consistent if each of its finite subclasses is.” In other words, if for each finite subset of schemata there is an interpretation which makes each of them true, then there is an interpretation which makes each of the schemata in their union true.
Since any collection of schemata is at most countably infinite in a regular language (one with a finite lexicon) it is a bit over exuberant to say “infinite class”, but, in fact it is not. Quine’s proof of the law of infinite conjunction makes clear how the law is indifferent to langauges with a potentially infinite language. Whereas Quine’s proof assumes an at most countably infinite lot of sentence letters, any language with a well ordered set of predications (that is concatenation of a predicate letter with a fixed number of variables, here we may say tuples of a well ordered set of variables and use the methods of a free logic like Quine advocates for, or contemplate only those predications of a predicate letter with tuples of constants the set of which is well ordered as well) admits of a similar construction as Quine’s by transfinite induction.
What does all of this have to do with Hilbert and Bernay’s completeness? All of these problems are sidestepped in that one way or another they end up working in a way equivalent to mechanical substitution (as much as mechanical substitution fits in their theory of arithmetic, which if it doesn’t then there are bigger problems to solve). The reasoning is a bit remote because it requires that you trace your work through the theory whose predicate logic gives you definitions of terms like ‘finite’, ‘countably infinite’, ‘uncountably infinite’, ‘transfinite induction’ etc. This is something that is rarely if ever done except perhaps by Hilbert and Bernay’s. I am told that this is actually done often in automated proof systems, but I have yet to find an explanation of such that I can grasp.
Quine’s exposition ends with a proof of Hilbert Bernay’s completeness as a special case of the construction of an arithmetic interpretation for any countably infinite class of consistent schemata. In this form it happens to be a proof of completeness of Quine’s main method that is “constructive”.
The principle that
primes these preliminary steps of Quine’s proof:
if the set of schemata is finite then it is extended to a countably infinite set by including redundantly equivalent reletterings of their quantifiers
each schemata is interchanged with its equivalent universal closure
each schemata is interchanged with its equivlaent prenex form (where all quantifiers are out infront of a truth function of its component predications)
each schemata is interchanged by its equivalent alphabetized relettering so that no two quantifiers of any such schemata are identical.
The last step imposes an overt order upon the schemata: two appropriately relettered schema are such that the letters of the quantifiers of the one are all alphabetically prior to those of the other. If the first schema does not begin with a universal quantifier then interchange it with an equivalent one that does and reletter each of the subsequent schema appropriately i.e. so that the newly introduced schema remains the first even after the relettering (e.g. prepend the universal quantifier whose variable is a letter alphabetically after each of the letters of the other quantifiers in it and reletter the second, third, fourth, .. schema accordingly).
Supplemental definitions:
The prime power product of a sequence of n+1 numbers is the product of the n-th prime to the power of the n-th item of the sequence with the prime power product of the first n numbers of the sequence. The prime power product of the empty sequence is one.
Each existential of a numeric substitution sequence (NSS) of a concatenation of quantifiers is assigned the prime power product of the count of its prior existentials concatenated with the sequence of values assigned to its prior universals. The values of universals index NSSs lexicographically.
The n-th numeral substitution of the first m schema is the substitution, in the conjunction of their nonquantificational parts, of its variables for the corresponding numerals of the n-th numeric substitution sequence of the concatenation of its quantificational parts. The n-th numeral conjunction of the same is the conjunction of their first n numeral substitutions.
A numeral predication is the concatenation of an n place lexical predicates with n numerals. An ordering of the lexical predicates together with the lexicographic ordering of concatenations of numerals orders the numeral predications.
Each numeral conjunction is a truth function of numeral predications.
An assignment of truth values to the first n numeral predications is dirty if some numeral conjunction reduces to falsity under it, otherwise it’s clean.
The n-th numeral predication is assigned truth if and only if there is an earliest clean assignment that assigns truth to it.
Numeral predications are assigned truth under the numeral interpretation if and only if there is an earliest clean assignment that assigns truth to them.
Each schema in a consistent set is assigned truth under the numeral interpretation.
It is definitely well past the time when I should have written about the fundamental contingencies of mathematics. These are the things that you have to worry about if you’re trying to make an effective program of instruction in mathematical behavior. This is not a program of instruction though. This is just a description of the contingencies. So what are contingencies and what is a program of instruction (because the way that I’ve used them seems to have a formal or special meaning, because they do!).
A contingencie is a consequence from a response on an occasion. The simplest and most classic example is a door opening from a push on a button or, to get even more ancient and classical, food from a tap on a disc. More complex examples are 1) an electron from a photon on an atom 2) an increased heart rate from a kiss on a cheek 3) metabolism from sunlight on a plant
and so on. These are way more complex than they might seem. Contingencies are everywhere and they happen on all sorts of time scales e.g. in nanoseconds or eons! The contingencies of mathematics that are relevant to an effective program of instruction are just those that shape up mathematical behavior (and all the thinking and feeling that go into it).
I don’t want to spend too much time on the abstract principles of contingency analysis, but I do want to make it clear that a program of instruction is just a collection of contingencies of reinforcement. The things that we do when we do math are responses the fit into the contingencies of mathematics.
Almost all of the fundamental contingencies of mathematics can be organized from the contingencies of counting. Broadly, the responses in the contingencies of mathematics are constructed verbal behavior. Counting is the canonical example of a repertoire of constructing effective verbal behavior under the control of the contingencies of mathematics.
It is difficult, if not impossible, to explain the contingencies of mathematics without assembling the definitions of the contingencies of verbal behavior generally. Mathematical behavior is often, if not exclusively, a very special kind of verbal behavior on par, if not inseperable from, scientific and logical behavior.
Verbal behavior is a special kind of social behavior. Social behavior is a special kind of operant behavior and operant behavior is behavior whose response rates are changed by their consequences. There are more detailed and technical definitions for each of these terms. These technical definitions take into account the contributions of natural, operant, and social selection by consequences. A bit of behaving biology is the result of natural selection. Operant and social selection are special consequences of natural selection. The contingencies of survival do the selecting on biological scales and the contingencies of reinforcement do the selecting on behavioral scales. There are not so many technical terms for the social scale because social behavior and the practices they compose have hardly undergone an experimental analysis.
Social behavior, like any other behavior, can be analyzed with respect to the ontogenic and phylogenic contingencies that selected it. When we distinguish between operant and respondent behavior where respondant behavior is such that its responses almost always follow their characteristic occassion (e.g. a kick of the leg from a tap on the knee) we are making a distinction between the contributions of operaent and phylogenic contingencies. Pavlov showed that the stimulus characteristic of a response in respondent behavior can be pair with another which was not present during the evolution of the bit of biology behaving respondently. A full description of the respondent behavior of an organism thus includes the history of consequences during the life time of the organism and the life time of the species.
Since there is little within the vernacular to rely upon when making the distinctions necessary to deal well with behavior, there is little to be gained from leaning on metaphors. This is especially the case when solving those problems which seem to be most pressing at the moment. There are two options available to the expositor: be sufficiently vague, that is sufficiently imprecise, so that there is no mistake made in the descriptions provided, or assemble a carefully constructed and interlocking system of technical terms for making the exact distinctions uncovered by an experimental analysis.
My methods have always tended to emphasize the technical terms because these are the sharp lines which sketch out the boundary between our best science and our worst pseudoscience. While I am now allowing myself to be woefully imprecise, I am still ultimately only able to do so after constructing the tight technical terms that interlock into a network of logical strength.
While I have collected such technical terms elsewhere, I have not collected them in these notes, and that is a sadness that must be corrected before I can deal well with a public presentation of the contingencies of mathematics.
As with almost every other note there, it will take some time to ultimately accumulate what is necessary to make a short and punchy memo.
I still have yet to write a single memo. This is not because they are not forthcoming, it is rather because they set a high bar that is worthy of their status. When people read a memo they only want to know what they need to know to deal with the problems which presently present themselves. There is no time nor no interest in anything more elaborate or discursive. These notes are the last thing that anyone looking for a memo wants to read.
I continue the work I do here because I have a history of finally assembling exactly what is needed in order to deal with the problems that consume me. These are sometimes problems that other people have solved, but they are now almost exclusively problems for which there are not yet any effective solutions (though many have made clean claims otherwise).
As a last comment, these notes are the hack and slash of an adventure. This is the jungle. This is the blizzard. This is the hurricane of trouble that one gets into when facing the world and its pains.
It occurred to me that I can solve the ‘rewind’ problem with complex gray words and include another key effect of slotted instructions utilized by Moore in the design of the GA144 and its individual chiplet computers.
push(); // more?
if(isTopNil()){ // no more, rewind
nil(); // prime the pump
while(isTopNil()){ drop();
push(); part(); swap(); pop(); pair(); push();
if(isTopNil()){ nil(); pair() }
else{ nil(); }}
drop(); pop(); pop();}
But this starts to get a bit heavy handed when inserted into the monolithic implementation of the programming environment. It also may not solve the problem I’m actually having. But something like it seems likely to be part of the solution I’m looking for.
These are notes on Quine’s 1954 “Interpretations of Sets of Conditions”. They are inspired by the recent popularity of Terrance Tao’s 2009 note on completeness and (sequential) compactness: https://terrytao.wordpress.com/2009/04/10/the-completeness-and-compactness-theorems-of-first-order-logic/.
It gives the step from Lowenheim and Skolem’s theorem that each consistent set of quantificational schema has a (true) numerical interpretation, i.e. each set of schema of predicate logic which has a true interpretation has an interpretation of its predicate letters in a theory of natural numbers, to the construction of such an interpretation (which is Hilibert and Bernay’s accomplishment).
“The case of finite S can be accommodated by supposing repetition, say with arbitrary reletter of bound variables, from the last schema onward.”
Simplifying assumptions are made to leverage the following principle: any conjunction of closed and prenex schema with distinct quantifiers is equivalent to the concatenation of their quantificational parts with the conjunction of their nonquantificational parts. (Generality is kept by universal closure equivalence and interchange of equivalents of alphabetization and passage.)
Two definitions are needed:
Assume the first schema begins with a universal quantifier so that appropriate Kleene sequences exist.
This is an administrative entry.
Hilbert-Bernay’s completeness, or its Lowenheim part, is found in volume 2 pages 234-253 of their ‘Grundlagen’. It has yet to be translated into English, and is one of the main reasons that I must learn German. As Quine says in the 4th edition of “Methods of Logic”: “a strengthened version, more simply proved, see Kleene, 1952, pg. 394, theorem 35”. This is Kleene’s famous “Introduction to Metamathematics”.
What Quine neglects to mention on that page is that an even more readable version can be found in his own poorly named paper “Interpretations of Sets of Conditions” from 1954. The title is not inaccurate. It is poorly named because unless you know exactly what he’s talking about, ‘interpretation’ is the crux, it does not give a hint as to what is to be shown. It is perhaps a consequence of merely exposing an old argument to a newer crowd that he gave it this drab title. Whatever his reasons, we still live in a world that has yet to embrace the import of Hilbert and Bernay’s work even a century later.
B. F. Skinner made great use of the history of the cure to scurvy whenever he mentioned the metaphorical distance between discovering and adoption e.g. 200 years from citrus curing scurvy being recorded on captains books to requiring that ships sail with citrus.
I must show how Hilbert-Bernay’s completeness works. At the same time I must also complete two other projects that have appeared here lately (not to mention all the others that are in the works but which I always planned to take a very long time in between the emergence of relevant projects such as these). The first is the notes on Brad Nelson’s colorForth bootstrapped with C and Moore’s uhdForth. The second is the contingencies of mathematics for the design of a program of instruction in math (and logic).
I can simplify the colorForth and uhdForth tasks by first completing the introduction of compound gray words to my programming environment (aka the programmable program) and by reducing all loops to the one loop of the scheduler. This is because the logic of Brad and Chuck’s methods are bound up with the same problems and it is easier to interpret their work when I’ve done more of my own.
Hilbert and Bernay’s completeness has come up because an old post of Terry Tao’s https://terrytao.wordpress.com/2009/04/10/the-completeness-and-compactness-theorems-of-first-order-logic/ stumbled upon what he calls ‘sequential compactness’ and which is substantially Quine’s ‘law of infinite conjucntion’. They are both about the existance of interpretations of schema of predicate logic. My contribution is to be that the shift from Hilbert and Bernay’s completeness in terms of the arithmetic of natural numbers to the arithmetic of ordered pairs may greatly simplify the links between compactness, completeness, and the scope and limits of consistency i.e. validity.
Finally, there is the extension of all of these results to predicate functor logic.
Since I am most interest in Hilbert and Bernays completeness at this moment, I shall begin by taking notes on Quine’s “Interpretation of Sets of Conditions”.
This is hopefully going to be a fun entry. I’ve been trying to see the clearest way of introducing conditional execution to the programmable program. It is available because of the gray word ‘select’ which is a primitive word which drops the top and third from top item from the stack if the top is nil and drops the top two items from the stack otherwise.
Conditional execution remains one of those pesky problems that has yet to be properly resolved. Its connections with partial, general, and primitive recursion are often poorly illustrated when it comes to proper programming languages. Some of the only concrete examples which deal with the realities of modern programming while bringing in some of the mathematics that comes along with it are still those written by the ancients, e.g. Meyer and Richie’s “Complexity of Loop Programs” https://people.csail.mit.edu/meyer/meyer-ritchie.pdf or Uwe Schöning’s “Theoretische Informatik – kurz gefasst (5th edition)” which introduces ‘While Programs’ and ‘Goto Programs’ in Chapter 2 (notes on which can be found here https://ai.dmi.unibas.ch/_files/teaching/fs16/theo/slides/theory-d03.pdf).
It should infuriate modern programmers that less has been done to bring math, and, most importantly, predicate logic to the problems of programming. It is clear enough that the practices of programming are human practices, that programmers have a repertoire of programming selected and maintained by a language as the reinforcing practices of a verbal community, but, for reasons that escape me, it is not equally clear that the practices of math and logic are equally human practices and are just as immediately relevant to the problems of programming in the most practical of realms. The belief is that logic and math are abstract and the practices of programming are concrete. This is naive and shallow.
A few things occurred to me so I will start by listing them and then look at the list to see which one I should write my way through first:
swapping out the space bar for the comma key so that pressing space constructs a compound name that can then be compiled as a compound gray word by pressing comma
make every key press a blue word and thus eliminate colors from the interpreter so that every keypress is run as code e.g. pressing the ‘a’ button sends off a search for the matching red word and, as part of the definition of the interpreter, pushes the tree quotation of an ‘a’ atop the top item of the stack. This would make buttons for ‘coloring’ programmable as well. This is something to seriously consider because it is very like what I imagined programming would be like when I was a five or six year old playing pretend.
the list can be organized so that the item at the top of the list is a list of mands, the items second and third from the top can be the program that is currently being played by the machine, and the space above that is mostly reserved for paused programs.
is there a way to get the keys for coloring things to be at the end of the alphabet without making it so hard to hand compile the interpreter?
Just how important is the op that compiles an op that compiles an op? Introducing it before introducing conditionals is really helpful. Compare and contrast with Brad Nelson’s methods.
What about the internal operations and programming up a forrest of trees?
What about introducing compound gray words, slotted instructions, and introducing things that way?
What about the elimination of the basic command ‘mand’ for ‘set’ and ‘get’ to self?
The maxim of minimum mutilation suggests that I start by discovering how to introduce conditionals with ‘if.’, ‘else.’, and ‘then.’ words. I’ll start here because it leads into defining a word that compiles a word that compiles a word.
Since the programs that I’m now contemplating are in my programming environment, I can write the code as I do when I think about the things I’ve only shown thus far in javascript. I must emphasize that the design of my programming environment is such that there should be little difference between what it ultimately looks like to program in the native language and what it looks like to program in my language. This can not be entirely accurate ever because few languages permit the kind of overt control afforded by mine and those like mine.
Before programming conditionals, I shall start with a simpler word that thus far I have gone ahead and introduced as a gray word i.e. ‘return’. And, before that, let me do a once over the monolithic version of my interpreter.
var Nil = new Object(); // the empty pair
function isNil(p){return Object.is(p, Nil);} // empty pair?
function Pair(l,r){return new Array(l,r);} // allot pair
function Left(p){if(isNil(p)) return Nil; else return p[0];} // the left part of
function Right(p){if(isNil(p)) return Nil; else return p[1];} // the right part of
var T=Nil; // the tree
function isTopNil(){return isNil(Right(Left(T)));}
function dup(){T=Pair(Pair(Left(T), Right(Left(T))), Right(T));}
function drop(){T=Pair(Left(Left(T)), Right(T));}
function pop(){T=Pair(Left(Left(T)), Pair(Right(Left(T)), Right(T)));}
function push(){T=Pair(Pair(Left(T), Left(Right(T))), Right(Right(T)));}
function swap(){T=Pair(Pair(Pair(Left(Left(Left(T))),
Right(Left(T))), Right(Left(Left(T)))), Right(T));}
function nil(){T=Pair(Pair(Left(T), Nil), Right(T));}
function pair(){T=Pair(Pair(Left(Left(Left(T))),
Pair(Right(Left(Left(T))), Right(Left(T)))), Right(T));}
function part(){T=Pair(Pair(Pair(Left(Left(T)),
Left(Right(Left(T)))), Right(Right(Left(T)))), Right(T));}
function select(){if(isTopNil()){drop(); swap(); drop();}else{drop(); drop();}}
onkeydown=e=>{nil();for(let k=0; // prime the pump
e.key!='.;,:0123456789abcdefghijklmnopqrstuvwxyz'[k++]; // key not k-th char of abc?
nil(), pair())if(k==40)return drop(); // yes. no.
dup(); part(); drop(); part(); drop(); part(); drop(); // not letter?
if(isTopNil()){ drop(); // not letter
pair(); dup(); pop(); pair(); // compile letter and word
push(); part(); // blue?
if(isTopNil()){// blue
nil(); pair(); pair(); // greenify
swap(); pop(); swap(); nil(); //prime the pump
while(--limit && isTopNil()){ drop(); // execute
part(); push(); swap(); pair(); pop(); pop(); // fore
part(); // blue?
if(isTopNil()){ drop(); drop(); } // blue
else{ part(); drop(); // not blue, green?
if(isTopNil()){ // green
nil(); pair(); nil(); pair(); nil(); pair(); pair(); // redify
dup(); push(); dup(); push(); dup(); pop(); swap(); pop(); // pull program
nil(); // prime the pump
while(--limit && isTopNil()){ drop(); // find match
part(); swap(); pop(); pair(); dup(); pop(); // back
part(); swap(); drop(); // next word
nil(); dup(); pop(); // id? prime the pump
while(--limit && isTopNil()){ drop();
if(isTopNil()){ drop(); // is other nil?
if(isTopNil()){ drop(); push();// both nil, empty rest?
if(isTopNil()){ dup(); pop(); dup(); pair(); }// empty rest means id
else{ part(); part(); pop(); nil(); }} // rest id?
else{ push(); drop(); dup(); pop(); }} // not both nil means not id
else{ swap(); // not nil, is other?
if(isTopNil()){ nil(); }// nil means not id
else{ part(); push(); pair(); pop(); swap();
part(); push(); pair(); pop(); nil(); }}} // parts id?
drop(); push(); // match found?
if(isTopNil()){ drop(); } // match found
else{ drop(); dup(); push(); push(); // not a match, more?
if(isTopNil()){ drop(); drop(); drop();} // no more
else{ nil(); }}} // more
drop();}
else{ part(); drop(); // not green, gray?
if(isTopNil()){ drop(); part(); swap(); drop(); // gray, get op
part(); drop(); part(); drop(); part(); drop(); part(); drop(); // dup?
if(isTopNil()){ drop(); dup(); } // dup
else{ part(); drop(); // not dup, drop?
if(isTopNil()){ drop(); drop(); } // drop
else{ part(); drop(); // not drop, pop?
if(isTopNil()){ drop(); // pop
push(); swap(); push(); swap(); pop(); pop(); pop();}
else{ part(); drop(); // not pop, push?
if(isTopNil()){ drop(); // push
push(); push(); push(); swap(); pop(); swap(); pop();}
else{ part(); drop(); // not push, swap?
if(isTopNil()){ drop(); swap(); } // swap
else{ part(); drop(); // not swap, nil?
if(isTopNil()){ drop(); nil(); } // nil
else{ part(); drop(); // pair?
if(isTopNil()){ drop(); pair(); } // pair
else{ part(); drop(); // not pair, part?
if(isTopNil()){ drop(); part(); } // part
else{ part(); drop(); // not part, select?
if(isTopNil()){ drop(); sel(); } // select
else{ part(); drop(); // not select, return?
if(isTopNil()){ drop(); push(); push(); drop(); drop(); } // return
else{ drop(); }}}}}}}}}}} // not return.
else{drop(); drop();}}} // not gray.
push(); // no next word?
if(isTopNil()){ push();} // no next word
else{ dup(); part(); swap(); drop(); swap(); nil();}}}// next word
else{ drop(); drop(); } // not blue
nil();} // new empty word
else{ drop(); pair();}}// is letter, compile it.
I’ve gone ahead and left ‘return’ in as a gray word. When a gray word spelled by the single letter ‘9’ is hit upon by the executor it executes two pushes followed by two drops. This has the result of eliminating the program that is currently being played by the executer and playing the most recently paused program.
The simplest way to eliminate return as a gray word and add it as a defined word that can be used as a green word in all future definitions is to carefully consider what the paused programs look like when a green ‘return’ is confronted by the executer.
When any green word is hit, in this case it’s a green ‘return’, the program that is currently playing is paused and duplicated. It is this duplicate that is search through by going back one word at a time.
Note, the paused program is paused after it has been moved one step foreward and before the next word to be executed has been extracted from it. So, technically, a step of searching is wasted by checking that the word we already know is a green ‘return’ is not a matching red ‘return’.
Once the matching red ‘return’ is found in the duplicate of the paused program it is then played.
Alas, here is a problem that I did not foresee: there does not appear to be any way of programming ‘return’ since the gray ‘push’ does not do the same thing as the ‘push();’ of javascript.
It is not entirely accurate for me to say that I did not foresee this. I’ve been stressing the importance of implementing the virtual machine or implementing complex gray words. Here I am forced to introduce complex gray words if I wish to introduce ‘return’ and other execution control words as defined rather than primitive.
Compound gray words work like a little interpreter within the interpreter but with a little difference. There is a local place which is inaccessible anywhere else and which holds onto a list of gray words that are to be executed all together before anything else happens. When dealing with bitstrings rather than binary trees, this local place is called a register and this particular register has enough room for multiple graywords to fit within it in a list.
Since I’m not working with bits right now, I’m able to make arbitrarily long compound gray words. One way of carrying this back over to bitstrings is that when slotted instructions are ‘well packed’ they function as if they were all packed into one big bitstring. This isn’t quite true, but it’s true enough.
Complex gray words also explain why there is that strange operation at the beginning of executing gray words which ‘gets the op name’ as a single letter. The way I’ve implemented execution of gray words is the degenerate case of complex gray words with only one instruction per gray word. These instructions are often called microinstructions when talking about the design of digital machines.
Interestingly, I plan on getting away with something like the scheme I’ve already shown fails when introducing ‘return’ as a defined word. I’m going to use the top of the list to keep the current state of the complex gray word to be executed and make it so that gray ‘push’ and ‘pop’ are not the same as ‘push();’ and ‘pop();’. I emphasize that this is planned, but may not work out.
So, almost all of the same code for executing gray words can be kept but must now be wrapped in yet another while loop (making the total number of while loops four, or five if you count the for loop which is technically an external input tool).
if(isTopNil()){drop(); // gray.
pop(); nil(); // prime the pump
while(isTopNil()){
push(); part(); pop(); // get next gray
part(); drop(); part(); drop();
part(); drop(); part(); drop(); // dup?
if(isTopNil()){ drop(); dup(); } // dup
else{ part(); drop(); // not dup, drop?
if(isTopNil()){ drop(); drop(); } // drop
else{ part(); drop(); // not drop, pop?
if(isTopNil()){ drop(); push(); swap(); pop(); pop();} // pop
else{ part(); drop(); // not pop, push?
if(isTopNil()){ drop(); push(); push(); swap(); pop();} // push
else{ part(); drop(); // not push, swap?
if(isTopNil()){ drop(); swap(); } // swap
else{ part(); drop(); // not swap, nil?
if(isTopNil()){ drop(); nil(); } // nil
else{ part(); drop(); // pair?
if(isTopNil()){ drop(); pair(); } // pair
else{ part(); drop(); // not pair, part?
if(isTopNil()){ drop(); part(); } // part
else{ part(); drop(); // not part, select?
if(isTopNil()){ drop(); sel(); } // select
else{ drop(); }}}}}}}}}
push(); // more?
if(isTopNil()){nil(); pair();} // no more.
else{ pop(); nil(); }} // more.
drop();}
That should be all that is needed (though it will need to be debugged as all things do).
Next is changing the interpreter so that it can compile compound gray words. It is here that the idea of letting key presses all be like blue words comes forward. I shall replace the comma with the space character. I’m going to also do an experiment which is half way to testing what it would be like for each key to execute as a blue word and solving the problem of assembling compound gray words.
When a space is put atop the stack and the editor is playing then it drops the space letter and checks to see if the top is nil or not. If it is nil then it pairs that nil to the second from top (coloring it gray) and pairs (compiles) this gray word to the program being edited. Otherwise it pairs the top two (compiles a slot of the compound gray word being edited) and pushes a nil atop the stack.
if(isTopNil()){ // space.
drop();
if(isTopNil()){ pair(); pair(); } // compile gray word
else{ pair(); nil(); }} // compile slot, new slot
This way of doing things adds a lot of strange behavior to the whole setup. It does make it possible to edit and compile subprograms within programs, but this is not its intended effect. I will stick with how I’m doing things at this moment and then worry about simplifying later.
Oh! It just occurred to me that I do not need to add anything about this space bar stuff to the code. Complex gray words can already be compiled using the same method I’ve been using this whole time. Since each gray word is spelled by a single letter, then spelling out a complex gray word is just a list of letters i.e. it’s just a gray word with more than one letter in its name. The only thing that needs to be changed is the little gray interpreter.
I also needt to add some sort of mechanism for ‘rewinding’ because compiling gray words by this method, or perhaps any method consistent with the monolithic implementation, collects them in the reverse of the order they need to be executed in.
This may be a fundamental problem of any slotted methods. There is probably a way to use the operations of ‘finding’ to automatically compile the ‘rewinded’ complex gray word. But, since this is the first thing that occurred to me I’d not bet on it being the best way of doing things.
The heterological paradox inspires the conclusion that there is no consistent theory of truth, or, better, denotation. It is implied by Quine’s principle of concretion for predicate abstracts and the principle of disquotation (which is the crux of inconsistency).
While it has been clear to me for some time that Hilbert and Bernay’s completeness theorem is the passage through which incompleteness follows from the inconsistency of a disquotational theory of denotation, I have not made it clear.
I shall not make it clear here either, but leave a link to this paper which can be used to triangulate with the remarks I’ve made in this note and others.
This begin a series of notes on the following links (mostly to YouTube videos) about Forth, colorForth, and uhdForth. I’m writing them because there is not much out there which takes all these threads of history and ties them to their logical origins in Quine’s predicate functor analysis of the crossreferential mechanics of variables. Here are the links:
These are notes on how Chuck Moore’s uhdForth works. Perhaps the most informative explanation of how Moore’s uhdForth works is in a video that has severe audio issues. I shall first explain what is covered in that video.
The video is from 2021. It is one of Moore’s annual “Fireside Chats”. The video can be watched at the following link, but on that day Chuck’s microphone makes it almost impossible to hear him clearly enough to know what he is saying: https://www.youtube.com/watch?v=BpsXyB2WsUw.
The entire presentation is framed around solving the problem of drawing each character of the alphabet of uhdForth to the screen: “It’s the biggest ‘OK’ I’ve ever drawn”. It is not introduced as “the newest version of Forth” because Chuck has a thing about solving problems rather than selling solutions. He also doesn’t call it ‘the alphabet’ but rather ‘the character set’.
The character set is drawn in three different sizes: normal, six times larger, and ten times larger. Chuck groups them as follows:
Though he does not say it yet, these groupings and their ordered are an essential part of how parsing and interpretation work in uhdForth.
“My goal in designing this character set is to make best use of similarities between characters.” 2:10
“These characters are made up of three kinds of elements: rectangles, slanted lines, and ellispes.” 2:44
What Moore neglects to mention is that he’s summarizing the practice of technical lettering, which, until recently, was the prevailing method by which technical drawings were marked up to be read. Today, if you search for ‘technical lettering’ you may struggle to find a simple and complete description of what people were doing for decades before computer aided design flooded the world (for better and for worse).
To learn about technical drawing and technical lettering, look for books on ‘descriptive geometry’. There you will find that letters are taught by pointing up similiarities between their shapes in precisely the way that Moore goes on to describe.
Where Moore’s method deviates from those of earlier draftsman is that instead of drawing with pen and paper he draws with pixels using Bresenham’s algorithm(s). The best explanation that gets you going immediately with the simplest code can be found here http://members.chello.at/~easyfilter/bresenham.html.
The logic behind the construction of Bresenham’s algorithm is far more important than the algorithm itself and is completely explained in Bresenham’s original paper here: https://web.archive.org/web/20080528040104/http://www.research.ibm.com/journal/sj/041/ibmsjIVRIC.pdf. It is an important example of what has largely been lost from our rapid change from a largely analoge to digital world.
Like Moore’ I’ve contemplated the simplest methods for drawing with a programmable machine like a computer. The only method of technical lettering that may be simpler than Bresenham’s is the method of ray marching signed distance functions. Since this is beyond the scope of the video, I shall not mention it further.
Moore breaks the drawing of ellipses into those that are the result of a horizontal or vertical dilation of a circle. Thus, he does not need to implement Bresenham’s algorithm for drawing ellipses. He starts with the much simpler algorithm of drawing Bresenham circles, which have an eight fold radial symmetry, and scaling along the horizontal or verticle axis.
For those familiar with graphical algorithms, this method of scaling is immediately evident when you look at how each character with a rounded part gets drawn: the curved lines are thin at the top and bottom of the curve in a character and fat around their midsection. This is one of many artifacts that the intreped graphics programmer discovers as they begin to implement everything from scratch on their own and are faced with reliving the errors of graphics programmers past. There should be a gallary somewhere of classical graphic errors and their implementation.
Moore calls a horizontally scaled Bresenham circle an oblate ellipse and a vertically scaled one a prolate ellipse.
At 4:40 he shows the first screen of uhdForth code publicly seen. It is profound in the most subdued way possible. My writing such things may appear as hyperbole (and perhaps there is a bit of it in there), but the number of consequential things on display is astounding.
Following a pattern set way back in the origins of Forth, code is displayed in blocks. Rather than navigating the short or long term memory of a machine by using the metaphor of a filing system, as if there were folders within folders and documents down at the bottom, the block system breaks memory into equal sized chunks whose size tends to depend on that of the whole memory.
I do not know what the size of blocks are in the version of uhdForth that Chuck is showing.
The code for drawing the characters from ‘0’ to ‘o’ (a lowercase o) is displayed by typing in what I’m going to call the ‘command line’. The command line is at the top of the screen and a red period is used to show where the next character keyed in will show on screen. Moore types ‘1’, ‘4’, and ‘3’, and it shows ‘143’ in white in the upper left hand corner of the screen.
A space is put between the ‘3’ and the next white word typed ‘list’. I’m pretty sure that a space is inserted when Chuck presses the space bar on his keyboard, but he has been known to make neat one handed keyboards that function contextually like modern iPhone keyboards. When you press the space bar (or whatever corresponding button Chuck uses) after writing a white word, that word is executed immediately. This is only true of white words that are typed in the command line.
When white words are placed within a block of code they are executed as they are hit upon by the interpreter/compiler.
We get a glimpse at how a white word typed can fail to be executed. After typing ‘143’ which is the number of the block of code he plans to show, he types ‘list’, presses the space bar (or perhaps enter), and then the word ‘list’ is spelled out again but this time in red.
One of the reasons that I was able to understand all that was happening in Moore’s presentation is because we are both working on a similar kind of problem: the search for timeless simplicity. The moment I saw that red ‘list’ I knew what was going on even though there was no way I could really know for sure. The white word ‘list’ had sent the interpreter off on a search for a red ‘list’ in one of the blocks of code that has already been compiled. If it is not found, when executed from the command line, the red word is appended to the end of what is already shown on the command line.
If this did not happen, then you would not know that the word you’ve typed was not found and thus not executed. This silent failure would be problematic for any program which when executing does not change any pixels on the screen and hence has no effect that you can readily see and respond appropriately to.
There is also one other subtle thing that can be learned from what Chuck types to show the first block of uhdForth code. After the red ‘list’ appears, the red dot is already a space away from the end of the red ‘list’. He types ‘editor’ and we hear the sound of him pressing the spacebar (by this time I’m almost certain it must be the space bar because it is the loudest of the keyboard keys and he hits it triumphantly). After he is heard pressing the space bar, the red dot is still at the end of the white word ‘editor’. After a very brief moment, the block of code finally displays. At that moment, the red dot that was next to the white ‘editor’ moves a space away. So that red dot gives the programmer a bit of feedback: if it is still next to the white word last ended with a press of the space bar then the machine has not released control to the command line.
Given how I would do things, I can make an educated guess that typing the white word ‘143’ executes the decimal numeral parser and puts the corresponding machine word onto the data stack. Typing a white ‘editor’ in the command line probably compiles and executes the code which composes the editing application. It expects the top of the data stack to be the number of the block of code to be edited.
The reason I very strongly suspect that a white ‘editor’ in the command line compiles then executes code is because it takes longer to open the fist block than it does any subsequent blocks shown by typing the number of the block and executing ‘link’.
While I have not yet made it to the five minute mark in this video, I am not discouraged. I have watched all of these videos enough times to have the big picture in view. The big picture is that just as there was a great reduction in work that must be done to complete arithmetical tasks by shifting from roman numeral methods to place value methods, there is a similar reduction in work got from shifting from methods that use variables and those that use predicate functors. The only thing confusing about this description of the big picture is that few people know what Quine’s predicate functors are, what they do, and how they are different from the method of combinators (and the methods of cylindrical algebras and algebraic logics). I am as yet unable to eliminate this confusion from the world.
My implementation of the programmable program, i.e. my programmable environment, in javascript is in a stable state. Having put all the pieces together into a single monolithic javascript definition, I’ll now go about factoring them back into their pieces and then move onto the so called ‘internal operations’.
I expect to find that the defined javascript functions are just an elaborate way of saying the same things that can be learned from simply reading the comments. Only this little experiment will reveal that to me.
The beginning definitions are the same. I still have delegated control over allocation to javascript and not made it explicit.
// Native Helper Functions (without explicit garbage collection)
var Nil = new Object(); // the empty pair
function isNil(p){return Object.is(p, Nil);} // empty pair?
function Pair(l,r){return new Array(l,r);} // allot pair
function Left(p){if(isNil(p)) return Nil; else return p[0];} // the left part of
function Right(p){if(isNil(p)) return Nil; else return p[1];} // the right part of
var T=Nil; // the tree
function isTopNil(){return isNil(Right(Left(T)));} // top of tree nil?
// basic operations on the tree
function dup(){T=Pair(Pair(Left(T), Right(Left(T))), Right(T));}
function drop(){T=Pair(Left(Left(T)), Right(T));}
function pop(){T=Pair(Left(Left(T)), Pair(Right(Left(T)), Right(T)));}
function push(){T=Pair(Pair(Left(T), Left(Right(T))), Right(Right(T)));}
function swap(){T=Pair(Pair(Pair(Left(Left(Left(T))),
Right(Left(T))), Right(Left(Left(T)))), Right(T));}
function nil(){T=Pair(Pair(Left(T), Nil), Right(T));}
function pair(){T=Pair(Pair(Left(Left(Left(T))),
Pair(Right(Left(Left(T))), Right(Left(T)))), Right(T));}
function part(){T=Pair(Pair(Pair(Left(Left(T)),
Left(Right(Left(T)))), Right(Right(Left(T)))), Right(T));}
function select(){if(isTopNil()){drop(); swap(); drop();}else{drop(); drop();}}
The next function that can be factored out is ‘id’ which replaces the top two items on the stack with nil if they are identical and with something not nil if they aren’t.
function id(){ nil(); dup(); pop(); // prime the pump
while(isTopNil()){ drop();
if(isTopNil()){ drop(); // is other nil?
if(isTopNil()){ drop(); push();// both nil, empty rest?
if(isTopNil()){ dup(); pop(); dup(); pair(); }// empty rest means id
else{ part(); part(); pop(); nil(); }} // rest id?
else{ push(); drop(); dup(); pop(); }} // not both nil means not id
else{ swap(); // not nil, is other?
if(isTopNil()){ nil(); }// nil means not id
else{ part(); push(); pair(); pop(); swap();
part(); push(); pair(); pop(); nil(); }}} // parts id?
drop(); push();
}
The next function is ‘find’. Call the top item on the stack ‘the back stack’ and the second from top ‘teh fore stack’. The function named ‘find’, searches through the back stack for an item that matches the item presented to ‘find’ as third from the top. It searches by popping the top of the back stack, pushing it onto the fore stack while duplicating it and then comparing it to the search item. It does this until a matching item is found or until the back stack is empty.
function find(){
pop(); pop(); dup(); push(); push(); nil(); // prime the pump
while(isTopNil()){ drop();
part(); swap(); pop(); pair(); dup(); pop(); // back
part(); swap(); drop(); // next item
id();
if(isTopNil()){ drop(); } // match found
else{ drop(); dup(); push(); push(); // not a match, more?
if(isTopNil()){ pop(); pop(); drop(); } // no more
else{ nil(); }}}
drop(); push(); push();
}
Here it is already clear that a handful of fiddling operations are required to extricate ‘find’ from its tight fit in the context of the programmable program (that is if you include the details of ‘id’ that are obscured by the ‘id()’).
In the past I’ve also factored out ‘back’ and ‘fore’ as follows:
function back(){ part(); swap(); pop(); pair(); push();}
function fore(){ pop(); part(); push(); swap(); pair();}
Which, if so factored, would change the definition of ‘find’ to
function find(){
pop(); pop(); dup(); push(); push(); nil(); // prime the pump
while(isTopNil()){ drop();
back();
pop(); dup(); pop();
part(); swap(); drop(); // next item
id();
if(isTopNil()){ drop(); } // match found
else{ drop(); dup(); push(); push(); // not a match, more?
if(isTopNil()){ pop(); pop(); drop(); } // no more
else{ nil(); }}} // more
drop(); push(); push();
}
In fact, the entire design of ‘find’ would be changed if it was contemplated on its own e.g. the item being searched for would probably be thought to be the item at the top of the stack so that ‘find might look like this
function find(){ pop(); nil(); // prime the pump
while(isTopNil()){ drop();
dup(); part(); swap(); drop(); // next item
push(); dup(); pop(); // pull
id();
if(isTopNil()){ drop(); push();} // found
else{ drop(); back();
if(isTopNil){ push(); } // no more
else{ nil() }}} // more
drop();
}
I go through this all to make the point that factoring things out into separate functions doesn’t simplify things in the way that people are told or commonly think. Hasty conclusions have been made in many directions and it is only in following Knuth’s metamachine methods that anyone can really start to pin down what counts as simplification or complexification.
In general, factoring out into a function introduces complexity into the pre and post script encompassing the context within which the defined function is invoked. This goes unnoticed in most modern programming languages because using variables, like using roman numerals to do arithmetic, obscures the recombic operations which are ultimately called for to accomplish the full crossreferential recurrence of variables.
Note, the costs can be obscured in a language like the one contemplated here as well. Simply introduce operations that make make local red words and green words as needed. I’ll have to explain this more later, but it comes out to stuff like LISP’s ‘let’. To get at the logic of it all requires predicate abstracts as these are the method by which variables may dealt with systematically as an addition to a predicate functor logic. Predicate abstracts are the logical origin of all problems of scope in programming and in any theories with a predicate logic.
The next function to be defined is ‘green’ which is what hapens when a green word is come upon while executing a program. The green word is colored red, and that word is looked for in a duplicate of the current program. It is tempting to name this function ‘call’ because it does something like calling an instruction.
This function is where I see much less value in ‘factoring it out’ rather than simply introducing effective comments. It is easy to make simplifying assumptions that later shoot you in the foot e.g. I’ll assume that the spelling of a word is atop the stack as a list of letters. Then ‘green’ forms the red word, pulls down a duplicate of the curretn program, finds the matching red word in the program (if there is one). As defined in the larger program, if a match is not found, execution just returns to where it left off.
function redify(){nil(); nil(); pair(); nil(); pair(); nil(); pair(); pair();}
function pull2(){push(); dup(); push(); dup(); pop(); swap(); pop();}
function green(){
redify();
pull2(); // pull program
find();
if(isTopNil()){ drop(); drop();} // not found continue
else{ pop(); pop();} // found, "call"
}
The next function to define then is ‘gray’. It is like a dispatcher, it looks at the top item and if it is the letter of a gray operation then that operation is taken. These are the words that are primitive to the programmable program.
function gray(){
part(); drop(); part(); drop(); part(); drop(); part(); drop(); // dup?
if(isTopNil()){ drop(); dup(); } // dup
else{ part(); drop(); // not dup, drop?
if(isTopNil()){ drop(); drop(); } // drop
else{ part(); drop(); // not drop, pop?
if(isTopNil()){ drop(); // pop
push(); swap(); push(); swap(); pop(); pop(); pop();}
else{ part(); drop(); // not pop, push?
if(isTopNil()){ drop(); // push
push(); push(); push(); swap(); pop(); swap(); pop();}
else{ part(); drop(); // not push, swap?
if(isTopNil()){ drop(); swap(); } // swap
else{ part(); drop(); // not swap, nil?
if(isTopNil()){ drop(); nil(); } // nil
else{ part(); drop(); // pair?
if(isTopNil()){ drop(); pair(); } // pair
else{ part(); drop(); // not pair, part?
if(isTopNil()){ drop(); part(); } // part
else{ part(); drop(); // not part, select?
if(isTopNil()){ drop(); sel(); } // select
else{ part(); drop(); // not select, return?
if(isTopNil()){ drop(); push(); push(); drop(); drop(); } // return
else{ drop(); }}}}}}}}}} // not the letter of a gray word
}
While I’ve included ‘return’ as one of the gray words, it is not necessary. Such an operation can be defined and used as a green word. I’ve just avoided doing that here because it adds complexity that doesn’t pay its way yet.
The next word is then ‘blue’. This might be called ‘execute’ in another language. It expects the top item to spell out a word and below that is the back and then the fore of the program the matching green word is going to start execution within. That is, the top is colored green, and then executed as a green word. The second and third item from the top are the back and fore of the program the green word is executed within.
function greenify(){ nil(); nil(); pair(); pair();}
function blue(){
greenify();
swap(); pop(); swap(); push(); nil(); // prime the pump
while(isTopNil){
fore(); pop(); pop();
part(); // is blue?
if(isTopNil()){ drop(); drop();}
else{ part(); drop() // not blue, green?
if(isTopNil()){ green(); }
else{ part(); drop(); // not green, gray?
if(isTopNil()){ drop(); part(); swap(); drop(); gray(); } // gray, get op
else{ drop(); drop();}}}
push(); // no next word?
if(isTopNil()){ push(); } // no next word
else{ dup(); pop(); part(); swap(); drop(); nil();}}}}} // next word
Finally comes the editor (which is hardly what it is, since it doesn’t even include “backspace”).
function edit(){ dup(); part(); drop(); part(); drop(); part(); drop(); // not letter?
if(isTopNil()){ drop(); // not letter
pair(); dup(); pop(); pair(); // compile letter and word
push(); part(); // blue?
if(isTopNil()){ blue(); } // blue
else{ drop(); drop(); } // not blue
nil();} // new empty word
else{ drop(); pair(); }} // is letter, compile it
But this not quite everything. Input comes first. This is about the limit of the abstract model that has been adopted here. One which lets the javascript event system take care of scheduling when to do what based on which key is pressed.
It takes the name of a key pressed and finds its location in “the alphabet” and then uses that location to construct a binary tree corresponding to it.
var abc = '.;,:0123456789abcdefghijklmnopqrstuvwxyz'; // the alphabet
function input(key){var k = abc.length; while(k-- && abc[k]!=key); // where in abc?
if(k<0) return; // not in abc
nil(); while(k--){nil(); pair();}} // quote/construct as tree
In past programs I have gone further and linked up everything to the ‘onkeydown’ event handler. It would look just like this:
onkeydown = function(e){ input(e.key); }
But I’ll leave that out because I’m also leaving out the all the elaborate methods I have used in the past to view the state of “the tree” after a key press has been tumbled through all the above mechanisms.
In the next note I’ll explain the ‘internal operations’ which do for the top of the tree what each basic operation does for “the tree”. This begins the move to abstract virtual machines.
Here is all the code together in a program that has not been debugged:
// Native Helper Functions (without explicit garbage collection)
var Nil = new Object(); // the empty pair
function isNil(p){return Object.is(p, Nil);} // empty pair?
function Pair(l,r){return new Array(l,r);} // allot pair
function Left(p){if(isNil(p)) return Nil; else return p[0];} // the left part of
function Right(p){if(isNil(p)) return Nil; else return p[1];} // the right part of
var T=Nil; // the tree
function isTopNil(){return isNil(Right(Left(T)));} // top of tree nil?
// basic operations on the tree
function dup(){T=Pair(Pair(Left(T), Right(Left(T))), Right(T));}
function drop(){T=Pair(Left(Left(T)), Right(T));}
function pop(){T=Pair(Left(Left(T)), Pair(Right(Left(T)), Right(T)));}
function push(){T=Pair(Pair(Left(T), Left(Right(T))), Right(Right(T)));}
function swap(){T=Pair(Pair(Pair(Left(Left(Left(T))),
Right(Left(T))), Right(Left(Left(T)))), Right(T));}
function nil(){T=Pair(Pair(Left(T), Nil), Right(T));}
function pair(){T=Pair(Pair(Left(Left(Left(T))),
Pair(Right(Left(Left(T))), Right(Left(T)))), Right(T));}
function part(){T=Pair(Pair(Pair(Left(Left(T)),
Left(Right(Left(T)))), Right(Right(Left(T)))), Right(T));}
function select(){if(isTopNil()){drop(); swap(); drop();}else{drop(); drop();}}
function id(){ nil(); dup(); pop(); // prime the pump
while(isTopNil()){ drop();
if(isTopNil()){ drop(); // is other nil?
if(isTopNil()){ drop(); push();// both nil, empty rest?
if(isTopNil()){ dup(); pop(); dup(); pair(); }// empty rest means id
else{ part(); part(); pop(); nil(); }} // rest id?
else{ push(); drop(); dup(); pop(); }} // not both nil means not id
else{ swap(); // not nil, is other?
if(isTopNil()){ nil(); }// nil means not id
else{ part(); push(); pair(); pop(); swap();
part(); push(); pair(); pop(); nil(); }}} // parts id?
drop(); push();}
function back(){ part(); swap(); pop(); pair(); push();}
function find(){ pop(); pop(); dup(); push(); push(); nil(); // prime the pump
while(isTopNil()){ drop();
back();
pop(); dup(); pop(); part(); swap(); drop(); // next item
id();
if(isTopNil()){ drop(); } // match found
else{ drop(); dup(); push(); push(); // not a match, more?
if(isTopNil()){ pop(); pop(); drop(); } // no more
else{ nil(); }}} // more
drop(); push(); push();}
function redify(){nil(); pair(); nil(); pair(); nil(); pair(); pair();}
function pull2(){push(); dup(); push(); dup(); pop(); swap(); pop();} // pull program
function green(){ redify(); pull2(); find();
if(isTopNil()){ drop(); drop();} // not found continue
else{ pop(); pop();}} // found, "call"
function gray(){
part(); drop(); part(); drop();
part(); drop(); part(); drop(); // dup?
if(isTopNil()){ drop(); dup(); } // dup
else{ part(); drop(); // not dup, drop?
if(isTopNil()){ drop(); drop(); } // drop
else{ part(); drop(); // not drop, pop?
if(isTopNil()){ drop(); // pop
push(); swap(); push(); swap(); pop(); pop(); pop();}
else{ part(); drop(); // not pop, push?
if(isTopNil()){ drop(); // push
push(); push(); push(); swap(); pop(); swap(); pop();}
else{ part(); drop(); // not push, swap?
if(isTopNil()){ drop(); swap(); } // swap
else{ part(); drop(); // not swap, nil?
if(isTopNil()){ drop(); nil(); } // nil
else{ part(); drop(); // pair?
if(isTopNil()){ drop(); pair(); } // pair
else{ part(); drop(); // not pair, part?
if(isTopNil()){ drop(); part(); } // part
else{ part(); drop(); // not part, select?
if(isTopNil()){ drop(); sel(); } // select
else{ part(); drop(); // not select, return?
if(isTopNil()){ drop(); push(); push(); drop(); drop(); } // return
else{ drop(); }}}}}}}}}}}
function greenify(){ nil(); nil(); pair(); pair(); }
function fore(){ pop(); part(); push(); swap(); pair();}
function blue(){ greenify();
swap(); pop(); swap(); push(); nil(); // prime the pump
while(isTopNil){ // execute
fore(); pop(); pop(); // set up for next word
part(); // is blue?
if(isTopNil()){ drop(); drop();}
else{ part(); drop() // not blue, green?
if(isTopNil()){ green(); }
else{ part(); drop(); // not green, gray?
if(isTopNil()){ drop(); part(); swap(); drop(); gray(); } // gray, get op
else{ drop(); drop();}}}
push(); // no next word?
if(isTopNil()){ push(); } // no next word
else{ dup(); pop(); part(); swap(); drop(); nil(); }}} // next word
function edit(){ dup(); part(); drop(); part(); drop(); part(); drop(); // not letter?
if(isTopNil()){ drop(); // not letter
pair(); dup(); pop(); pair(); // compile letter and word
push(); part(); // blue?
if(isTopNil()){ blue(); } // blue
else{ drop(); drop(); } // not blue
nil(); } // new empty word
else{ drop(); pair(); }} // is letter, compile it
var abc = '.;,:0123456789abcdefghijklmnopqrstuvwxyz'; // the alphabet
function input(key){var k = abc.length; while(k-- && abc[k]!=key); // where in abc?
if(k<0) return; // not in abc
nil(); while(k--){nil(); pair();}} // quote/construct as tree
I now have the javascript implementation of my programmable environment logically debugged. Here it is:
// Native Helper Functions (without explicit garbage collection)
var Nil = new Object(); // the empty pair
function isNil(p){return Object.is(p, Nil);} // empty pair?
function Pair(l,r){return new Array(l,r);} // allot pair
function Left(p){if(isNil(p)) return Nil; else return p[0];} // the left part of
function Right(p){if(isNil(p)) return Nil; else return p[1];} // the right part of
var T=Nil; // the tree
function isTopNil(){return isNil(Right(Left(T)));} // top of tree nil?
// basic tree operations
function dup(){T=Pair(Pair(Left(T), Right(Left(T))), Right(T));}
function drop(){T=Pair(Left(Left(T)), Right(T));}
function pop(){T=Pair(Left(Left(T)), Pair(Right(Left(T)), Right(T)));}
function push(){T=Pair(Pair(Left(T), Left(Right(T))), Right(Right(T)));}
function swap(){T=Pair(Pair(Pair(Left(Left(Left(T))),
Right(Left(T))), Right(Left(Left(T)))), Right(T));}
function nil(){T=Pair(Pair(Left(T), Nil), Right(T));}
function pair(){T=Pair(Pair(Left(Left(Left(T))),
Pair(Right(Left(Left(T))), Right(Left(T)))), Right(T));}
function part(){T=Pair(Pair(Pair(Left(Left(T)),
Left(Right(Left(T)))), Right(Right(Left(T)))), Right(T));}
function select(){if(isTopNil()){drop(); swap(); drop();}else{drop(); drop();}}
// editer = interpreter + compiler + executor
var abc = '.;,:0123456789abcdefghijklmnopqrstuvwxyz'; // the alphabet
onkeydown=e=>{switch(e.key){
case'Shift': case'Enter': case'Control': case'Alt': case'Tab': case'CapsLock':return;
default: var k=abc.indexOf(e.key); if(k<0)k=abc.length; nil(); while(k--){nil(); pair();}; // input
dup(); part(); drop(); part(); drop(); part(); drop(); // not letter?
if(isTopNil()){ drop(); // not letter
pair(); dup(); pop(); pair(); // compile letter and word
push(); part(); // blue?
if(isTopNil()){// blue
nil(); pair(); pair(); // greenify
swap(); pop(); swap(); nil(); //prime the pump
while(--limit && isTopNil()){ drop(); // execute
part(); push(); swap(); pair(); pop(); pop(); // fore
part(); // blue?
if(isTopNil()){ drop(); drop(); } // blue
else{ part(); drop(); // not blue, green?
if(isTopNil()){ // green
nil(); pair(); nil(); pair(); nil(); pair(); pair(); // redify
dup(); push(); dup(); push(); dup(); pop(); swap(); pop(); // pull program
nil(); // prime the pump
while(--limit && isTopNil()){ drop(); // find match
part(); swap(); pop(); pair(); dup(); pop(); // back
part(); swap(); drop(); // next word
nil(); dup(); pop(); // prime the pump
while(--limit && isTopNil()){ drop(); // id
if(isTopNil()){ drop(); // is other nil?
if(isTopNil()){ drop(); push();// both nil, empty rest?
if(isTopNil()){ dup(); pop(); dup(); pair(); }// empty rest means id
else{ part(); part(); pop(); nil(); }} // rest id?
else{ push(); drop(); dup(); pop(); }} // not both nil means not id
else{ swap(); // not nil, is other?
if(isTopNil()){ nil(); }// nil means not id
else{ part(); push(); pair(); pop(); swap();
part(); push(); pair(); pop(); nil(); }}} // parts id?
drop(); push(); // match found?
if(isTopNil()){ drop(); } // match found
else{ drop(); dup(); push(); push(); // not a match, more?
if(isTopNil()){ drop(); drop(); drop();} // no more
else{ nil(); }}} // more
drop();}
else{ part(); drop(); // not green, gray?
if(isTopNil()){ drop(); part(); swap(); drop(); // gray, get op
part(); drop(); part(); drop(); part(); drop(); part(); drop(); // dup?
if(isTopNil()){ drop(); dup(); } // dup
else{ part(); drop(); // not dup, drop?
if(isTopNil()){ drop(); drop(); } // drop
else{ part(); drop(); // not drop, pop?
if(isTopNil()){ drop(); // pop
push(); swap(); push(); swap(); pop(); pop(); pop();}
else{ part(); drop(); // not pop, push?
if(isTopNil()){ drop(); // push
push(); push(); push(); swap(); pop(); swap(); pop();}
else{ part(); drop(); // not push, swap?
if(isTopNil()){ drop(); swap(); } // swap
else{ part(); drop(); // not swap, nil?
if(isTopNil()){ drop(); nil(); } // nil
else{ part(); drop(); // pair?
if(isTopNil()){ drop(); pair(); } // pair
else{ part(); drop(); // not pair, part?
if(isTopNil()){ drop(); part(); } // part
else{ part(); drop(); // not part, select?
if(isTopNil()){ drop(); sel(); } // select
else{ part(); drop(); // not select, return?
if(isTopNil()){ drop(); push(); push(); drop(); drop(); } // return
else{ drop(); }}}}}}}}}}} // not return.
else{drop(); drop();}}} // not gray.
push(); // next word?
if(isTopNil()){ push();} // no next word
else{ dup(); part(); swap(); drop(); swap(); nil();}}}// next word
else{ drop(); drop(); } // not blue
nil();} // new empty word
else{ drop(); pair();}// is letter, compile it.
show(); break;}}
function ln(p){return isNil(p)?0:1+ln(Left(p));} // length (or height) of stack
function a(n){return abc.length<=n?'?':abc.at(n);} // nth letter of abc
function letter(p){return a(ln(p));} // letter of stack height
function letters(p){return isNil(p)?'#':letters(Left(p))+letter(Right(p));} // letters of stack
function lwords(p){return isNil(p)?'#':lwords(Left(p))+letters(Right(p));} // words stacked last top
function rwords(p){return isNil(p)?'#':letters(Right(p))+rwords(Left(p));} // words stack first top
function prog(){return '{'+lwords(Right(Left(Left((T)))))
+'['+letters(Right(Left(T)))+']'+lwords(Right(Left(Left(Left(T)))))+'}';}// edited program
document.body.style.font='12pt monospace';document.body.style.whiteSpace='pre';
var limit=1<<10; // loop limit
function show(s=''){document.body.textContent+=`\t${limit} ${prog()} ${s}\n`;}
function bin(p){return isNil(p)?'0':bin(Left(p))+bin(Right(p))+'1';} // bitree -> bistring
function par(p){return isNil(p)?'()':'('+par(Left(p))+' '+par(Right(p))+')';} // bitree -> parens
function lbin(p){return isNil(p)?'':bin(Right(p))+'\n'+lbin(Left(p));} // bistring stack
function rbin(p){return isNil(p)?'':rbin(Right(p))+'\n'+bin(Left(p));} // bistring list
function bins(p,s=''){return rbin(Right(p))+'\n-'+s+'\n'+lbin(Left(p));} // bistring tree
This newest version is special in a number of ways. The most important is that it uses only the primitive operations of dup, drop, push, pop, swap, nil, pair, and part. It includes ‘select’ at the end of the list of primitive operations because ‘select’ is implicit in the use of the ‘if’, ‘else’, and ‘while’s of javascript. This will be easier to explain once I elminated all ‘if’, ‘else’, and ‘while’ for a single javascript ‘while’ (that of the scheduler).
While there are comments, they are not final. They do not strongly suggest what is going on, or why it is happening. This is a problem with any well made machine that has ultimately lost touch with the environments from which its key components were selected.
Although I claim that the program is debugged, I have not given a clear and exact definition of just what debugging entails. It is a common word among programmers. The question is what consequences come from it?
I often say “I’ve got the logic right but the code may be wrong.” How can this be? Or, what does this mean? For some time I have felt that I’ve got the logic of my little programmable environment ‘right’ or ‘worked out’. The moment I put the finishing touches on my first description of what I then called “the interpreter algorithm” and what I now call “the programmable program” I knew I had got the logic right.
Only a fool can be so certain of anything, and while it may take a fool to finish anything, it only furthers foolishness to leave what the fool finished undisturbed or unexamined.
So, what could it mean to say that I’ve got the logic right? It can mean that if I was prompted to write out the predicate logic of the theory within which my programmable program is designated by a singular description, then I am very likely to do just that.
Perhaps I would do better to say “I’ve got the science right” or perhaps even “I’ve got the experiments right”. Perhaps even better still to say “I’ve go the right equipment” or “I’m in the right laboratory.”
Perhaps I can only say “What I’ve said points up relevant contingencies.” For, what I’ve said does not point at or to them: they may be part of spacetime but they can never be here and now. Perhaps I can say they ‘point out’ rather than ‘point up’.
I must leave an examination of the practice of programming for another time. This present speculation is weak and amounts to an excuse or escape.
What’s next?
I have drastically changed the formatting to make it easier to read. I fixed a few bugs, but it still does not work as expected.
var Nil = new Object();
function isNil(p){return Object.is(p, Nil);}
function Pair(l,r){return new Array(l,r);}
function Left(p){if(isNil(p)) return Nil; else return p[0];}
function Right(p){if(isNil(p)) return Nil; else return p[1];}
var T=Nil;
function isTopNil(){return isNil(Right(Left(T)));}
function dup(){T=Pair(Pair(Left(T), Right(Left(T))), Right(T));}
function drop(){T=Pair(Left(Left(T)), Right(T));}
function pop(){T=Pair(Left(Left(T)), Pair(Right(Left(T)), Right(T)));}
function push(){T=Pair(Pair(Left(T), Left(Right(T))), Right(Right(T)));}
function swap(){T=Pair(Pair(Pair(Left(Left(Left(T))),
Right(Left(T))), Right(Left(Left(T)))), Right(T));}
function nil(){T=Pair(Pair(Left(T), Nil), Right(T));}
function pair(){T=Pair(Pair(Left(Left(Left(T))),
Pair(Right(Left(Left(T))), Right(Left(T)))), Right(T));}
function part(){T=Pair(Pair(Pair(Left(Left(T)),
Left(Right(Left(T)))), Right(Right(Left(T)))), Right(T));}
function select(){if(isTopNil()){drop(); swap(); drop();}else{drop(); drop();}}
var abc = '.;,:0123456789abcdefghijklmnopqrstuvwxyz', limit=1<<10;
onkeydown=e=>{switch(e.key){
case'Shift': case'Enter': case'Control': case'Alt': case'Tab': case'CapsLock':return;
default: var k=abc.indexOf(e.key); if(k<0)k=abc.length; nil(); while(k--){nil(); pair();}; // input
dup(); part(); drop(); part(); drop(); part(); drop(); // not letter?
if(isTopNil()){ drop(); // not letter
pair(); dup(); pop(); pair(); // compile letter and word
push(); part(); // blue?
if(isTopNil()){// blue
nil(); pair(); pair(); // greenify
swap(); pop(); swap(); nil(); //prime the pump
while(--limit && isTopNil()){ drop(); // execute
part(); push(); swap(); pair(); pop(); pop(); // fore
part(); // blue?
if(isTopNil()){ drop(); drop(); } // blue
else{ part(); drop(); // not blue, green?
if(isTopNil()){ // green
nil(); pair(); nil(); pair(); nil(); pair(); pair(); // redify
dup(); push(); dup(); push(); dup(); pop(); swap(); pop(); // pull program
nil(); // prime the pump
while(--limit && isTopNil()){ drop();
part(); swap(); pop(); pair(); dup(); pop(); // back
part(); swap(); drop(); // next word
nil(); dup(); pop(); // id? prime the pump
while(--limit && isTopNil()){ drop();
if(isTopNil()){ drop(); // is other nil?
if(isTopNil()){ drop(); push();// both nil, empty rest?
if(isTopNil()){ dup(); pop(); dup(); pair(); }// empty rest means id
else{ part(); part(); pop(); nil(); }} // rest id?
else{ push(); drop(); dup(); pop(); }} // not both nil means not id
else{ swap(); // not nil, is other?
if(isTopNil()){ drop(); push(); drop(); dup(); pop(); }// nil means not id
else{ part(); push(); pair(); pop(); swap();
part(); push(); pair(); pop(); nil(); }}} // parts id?
drop(); push(); // match found?
if(isTopNil()){ drop(); } // match found
else{ drop(); dup(); push(); // not a match, more?
if(isTopNil()){ pop(); drop(); }}} // no more.
drop(); drop();}
else{ part(); drop(); // not green, gray?
if(isTopNil()){ drop(); part(); drop(); part(); drop(); // gray, dup?
part(); drop(); part(); drop(); part(); drop(); // dup?
if(isTopNil()){ drop(); dup();} // dup
else{ part(); drop(); // not dup, drop?
if(isTopNil()){ drop(); drop(); } // drop
else{ part(); drop(); // not drop, pop?
if(isTopNil()){ drop(); // pop
push(); swap(); push(); swap(); pop(); pop(); pop();}
else{ part(); drop(); // not pop, push?
if(isTopNil()){ drop(); // push
push(); push(); push(); swap(); pop(); swap(); pop();}
else{ part(); drop(); // not push, swap?
if(isTopNil()){ drop(); swap(); } // swap
else{ part(); drop(); // not swap, nil?
if(isTopNil()){ drop(); nil(); } // nil
else{ part(); drop(); // pair?
if(isTopNil()){ drop(); pair(); } // pair
else{ part(); drop(); // not pair, part?
if(isTopNil()){ drop(); part(); } // part
else{ part(); drop(); // not part, select?
if(isTopNil()){ drop(); sel(); } // select
else{ part(); drop(); // not select, return?
if(isTopNil()){ drop(); push(); push(); drop(); drop();} // return
else{ drop(); }}}}}}}}}}} // not return.
else{drop(); drop();}}} // not gray.
push(); // next word?
if(isTopNil()){ push();} // no next word
else{ dup(); part(); swap(); drop(); swap(); }}}// next word
else{ drop(); drop(); } // not blue
nil();} // new empty word
else{ drop(); pair();}// is letter, compile it.
show(); break;}}
function ln(p){return isNil(p)?0:1+ln(Left(p));}
function a(n){return abc.length<=n?'?':abc.at(n);}
function letter(p){return a(ln(p));}
function letters(p){return isNil(p)?'#':letters(Left(p))+letter(Right(p));}
function lwords(p){return isNil(p)?'#':lwords(Left(p))+letters(Right(p));}
function rwords(p){return isNil(p)?'#':letters(Right(p))+rwords(Left(p));}
function prog(){return '{'+lwords(Right(Left(Left((T)))))
+'['+letters(Right(Left(T)))+']'+rwords(Right(Left(Left(Left(T)))))+'}';}
document.body.style.font='12pt monospace';
document.body.style.whiteSpace='pre';
function show(s=''){document.body.textContent+=`\t${limit} ${prog()} ${s}\n`;}
It should be no secret that all of this work on my programming environment is aimed at setting up an interlocking system of hardware and software that simplify the work of any programmer. I see my work as making the lathe of programming. It gets at the perfect balance between what can be acheived through those less than human parts of the world. It is designed to extend the reach of human behavior.
This iteration of the javascript for my programming environment eliminates all defined operations e.g. over, not, and, or. It has not been debugged and is unlike to work as expected in its current state.
var Nil = new Object();
function isNil(p){return Object.is(p, Nil);}
function Pair(l,r){return new Array(l,r);}
function Left(p){if(isNil(p)) return Nil; else return p[0];}
function Right(p){if(isNil(p)) return Nil; else return p[1];}
var T=Nil;
function isTopNil(){return isNil(Right(Left(T)));}
function dup(){T=Pair(Pair(Left(T), Right(Left(T))), Right(T));}
function drop(){T=Pair(Left(Left(T)), Right(T));}
function pop(){T=Pair(Left(Left(T)), Pair(Right(Left(T)), Right(T)));}
function push(){T=Pair(Pair(Left(T), Left(Right(T))), Right(Right(T)));}
function swap(){T=Pair(Pair(Pair(Left(Left(Left(T))),
Right(Left(T))), Right(Left(Left(T)))), Right(T));}
function nil(){T=Pair(Pair(Left(T), Nil), Right(T));}
function pair(){T=Pair(Pair(Left(Left(Left(T))),
Pair(Right(Left(Left(T))), Right(Left(T)))), Right(T));}
function part(){T=Pair(Pair(Pair(Left(Left(T)),
Left(Right(Left(T)))), Right(Right(Left(T)))), Right(T));}
function select(){if(isTopNil()){drop(); swap(); drop();}else{drop(); drop();}}
var limit=1<<10;
var abc = '.;:,0123456789abcdefghijklmnopqrstuvwxyz'
onkeydown=e=>{switch(e.key){
case'Shift': case'Enter': case'Control': case'Alt': case'Tab': case'CapsLock':return;
default:
var k=abc.indexOf(e.key); if(k<0)k=abc.length; nil(); while(k--){nil(); pair();}; // input
dup(); part(); drop(); part(); drop(); part(); drop(); // not letter?
if(isTopNil()){ drop();
pair(); dup(); pop(); pair(); // compile letter and word
push(); part(); part(); // blue?
if(isTopNil()){// blue
nil(); pair(); pair(); // greenify
swap(); pop(); swap(); nil(); //prime the pump
while(isTopNil()){ drop(); // execute
part(); push(); swap(); pair(); pop(); pop(); // fore
part(); // blue?
if(isTopNil()){ drop(); drop(); // blue
}else{ // not blue
part(); drop(); // green?
if(isTopNil()){ drop(); // green
nil(); nil(); pair(); nil(); pair(); nil(); pair(); pair(); // redify
dup(); push(); dup(); push(); dup(); pop(); swap(); pop(); // pull program
nil(); // prime the pump
while(isTopNil()){ drop();
part(); swap(); pop(); pair(); dup(); pop(); // back
part(); swap(); drop(); // next word
nil(); dup(); pop(); // id? prime the pump
while(isTopNil()){ drop();
if(isTopNil()){ drop(); // is other nil?
if(isTopNil()){ drop(); // both nil
push(); // empty rest?
if(isTopNil()){ // empty rest means id
dup(); pop(); dup(); pair();
}else{ // rest id?
part(); part(); pop(); nil();
}
}else{ // not both nil means not id
push(); drop(); dup(); pop();
}
}else{ // not nil
swap(); // is other not nil?
if(isTopNil()){ drop(); // nil means not id
push(); drop(); dup(); pop();
}else{ // parts id?
part(); push(); pair(); pop();
swap();
part(); push(); pair(); pop();
nil();
}
}
}
drop(); push(); // match found?
if(isTopNil()){ drop(); // match found
}else{ drop(); // not a match
dup(); push(); // more?
if(isTopNil()){ // no more.
pop(); drop();
}
}
}
drop(); drop();
}else{ // not green
part(); drop(); // gray?
if(isTopNil()){ drop() // gray
part(); drop(); part(); drop(); part(); drop();
part(); drop(); part(); drop(); if(isTopNil()){ drop(); // dup?
dup(); }else{ part(); drop(); if(isTopNil()){ drop(); // drop?
drop(); }else{ part(); drop(); if(isTopNil()){ drop(); // pop?
push(); swap(); push(); swap(); pop(); pop(); pop();
}else{ part(); drop(); if(isTopNil()){ drop(); // push?
push(); push(); push(); swap(); pop(); swap(); pop();
}else{ part(); drop(); if(isTopNil()){ drop(); // swap?
swap(); }else{ part(); drop(); if(isTopNil()){ drop(); // nil?
nil(); }else{ part(); drop(); if(isTopNil()){ drop(); // pair?
pair(); }else{ part(); drop(); if(isTopNil()){ drop(); // part?
part(); }else{ part(); drop(); if(isTopNil()){ drop(); // select?
sel(); }else{ part(); drop(); if(isTopNil()){ drop(); // return?
push(); push(); drop(); drop();
}else{ drop(); }}}}}}}}}}
}else{drop(); drop();}}}
push(); // next word?
if(isTopNil()){ // no next word
push();
}else{
dup(); part(); swap(); drop(); swap(); // next word
}
}
}else{ // not blue
drop(); drop();
}
nil(); // new empty word
}else{ drop(); // is letter
pair(); // compile letter
}
show(); break;}}
function ln(p){return isNil(p)?0:1+ln(Left(p));}
function a(n){return abc.length<=n?'?':abc.at(n);}
function letter(p){return a(ln(p));}
function letters(p){return isNil(p)?'#':letters(Left(p))+letter(Right(p));}
function lwords(p){return isNil(p)?'#':lwords(Right(p))+letters(Left(p));}
function rwords(p){return isNil(p)?'#':letters(Left(p))+rwords(Right(p));}
function prog(){return '{'+lwords(Right(Left(Left((T)))))
+'['+letters(Right(Left(T)))+']'+lwords(Right(Left(Left(Left(T)))))+'}';}
document.body.style.font='12pt monospace';
document.body.style.whiteSpace='pre';
function show(s=''){document.body.textContent+=`\t${limit} ${prog()} ${s}\n`;}
Among the things that I think people should know are the methods of logic. Anyone who can be taught up to and including Quine’s main method of proof is ready to deal with the rest of science in a profoundly more effective way than anyone who does not know the same.
Technically, you need to know a bit more than Quine’s main method. Once you know how Quine’s main method works it becomes very easy to show that other methods of proof that are more familiar are also effective. The reason I do not put any emphasis on the more familiar methods of proof is that their familiarity is a distraction. Quine’s main method of proof is as ‘natural’ as the more familiar methods of natural deduction that are popular among modern mathematical logicians (I dare not give ‘computer science’ a name because those who fly under that flag would rather be called mathematical logicians or outright programmers).
For the life of me, I do not know why Quine’s main method is not well known. Even Goldfarb’s book on “deductive logic” includes Quine’s main method without naming it as such! What’s perhaps most startling: Quine’s main method corresponds to the oldest method of arguments on display in Plato’s Socratic dialogues. It goes like this. You say that such and such is true of something. Then, let ‘x’ be the name of such a thing. Now you also say that such and such is true of everything. So it must also be true of x in particular. This method proceeds until a contradiction is reach. Upon reaching such a contradiction it asserted that what we started with was inconsistent and that something has to be wrong with our premises.
Now, there is a technical bit to Quine’s main method that makes it not correspond precisely to Socrate’s method. It is the problem of taking what is said and getting all of its ‘everythings’ and ‘somethings’ out to the front of each premise. I believe it was Dreben who showed that this ‘prenex form’, as it is called, is unnecessary. Socrates’s method is enough.
Socrates’s method is a bit larger than I’ve sketched out. What I described is really the second half of his method. That is, once premises have been uncovered, it is our job to look for their ultimate inconsistency. The first half of Socrate’s method (perhaps we should properly call it Plato’s method because it is clear enough that Plato is speaking through the dialogues he wrote in which Socrates stars) is uncovering the premises to the theories that people have left unexamined.
It has been a while since I read all of Plato’s dialogues and I am uncertain if Socrates noticed that if the negation of a premise is inconsistent then that means the premise itself is valid. Valid premises can be dropped from any theory: they are ‘purely logical’ in that there is no conclusion which is lost from dropping a valid premise in a logical argument.
For better or worse I’m leaving more scraps here than ever. I’m just going to let more things hang out there without resolution. The world is unresolved.
I’m going to introduce a new implementation of my programming environment that decouples the io from the vm. It starts with the scheduler:
var scheduled = [];
var following = false;
function schedule(it){
scheduled.push(it);
if(following) return;
following = true;
while(scheduled.length) scheduled.shift()();
following = false;
}
Scheduled actions are pushed onto the end of a list and if there isn’t already some function following the schedule, then the following flag is raised and while there are still actions scheduled the the item at the front of the list is shifted off and executed. When there are no more scheduled items the following flag is lowered.
The method of pushing new items onto the end of the schedule and shifting them off the front of the schedule occurs in so many different contexts that it is given an abstract name: queue. People queue in lines at a theme park. The first person in is also the first person out of the queue. Instead of ‘push’ and ‘shift’, we say ‘queue’ and ‘dequeue’ to emphasize the name for the abstract methods of queuing.
It just so happens that saying the letter ‘q’ sounds like ‘queue’ and saying the letters ‘dq’ sounds like ‘dequeue’. These are abbreviations I often use because there is also another abbreviation that answers a question commonly asked of items that may or may not be in a queue: are they queued? The abbreviation being ‘qd’. I won’t use these abbreviations here, but I’ve used them often enough elsewhere that I felt like I need to explain myself somewhere.
A terminal is a keyboard with a screen. A keyboard works a keypress at a time. When a key is pressed, its name is put on a queue. I’ll call this queue ‘keyspressed’.
var keyspressed=[];
In a modern internet browser the javascript function ‘onkeydown’ is executed when focus is on a browser window and a key is pressed on the keyboard. It is executed with an argument called ‘a keydown event’. Such events are complicated because they are built to accomodate the entire operation of the browser with respect to a given website. Thankfully, all this complexity can be avoided by knowing that if the argument variable of the function onkeydown is set to is ‘e’ then the name of the key pressed is ‘e.key’.
Whenever I set up a keyboard thing with javascript I always end up writing something like this
onkeydown = function(e){
console.log(e.key);
}
This makes it so that if I put this code into a script and open to ‘about:blank’ and run the script then the name of each key I press after navigating to the window is logged into the javascript console. I end up doing this because I often forget the names of the modifier keys like ‘Shift’ or ‘Enter’. Just as often, I forget that they are sent through the onkeydown event handler. There are more formal ways of dealing with events, that’s where the phrase ‘event handler’ comes from, but I just go straight to ‘onkeydown =’ and then set it equal to a function I’ve defined. It is quick, short, simple, and gets things moving without having to dig into javascript lore.
A good language works without its lore.
For reasons that are beyond me, it is often necessary to include the following code if you want to intercept all of the keys that you might press on the keyboard e.g. ‘F1’, ‘F2’, and so on.
onkeydown = function(e){
e.preventDefault();
console.log(e.key);
I am told that adding ‘.preventDefault()’ to the onkeydown event stops the browser from sending the key pressed to the operating system (or some other program). Usually, pressing function keys does something with the browser like open a new tab. Preventing default actions stops this from happening. Stuff like pressing the key labeld ‘NumLock’ seems to still do what it always does. There are a lot of weird inconsistencies with input and output methods that I am told are largely beyond the control of us mere mortals.
The next thing that almost every keyboard program I’ve written does is to put the ‘e.key’ into a switch statement and then list the cases where a modifier key like Shift, Control, or Alt are pressed so that they immediately return out of the onkeydown function and nothing happens.
onkeydown = function(e){
e.preventDefault();
switch(e.key){
case'Shift':
case'Enter':
case'Control':
case'Alt':
case'Tab':
case'CapsLock': return;
}
console.log(e.key);
}
Note, pressing ‘shift’ still works to capitlize letters. For reasons that are beyond me, capital letters are sent as the name of the key pressed while holding down shift. Similar things are true in general in a way that I find fundamentally confusing. Instead of using a switch statement as I often do, there is another method I’ve been using that is growing on me. It starts by establishing an alphabet and returning if the result of ‘e.key’ does not belong to the alphabet. Here is a concrete example of what I mean:
var abc = '.;:,0123456789abcdefghijklmnopqrstuvwxyz '
onkeydown = function(e){
e.preventDefault();
var abcIndex = abc.indexOf(e.key);
if( abcIndex < 0) return;
console.log(e.key);
}
This tends to work out well because setting up your own alphabetical order often simplifies the programming environment. Since not every language has a procedure like ‘indexOf’ it can be implemented with a while loop as follows:
var abcIndex = abc.length;
while( abcIndex-- && abc[abcIndex] != e.key);
There are some alternatives that do the same thing:
var abcIndex = abc.length;
while(abcIndex >= 0){
if(abc[abcIndex] == e.key) break;
--abcIndex;
}
var abcIndex = 0;
while(abcIndex < abc.length){
if(abc[abcIndex] == e.key) break;
++abcIndex;
}
var abcIndex = abc.length;
while(abcIndex--)
if(abc[abcIndex]==e.key) break;
Whichever way works.
The basic operations: dup, drop, push, pop, swap, nil, pair, part, get, set, mand. I’ll allow them to be input at the keyboard in native javascript strings.
var ops=['dup', 'drop', 'push', 'pop', 'swap', 'nil', 'pair', 'part', 'get', 'set', 'mand']
code.split(' ').map(w=>ops.indexOf(w))
I was unable to finish the thinking that started here. It wasn’t moving anything forward. It was just rehashing things I’ve done before but in a worse way. I can really only work on something old if there is something new that I have to say about it, even if it is really sublte, as long as it’s new I can usually get myself to finish what I started.
In some rare cases I can get myself to paint a clearer picture of what I’ve already done before, but this is not an example of that.
My plan is to think about this in my wonderful rocking chair and see if there isn’t something that requires me to finish this in order to figure it out. Otherwise I’ll keep looking for something newly simplifying to add here.
What should everyone know? Is there such a thing as “what everyone should know”? Maybe a similar question is easier to answer: what are the fewest things that someone needs to know in order to live a good life? Rather than focusing on the problem of specifying “good” like the philosophers of days yore, let us focus on the stuff that people do that helps them out the most throughout their life and work.
Should work be included in this contemplation? A lot of work doesn’t seem to really be about doing well in life. Anyone who has had a job when they were younger knows just how dumb work can be. Talk to enough people and you start to get the idea that jobs are really just the only thing they know how to do in order to live in the world. Are jobs the only things that make it possible to live in this world?
What is “this world”? Someone who lives in India may seem to be in a different world from someone who lives in Japan or Argentina. The Germany of years ago is not the Germany of today. They are practically different worlds.
History seems to be an important part of living well. If you don’t know about the history of Germany, Argentina, Scotland, Iran, and all the other parts of the surface of the Earth that fought to name, then how are you going to know the ways that people have lived and if you’re living well relative to that?
I’m not aiming here to write some self contained exposition or argument. I’m writing here to discover what I have to say about the things I rarely speak about publicly. It’s a shame that I haven’t spent more time explaining my thinking about the world in this public way. So much time has been spent on explaining what I’ve uncoverd about logic and programming. What a narrow slice.
To make this problem even more concrete I can ask “What would I want my kids to know?” or “What do I want my nephews to know?”. I have a lot of nephews, and they don’t have access to the same kind of education that I had when I was their age. It’s important to remember that I had no choice in what family I was born into. There was no me to be born, much less a me that could choose to be born here rather than Tibet, now rather than the middle ages.
There is something helpful about being born here and now. This being my only access point to effective action really limits what I have to contemplate. Even if I take all of history into account, there is nothing I can do to change the past. I can in a way change the past by perhaps uncovering something about it that is inconsistent with what we knew of the past in the past. Such details only becloud the problem to solve.
I’ve already listed in the first hint the major divisions of the world as I’ve uncovered them. They are logical, mathematical, physical, chemical, biological, behavioral, and cultural. The behavioral parts of the world are the ones that we are most often in contact with. We are always in the presence of a behaving organism, our self. The people around us behave. What we are likely to do is effected by tons of people whose behavior can have consequences for whole populations of people. War on any scale pits people against people, behavior against behavior.
I’d like for everyone in the world to avoid war as often as possible. It’s not even clear if war is a necessary part of a world filled with meaningful life and work.
One relevant problem here is that I only have my contact with the world to draw upon when examining the scope of what people are more or less likely to do. I may read a book written to report on how one person lived an exceptional life, be it autobiography or biography. But my contact with this reports is not independent of my personal history. It is a sliver of a sliver of the events behing reported. Who knows what is accurate and what is inaccurate? As history slips away our grasp of the moment is lost. We scratch marks on stones and throw them into the sea of time. The wash up, or some of them do, and then fewer still are picked up, even fewer still are examined and shared with the world.
The internet was suppose to connect us with worlds of knowledge that were nothing like the world we had come to know before the internet. Did this happen? Is the world we are in touch with that different from the one we once knew before there were portals to every person in our pocket?
While I may say that logic and math have helped me live a meaningful life, how do I know that? An uncontemplated life is worse than an unexamined one.
These disorganized thoughts do not self assemble. They have to be put together. Collected, organized, and revised.
One thing the internet has helped more and more people to understand is that there is more to know than can be known. Our lives are shaped more by what we do not know than by what we know. There is no preperation that will tame the detailed knowledge that has been accumulated by even the tiniest of the sciences, much less all the other things that go on unseen on the surface of the Earth.
This state of affairs is not new. Throughout history there has been more to know than could ever be known. The question looms as to whether it is worth knowing this thing rather than that thing. How do we pick and choose what to know and what not to know? Do we even do that? Can we pick and choose what we know and what we don’t know? How would we know?
Most people who read whatever this note has become should be rolling their eyes. These are the kind of questions that are basically nonsense. We don’t have time to think about such things. There’s more to do than contemplate what we might otherwise do. Effective action happens now, not after some protracted contemplation.
It is strange that I can write about these things without having any specific part of my educational history that hash dealt directly with these problems. Education is supposed to make a better person. The educated person’s behavior is more effective than the uneducated person’s behavior in almost every way. (Is this true?)
This is where people tend to find religion as far as I’ve been able to see from reading on history. Philosophy was supposed to help answer questions like those brought up here. It doesn’t seem to have made much progress or come to any helpful conclusions. There is that old story that it is supposed to help us doubt that we have come to the conclusions presented by past people. A sort of generational hint that settling for a way of life is a fast way to living a sad life.
Crack open a religoius text and you’ll find stories about people living and working together. You’ll learn about how they dealt with disagreement and how they came to blows or bonded over shared agreeement. The world intervened through gods or through people under the control of gods. The things we didn’t know about the world were things that the gods knew and we coudln’t know the gods as we might know each other. Philosophy doubted that we even knew ourselves. Later, the behaviorist showed us just how little we know of ourselves. The one may have preceeded the other, but was it knowledge?
Back to the concrete question: what do my nephews need to know in order to live a good life? I can weaken it a bit by saying “What do I wish my nephews knew?” or keep it self contained by asking “What do I wish I knew when I was their age?” or perhaps even “What do I wish I knew now?”. Perhaps a simpler yet still complex question is “What do I know?” Can I answer that question and some how get at an answer to the question “What do I wish I knew?” or “What do I want to know?” or “What can I know?” or “What am I likely to know?”. After enough of these questions a common one appears: why does any of this matter? What do any of these questions about knowledge have to do with my daily life or the life and work of my nephews?
Why should I care? Why do I care? Do I care? These questions are all expressions of doubt, if they are expressions at all.
I don’t think that there is anything wrong with fumbling like this in public. The world is like a blizzard. Its like a foggy hailstorm. Things are pelting me from all directions, they are what we call thoughts and feelings.
Looming behind all of this are the scientific methods that I have been taught e.g. logic and radical behaviorism. These social practices have taught me to approach all these questions in an entirely different way than how I’ve approached them thus far.
Logic aims me at the basic predicates which are compounded into the premises of a theory of knowledge, a theory of the good life, a theory of teaching and learning.
The problem with the methods of logic and radical behaviorism is that they are not familiar to most people. If I am writing for soemone other than myself then I can not take logic and radical behaviorism as going concerns. My nephews don’t know much about logic and radical behaviorism besides whatever I’ve brought up whenever they visit. Sometimes they listen with interest to what I have to say, but most of the time they just laugh it off as “mean” uncle John being “mean” uncle John. It is a term of endearment that I’m pretty proud of. Nothing is funnier than getting a heartfelt hug while being told “I love you mean uncle John.”
If I start writing about what can be known about logic, math, or physics, then I’ll never answer the bigger question about what my nephews need to know in order to live and work well in this ever changing world. Why does anything I might have to say about living and woroking well in the world even matter. I don’t live or work well in the world!
I may know some things that not many people know, but that doesn’t translate into living and working well. I’m puzzled by the mere question “How are you doing?” or “Are things going well for you?” I don’t interpret the things that occur to me in ways that lend themselves to quick, simple, and accurate answers to such questions. My writing here is itself clear enough evidence of that.
I forgot to put on socks and now my feet are cold.
I’ve been thinking a lot about my programming environment over the past few days and haven’t had any time to sit down and write about it. Sadly, this means that some of what I’ve thought may have been lost. I don’t have a perfect memory and I don’t wish to have one. My memory is already good enough to get me in trouble. The less trouble that comes my way the happier I’ll be.
There are two directions I can proceed with my thinking: down and up. Thinking down gets me closer to what is the “virtual machine” or “abstract engine” running each of the little mechanisms that make up my programmable environment. Thinking up takes me closer to the realm of “metacircular evaluators” or writing an interpreter that can interpret itself (as far as that actually goes, more on this later).
Both of these directions are informative. Going down simplifies one way and going up simplifies in another way. Strangely, these two ways are complementary or are perhaps even two perspectives on one way. Going down uncovers what basic operations are most effective, and these basic operations contribute greatly to how an interpreter that interprets itself is designed.
To say “an interpreter that interprets itself” is to say that you can write a program that runs itself as a program. How it runs itself is ultimately explained by the abstract engine churning away either as a concrete instance in the real world or as some computational essence (of which I have nothing to say).
The large single function that I keep working on is what most people would call the interpreter of my programming environment. It takes a key press at a time and based on whether it is a letter or punctuation does the appropriate thing. The punctuation is treated abstractly by me as ‘color’. Words have a spelling and a color. I could just have well say that words have a spelling and a punctuation, but that doesn’t get at the abstraction I’m aiming at.
I found a better way of putting it all together that hopefully simplifies everything.
These are fragmentary notes and do not appear to make complete sense. I’m leaving them here as I’ve left others. I just want people to see my thinking as accurately as possible.
The programmable program. Programs are lists of words. Words have spellings and colors. Spellings are lists of letters and colors are gray, red, green, or blue. A word is made by spelling it out and then coloring it. A program is edited by inserting words in it or deleting words from it.
When a word is colored blue the program executes in the following way. Starting from the blue word and going back word by a word a search begins. The search stops when a red word with the same spelling as the blue word is found. If no matching red word is found then exectution stops.
Coloring a word blue is like pressing play on the program. A search begins. It starts with the blue word and goes backwards word by word. It stops when a red word with the same spelling as the blue word is found, or when there are no more words left to search through. If no matching red word is found then execution stops and the programmer is free to edit the program as they see fit. Otherwise, execution begins from the matching red word and goes forwards word by word.
What happens when a word is executed depends on the color of the word. If a word is red or blue then nothing happens i.e. red and blue words do nothing.
If no matching red word is found then execution stops and the programmer is free to edit the program further.
If a matching red word is found, then execution
If the matching red word is found then execution moves forward word by word from there.
A program is made by navigating through the list of words to where you want to make edits and then inserting or deleting words from there. When a word is colored blue
I’ve got to sharpen my chisels. They have not been sharpened since they were bought by my father some years ago. I’m quite fond of Paul Seller’s websites and videos. They can all be accessed from his main one https://paulsellers.com/other-websites/.
Why chisels? Somehow I got it into my head that I needed to make a clock out of wood. Like a dope, I want to do it without power tools. Humans have been working with wood much longer than they’ve been working with electricity. My general interest in the history of technology has brought me closer to the problems solved by new technologies, and I want to be brought closer still.
A schedule is a list of things to do in a specific order. Once a schedule is made it may be changed depending on what turns up on it or depending on a change in circumstances. Programs are schedules of action. First do this, then do that, and if this other thing happens do this other other thing.
The general method of following a schedule is to do the next thing on the list until the list is empty. In javascript, functions describe actions to take. If the name of a function is ‘act’ then you make a program that takes that action by appending a closed pair of parentheses ‘act()’. If a schedule is a list of actions, i.e. functions, then following the schedule means taking each action in the list until the list is empty.
In javascript, the schedule is defined first as an empty list:
var scheduled= new Array();
To schedule something, we push it onto the end of the list e.g.
scheduled.push(it);
The first item of the list can be shifted off by
scheduled.shift()
So if the first item on the schedule is an action, taking that action is accomplished with
scheduled.shift()()
The two pairs of parentheses can be confusing but you can get rid of the confusion by introducing an intermediate variable like this
var act = scheduled.shift();
act();
This shows clearly enough that the first item on the schedule is shifted off the list and assigned to the variable named ‘act’ so that ‘act()’ takes the action shifted off the list.
All together, a schedule is followed by taking all the actions on the list from first to last until there are no more actions in the list. This is accomplished with a while loop as follows:
while(scheduled.length) scheduled.shift()();
If that’s too cryptic, details can be added to make it crystal clear:
while(scheduled.length > 0 ){
var act = scheduled.shift();
act();
}
The phrase “scheduled.length > 0” can be eliminated since shifting from the top of the list shortens the length by one and when the condition in a while loop is not zero it keeps going and stops when it is zero.
The only thing missing from our scheduler is a method of adding actions to the schedule. There is something else that needs to be considered: actions which schedule new actions! This is an important part of any effective plan that can be scheduled. You schedule a meeting to make a decision based on what is known at that time and usually schedule new things for a later time based on conclusions made during that meeting.
As we know all to well, there’s always a next meeting until the company folds at some unknown future date. So we have to leave open the possibility of new actions being added to the schedule at any time, even a time after which the schedule has already been emptied. This happens when a committee meets for a short time, comes to all the conclusions it can, and need not meet again, until suddenly something happens and the committee is suddenly brought into existence again.
So the action of scheduling has to work well with following the actions already scheduled. This is accomplished by making a flag that is raised when a schedule is being followed, and lowered when the schedule is empty. If the flag is up and you schedule a new action it gets added to the end of the list of actions scheduled and returns to following what was scheduled. If the flag is down and you schedule an action then after pushing the newly scheduled act onto the end of the list you start following the schedule.
This is accomplished as follows
var scheduled = [];
var following = false;
function schedule(it){
scheduled.push(it);
if(following) return;
following = true;
while(scheduled.length) scheduled.shift()();
following = false;
}
The flag named ‘following’ lets everone who wants to schedule things know that the schedule is being followed and that it doesn’t have to be started anew.
All together, this code is called a scheduler. It is the simplest of schedulers. Nothing is done to compare the action about to be scheduled with the actions that are already scheduled. If you work in a registrars office at a college then you’re constantly having to check how a schedule might be better designed to make sure everything gets done that can get done across the entire college e.g. making sure the students and teachers are free to be where they want to be when they need to be.
Most people think that the program of the scheduler has to be changed when these more sophistocated methods of scheduling are contemplated. This is just not true. The simple scheduler is all that is ever needed no matter how complex the method of scheduling is to be. This is accomplished by setting up actors whose actions are coordinated by scheduling actors as speakers and listeners that are part of complete verbal episodes.
Such verbal episodes of speaking and listening are scheduled just like any other acts, and as long as each actor can raise or lower a speaking or listening flag the problem of effective scheduling can be entirely offloaded onto the person writing the actors’s scripts.
I have done a lot of thinking over the past few days and have not done a good job of recording it here. This seems to be a pattern, and one which I would like to extinguish as soon as possible. The more thinking I do off the record the less likely I am to uncover better ways of thinking and doing.
One of the reasons that I haven’t recorded much is because the thinking has been weak. I have not seen to myself the path that takes me from where I was at to where it appears to me that I’m going. This is a general problem when planning, deciding, and solving problems.
My programming environment is set up with javascript for now. The transition from a monolithic interpreter to a miniature virtual machine looms. The change is to be the result of what I’ve called ‘slotted instructions’. These are what others call ‘microinstructions’. They are a key part of how modern computers work. A single instruction is composed of a sequence of microinstructions. The program pointer is used to fetch an instruction, that instruction is then decoded into a sequence of microinstructions. Those microinstructions are then executed by a process that almost completely mirrors the larger machines fetch-execute cycle.
A microinstruction pointer goes through each of the microinstructions and this orchestration ultimately controls the dance of electrons that make up the operations of the machine. As a very concrete example, most hardware has a main line that is connected to all the places where different electrical potentials are set up as in the classic 8 bit registers. A register is just a storage place whose contents can be rearranged to match those of another storage place or can be copied by the rearrangement of some other storage place to match its contents.
The main line is also called ‘the bus’. Places linked by the bus and open to its influence tend to end up with similarly arranged potentials. Technically, most of the storage places have an input side and an output side. When you connect one output to the input of another across the bus then the input place ends up matching the output place. This schedule of outputs to inputs across the bus is coordinated by microinstructions and these microinstructions are packed into machine instructions.
This packing process is actually a part of any effective programming environment. It is often treated as if it is only and solely a hardware problem, as if in an ideal world there would be no use for such complex or compound instructions. This is no more apparent than when introducing ‘control flow’ operations that change what operation is going to occur next based on something else changing for some other reason.
The work I can do on the javascript implementation of my programming environment is not as clear to me today as it has been in the past few days. Thankfully, I’m more willing today to give up on the precision that has crippled me when dealing with things outside the narrow worlds of predicate logic and behavior science.
The most interesting thing that occurred to me that solves a problem present when transitioning from the largely native methods of control by javascript to something more akin to the control simulated by a virtual machine, is conditional execution of this but not that part of a program.
In a past note I described this problem as making a machine that can choose what it does next based on what it is presented with now. I also said, perhaps somewhat earlier than that, that the operation of ‘mux’ which is now called ‘select’ and which selects one or the other of two items below the top of the stack depending on whether it is Nil or not, is inadequate to set up the kinds of conditional execution that makes programming so much easier than its nonconditional alternatives. Perhaps it was not so.
Something about the shift from a single item on the list containing the program to be executed (recall that it is like a tape that can be moved back and fore and that there is a top word such that each word before it is part of the back part of the tape and each word after it is part of the fore part of the tape) to the two item method that not only simplified compilation but also the rest of the basic operations like ‘find’ and ‘blue’, unlocked a clean enough way of implementing conditional execution precisely as this kind of metaphorical choice.
I’m so compelled by this method, that I may as well rename ‘select’ as ‘choose’ in a hope that it might eliminate ‘choice’ from its nasty use as a hallmark of autonomous humans.
Recall, while mediating a key press, my programming environment has the back of the current program at the front of the list and the fore of the program at the second from front of the list. The method I propose for dealing with conditional execution is to select one from two programs each of which ends in a return to the original program it was just compiled from.
I’ll go ahead, out on a limb, and take precisely those fragmentary note which occur to me while writing things with pad and pen. These notes do not always make sense outright and have to be edited at some later time to form that level of precision that a computer can work on.
What is needed are a few blue words. This is how all conditionals are actually implemented. The construction of a conditional is going to, for now, start with the blue ‘if’ i.e. with ‘if.’. This word has to set up a new blank program that words are going to be compiled into. Thus it has to invovle something like `nil(); nil(); pop(); pop();’ but note that now the ‘pop’s are primitive operations and not the gray pops that pair with the gray pushes and the gray return that are needed in order to assure that schedule of words to be executed are not garbled by stack manipuations. And, yet, it is careful garbling of that very sort which is called for here now.
I’ll sort out ‘if.’ soon. It appears to come with another important step in the evolution of these things: bringing compilation under programmable control.
If there is an ‘else.’ this sets up a new blank program with something similar to when ‘if.’ is input i.e. something again like ‘nil(); nil(); pop(); pop();’. Then ‘then.’ has to seal everything up back into the program from whence these conditional digressions came.
There is a general solution to problems posed by introducing programs like this: slotted instructions. The design of everything up to this point has been such that one word is taken at a time and the appropriate action is taken with respect to it without making a distinction between the program being executed and the program that was compiled.
In the standard design fo modern computers these distinctions are made by speaking of the instruction register and the program pointer and main memory. The way things are set up right now, there’s just a program pointer. Technically there are two: one points to the back of the program and the other points to the fore of the program. All the logic involved in setting up an instruction register is offloaded to the javascript environment. It is implicit in the way that functions are called and return when writing as, e.g., ‘push(); pop();’.
The method of slotted instructions makes overt the microinstruction pointer and microinstructions that make up most modern machines. Since there are no variables though, the kinds of microinstructions that can be constructed are often surprisingly powerful for how tiny they may other be. Effective microinstructions make it so that the fetch cycle of cpu operation can often been entire avoided. They are often expensive, relatively so. But, here I am uncovering them as a method of making complex primitive operations, primitive relative to the instructions of the machine and not relative to the microinstructions of the machine.
If this doesn’t make sense, that’s because it doesn’t yet make complete sense to me. I can see how it works to myself but it is not snappy like a carefully designed LEGO set.
Making way for such microinstructions without leaving the basic operations on the tree that have been set up thus far is easy enough. It continues the method of gray pushes and gray pops except that now gray words are themselves composed of a microprogram. In this case, something like a gray push is implemented by a microprogram as follows:
push, push, push, push,
swap, pop, swap, pop, swap, pop,
Note, that I’m only half imagining these as operations of a digital machine. It is helpful to do this, for me, and much of where it came from was studying the design of the F18A Technology Reference from GreenArrays https://www.greenarraychips.com/home/documents/greg/DB001-221113-F18A.pdf.
These things are neglected by modern chip designers who descend from the standard CMOS traditions because simplicity and programmatic countercontrol are rarely profitable. Metaphorically, computer chips are the APIs of hardware. The only difference, and it is a big deal, is that it was once harder for chip makers to know how people were using their API.
These micro programs give the programmer control over the execution of the machine as if each of its commands are coming from an external programmer and not from consulting some local record of a program to be executed.
An interesting way of implementing this just occurred to me. The key ‘,’ is the color gray, but it can just as well be the spelling of a blue word which is a micro compile (or something like that). Microcompilation would occur only with respect to gray words, but in this case, gray words would be only and solely the basic operations, they would not be a secondary definition as they are in the current state of the implementation of my programming environment.
This new method can be interpreted as defining ‘execute list’ where each item of the list is a primitive operation, which is now firmly a gray word. This is not unfamiliar to LISP programmers or FORTH programmers as they both have their own version of ‘execute a sequence of primitive operations’. In the FORTH case it is dropping into machine code, and in the LISP case it is evaluating a list of expressions.
Need a snack.
My methods of taking notes here has evolved as everything does. I used to spend a lot more time connecting together what I have already done here with what I’m doing in any given note. This has gone extinct recently. I’ve been working almost exclusively on implementing my programming environment in javascript.
It’s still not entirely clear to me why this is, but when I get hooked on something in this way I usually give it all I can. This is a consequence of my personal history with solving problems. You go with what grabs you because at that moment it is perhaps all that you are likely to ever contribute to the world. That moment of excitment and endurance is priceless. It can fuel you for years.
The edits I’ve been making to the implementation that has evolved thus far are tiny. If you are insensitive to such itty bitty things then it looks like I’m repeating the same things over and over again. So be it. I repeat many things. Usually the things repeated are the only things that sorta work e.g. I add things with the same results I’ve gotten since I was taught my addition and multiplication tables as a young lad.
One thing I’m unhappy with in these latest notes is that they only record a certain kind of work that I do. It is the work that has prevailed throughout my life. It is work which requires a kind of careful precision that is otherwise debilitating. It is also work which appears unavoidably necessary when it occurs. These are all very dangerous things, but the worse is not recording the other kinds of work that I do. These paragraphs are part of that other kind of work.
In the past few months I have been lucky enough to have a surprisingly comfortable wooden rocking chair from an Amish community. When I was younger I had time to dream. That time got stolen by the cultures that maintain this world filled with slavery, be it by the whip or wage. This new rocking chair has brought back the better parts of those early dreams. The whip and the wage are far from me now. I’m unable to articulate how much I am helped by things which are far beyond my control.
By putting on my noise canceling earmuffs and rocking in that Amish chair, I’ve gotten back some of what was lost in my youth. It is no surprise to me that William Wordsworth’s “Ode On Intimitations from Early Childhood” spoke to me even as a high school student. I didn’t know then just how strange a thing it is to believe in the beauty of your dreams.
But, I notice that my belief is not delusional in the same way that others are. As if they must convince themselves of the beauty of their dreams rather than that beauty simply occuring to them. This is my ego speaking now, surely. I’ll let myself speak of egos, dreams, beauty, and all that stuff that naive behaviorists and those who condemn psychology as a science of behavior say are entia non grata. These are my notes after all. They merely report on my as yet unnoticed mistakes.
Notes are as much for “streams of consciousness” as they are for that hard work that makes real things happen now.
Yesterday and today a number of things occurred to me with respect to the construction fo my programming environment in javascript:
how to deal with conditional execution (forward and backwards) without breaking the lattice structure i.e. without introducing graph theoretic cycles into the ‘data structure’ (perhaps one of my least favorite technical terms since it combines my hatred for ‘data’ and ‘structures’, not that they weren’t helpful stepping stones, but we are beyond all that junk).
By changing the method of compilation I can shorten everything about the implementation wtihout changing its shape. I will go into detail on this point because there is a lot to be learned from it.
I can use the notation I’ve adopted (where words end with ‘,;:.’) to explain things now that it has kinda settled down.
I’ve been so focused on this that I’ve neglected the other things that I’m working on e.g. woodworking (I got my sharpening stones) and my modern analysis of Russell and Whitehead’s “Principia Mathematica”.
I’m not looking for “maxwell’s equations of computing”. I’m looking for “the lathe of computing”.
I can include the slow steps from the javascript that is more on the javascript side than my programming environments side and vice versa inside the single ‘coin operated’ definition of ‘onkeydown’ and perhaps make further simplifications that tend to appear when you put everything together.
I’ll just push through as I always do. So staring with simplifying compilation of a word and going on to simplify the rest of the program. Starting from the definition of ‘onkeydown’:
onkeydown=e=>{
switch(e.key){
case'Shift': case'Enter': case'Control': case'Alt': case'Tab': case'CapsLock':return;
default: T=Pair(Pair(Left(T), list(e.key)), Right(T)); // input
dup(); part(); nip(); part(); nip(); part(); nip(); not(); // is letter?
if(isTopNil()){ drop(); pair(); // append to spelling
}else{ drop();
over(); over(); pair(); // make word
push(); part(); bury(); swap(); pair(); swap(); pair(); pop(); // compile word
part(); nip(); part(); nip(); not(); // blue?
if(isTopNil()){ drop();
nil(); nil(); pair(); pair(); //greenify
dup(); not(); // not empty word?
while(0<--limit && isTopNil()){ drop();
push(); part(); part(); bury(); pair(); swap(); pair(); pop(); // fore
part(); // gray?
if(isTopNil()){ drop();
part(); nip(); // get letter quoting gray word
part(); nip(); part(); nip(); part(); nip(); // left();left();left();
part(); nip(); if(isTopNil()){ drop(); dup(); }else{ // gray dup
part(); nip(); if(isTopNil()){ drop(); drop(); }else{ // gray drop
part(); nip(); if(isTopNil()){ drop(); push(); swap(); pop(); pop(); }else{ // gray pop
part(); nip(); if(isTopNil()){ drop(); push(); push(); swap(); pop(); }else{ // gray push
part(); nip(); if(isTopNil()){ drop(); swap(); }else{ // gray swap
part(); nip(); if(isTopNil()){ drop(); nil(); }else{ // gray nil
part(); nip(); if(isTopNil()){ drop(); pair();}else{ // gray pair
part(); nip(); if(isTopNil()){ drop(); part(); }else{ // gray part
part(); nip(); if(isTopNil()){ drop(); push(); drop(); }else{ // gray return
drop();}}}}}}}}}
}else{ part(); nip(); // green?
if(isTopNil()){ drop();
nil(); nil(); nil(); pair(); pair(); pair(); //redify();
push(); dup(); pop(); // get program to search
nil(); // prime the pump // beginning of find
while(0<--limit && isTopNil()){ drop();
part(); swap(); part(); unbury(); pair(); pair(); // back();
over(); over(); part(); drop(); part(); nip(); // over2();left();right();
nil(); pair(); pair(); pop(); // put trees in new list // beginning of id
nil();// prime the pump
while(0<--limit && isTopNil()){ drop();
push(); part(); part(); pop(); // get trees to check
over(); over(); and(); // both nil?
if(isTopNil()){ drop();
drop(); drop();// nils are identical
push(); dup(); pop(); not(); // more to check?
}else{ drop();
over(); over(); or(); // is one nil?
if(isTopNil()){ drop();
push();pair();pair();dup();pop(); // not identical
}else{ drop();
part(); unbury(); part(); unbury();// break into their left and right parts
push(); pair(); pair(); pair(); pair(); pop(); // put on list
nil();}}} // may be identical
drop(); push(); // end of id
over(); part(); drop(); or(); not();} // match or no more?
drop(); nip(); // end of find
pop();
}else{ drop(); drop();}}
push();dup();pop(); part(); drop(); part(); nip(); // get next word
dup(); not();} // not empty word?
drop(); drop(); // drop the not-Nil and the empty word // end of next
}else{ drop(); drop();}
nil();}
show();break;}}
In this version of the lathe I’ve paired up the back and fore of the program into a left and right part which mirror the organization of the entire tree itself. This seems to have aided me in completing this design, but it can be simplified. The simplification followed from taking a look at how ‘compile’ is defined
push(); part(); bury(); swap(); pair(); swap(); pair(); pop(); // compile word
That entire line can be transformed into just three basic operations:
push(); pair(); pop(); // compile word
This is easily accomplished by breaking apart the back and fore of the program so that they are simply two different items on the list. How to break them apart is revealed by following through from this simplest of definitions of compilation.
Since push, pair, pop, are all that it takes to compile a word to the program then the item at the front of the list is the back of the program and, presumably, the item second from front of the list is the fore of the program. This also sets the orientation of the back of the program: the further you go down the right hand side of the back of a program the further back into the program you go i.e. each word of the program is a left part of the item at the front of the list of the tree.
This is all so complicated only because such a simplification didn’t occur to me from the beginning. That is the problem with selection by consequences from variation as a principle: it just doesn’t work in the ways you’d expect based on how science tends to explain the world. Is science wrong? No, it’s just what you get when you “act with vigor without absolute certainty”.
This simplification leaves the orientation of the fore of the program unconstrained. The two operations which firmly fix that orientation are ‘back’ and ‘fore’. Since they only appear at one point in the entire program, there are only their two contexts that need to be contemplated when picking an orientation of the fore of a program.
The definition of ‘fore’ is
push(); part(); part(); bury(); pair(); swap(); pair(); pop(); // fore
the beginning ‘push()’ and the end ‘pop()’ are not technically part of the definition of ‘fore’. They follow the general pattern I’ve mentioned once or twice before: every program has a beginning, a middle, and an end, just like a good story, and often the middle repeats itself depending on what problem is being solved.
In this case, the middle can be changed to
part(); pop(); swap(); pair(); push();
that is, assuming the top two items on the stack are the program, the top being the fore of the program and the second from top being the back of the program. All together the new line reads
push(); push(); part(); pop(); swap(); pair(); pop(); // fore
which may not be exactly what was expected. Notice it is wonderfully simple in comparison to the previous one, being not only made entirely out of primitive operations but also one less operation overall.
A similar simplification occurs when examining ‘back’. It starts as
part(); swap(); part(); unbury(); pair(); pair(); // back();
but, is transformed into the opposite of the newly defined ‘fore’ as
pop(); part(); swap(); push(); pair(); // back()
The reason that ‘back’ is so much shorter than ‘fore’ is because it isn’t. It just so happens that at the point where ‘back’ occurrs it doesn’t need a beginning or an end.
To be clear, the definitions of fore and back can be sketched as
fore: part, pop, swap, pair, push,
back: pop, part, swap, push, pair,
and here the symmetry is clear enough.
There is then a simplification for getting the top word of the program. This is carried throughout. This also requires a change in how gray pushes, pops, and returns work.
Oops. I solved a lot of problems that I didn’t record solving e.g. changing the definitions of the javascript probe functions that show the current state of the program being edited, eliminating the occurrence of ‘over(); over();’ in the beginning of the definition of ‘green’. Here is the current state of things:
var Nil = new Object();
function isNil(p){return Object.is(p, Nil);}
function Pair(l,r){return new Array(l,r);}
function Left(p){if(isNil(p)) return Nil; else return p[0];}
function Right(p){if(isNil(p)) return Nil; else return p[1];}
var T=Nil;
function isTopNil(){return isNil(Right(Left(T)));}
function dup(){T=Pair(Pair(Left(T), Right(Left(T))), Right(T));}
function drop(){T=Pair(Left(Left(T)), Right(T));}
function pop(){T=Pair(Left(Left(T)), Pair(Right(Left(T)), Right(T)));}
function push(){T=Pair(Pair(Left(T), Left(Right(T))), Right(Right(T)));}
function swap(){T=Pair(Pair(Pair(Left(Left(Left(T))),
Right(Left(T))), Right(Left(Left(T)))), Right(T));}
function nil(){T=Pair(Pair(Left(T), Nil), Right(T));}
function pair(){T=Pair(Pair(Left(Left(Left(T))),
Pair(Right(Left(Left(T))), Right(Left(T)))), Right(T));}
function part(){T=Pair(Pair(Pair(Left(Left(T)),
Left(Right(Left(T)))), Right(Right(Left(T)))), Right(T));}
function nip(){pop(); drop(); push();}
function over(){pop(); dup(); push(); swap();}
function dig(){over(); pop(); nip();}
function bury(){dig(); dig(); push(); push();}
function unbury(){bury(); bury();}
function select(){if(isTopNil()){drop(); nip();}else{drop(); drop();}}
function not(){nil(); nil(); nil(); pair(); unbury(); select();}
function or(){dup(); select();}
function and(){not(); swap(); not(); or(); not();}
onkeydown=e=>{
switch(e.key){
case'Shift': case'Enter': case'Control': case'Alt': case'Tab': case'CapsLock':return;
default: T=Pair(Pair(Left(T), list(e.key)), Right(T)); // input
dup(); part(); nip(); part(); nip(); part(); nip(); not(); // is letter?
if(isTopNil()){ drop(); pair(); // append to spelling
}else{ drop();
over(); over(); pair(); // make word
push(); pair(); pop(); // compile word
part(); nip(); part(); nip(); not(); // blue?
if(isTopNil()){ drop();
nil(); nil(); pair(); pair(); //greenify
dup(); not(); // not empty word?
while(0<--limit && isTopNil()){ drop();
push(); push(); part(); pop(); swap(); pair(); pop(); // fore
part(); // gray?
if(isTopNil()){ drop();
part(); nip(); // get letter quoting gray word
part(); nip(); part(); nip(); part(); nip(); // left();left();left();
part(); nip(); if(isTopNil()){ drop(); dup(); }else{ // gray dup
part(); nip(); if(isTopNil()){ drop(); drop(); }else{ // gray drop
part(); nip(); if(isTopNil()){ drop();
push(); swap(); push(); swap(); pop(); pop(); pop(); }else{ // gray pop
part(); nip(); if(isTopNil()){ drop();
push(); push(); push(); swap(); pop(); swap(); pop(); }else{ // gray push
part(); nip(); if(isTopNil()){ drop(); swap(); }else{ // gray swap
part(); nip(); if(isTopNil()){ drop(); nil(); }else{ // gray nil
part(); nip(); if(isTopNil()){ drop(); pair();}else{ // gray pair
part(); nip(); if(isTopNil()){ drop(); part(); }else{ // gray part
part(); nip(); if(isTopNil()){ drop();
push(); push(); drop(); drop(); }else{ // gray return
drop();}}}}}}}}}
}else{ part(); nip(); // green?
if(isTopNil()){ drop();
nil(); nil(); nil(); pair(); pair(); pair(); //redify
push(); dup(); push(); dup(); pop(); swap(); pop(); // program to find in
pop(); pop(); nil(); // prime the pump // beginning of find
while(0<--limit && isTopNil()){ drop();
dup(); // word to find
push(); part(); swap(); push(); pair(); // back
pop(); dup(); pop(); part(); drop(); // get next word
nil(); pair(); pair(); pop(); // put trees in new list // beginning of id
nil();// prime the pump
while(0<--limit && isTopNil()){ drop();
push(); part(); part(); pop(); // get trees to check
over(); over(); and(); // both nil?
if(isTopNil()){ drop();
drop(); drop();// nils are identical
push(); dup(); pop(); not(); // more to check?
}else{ drop();
over(); over(); or(); // is one nil?
if(isTopNil()){ drop();
push();pair();pair();dup();pop(); // not identical
}else{ drop();
part(); unbury(); part(); unbury();// break into their left and right parts
push(); pair(); pair(); pair(); pair(); pop(); // put on list
nil();}}} // may be identical
drop(); push(); // end of id
push(); dup(); pop(); or(); not();} // match or no more?
drop(); drop(); // end of find
}else{ drop(); drop();}}
push(); dup(); pop(); part(); drop(); // get next word
dup(); not();} // not empty word?
drop(); drop(); // drop the not-Nil and the empty word // end of next
}else{ drop(); drop();}
nil();}
show();break;}}
var limit=1<<10;
var abc = ',;:.0123456789abcdefghijklmnopqrstuvwxyz'
function ln(p){return isNil(p)?0:1+ln(Right(p));}
function ls(n){return n<0?ls(abc.length):n?Pair(Nil,ls(n-1)):Nil;}
function n(a){return abc.indexOf(a);}
function a(n){return abc.length<=n?'?':abc.at(n);}
function list(a){return ls(n(a));}
function letter(p){return a(ln(p));}
function letters(p){return isNil(p)?'#':letters(Left(p))+letter(Right(p));}
function lists(s){return s==''?Nil:Pair(lists(s.slice(0,-1)),list(s.at(-1)));}
function lwords(p){return isNil(p)?'#':lwords(Right(p))+letters(Left(p));}
function rwords(p){return isNil(p)?'#':letters(Left(p))+rwords(Right(p));}
function prog(){return '{'+lwords(Left(Right(T)))
+'['+letters(Right(Left(T)))+']'+rwords(Left(Right(Right(T))))+'}';}
document.body.style.font='12pt monospace';
document.body.style.whiteSpace='pre';
function show(s=''){document.body.textContent+=`\t${limit} ${prog()} ${s}\n`;}
Somewhere while doing all this two things occurred to me that can further simplify everything: a) by quoting letters as the depth of a stack rather than the length of a list I can eliminate ‘nips’ in favor of ‘drops’ which are primitive and can eliminate the definition of ‘nip’ entirely without it being entirely arbitrary, and b) I should be able to eliminate unbury without making things too complicated. The benefits are huge in that they minimize the number of auxiliary terms that must even be spoken of when getting everything started.
I was able to eliminate nip, dig, bury, and unbury without simply replacing their later occurences by their definition. I was also able to program the input process as iterated execution of nils and pairs. I am too tired to do the same for the output, and the output has to be done differently anyway when it comes to dealing with that part inside my programming environment rather than half in and half out of javascript.
Here is the further simplified code. Note, this is what I could have started with if I had the foresight of all the experiments up to this point.
var Nil = new Object();
function isNil(p){return Object.is(p, Nil);}
function Pair(l,r){return new Array(l,r);}
function Left(p){if(isNil(p)) return Nil; else return p[0];}
function Right(p){if(isNil(p)) return Nil; else return p[1];}
var T=Nil;
function isTopNil(){return isNil(Right(Left(T)));}
function dup(){T=Pair(Pair(Left(T), Right(Left(T))), Right(T));}
function drop(){T=Pair(Left(Left(T)), Right(T));}
function pop(){T=Pair(Left(Left(T)), Pair(Right(Left(T)), Right(T)));}
function push(){T=Pair(Pair(Left(T), Left(Right(T))), Right(Right(T)));}
function swap(){T=Pair(Pair(Pair(Left(Left(Left(T))),
Right(Left(T))), Right(Left(Left(T)))), Right(T));}
function nil(){T=Pair(Pair(Left(T), Nil), Right(T));}
function pair(){T=Pair(Pair(Left(Left(Left(T))),
Pair(Right(Left(Left(T))), Right(Left(T)))), Right(T));}
function part(){T=Pair(Pair(Pair(Left(Left(T)),
Left(Right(Left(T)))), Right(Right(Left(T)))), Right(T));}
function over(){pop(); dup(); push(); swap();}
function select(){if(isTopNil()){drop(); swap(); drop();}else{drop(); drop();}}
function not(){pop(); nil(); nil(); nil(); pair(); push(); select();}
function or(){dup(); select();}
function and(){not(); swap(); not(); or(); not();}
var limit=1<<10;
var abc = ',;:.0123456789abcdefghijklmnopqrstuvwxyz'
onkeydown=e=>{
switch(e.key){
case'Shift': case'Enter': case'Control': case'Alt': case'Tab': case'CapsLock':return;
default: var k=abc.indexOf(e.key); if(k<0)k=abc.length;
nil(); while(k--){nil(); pair();}; // input
dup(); part(); drop(); part(); drop(); part(); drop(); not(); // is letter?
if(isTopNil()){ drop(); pair(); // append to spelling
}else{ drop();
over(); over(); pair(); // make word
push(); pair(); pop(); // compile word
part(); drop(); part(); drop(); not(); // blue?
if(isTopNil()){ drop();
nil(); nil(); pair(); pair(); //greenify
dup(); not(); // not empty word?
while(0<--limit && isTopNil()){ drop();
push(); push(); part(); pop(); swap(); pair(); pop(); // fore
part(); // gray?
if(isTopNil()){ drop();
part(); drop(); // get letter quoting gray word
part(); drop(); part(); drop(); part(); drop(); // left();left();left();
part(); drop(); if(isTopNil()){ drop(); dup(); }else{ // gray dup
part(); drop(); if(isTopNil()){ drop(); drop(); }else{ // gray drop
part(); drop(); if(isTopNil()){ drop();
push(); swap(); push(); swap(); pop(); pop(); pop(); }else{ // gray pop
part(); drop(); if(isTopNil()){ drop();
push(); push(); push(); swap(); pop(); swap(); pop(); }else{ // gray push
part(); drop(); if(isTopNil()){ drop(); swap(); }else{ // gray swap
part(); drop(); if(isTopNil()){ drop(); nil(); }else{ // gray nil
part(); drop(); if(isTopNil()){ drop(); pair();}else{ // gray pair
part(); drop(); if(isTopNil()){ drop(); part(); }else{ // gray part
part(); drop(); if(isTopNil()){ drop();
push(); push(); drop(); drop(); }else{ // gray return
drop();}}}}}}}}}
}else{ part(); drop(); // green?
if(isTopNil()){ drop();
nil(); nil(); pair(); nil(); pair(); pair(); //redify
push(); dup(); push(); dup(); pop(); swap(); pop(); // program to find in
pop(); pop(); nil(); // prime the pump // beginning of find
while(0<--limit && isTopNil()){ drop();
dup(); // word to find
push(); part(); swap(); push(); pair(); // back
pop(); dup(); pop(); part(); drop(); // get next word
nil(); pair(); pair(); pop(); // put trees in new list // beginning of id
nil();// prime the pump
while(0<--limit && isTopNil()){ drop();
push(); part(); part(); pop(); // get trees to check
over(); over(); and(); // both nil?
if(isTopNil()){ drop();
drop(); drop();// nils are identical
push(); dup(); pop(); not(); // more to check?
}else{ drop();
over(); over(); or(); // is one nil?
if(isTopNil()){ drop();
push();pair();pair();dup();pop(); // not identical
}else{ drop();
part(); pop(); swap(); part(); push();// break into their left and right parts
push(); pair(); pair(); pair(); pair(); pop(); // put on list
nil();}}} // may be identical
drop(); push(); // end of id
push(); dup(); pop(); or(); not();} // match or no more?
drop(); drop(); // end of find
}else{ drop(); drop();}}
push(); dup(); pop(); part(); drop(); // get next word
dup(); not();} // not empty word?
drop(); drop(); // drop the not-Nil and the empty word // end of next
}else{ drop(); drop();}
nil();}
show(); break;}}
function ln(p){return isNil(p)?0:1+ln(Left(p));}
function a(n){return abc.length<=n?'?':abc.at(n);}
function letter(p){return a(ln(p));}
function letters(p){return isNil(p)?'#':letters(Left(p))+letter(Right(p));}
function lwords(p){return isNil(p)?'#':lwords(Right(p))+letters(Left(p));}
function rwords(p){return isNil(p)?'#':letters(Left(p))+rwords(Right(p));}
function prog(){return '{'+lwords(Left(Right(T)))
+'['+letters(Right(Left(T)))+']'+rwords(Left(Right(Right(T))))+'}';}
document.body.style.font='12pt monospace';
document.body.style.whiteSpace='pre';
function show(s=''){document.body.textContent+=`\t${limit} ${prog()} ${s}\n`;}
I’m really happy with all the work that I’ve gotten done on explaining how my programming environment works as of late. THere are two things to do next. One is to write a new probe that shows the state of the whole tree so that I can test the execution of gray words to make sure they work correctly. The other is to work on putting everything into a single loop that, as far as I can see, is like running a virtual machine.
I’m going to go with the second of these options because it is the one that is most interesting to me right now. It also just so happens to be the one that involves the other essentially. The general principle is to expand the method of dealing with gray words and to supply the appropriate gray words to accomplish all that has been programmed in javascript thus far.
The method of programming the javascript that I’ve undertaken thus far is designed to aid in this significant step. The only thing that is missing from the operations that have been included as basic thus far is the gray operations that correspond to the javascript code ‘if(isTopNil()) xxx; else yyy;’ and ‘while(isTopNil()) xxx;’.
More debugging and rearranging. Here is where things are starting from, they are a bit different than where they eneded yesterday.
var Nil = new Object();
function isNil(p){return Object.is(p, Nil);}
function Pair(l,r){return new Array(l,r);}
function Left(p){if(isNil(p)) return Nil; else return p[0];}
function Right(p){if(isNil(p)) return Nil; else return p[1];}
var T=Nil;
function isTopNil(){return isNil(Right(Left(T)));}
function dup(){T=Pair(Pair(Left(T), Right(Left(T))), Right(T));}
function drop(){T=Pair(Left(Left(T)), Right(T));}
function pop(){T=Pair(Left(Left(T)), Pair(Right(Left(T)), Right(T)));}
function push(){T=Pair(Pair(Left(T), Left(Right(T))), Right(Right(T)));}
function swap(){T=Pair(Pair(Pair(Left(Left(Left(T))),
Right(Left(T))), Right(Left(Left(T)))), Right(T));}
function nil(){T=Pair(Pair(Left(T), Nil), Right(T));}
function pair(){T=Pair(Pair(Left(Left(Left(T))),
Pair(Right(Left(Left(T))), Right(Left(T)))), Right(T));}
function part(){T=Pair(Pair(Pair(Left(Left(T)),
Left(Right(Left(T)))), Right(Right(Left(T)))), Right(T));}
function nip(){pop(); drop(); push();}
function over(){pop(); dup(); push(); swap();}
function dig(){over(); pop(); nip();}
function bury(){dig(); dig(); push(); push();}
function unbury(){bury(); bury();}
function select(){if(isTopNil()){drop(); nip();}else{drop(); drop();}}
function not(){nil(); nil(); nil(); pair(); unbury(); select();}
function or(){dup(); select();}
function and(){not(); swap(); not(); or(); not();}
function green(){
nil(); nil(); nil(); pair(); pair(); pair(); // redify();
push(); dup(); pop(); // get program to search
nil(); // prime the pump // find
while(0<--limit && isTopNil()){ drop();
part(); swap(); part(); unbury(); pair(); pair(); // back();
over(); over(); part(); drop(); part(); nip(); // over2();left();right();
nil(); pair(); pair(); pop(); // put trees in new list // identical?
nil();// prime the pump
while(0<--limit && isTopNil()){ drop();
push(); part(); part(); pop(); // get trees to check
over(); over(); and(); // both nil?
if(isTopNil()){ drop();
drop(); drop();// nils are identical
push(); dup(); pop(); not(); // more to check?
}else{ drop();
over(); over(); or(); // is one nil?
if(isTopNil()){ drop();
push();pair();pair();dup();pop(); // not identical
}else{ drop();
part(); unbury(); part(); unbury();// break into their left and right parts
push(); pair(); pair(); pair(); pair(); pop(); // put on list
nil();}}} // may be identical
drop(); push(); // end of id
over(); part(); drop(); or(); not();} // match or no more?
drop(); nip(); // end of find
pop();}
onkeydown=e=>{
e.preventDefault();
switch(e.key){
case'Shift': case'Enter': case'Control':
case'Alt': case'Tab': case'CapsLock':return;
default: T=Pair(Pair(Left(T), list(e.key)), Right(T)); // input
dup(); part(); nip(); part(); nip(); part(); nip(); not(); // is letter?
if(isTopNil()){ drop(); pair();
}else{ drop();
over(); over(); pair();
push(); part(); bury(); swap(); pair(); swap(); pair(); pop(); // compile word
part(); nip(); part(); nip(); not(); // blue?
if(isTopNil()){ drop();
green();
push(); dup(); pop(); part(); drop(); part(); nip(); // get first word
dup(); not(); // not empty word?
while(0<--limit && isTopNil()){ drop();
push(); part(); part(); bury(); pair(); swap(); pair(); pop(); // fore
part(); // gray?
if(isTopNil()){ drop();
part(); nip(); // get letter quoting gray word
part(); nip(); part(); nip(); part(); nip(); // left();left();left();
part(); nip(); if(isTopNil()){ drop(); dup();}else{ // gray dup
part(); nip(); if(isTopNil()){ drop(); drop();}else{ // gray drop
part(); nip(); if(isTopNil()){ drop(); push(); swap(); pop(); pop(); }else{ // gray pop
part(); nip(); if(isTopNil()){ drop(); push(); push(); swap(); pop(); }else{ // gray push
part(); nip(); if(isTopNil()){ drop(); swap(); }else{ // gray swap
part(); nip(); if(isTopNil()){ drop(); nil(); }else{ // gray nil
part(); nip(); if(isTopNil()){ drop(); pair();}else{ // gray pair
part(); nip(); if(isTopNil()){ drop(); part(); }else{ // gray part
part(); nip(); if(isTopNil()){ drop(); push(); drop();}else{ // gray return
drop();}}}}}}}}}
}else{ part(); nip(); // green?
if(isTopNil()){ drop(); green();
}else{ drop(); drop();}}
push();dup();pop(); part(); drop(); part(); nip(); // get next word
dup(); not();} // not empty word?
drop(); drop(); // drop the not-Nil and the empty word // end of next
}else{ drop(); drop();}
nil();}
show();break;}}
var limit=1<<9;
var abc = ',;:.0123456789abcdefghijklmnopqrstuvwxyz'
function ln(p){return isNil(p)?0:1+ln(Right(p));}
function ls(n){return n<0?ls(abc.length):n?Pair(Nil,ls(n-1)):Nil;}
function n(a){return abc.indexOf(a);}
function a(n){return abc.length<=n?'?':abc.at(n);}
function list(a){return ls(n(a));}
function letter(p){return a(ln(p));}
function letters(p){return isNil(p)?'#':letters(Left(p))+letter(Right(p));}
function lists(s){return s==''?Nil:Pair(lists(s.slice(0,-1)),list(s.at(-1)));}
function lprog(p){return isNil(p)?'#':lprog(Left(p))+letters(Right(p));}
function rprog(p){return isNil(p)?'#':letters(Left(p))+rprog(Right(p));}
function prog(){return '{'+lprog(Left(Left(Right(T))))
+'['+letters(Right(Left(T)))+']'+rprog(Right(Left(Right(T))))+'}';}
document.body.style.font='12pt monospace';
document.body.style.whiteSpace='pre';
document.body.style.color='white';
document.body.style.background='gray';
function show(s=''){document.body.textContent+=`\t${limit} ${prog()} ${s}\n`;}
A beautiful solution to the problem of there being two places where the green code was called occurred to me in the shower. Because of the way that I’ve set up the ‘next’ loop the spelling of the blue word is still atop the stack before the loop begins and the loop exepects the next word (color and spelling) to be atop the stack. Thus I can color the spelling green and let the ‘next’ loop take care of it without having to insert anything new into the program being run.
part(); nip(); part(); nip(); not(); // blue?
if(isTopNil()){ drop();
green();
push(); dup(); pop(); part(); drop(); part(); nip(); // get first word
dup(); not(); // not empty word?
to this
part(); nip(); part(); nip(); not(); // blue?
if(isTopNil()){ drop();
nil(); nil(); pair(); pair(); //greenify
dup(); not(); // not empty word?
-2 {##a:#0,[#]#6,#8,#a.##b:#a;#a;#8,#b.##}
was output. So either it didn’t work or the value of limit was just too small. I’ll start by increasing limit and try everything again.
191 {##a:#0,#6,#8,#a.##b:#a;#a;#8,#b.#[#]#}
var Nil = new Object();
function isNil(p){return Object.is(p, Nil);}
function Pair(l,r){return new Array(l,r);}
function Left(p){if(isNil(p)) return Nil; else return p[0];}
function Right(p){if(isNil(p)) return Nil; else return p[1];}
var T=Nil;
function isTopNil(){return isNil(Right(Left(T)));}
function dup(){T=Pair(Pair(Left(T), Right(Left(T))), Right(T));}
function drop(){T=Pair(Left(Left(T)), Right(T));}
function pop(){T=Pair(Left(Left(T)), Pair(Right(Left(T)), Right(T)));}
function push(){T=Pair(Pair(Left(T), Left(Right(T))), Right(Right(T)));}
function swap(){T=Pair(Pair(Pair(Left(Left(Left(T))),
Right(Left(T))), Right(Left(Left(T)))), Right(T));}
function nil(){T=Pair(Pair(Left(T), Nil), Right(T));}
function pair(){T=Pair(Pair(Left(Left(Left(T))),
Pair(Right(Left(Left(T))), Right(Left(T)))), Right(T));}
function part(){T=Pair(Pair(Pair(Left(Left(T)),
Left(Right(Left(T)))), Right(Right(Left(T)))), Right(T));}
function nip(){pop(); drop(); push();}
function over(){pop(); dup(); push(); swap();}
function dig(){over(); pop(); nip();}
function bury(){dig(); dig(); push(); push();}
function unbury(){bury(); bury();}
function select(){if(isTopNil()){drop(); nip();}else{drop(); drop();}}
function not(){nil(); nil(); nil(); pair(); unbury(); select();}
function or(){dup(); select();}
function and(){not(); swap(); not(); or(); not();}
onkeydown=e=>{
switch(e.key){
case'Shift': case'Enter': case'Control': case'Alt': case'Tab': case'CapsLock':return;
default: T=Pair(Pair(Left(T), list(e.key)), Right(T)); // input
dup(); part(); nip(); part(); nip(); part(); nip(); not(); // is letter?
if(isTopNil()){ drop(); pair();
}else{ drop();
over(); over(); pair();
push(); part(); bury(); swap(); pair(); swap(); pair(); pop(); // compile word
part(); nip(); part(); nip(); not(); // blue?
if(isTopNil()){ drop();
nil(); nil(); pair(); pair(); //greenify
dup(); not(); // not empty word?
while(0<--limit && isTopNil()){ drop();
push(); part(); part(); bury(); pair(); swap(); pair(); pop(); // fore
part(); // gray?
if(isTopNil()){ drop();
part(); nip(); // get letter quoting gray word
part(); nip(); part(); nip(); part(); nip(); // left();left();left();
part(); nip(); if(isTopNil()){ drop(); dup();}else{ // gray dup
part(); nip(); if(isTopNil()){ drop(); drop();}else{ // gray drop
part(); nip(); if(isTopNil()){ drop(); push(); swap(); pop(); pop(); }else{ // gray pop
part(); nip(); if(isTopNil()){ drop(); push(); push(); swap(); pop(); }else{ // gray push
part(); nip(); if(isTopNil()){ drop(); swap(); }else{ // gray swap
part(); nip(); if(isTopNil()){ drop(); nil(); }else{ // gray nil
part(); nip(); if(isTopNil()){ drop(); pair();}else{ // gray pair
part(); nip(); if(isTopNil()){ drop(); part(); }else{ // gray part
part(); nip(); if(isTopNil()){ drop(); push(); drop();}else{ // gray return
drop();}}}}}}}}}
}else{ part(); nip(); // green?
if(isTopNil()){ drop();
nil(); nil(); nil(); pair(); pair(); pair(); // beginning of green //redify();
push(); dup(); pop(); // get program to search
nil(); // prime the pump // beginning of find
while(0<--limit && isTopNil()){ drop();
part(); swap(); part(); unbury(); pair(); pair(); // back();
over(); over(); part(); drop(); part(); nip(); // over2();left();right();
nil(); pair(); pair(); pop(); // put trees in new list // beginning of id
nil();// prime the pump
while(0<--limit && isTopNil()){ drop();
push(); part(); part(); pop(); // get trees to check
over(); over(); and(); // both nil?
if(isTopNil()){ drop();
drop(); drop();// nils are identical
push(); dup(); pop(); not(); // more to check?
}else{ drop();
over(); over(); or(); // is one nil?
if(isTopNil()){ drop();
push();pair();pair();dup();pop(); // not identical
}else{ drop();
part(); unbury(); part(); unbury();// break into their left and right parts
push(); pair(); pair(); pair(); pair(); pop(); // put on list
nil();}}} // may be identical
drop(); push(); // end of id
over(); part(); drop(); or(); not();} // match or no more?
drop(); nip(); // end of find
pop(); // end of green
}else{ drop(); drop();}}
push();dup();pop(); part(); drop(); part(); nip(); // get next word
dup(); not();} // not empty word?
drop(); drop(); // drop the not-Nil and the empty word // end of next
}else{ drop(); drop();}
nil();}
show();break;}}
var limit=1<<10;
var abc = ',;:.0123456789abcdefghijklmnopqrstuvwxyz'
function ln(p){return isNil(p)?0:1+ln(Right(p));}
function ls(n){return n<0?ls(abc.length):n?Pair(Nil,ls(n-1)):Nil;}
function n(a){return abc.indexOf(a);}
function a(n){return abc.length<=n?'?':abc.at(n);}
function list(a){return ls(n(a));}
function letter(p){return a(ln(p));}
function letters(p){return isNil(p)?'#':letters(Left(p))+letter(Right(p));}
function lists(s){return s==''?Nil:Pair(lists(s.slice(0,-1)),list(s.at(-1)));}
function lprog(p){return isNil(p)?'#':lprog(Left(p))+letters(Right(p));}
function rprog(p){return isNil(p)?'#':letters(Left(p))+rprog(Right(p));}
function prog(){return '{'+lprog(Left(Left(Right(T))))
+'['+letters(Right(Left(T)))+']'+rprog(Right(Left(Right(T))))+'}';}
document.body.style.font='12pt monospace';
document.body.style.whiteSpace='pre';
function show(s=''){document.body.textContent+=`\t${limit} ${prog()} ${s}\n`;}
Time for some lunch. Then either debugging gray words (they seem to work, but do they?) or reducing the three while loops into one (this is also called virtualization).
Another debugging session. Time is not always easy to get, but when I’ve got it I try to put it to use as best I can. But, in the end, I do not want to be the best john, I just want o be the accurate John.
I went ahead and eliminated the separate definition of ‘edit’ and put it inside the definition of the javascript ‘onkeydown’ property of the window.
var Nil = new Object();
function isNil(p){return Object.is(p, Nil);}
function Pair(l,r){return new Array(l,r);}
function Left(p){if(isNil(p)) return Nil; else return p[0];}
function Right(p){if(isNil(p)) return Nil; else return p[1];}
var T=Nil;
function isTopNil(){return isNil(Right(Left(T)));}
function dup(){T=Pair(Pair(Left(T), Right(Left(T))), Right(T));}
function drop(){T=Pair(Left(Left(T)), Right(T));}
function pop(){T=Pair(Left(Left(T)), Pair(Right(Left(T)), Right(T)));}
function push(){T=Pair(Pair(Left(T), Left(Right(T))), Right(Right(T)));}
function swap(){T=Pair(Pair(Pair(Left(Left(Left(T))),
Right(Left(T))), Right(Left(Left(T)))), Right(T));}
function nil(){T=Pair(Pair(Left(T), Nil), Right(T));}
function pair(){T=Pair(Pair(Left(Left(Left(T))),
Pair(Right(Left(Left(T))), Right(Left(T)))), Right(T));}
function part(){T=Pair(Pair(Pair(Left(Left(T)),
Left(Right(Left(T)))), Right(Right(Left(T)))), Right(T));}
function nip(){pop(); drop(); push();}
function over(){pop(); dup(); push(); swap();}
function dig(){over(); pop(); nip();}
function bury(){dig(); dig(); push(); push();}
function unbury(){bury(); bury();}
function select(){if(isTopNil()){drop(); nip();}else{drop(); drop();}}
function not(){nil(); nil(); nil(); pair(); unbury(); select();}
function or(){dup(); select();}
function and(){not(); swap(); not(); or(); not();}
function green(){
nil(); nil(); nil(); pair(); pair(); pair(); // redify
push();
dup();
pop();
// find
nil(); // prime the pump
while(0<--limit && isTopNil()){
// setup this iteration
drop();
part(); swap(); part(); unbury(); pair(); pair(); //back
// match?
// get top of program
over(); over();
part(); drop(); // left
part(); nip(); // right
// id
// make a new empty list
nil();
// put the trees at the front of it
pair(); pair();
// pop it out of the way for later
pop();
// prime the pump
nil();
// while the list is not empty
// and the top two items are not nil
while(0<--limit && isTopNil()){
drop();
// get the top two trees from the list and pop it out of the way for later
push();
part(); part();
pop();
// are they both nil?
over(); over(); and();
// if they're both nil
if(isTopNil()){
drop();
// drop them: they're identical
drop(); drop();
// is the list nil?
push();
dup();
pop();
not();
// if they're not both nil
}else{
drop();
// is one of them nil?
over(); over(); or();
// if one is nil
if(isTopNil()){
drop();
// then the other one isn't
// they're not identical
// put them back on the list
push();
pair(); pair();
// the list is not nil! It's over.
dup();
pop();
// if they're both not nil
}else{
drop();
// break into their left and right parts
part(); unbury(); part(); unbury();
// put them on the list
push();
pair(); pair(); pair(); pair();
pop();
nil();
}}}
drop();
push();
// end of id
// no more?
over();
part(); drop(); // left
or(); // match or no more?
not();
}
drop();
nip();
// end of find
pop();
}
var abc = ',;:.0123456789abcdefghijklmnopqrstuvwxyz '
function ln(p){return isNil(p)?0:1+ln(Right(p));}
function ls(n){return n<0?ls(abc.length):n?Pair(Nil,ls(n-1)):Nil;}
function n(a){return abc.indexOf(a);}
function a(n){return abc.length<=n?'?':abc.at(n);}
function list(a){return ls(n(a));}
function letter(p){return a(ln(p));}
function letters(p){return isNil(p)?'':letters(Left(p))+letter(Right(p));}
function lists(s){return s==''?Nil:Pair(lists(s.slice(0,-1)),list(s.at(-1)));}
function lprog(p){return isNil(p)?'':lprog(Left(p))+' '+letters(Right(p));}
function rprog(p){return isNil(p)?'':letters(Left(p))+' '+rprog(Right(p));}
function prog(){
return lprog(Left(Left(Right(T))))
+' [ '+letters(Right(Left(T)))+' ] '
+rprog(Right(Left(Right(T))));
}
var limit=1<<9;
onkeydown=e=>{
e.preventDefault();
//console.log(e.key);
switch(e.key){
case'Shift':
case'Enter':
case'Control':
case'Alt':
case'Tab':
case'CapsLock':return;
default:
// input
T=Pair(Pair(Left(T), list(e.key)), Right(T));
// edit
dup();
// is letter?
part(); nip(); // right
part(); nip(); // right
part(); nip(); // right
not();
if(isTopNil()){
drop();
pair();
}else{
drop();
// compile word
over(); over();
pair();
push();
part(); bury();
swap(); pair();
swap(); pair();
pop();
// blue?
part(); nip(); // right
part(); nip(); // right
not();
if(isTopNil()){
drop();
green();
// next
// get first word
push();
dup();
part(); part(); bury(); pair(); swap(); pair(); // fore
pop();
part(); drop(); // left
part(); nip(); // right
dup();
not();
// while next word isn't nil
while(0<--limit && isTopNil()){
drop();
// get color of word
part();
// if color is gray
if(isTopNil()){ // gray
console.log('gray')
drop();
part(); nip(); // get letter quoting gray word
part(); nip(); part(); nip(); part(); nip();
part(); nip(); if(isTopNil()){ drop(); dup(); console.log('dup') }else{ // gray dup
part(); nip(); if(isTopNil()){ drop(); drop();}else{ // gray drop
part(); nip(); if(isTopNil()){ drop(); push(); swap(); pop(); pop(); }else{ // gray pop
part(); nip(); if(isTopNil()){ drop(); push(); push(); swap(); pop(); }else{ // gray push
part(); nip(); if(isTopNil()){ drop(); swap(); }else{ // gray swap
part(); nip(); if(isTopNil()){ drop(); nil(); }else{ // gray nil
part(); nip(); if(isTopNil()){ drop(); pair(); console.log('pair')}else{ // gray pair
part(); nip(); if(isTopNil()){ drop(); part(); }else{ // gray part
part(); nip(); if(isTopNil()){ drop(); push(); drop(); console.log('ret')}else{ // gray return
drop();}}}}}}}}}
// if color is not gray
}else{
// if color is green
part(); nip(); // right
if(isTopNil()){ // green
drop();
green();
// if color is not green
}else{
drop();
drop();
}}
// get next word
push();
dup();
part(); part(); bury(); pair(); swap(); pair(); // fore
pop();
// get top of program stack
part(); drop(); // left
part(); nip(); // right
dup();
// is it empty?
not();
}
drop();
// drop empty word atop stack of T
drop();
// drop the two empty words atop program stack
console.log('here')
//push(); part(); swap(); part(); part(); drop(); drop(); swap(); pair(); pop();
// end of next
}else{
drop(); drop();
}
nil();
}
// end of edit
console.log(prog());
console.log('limit: ', limit);
break;
}
};
There is a way to make only a single place where ‘green’ is executed. It involves making an edited version of the executing program that uses an empty word before the relevant green word that corresponds to the blue one. Then, this can be dropped or edited out. There are surely some restrictions on how this might work which I have yet to contemplate. Before I do anything like this I’m going to finish debugging the present version of the program.
First, I have to look back on the last note to see where I left off.
I was testing a string like ‘0:test:0,6,8,test.’.
I must have been sleepy when I was debugging last night. It is clear to me today that there should not be a space character in the alphabet. This may reveal something about where the problem is becuase I have to javascript side of things print a question mark whenever there is a list whose length doesn’t quote a letter of the alphabet.
That didn’t change the output from the test keypresses ‘a:0,6,8,a.’. I’m going to get rid of any spaces in the program that prints the javascript string that quotes the state of the program.
I went ahead and eliminated all spaces and made it so that instead of using an empty string whenever it hits Nil, now everything returns a pound sign. Now what is displayed is consistent with what I see upon examining T in the console.
512 "{#[#a]#}"
512 "{##a:[#]#}"
512 "{##a:[#0]#}"
512 "{##a:#0,[#]#}"
512 "{##a:#0,[#6]#}"
512 "{##a:#0,#6,[#]#}"
512 "{##a:#0,#6,[#8]#}"
512 "{##a:#0,#6,#8,[#]#}"
512 "{##a:#0,#6,#8,[#a]#}"
402 "{##a:#0,[#]#6,#8,#a.#}"
401 "{##a:#0,#6,[#]#8,#a.#}"
400 "{##a:#0,#6,#8,[#]#a.#}"
399 "{##a:#0,#6,#8,#a.[#]#}"
398 "{##a:#0,#6,#8,#a.#[#]#}"
397 "{##a:#0,#6,#8,#a.##[#,]#}"
396 "{##a:#0,#6,#8,#a.##[#]#}"
Notice that I have not gone the full length to show the intermediate state of the rest of the tree. This reveals conspicuously that the bug is in the loop of ‘next’.
document.body.style.font='12pt monospace';
document.body.style.whiteSpace='pre';
document.body.style.color='white';
document.body.style.background='gray';
function show(s=''){document.body.textContent+=`${limit} ${prog()} ${s}\n`;}
512 {##a:#0,#6,#8,[#a]#}
402 {##a:#0,[#]#6,#8,#a.#} before next
401 {##a:#0,#6,[#]#8,#a.#} end of next loop
400 {##a:#0,#6,#8,[#]#a.#} end of next loop
399 {##a:#0,#6,#8,#a.[#]#} end of next loop
398 {##a:#0,#6,#8,#a.#[#]#} end of next loop
397 {##a:#0,#6,#8,#a.##[#,]#} end of next loop
396 {##a:#0,#6,#8,#a.##[#]#}
512 {##a:#0,#6,#8,[#a]#}
402 {##a:[#]#0,#6,#8,#a.#} before next
401 {##a:#0,[#]#6,#8,#a.#} end of next loop
400 {##a:#0,#6,[#]#8,#a.#} end of next loop
399 {##a:#0,#6,#8,[#]#a.#} end of next loop
398 {##a:#0,#6,#8,#a.[#]#} end of next loop
397 {##a:#0,#6,#8,#a.#[#,]#} end of next loop
396 {##a:#0,#6,#8,#a.#[#]#}
So that got rid of one of the Nils. The correct way to deal with this is check if the rest of the program is Nil or if the current word is nil. That is, first check if the current word is nil… wait.
512 {##a:#0,#8,#b:#a;#a;#8,[#b]#}
371 {##a:#0,#8,#b:#a;[#]#a;#8,#b.#} end of next loop
370 {##a:#0,#8,#b:#a;#a;[#a]#8,#b.#} going to run green
215 {##a:[#]#0,#8,#b:#a;#a;#8,#b.#} end of next loop
214 {##a:#0,[#]#8,#b:#a;#a;#8,#b.#} end of next loop
213 {##a:#0,#8,[#]#b:#a;#a;#8,#b.#} end of next loop
212 {##a:#0,#8,#b:#a;#a;[#]#8,#b.#} end of next loop
211 {##a:#0,#8,#b:#a;#a;#8,[#a]#b.#} going to run green
19 {##a:[#]#0,#8,#b:#a;#a;#8,#b.#} end of next loop
18 {##a:#0,[#]#8,#b:#a;#a;#8,#b.#} end of next loop
17 {##a:#0,#8,[#]#b:#a;#a;#8,#b.#} end of next loop
16 {##a:#0,#8,#b:#a;#a;#8,[#]#b.#} end of next loop
15 {##a:#0,#8,#b:#a;#a;#8,#b.[#]#} end of next loop
14 {##a:#0,#8,#b:#a;#a;#8,#b.#[#,]#} end of next loop
13 {##a:#0,#8,#b:#a;#a;#8,#b.#[#]#}
Hurray! Everything appears to be working pretty pretty .. pretty good.
var Nil = new Object();
function isNil(p){return Object.is(p, Nil);}
function Pair(l,r){return new Array(l,r);}
function Left(p){if(isNil(p)) return Nil; else return p[0];}
function Right(p){if(isNil(p)) return Nil; else return p[1];}
var T=Nil;
function isTopNil(){return isNil(Right(Left(T)));}
function dup(){T=Pair(Pair(Left(T), Right(Left(T))), Right(T));}
function drop(){T=Pair(Left(Left(T)), Right(T));}
function pop(){T=Pair(Left(Left(T)), Pair(Right(Left(T)), Right(T)));}
function push(){T=Pair(Pair(Left(T), Left(Right(T))), Right(Right(T)));}
function swap(){T=Pair(Pair(Pair(Left(Left(Left(T))),
Right(Left(T))), Right(Left(Left(T)))), Right(T));}
function nil(){T=Pair(Pair(Left(T), Nil), Right(T));}
function pair(){T=Pair(Pair(Left(Left(Left(T))),
Pair(Right(Left(Left(T))), Right(Left(T)))), Right(T));}
function part(){T=Pair(Pair(Pair(Left(Left(T)),
Left(Right(Left(T)))), Right(Right(Left(T)))), Right(T));}
function nip(){pop(); drop(); push();}
function over(){pop(); dup(); push(); swap();}
function dig(){over(); pop(); nip();}
function bury(){dig(); dig(); push(); push();}
function unbury(){bury(); bury();}
function select(){if(isTopNil()){drop(); nip();}else{drop(); drop();}}
function not(){nil(); nil(); nil(); pair(); unbury(); select();}
function or(){dup(); select();}
function and(){not(); swap(); not(); or(); not();}
function green(){
nil(); nil(); nil(); pair(); pair(); pair(); // redify();
push(); dup(); pop(); // get program to search
// find
nil(); // prime the pump
while(0<--limit && isTopNil()){
drop();
part(); swap(); part(); unbury(); pair(); pair(); // back();
over(); over(); part(); drop(); part(); nip(); // over2();left();right();
// identical?
nil(); pair(); pair(); pop(); // put trees in new list
nil();// prime the pump
while(0<--limit && isTopNil()){
drop();
push(); part(); part(); pop(); // get trees to check
over(); over(); and(); // both nil?
if(isTopNil()){
drop();
drop(); drop();// nils are identical
push(); dup(); pop(); not(); // more to check?
}else{
drop();
over(); over(); or(); // is one nil?
if(isTopNil()){
drop();
push();pair();pair();dup();pop(); // not identical
}else{
drop();
// break into their left and right parts and put on list
part(); unbury(); part(); unbury();
push(); pair(); pair(); pair(); pair(); pop();
nil();}}} // may be identical
drop(); push();
// end of id
over(); part(); drop(); or(); not();} // match or no more?
drop();
nip();
// end of find
pop();}
var abc = ',;:.0123456789abcdefghijklmnopqrstuvwxyz'
function ln(p){return isNil(p)?0:1+ln(Right(p));}
function ls(n){return n<0?ls(abc.length):n?Pair(Nil,ls(n-1)):Nil;}
function n(a){return abc.indexOf(a);}
function a(n){return abc.length<=n?'?':abc.at(n);}
function list(a){return ls(n(a));}
function letter(p){return a(ln(p));}
function letters(p){return isNil(p)?'#':letters(Left(p))+letter(Right(p));}
function lists(s){return s==''?Nil:Pair(lists(s.slice(0,-1)),list(s.at(-1)));}
function lprog(p){return isNil(p)?'#':lprog(Left(p))+letters(Right(p));}
function rprog(p){return isNil(p)?'#':letters(Left(p))+rprog(Right(p));}
function prog(){return '{'+lprog(Left(Left(Right(T))))
+'['+letters(Right(Left(T)))+']'+rprog(Right(Left(Right(T))))+'}';}
var limit=1<<9;
onkeydown=e=>{
e.preventDefault();
switch(e.key){
case'Shift': case'Enter': case'Control':
case'Alt': case'Tab': case'CapsLock':return;
default: T=Pair(Pair(Left(T), list(e.key)), Right(T)); // input
// is letter?
dup();
part(); nip(); part(); nip(); part(); nip(); // right();right();right();
not();
if(isTopNil()){
drop(); pair();
}else{
drop();
// compile word
over(); over();
pair(); push(); part(); bury();
swap(); pair(); swap(); pair();
pop();
// blue?
part(); nip(); part(); nip(); // right(); right();
not();
if(isTopNil()){
drop();
green();
push(); dup(); pop(); part(); drop(); part(); nip(); // get first word
dup(); not(); // not empty word?
while(0<--limit && isTopNil()){
drop();
// move program fore
push();
part(); part(); bury(); pair(); swap(); pair(); // fore
pop();
part(); // gray?
if(isTopNil()){
drop();
part(); nip(); // get letter quoting gray word
part(); nip(); part(); nip(); part(); nip(); // left();left();left();
part(); nip(); if(isTopNil()){ drop(); dup();}else{ // gray dup
part(); nip(); if(isTopNil()){ drop(); drop();}else{ // gray drop
part(); nip(); if(isTopNil()){ drop(); push(); swap(); pop(); pop(); }else{ // gray pop
part(); nip(); if(isTopNil()){ drop(); push(); push(); swap(); pop(); }else{ // gray push
part(); nip(); if(isTopNil()){ drop(); swap(); }else{ // gray swap
part(); nip(); if(isTopNil()){ drop(); nil(); }else{ // gray nil
part(); nip(); if(isTopNil()){ drop(); pair();}else{ // gray pair
part(); nip(); if(isTopNil()){ drop(); part(); }else{ // gray part
part(); nip(); if(isTopNil()){ drop(); push(); drop();}else{ // gray return
drop();}}}}}}}}}
}else{
part(); nip(); // green?
if(isTopNil()){
drop(); green();
}else{
drop(); drop();}}
push();dup();pop(); part(); drop(); part(); nip(); // get next word
dup(); not();} // not empty word?
drop(); drop(); // drop the not-Nil and the empty word
// end of next
}else{
drop(); drop();}
nil();}
show();break;}}
document.body.style.font='12pt monospace';
document.body.style.whiteSpace='pre';
document.body.style.color='white';
document.body.style.background='gray';
function show(s=''){document.body.textContent+=`\t${limit} ${prog()} ${s}\n`;}
Belly full. Time to debug. Starting from the end of last note.
Step one. Make some decaf coffee. Step two. Read where I left off.
OK, hot water is ready to boil.
Two things occurred to me the moment I looked back on where I left off in the last note: a) should “not finding” a word stop execution? b) the operation of execution is ‘pop’ and the operation of exit is a ‘push’ followed by a ‘drop’.
While I was writting that, it occurred to me that I haven’t ested a green word in the process of execution (though the blue words already use the same green code, so the only problem that could occur is in the design of the ‘next’ part of the definition of ‘edit’).
Yet another thing occurred to me: I would like to evolve the design of everything up to this point so that I do not need so many foreign probes to check that everything has been set up well.
// if color is gray
if(isTopNil()){ // gray
drop();
part(); nip(); part(); nip(); part(); nip(); part(); nip();
part(); nip(); if(isTopNil()){ drop(); dup(); }else{ // gray dup
part(); nip(); if(isTopNil()){ drop(); drop();}else{ // gray drop
part(); nip(); if(isTopNil()){ drop(); push(); swap(); pop(); pop(); }else{ // gray pop
part(); nip(); if(isTopNil()){ drop(); push(); push(); swap(); pop(); }else{ // gray push
part(); nip(); if(isTopNil()){ drop(); swap(); }else{ // gray swap
part(); nip(); if(isTopNil()){ drop(); nil(); }else{ // gray nil
part(); nip(); if(isTopNil()){ drop(); pair(); }else{ // gray pair
part(); nip(); if(isTopNil()){ drop(); part(); }else{ // gray part
part(); nip(); if(isTopNil()){ drop(); push(); drop(); }else{ // gray return
drop();}}}}}}}}}
The reason to put a return in is to see if everything works out as expected when execution doesn’t abruptly end.
I tried ‘0:test:0,6,8,test.’ as the test input. This didn’t work as expected and it appears as if there is one more ‘part();nip()’ in the beginning of ‘gray’.
Oh! I put some ‘console.log’s in the ‘gray’ part to see if that part of the code was reached. It was. Then it occurred to me that ‘part(); nip()’ was from an earlier version where I was thinking that words were spelled as lists of letters. This would make sense, but makes as much sense as spelling out words as stacks. So, ‘part(); nip();’ must be replaced by ‘part(); drop();’.
Oh. Gosh, I’m glad I tried to see this all to myself. So there is a bit of a tricky thing going on and that is part of how I got everything confused. The primitive gray words contemplated here are spelled out by a stack with a single item on it which is a list whose length quotes one of the letters in the string ‘012345678’. So the first, ‘part(); nip();’ gets that letter off the stack.
}
drop();
// end of next
This only drops the result of ‘not’ and does not then drop the empty word itself. This may account for all of the Nils atop the stack (and also for why I shouldn’t test just the gray-dup).
Hurray! It worked as expected. While I haven’t tested green words yet, I want to work on the end of ‘next’ still. It stops after it has already moved an empty word onto the stack of the program as a side effect of how parting a Nil works.
Oh! Ha. The space is from the program I use to view the state of the tree.
So one of the spaces is from the view program but the others are from Nils atop the program stack.
var Nil = new Object();
function isNil(p){return Object.is(p, Nil);}
function Pair(l,r){return new Array(l,r);}
function Left(p){if(isNil(p)) return Nil; else return p[0];}
function Right(p){if(isNil(p)) return Nil; else return p[1];}
var T=Nil;
function isTopNil(){return isNil(Right(Left(T)));}
function dup(){T=Pair(Pair(Left(T), Right(Left(T))), Right(T));}
function drop(){T=Pair(Left(Left(T)), Right(T));}
function pop(){T=Pair(Left(Left(T)), Pair(Right(Left(T)), Right(T)));}
function push(){T=Pair(Pair(Left(T), Left(Right(T))), Right(Right(T)));}
function swap(){T=Pair(Pair(Pair(Left(Left(Left(T))),
Right(Left(T))), Right(Left(Left(T)))), Right(T));}
function nil(){T=Pair(Pair(Left(T), Nil), Right(T));}
function pair(){T=Pair(Pair(Left(Left(Left(T))),
Pair(Right(Left(Left(T))), Right(Left(T)))), Right(T));}
function part(){T=Pair(Pair(Pair(Left(Left(T)),
Left(Right(Left(T)))), Right(Right(Left(T)))), Right(T));}
function nip(){pop(); drop(); push();}
function over(){pop(); dup(); push(); swap();}
function dig(){over(); pop(); nip();}
function bury(){dig(); dig(); push(); push();}
function unbury(){bury(); bury();}
function select(){if(isTopNil()){drop(); nip();}else{drop(); drop();}}
function not(){nil(); nil(); nil(); pair(); unbury(); select();}
function or(){dup(); select();}
function and(){not(); swap(); not(); or(); not();}
function green(){
nil(); nil(); nil(); pair(); pair(); pair(); // redify
push();
dup();
pop();
// find
nil(); // prime the pump
while(0<--limit && isTopNil()){
// setup this iteration
drop();
part(); swap(); part(); unbury(); pair(); pair(); //back
// match?
// get top of program
over(); over();
part(); drop(); // left
part(); nip(); // right
// id
// make a new empty list
nil();
// put the trees at the front of it
pair(); pair();
// pop it out of the way for later
pop();
// prime the pump
nil();
// while the list is not empty
// and the top two items are not nil
while(0<--limit && isTopNil()){
drop();
// get the top two trees from the list and pop it out of the way for later
push();
part(); part();
pop();
// are they both nil?
over(); over(); and();
// if they're both nil
if(isTopNil()){
drop();
// drop them: they're identical
drop(); drop();
// is the list nil?
push();
dup();
pop();
not();
// if they're not both nil
}else{
drop();
// is one of them nil?
over(); over(); or();
// if one is nil
if(isTopNil()){
drop();
// then the other one isn't
// they're not identical
// put them back on the list
push();
pair(); pair();
// the list is not nil! It's over.
dup();
pop();
// if they're both not nil
}else{
drop();
// break into their left and right parts
part(); unbury(); part(); unbury();
// put them on the list
push();
pair(); pair(); pair(); pair();
pop();
nil();
}}}
drop();
push();
// end of id
// no more?
over();
part(); drop(); // left
or(); // match or no more?
not();
}
drop();
nip();
// end of find
pop();
}
function edit(){
dup();
// is letter?
part(); nip(); // right
part(); nip(); // right
part(); nip(); // right
not();
if(isTopNil()){
drop();
pair();
}else{
drop();
// compile word
over(); over();
pair();
push();
part(); bury();
swap(); pair();
swap(); pair();
pop();
// blue?
part(); nip(); // right
part(); nip(); // right
not();
if(isTopNil()){
drop();
green();
// next
// get first word
push();
dup();
part(); part(); bury(); pair(); swap(); pair(); // fore
pop();
part(); drop(); // left
part(); nip(); // right
dup();
not();
// while next word isn't nil
while(0<--limit && isTopNil()){
drop();
// get color of word
part();
// if color is gray
if(isTopNil()){ // gray
console.log('gray')
drop();
part(); nip(); // get letter quoting gray word
part(); nip(); part(); nip(); part(); nip();
part(); nip(); if(isTopNil()){ drop(); dup(); console.log('dup') }else{ // gray dup
part(); nip(); if(isTopNil()){ drop(); drop();}else{ // gray drop
part(); nip(); if(isTopNil()){ drop(); push(); swap(); pop(); pop(); }else{ // gray pop
part(); nip(); if(isTopNil()){ drop(); push(); push(); swap(); pop(); }else{ // gray push
part(); nip(); if(isTopNil()){ drop(); swap(); }else{ // gray swap
part(); nip(); if(isTopNil()){ drop(); nil(); }else{ // gray nil
part(); nip(); if(isTopNil()){ drop(); pair(); console.log('pair')}else{ // gray pair
part(); nip(); if(isTopNil()){ drop(); part(); }else{ // gray part
part(); nip(); if(isTopNil()){ drop(); push(); drop(); console.log('ret')}else{ // gray return
drop();}}}}}}}}}
// if color is not gray
}else{
// if color is green
part(); nip(); // right
if(isTopNil()){ // green
drop();
green();
// if color is not green
}else{
drop();
drop();
}}
// get next word
push();
dup();
part(); part(); bury(); pair(); swap(); pair(); // fore
pop();
// get top of program stack
part(); drop(); // left
part(); nip(); // right
dup();
// is it empty?
not();
}
drop();
// drop empty word atop stack of T
drop();
// drop the two empty words atop program stack
console.log('here')
//push(); part(); swap(); part(); part(); drop(); drop(); swap(); pair(); pop();
// end of next
}else{
drop(); drop();
}
nil();
}}
var abc = ',;:.0123456789abcdefghijklmnopqrstuvwxyz '
function ln(p){return isNil(p)?0:1+ln(Right(p));}
function ls(n){return n<0?ls(abc.length):n?Pair(Nil,ls(n-1)):Nil;}
function n(a){return abc.indexOf(a);}
function a(n){return abc.length<=n?'?':abc.at(n);}
function list(a){return ls(n(a));}
function letter(p){return a(ln(p));}
function letters(p){return isNil(p)?'':letters(Left(p))+letter(Right(p));}
function lists(s){return s==''?Nil:Pair(lists(s.slice(0,-1)),list(s.at(-1)));}
function lprog(p){return isNil(p)?'':lprog(Left(p))+' '+letters(Right(p));}
function rprog(p){return isNil(p)?'':letters(Left(p))+' '+rprog(Right(p));}
function prog(){
return lprog(Left(Left(Right(T))))
+' [ '+letters(Right(Left(T)))+' ] '
+rprog(Right(Left(Right(T))));
}
var limit=1<<9;
onkeydown=e=>{
e.preventDefault();
//console.log(e.key);
switch(e.key){
case'Shift':
case'Enter':
case'Control':
case'Alt':
case'Tab':
case'CapsLock':return;
default:
T=Pair(Pair(Left(T), list(e.key)), Right(T));
edit();
console.log(prog());
console.log('limit: ', limit);
break;
}
};
Well, going to lunch turned into going to stores and doing other chores. This is often how it goes. A thirty minute lunch turns into a chore fest. Back to debugging.
Here is the current state of the program:
var Nil = new Object();
function isNil(p){return Object.is(p, Nil);}
function Pair(l,r){return new Array(l,r);}
function Left(p){if(isNil(p)) return Nil; else return p[0];}
function Right(p){if(isNil(p)) return Nil; else return p[1];}
var T=Nil;
function isTopNil(){return isNil(Right(Left(T)));}
function drop(){T=Pair(Left(Left(T)), Right(T));}
function dup(){T=Pair(Pair(Left(T), Right(Left(T))), Right(T));}
function pop(){T=Pair(Left(Left(T)), Pair(Right(Left(T)), Right(T)));}
function push(){T=Pair(Pair(Left(T), Left(Right(T))), Right(Right(T)));}
function swap(){T=Pair(Pair(Pair(Left(Left(Left(T))),
Right(Left(T))), Right(Left(Left(T)))), Right(T));}
function nil(){T=Pair(Pair(Left(T), Nil), Right(T));}
function pair(){T=Pair(Pair(Left(Left(Left(T))),
Pair(Right(Left(Left(T))), Right(Left(T)))), Right(T));}
function part(){T=Pair(Pair(Pair(Left(Left(T)),
Left(Right(Left(T)))), Right(Right(Left(T)))), Right(T));}
function nip(){pop(); drop(); push();}
function over(){pop(); dup(); push(); swap();}
function dig(){over(); pop(); nip();}
function bury(){dig(); dig(); push(); push();}
function unbury(){bury(); bury();}
function select(){
if(isTopNil()){
drop(); nip();
}else{
drop(); drop();
}}
function not(){
nil(); nil(); nil(); pair();
unbury(); select();
}
function or(){
dup(); select();
}
function and(){
not(); swap(); not();
or(); not();
}
function back(){
part();
swap(); part();
bury();
pair(); pair();
}
function fore(){
part(); part();
bury();
pair();
swap(); pair();
}
function green(){
nil(); nil(); nil(); pair(); pair(); pair(); // redify
push();
dup();
fore();
pop();
// find
nil(); // prime the pump
while(0<--limit && isTopNil()){
// setup this iteration
drop();
back();
// match?
// get top of program
over(); over();
part(); drop(); // left
part(); nip(); // right
// id
// make a new empty list
nil();
// put the trees at the front of it
pair(); pair();
// prime the pump
nil();
// while the list is not empty
// and the top two items are not nil
while(0<--limit && isTopNil()){
drop();
// get the top two trees from the list
push();
part(); part();
pop();
// are they both nil?
over(); over(); and();
// if they're both nil
if(isTopNil()){
drop();
// drop them: they're identical
drop(); drop();
// is the list nil?
push();
dup();
pop();
// if they're not both nil
}else{
drop();
// is one of them nil?
over(); over(); or();
// if one is nil
if(isTopNil()){
drop();
// then the other one isn't
// they're not identical
// put them back on the list
push();
pair(); pair();
// the list is not nil! It's over.
dup();
pop();
// if they're both not nil
}else{
drop();
// break into their left and right parts
part(); unbury(); part(); unbury();
// put them on the list
push();
pair(); pair(); pair(); pair();
pop();
nil();
}}}
drop();
push();
// end of id
// no more?
over();
part(); drop(); // left
or(); // match or no more?
not();
}
drop();
nip();
// end of find
pop();
}
function edit(){
dup();
// is letter?
part(); nip(); // right
part(); nip(); // right
part(); nip(); // right
not();
if(isTopNil()){
drop();
pair();
}else{
drop();
// compile word
over(); over();
pair();
push();
part(); bury();
swap(); pair();
swap(); pair();
pop();
// blue?
part(); nip(); // right
part(); nip(); // right
not();
if(isTopNil()){
drop();
green();
// next
// get first word
push();
dup();
fore();
pop();
part(); drop(); // left
part(); nip(); // right
dup();
not();
// while next word isn't nil
while(0<--limit && isTopNil()){
drop();
// get color of word
part();
// if color is gray
if(isTopNil()){ // gray
drop();
part(); nip(); part(); nip(); part(); nip(); part(); nip();
part(); nip(); if(isTopNil()){ drop(); dup(); }else{
part(); nip(); if(isTopNil()){ drop(); drop();}else{
part(); nip(); if(isTopNil()){ drop(); push(); swap(); pop(); pop(); }else{ // gray pop
part(); nip(); if(isTopNil()){ drop(); push(); push(); swap(); pop(); }else{ // gray push
part(); nip(); if(isTopNil()){ drop(); swap(); }else{
part(); nip(); if(isTopNil()){ drop(); nil(); }else{
part(); nip(); if(isTopNil()){ drop(); pair(); }else{
part(); nip(); if(isTopNil()){ drop(); part(); }else{
drop();}}}}}}}}
// if color is not gray
}else{
// if color is green
part(); nip(); // right
if(isTopNil()){ // green
drop();
green();
// if color is not green
}else{
drop();
drop();
}}
// get next word
push();
dup();
fore();
pop();
part(); drop(); // left
part(); nip(); // right
dup();
not();
}
drop();
// end of next
}else{
drop(); drop();
}
nil();
}}
var abc = ',;:.0123456789abcdefghijklmnopqrstuvwxyz '
function ln(p){return isNil(p)?0:1+ln(Right(p));}
function ls(n){return n<0?ls(abc.length):n?Pair(Nil,ls(n-1)):Nil;}
function n(a){return abc.indexOf(a);}
function a(n){return abc.length<=n?'?':abc.at(n);}
function list(a){return ls(n(a));}
function letter(p){return a(ln(p));}
function letters(p){return isNil(p)?'':letters(Left(p))+letter(Right(p));}
function lists(s){return s==''?Nil:Pair(lists(s.slice(0,-1)),list(s.at(-1)));}
function lprog(p){return isNil(p)?'':lprog(Left(p))+' '+letters(Right(p));}
function rprog(p){return isNil(p)?'':letters(Left(p))+' '+rprog(Right(p));}
function prog(){
return lprog(Left(Left(Right(T))))
+' [ '+letters(Right(Left(T)))+' ] '
+rprog(Right(Left(Right(T))));
}
var limit=100;
onkeydown=e=>{
e.preventDefault();
//console.log(e.key);
switch(e.key){
case'Shift':
case'Enter':
case'Control':
case'Alt':
case'Tab':
case'CapsLock':return;
default:
T=Pair(Pair(Left(T), list(e.key)), Right(T));
edit();
console.log(prog());
console.log('limit: ', limit);
break;
}
};
I left off debugging the definition of ‘green’.
I just looked at the code to see exactly where I left off and I’m glad that I did becuase 1) it is not in the definition of ‘green’ and 2) I would have missed it for a while if I had made the assumption that ‘green’ is where the problem was.
The problem is right after compilation where it is checked if the word compiled is blue. The test keypresses are ‘a:0,a.’ which should duplicate the top of the stack. Right now it goes off into an endless loop that is only tamed by introducing a limit on the number of times, globally, the bodies of while loops can be executed.
I’m missing a ‘drop();’ after the conditional that checks for blue. Right now, it leaves the color blue as the top of the stack. The problem is that the definition of ‘green’ expects the top of the stack to spell out the matching red word to be found. I’ll add the missing drop and run the same test case to see what happens.
Oops! I totally forgot that the color isn’t duplicated before it is checked to see if it is blue. This means that one drop is good enough to get rid of the Nil at the top and leave the spelling of the red word whose match is to be found.
There are two ways to approach debugging the problem that presents itself: a) continue carefully reading through the operations and double checking that they do what I expected, and b) setting up some more probes on the javascript side so that I can see the whole state of the tree after each of the basic operations.
Side note entirely unrelated to debugging: I just got my diamond sharpening blocks! Now I can set up my sharpening system and start making some boxes.
I’ll avoid making new probes until I look over the program and see if I can spot where I went wrong.
nil(); nil(); nil(); pair(); pair(); pair(); // redify
push();
dup();
fore();
pop();
nil(); nil(); nil(); pair(); pair(); pair(); // redify
push();
dup();
part(); part(); bury(); pair(); swap(); pair(); // fore
pop();
Redify works in that it takes the top of the stack as the spelling of a word and colors it red by pairing the spelling with ‘(() ())’ the list of two items each of which is Nil.
The ‘push();’ brings down the program from the front of the list.
‘dup();’ makes a copy of where the program is right now. This is where the search will proceed from. For some reason there is a ‘fore()’ here, expanded to “part(); part(); bury(); pair(); swap(); pair(); // fore”, which, as far as I can see, serves no purpose. In the regular course of execution, that is through the ‘next’ part of ‘edit’, the program is already shifted to the next word when it’s searching for the last matching red. Notice that this wastes a step i.e. it will check if the red word is identical to the green word but we already know it isn’t. This is yet another example of boundary value problems.
So, the ‘fore’ can be removed entirely. This leaves only two occurrences of ‘fore’. They are both part of the ‘next’ part of ‘edit’. The first one can probably be eliminated, but I don’t know how yet.
Now to test the same code after making this little change: ‘a:0,a.’. That still didn’t work.
part(); swap(); part(); unbury(); pair(); pair(); //back
This makes me suspect that ‘fore’ may be defined incorrectly as well. No it’s ok.
a: 0, a., 9 [ ]
limit: -2
// get top of program
over(); over();
part(); drop(); // left
part(); nip(); // right
It does what the comments say it should. Though, it doesn’t quite explain why there is an ‘over(); over();’ before getting the top of the program. The explanation is that after this comparison between the red word and the top of the program is over the program is turned back and checked from there.
Next is the beginning of ‘id’. OH, gosh. I have forgotten to mention something important that has been lurking for a long long while. Every program in my environment has a beginning a middle and an end. Usually the middle is repeated until some condition is met. The beginning is the set up, the end is the break down. This makes programming more story like. It happens naturally when you’re solving problems that they turn into stories with a beginning middle (that often repeats) and an end. That they are complete stories is kinda a cool unexpected side effect of factoring things this way.
// id
// make a new empty list
nil();
// put the trees at the front of it
pair(); pair();
// prime the pump
nil();
// while the list is not empty
// and the top two items are not nil
// id
// make a new empty list
nil();
// put the trees at the front of it
pair(); pair();
pop();
// prime the pump
nil();
// while the list is not empty
// and the top two items are not nil
Now to test everything again. And, it’s a pint of Guiness time. So smooth.
a: [ ] 0, a.
limit: 85
Hopefully a little Guiness won’t mess me up while debugging.
I went ahead and checked the value of ‘T’ in the console to see if a dup was executed or not. Happily it was! The problem seems to be something about getting to the next operation. Though, since I haven’t implemented a ‘return’ command, there’s nothing surprising there.
I’m uncertain if execution occurred in the ‘next’ loop or not. I’ll put a ‘console.log’ in there and see if it gets there (after starting from scratch and using the same test string).
And I have sadly demonstrated that it does not reach the middle of ‘next’. This is consistent with what I’m seeing e.g. the Nil atop the stack is probably from an error later in the definition of ‘edit’.
0; [ ] a: 0, a.
limit: 80
while(0<--limit && isTopNil()){
console.log('beginning of middle of id');
drop();
// get the top two trees from the list and pop it out of the way for later
push();
part(); part();
pop();
// are they both nil?
over(); over(); and();
// if they're both nil
if(isTopNil()){
drop();
// drop them: they're identical
drop(); drop();
// is the list nil?
push();
dup();
pop();
// if they're not both nil
}else{
There should be a ‘not’ after that last ‘pop();’ so that the loop continues if the list of subtrees to visit and compare is nonempty. Right now, if it is nonempty it stop checking for identity and pushes that non empty list onto the top fo the list.
b; a: 0, a. [ ]
limit: 42
0; a: 0, a. [ ] a.
limit: 95
var Nil = new Object();
function isNil(p){return Object.is(p, Nil);}
function Pair(l,r){return new Array(l,r);}
function Left(p){if(isNil(p)) return Nil; else return p[0];}
function Right(p){if(isNil(p)) return Nil; else return p[1];}
var T=Nil;
function isTopNil(){return isNil(Right(Left(T)));}
function drop(){T=Pair(Left(Left(T)), Right(T));}
function dup(){T=Pair(Pair(Left(T), Right(Left(T))), Right(T));}
function pop(){T=Pair(Left(Left(T)), Pair(Right(Left(T)), Right(T)));}
function push(){T=Pair(Pair(Left(T), Left(Right(T))), Right(Right(T)));}
function swap(){T=Pair(Pair(Pair(Left(Left(Left(T))),
Right(Left(T))), Right(Left(Left(T)))), Right(T));}
function nil(){T=Pair(Pair(Left(T), Nil), Right(T));}
function pair(){T=Pair(Pair(Left(Left(Left(T))),
Pair(Right(Left(Left(T))), Right(Left(T)))), Right(T));}
function part(){T=Pair(Pair(Pair(Left(Left(T)),
Left(Right(Left(T)))), Right(Right(Left(T)))), Right(T));}
function nip(){pop(); drop(); push();}
function over(){pop(); dup(); push(); swap();}
function dig(){over(); pop(); nip();}
function bury(){dig(); dig(); push(); push();}
function unbury(){bury(); bury();}
function select(){
if(isTopNil()){
drop(); nip();
}else{
drop(); drop();
}}
function not(){
nil(); nil(); nil(); pair();
unbury(); select();
}
function or(){
dup(); select();
}
function and(){
not(); swap(); not();
or(); not();
}
function green(){
nil(); nil(); nil(); pair(); pair(); pair(); // redify
push();
dup();
pop();
// find
nil(); // prime the pump
while(0<--limit && isTopNil()){
// setup this iteration
drop();
part(); swap(); part(); unbury(); pair(); pair(); //back
// match?
// get top of program
over(); over();
part(); drop(); // left
part(); nip(); // right
// id
// make a new empty list
nil();
// put the trees at the front of it
pair(); pair();
// pop it out of the way for later
pop();
// prime the pump
nil();
// while the list is not empty
// and the top two items are not nil
while(0<--limit && isTopNil()){
console.log('beginning of middle of id');
drop();
// get the top two trees from the list and pop it out of the way for later
push();
part(); part();
pop();
// are they both nil?
over(); over(); and();
// if they're both nil
if(isTopNil()){
drop();
// drop them: they're identical
drop(); drop();
// is the list nil?
push();
dup();
pop();
not();
// if they're not both nil
}else{
drop();
// is one of them nil?
over(); over(); or();
// if one is nil
if(isTopNil()){
drop();
// then the other one isn't
// they're not identical
// put them back on the list
push();
pair(); pair();
// the list is not nil! It's over.
dup();
pop();
// if they're both not nil
}else{
drop();
// break into their left and right parts
part(); unbury(); part(); unbury();
// put them on the list
push();
pair(); pair(); pair(); pair();
pop();
nil();
}}}
drop();
push();
// end of id
// no more?
over();
part(); drop(); // left
or(); // match or no more?
not();
}
drop();
nip();
// end of find
pop();
console.log('end of green')
}
function edit(){
dup();
// is letter?
part(); nip(); // right
part(); nip(); // right
part(); nip(); // right
not();
if(isTopNil()){
drop();
pair();
}else{
drop();
// compile word
over(); over();
pair();
push();
part(); bury();
swap(); pair();
swap(); pair();
pop();
// blue?
part(); nip(); // right
part(); nip(); // right
not();
if(isTopNil()){
drop();
green();
// next
// get first word
push();
dup();
part(); part(); bury(); pair(); swap(); pair(); // fore
pop();
part(); drop(); // left
part(); nip(); // right
dup();
console.log(letter(Right(Left(T))))
not();
console.log('in the beginning of next');
// while next word isn't nil
while(0<--limit && isTopNil()){
console.log('in the middle of next')
drop();
// get color of word
part();
// if color is gray
if(isTopNil()){ // gray
drop();
part(); nip(); part(); nip(); part(); nip(); part(); nip();
part(); nip(); if(isTopNil()){ drop(); dup(); }else{
part(); nip(); if(isTopNil()){ drop(); drop();}else{
part(); nip(); if(isTopNil()){ drop(); push(); swap(); pop(); pop(); }else{ // gray pop
part(); nip(); if(isTopNil()){ drop(); push(); push(); swap(); pop(); }else{ // gray push
part(); nip(); if(isTopNil()){ drop(); swap(); }else{
part(); nip(); if(isTopNil()){ drop(); nil(); }else{
part(); nip(); if(isTopNil()){ drop(); pair(); }else{
part(); nip(); if(isTopNil()){ drop(); part(); }else{
drop();}}}}}}}}
// if color is not gray
}else{
// if color is green
part(); nip(); // right
if(isTopNil()){ // green
drop();
green();
// if color is not green
}else{
console.log('the next word was red or blue')
drop();
drop();
}}
// get next word
push();
dup();
part(); part(); bury(); pair(); swap(); pair(); // fore
pop();
part(); drop(); // left
part(); nip(); // right
dup();
not();
}
drop();
// end of next
}else{
drop(); drop();
}
nil();
}}
var abc = ',;:.0123456789abcdefghijklmnopqrstuvwxyz '
function ln(p){return isNil(p)?0:1+ln(Right(p));}
function ls(n){return n<0?ls(abc.length):n?Pair(Nil,ls(n-1)):Nil;}
function n(a){return abc.indexOf(a);}
function a(n){return abc.length<=n?'?':abc.at(n);}
function list(a){return ls(n(a));}
function letter(p){return a(ln(p));}
function letters(p){return isNil(p)?'':letters(Left(p))+letter(Right(p));}
function lists(s){return s==''?Nil:Pair(lists(s.slice(0,-1)),list(s.at(-1)));}
function lprog(p){return isNil(p)?'':lprog(Left(p))+' '+letters(Right(p));}
function rprog(p){return isNil(p)?'':letters(Left(p))+' '+rprog(Right(p));}
function prog(){
return lprog(Left(Left(Right(T))))
+' [ '+letters(Right(Left(T)))+' ] '
+rprog(Right(Left(Right(T))));
}
var limit=1<<9;
onkeydown=e=>{
e.preventDefault();
//console.log(e.key);
switch(e.key){
case'Shift':
case'Enter':
case'Control':
case'Alt':
case'Tab':
case'CapsLock':return;
default:
T=Pair(Pair(Left(T), list(e.key)), Right(T));
edit();
console.log(prog());
console.log('limit: ', limit);
break;
}
};
Oops, I forgot to record my work I was doing on my programming environment. This happens from time to time and when it happens I usually lose some potentially valuable records of the kind of thinking that matters most.
I’m going over the program now that I’ve got most of the logic where I want it. What this means is that I have pinned down how I expect each part of the program to operate when executed. This does not mean that each part actually operates in this way. When you set up a calendar or schedule of events, there are always other things that happen that you didn’t take into account, or errors that you made which are revealed to you once the programmed events actually occur.
The design of the programming environment makes it easy to debug. Probes from the javascript environment are constructed that let me examine the state of my programming environment. Here they are:
var abc = ',;:.0123456789abcdefghijklmnopqrstuvwxyz '
function ln(p){return isNil(p)?0:1+ln(Right(p));}
function ls(n){return n<0?ls(abc.length):n?Pair(Nil,ls(n-1)):Nil;}
function n(a){return abc.indexOf(a);}
function a(n){return abc.length<=n?'?':abc.at(n);}
function list(a){return ls(n(a));}
function letter(p){return a(ln(p));}
function letters(p){return isNil(p)?'':letters(Left(p))+letter(Right(p));}
function lists(s){return s==''?Nil:Pair(lists(s.slice(0,-1)),list(s.at(-1)));}
function lprog(p){return isNil(p)?'':lprog(Left(p))+' '+letters(Right(p));}
function rprog(p){return isNil(p)?'':letters(Left(p))+' '+rprog(Right(p));}
function prog(){
return lprog(Left(Left(Right(T))))
+' [ '+letters(Right(Left(T)))+' ] '
+rprog(Right(Left(Right(T))));
}
onkeydown=e=>{
e.preventDefault();
//console.log(e.key);
switch(e.key){
case'Shift':
case'Enter':
case'Control':
case'Alt':
case'Tab':
case'CapsLock':return;
default:
T=Pair(Pair(Left(T), list(e.key)), Right(T));
edit();
console.log(prog());
break;
}
};
An alphabet is set up in the javascript environment so that keypresses are quoted in my programming environment as the length of a list. Lists of these lists are then quoted in the javascript environment as javscript strings and these are manipulated so that the compiled program is shown with the word currently being edited enclosed in square brackets where it will occur in the program if compiled.
Here are some bugs that I found and fixed prior to starting this note here:
‘swap’ was defined incorrectly. I uncovered this when I pressed one of the color keys, e.g. ‘,’, and saw that the word being constructed was not compiled into the program being constructed. Testing of ‘not’ revealed that it was part of the problem. I then examined each step of execution that goes into ‘not’ and looked at the output of ‘T’ in the javascript console to see that ‘unbury’ did not work as expected. Since the definition of ‘unbury’ was as it should be (in my judgement), I went straight to the basic operations. There was a ‘Right’ where there neededed to be a ‘Left’ in the definition of ‘swap’.
‘isNilTop’ instead of ‘isTopNil’: I permuted the name and solving this problem was a matter of a simple find and replace.
Now I’m all caught up with what I did without taking notes.
Here is the sequence of events that occurred while I debugged the program:
I run the javascript as a script snippet in a Chrome browser on the page ‘about:blank’ by opening the ‘developer tools’ and pressing ‘Control’ and ‘Enter’ after left clicking into the specific snippet. Cursor focus is then moved to the console. I then left click onto the window presented by having navigated alreayd to ‘about:blank’. Once cursor focus is on the window, I type on they keyboard and can see the output from ‘onkeypress’ event firing in the console.
After pressing the ‘a’ key, I then press the comma ‘,’ key. The expected output is a, [ ] but the output a get is [ ]. This means that there is something wrong with the compile part of ‘edit’ or with my method of viewing the program and word under construction.
I shall press ‘a’ and then ‘,’ after having restarted everything and then directly examine ‘T’ in the console. It is presented as nested arrays whose depths can be explored by clicking on little triangles that drop down the object notation for the indexed items of an array (with further triangles for compound objects that are not Nil).
My method of viewing the program and word under construction is consistent with what the console shows me of T. The list of T is empty and hence there are no letters to the left of ‘[’ or to the right of ‘]’ in the string printed to the console. The top of T is also Nil, and this is consistent with the empty space between ‘[’ and ‘]’ in the output to the console.
So, now I strongly suspect that ‘compile’ is where the problem has occurred. But, I’ll have to examine T further in order to see if that is the case. I shall do that now and report on what I did after.
Everything seems to be good up to the point in the compile code after which ‘over(); over();’ is executed then it looks like the program doesn’t reflect what I saw to myself when I first programmed it.
As I mentioned earlier, I have the logic straightened out even if the program doesn’t reflect that. At least, I would bet on having the logic straightened out. I’ll describe how it should work now. The description is of a contingency. That is what programmable machines mediate: contingencies as consequences from responses on occasions. The occasion here is one where the top item of T is the color of the word to be compiled and the second from top is how to spell that word.
They are needed in their present relationship for when I check to see if the word just compiled is blue or not (and hence whether they need to be executed as if they were a green word). Thus, ‘over(); over();’ creates duplicates of them that are to be part of compilation. The top two items are ‘pair();’ -ed because a compiled word is a pair whose left part spells it out and whose right part is its color.
The program under construction is ‘push();’ -ed from the front of the list to the top of the stack. It must then be ‘part();’ -ed so that the word being compiled can be put atop the stack of the program under construction. So far so good.
The top of the stack is the list of the program being edited and must be ‘bury();’ -ed to get at both the stack of the program and the word to be put atop it.
Now, the top of the stack is the stack of the program and the second from top is the word to be compiled. They are ‘swap();’ -ed and ‘pair();’ -ed which pushes the word onto the stack of the program. Now the stack and list of the program are put back together by another ‘swap();’ and ‘pair();’.
Aha! This is where the bug is. The next step, in what I saw to myself, is that the program is popped so that if the word compiled is blue then all that stuff can happen next. I did not put a ‘pop();’! I’ll put it there now and run the same test that started me off on this bug search.
Hurray! I pressed ‘a’ and then ‘,’ keys and the output to the console was a, [ ]. I haven’t included any leading or trailing spaces in my quotation of the console output because they are not relevant to the bug hunt.
Now, I’m going to continue testing compilation by checking that each key which corresponds to a color produces a compiled word. The order of ‘.;:.’ at the beginning of the definition of the alphabet shows that ‘.’ should be gray, ‘;’ should be green, ‘:’ should be red, and ‘.’ should be blue. Notice that the colors are treated as punctuation in this situation.
I pressed the following sequence of keys: ‘a,b;c:d.f0’. The output was ` a, b; c: d. [ f0 ] ` (now I have included the surrounding spaces just in case) and is as expected. This output shows that pressing the key ‘0’ does not start compilation but pressing the key ‘.’ does. This is a boundary that has to be carefully tested as all boundaries do because they often cause ‘boundary problems’. Most boundary problems are solved when dealing with the logic of a program by forcing every boundary to be clopen.
Now, typing ‘d.’ should have been treated like compiling a blue word.
Oh, I forgot to present the program that I was correcting. Well, that program is gone. So I can show the program that I am examining at this moment:
var Nil = new Object();
function isNil(p){return Object.is(p, Nil);}
function Pair(l,r){return new Array(l,r);}
function Left(p){if(isNil(p)) return Nil; else return p[0];}
function Right(p){if(isNil(p)) return Nil; else return p[1];}
var T=Nil;
function isTopNil(){return isNil(Right(Left(T)));}
function drop(){T=Pair(Left(Left(T)), Right(T));}
function dup(){T=Pair(Pair(Left(T), Right(Left(T))), Right(T));}
function pop(){T=Pair(Left(Left(T)), Pair(Right(Left(T)), Right(T)));}
function push(){T=Pair(Pair(Left(T), Left(Right(T))), Right(Right(T)));}
function swap(){T=Pair(Pair(Pair(Left(Left(Left(T))),
Right(Left(T))), Right(Left(Left(T)))), Right(T));}
function nil(){T=Pair(Pair(Left(T), Nil), Right(T));}
function pair(){T=Pair(Pair(Left(Left(Left(T))),
Pair(Right(Left(Left(T))), Right(Left(T)))), Right(T));}
function part(){T=Pair(Pair(Pair(Left(Left(T)),
Left(Right(Left(T)))), Right(Right(Left(T)))), Right(T));}
function nip(){pop(); drop(); push();}
function over(){pop(); dup(); push(); swap();}
function dig(){over(); pop(); nip();}
function bury(){dig(); dig(); push(); push();}
function unbury(){bury(); bury();}
function select(){
if(isTopNil()){
drop(); nip();
}else{
drop(); drop();
}}
function not(){
nil(); nil(); nil(); pair();
unbury(); select();
}
function or(){
dup(); select();
}
function and(){
not(); swap(); not();
or(); not();
}
function back(){
part();
swap(); part();
bury();
pair(); pair();
}
function fore(){
part(); part();
bury();
pair();
swap(); pair();
}
function green(){
nil(); nil(); nil(); pair(); pair(); pair(); // redify
push();
dup();
fore();
pop();
// find
nil(); // prime the pump
while(isTopNil()){
// setup this iteration
drop();
back();
// match?
over(); over();
part(); drop(); // left
part(); nip(); // right
// id
// make a new empty list
nil();
// put the trees at the front of it
pair(); pair();
// prime the pump
nil();
// while the list is not empty
// and the top two items are not nil
while(isTopNil()){
drop();
// get the top two trees from the list
push();
part(); part();
pop();
// are they both nil?
over(); over(); and();
// if they're both nil
if(isTopNil()){
drop();
// drop them: they're identical
drop(); drop();
// is the list nil?
push();
dup();
pop();
// if they're not both nil
}else{
drop();
// is one of them nil?
over(); over(); or();
// if one is nil
if(isTopNil()){
drop();
// then the other one isn't
// they're not identical
// put them back on the list
push();
pair(); pair();
// the list is not nil! It's over.
dup();
pop();
// if they're both not nil
}else{
drop();
// break into their left and right parts
part(); unbury(); part(); unbury();
// put them on the list
push();
pair(); pair(); pair(); pair();
pop();
nil();
}}}
drop();
push();
// end of id
// no more?
over();
part(); drop(); // left
or(); // match or no more?
not();
}
drop();
nip();
// end of find
pop();
}
function edit(){
dup();
// is letter?
part(); nip(); // right
part(); nip(); // right
part(); nip(); // right
not();
if(isTopNil()){
drop();
pair();
}else{
drop();
// compile word
over(); over();
pair();
push();
part(); bury();
swap(); pair();
swap(); pair();
pop();
// blue?
part(); nip(); // right
part(); nip(); // right
part(); nip(); // right
not();
if(isTopNil()){
drop();
green();
// next
// get first word
push();
dup();
fore();
pop();
part(); drop(); // left
part(); nip(); // right
dup();
not();
// while next word isn't nil
while(isTopNil()){
drop();
// get color of word
part();
// if color is gray
if(isTopNil()){ // gray
drop();
part(); nip(); part(); nip(); part(); nip(); part(); nip();
part(); nip(); if(isTopNil()){ drop(); dup(); }else{
part(); nip(); if(isTopNil()){ drop(); drop();}else{
part(); nip(); if(isTopNil()){ drop(); push(); swap(); pop(); pop(); }else{
part(); nip(); if(isTopNil()){ drop(); push(); push(); swap(); pop(); }else{
part(); nip(); if(isTopNil()){ drop(); swap(); }else{
part(); nip(); if(isTopNil()){ drop(); nil(); }else{
part(); nip(); if(isTopNil()){ drop(); pair(); }else{
part(); nip(); if(isTopNil()){ drop(); part(); }else{
drop();}}}}}}}}
// if color is not gray
}else{
// if color is green
part(); nip(); // right
if(iNilTop()){ // green
drop();
green();
// if color is not green
}else{
drop();
drop();
}}
// get next word
push();
dup();
fore();
pop();
part(); drop(); // left
part(); nip(); // right
dup();
not();
}
drop();
// end of next
}else{
drop(); drop();
}
nil();
}}
var abc = ',;:.0123456789abcdefghijklmnopqrstuvwxyz '
function ln(p){return isNil(p)?0:1+ln(Right(p));}
function ls(n){return n<0?ls(abc.length):n?Pair(Nil,ls(n-1)):Nil;}
function n(a){return abc.indexOf(a);}
function a(n){return abc.length<=n?'?':abc.at(n);}
function list(a){return ls(n(a));}
function letter(p){return a(ln(p));}
function letters(p){return isNil(p)?'':letters(Left(p))+letter(Right(p));}
function lists(s){return s==''?Nil:Pair(lists(s.slice(0,-1)),list(s.at(-1)));}
function lprog(p){return isNil(p)?'':lprog(Left(p))+' '+letters(Right(p));}
function rprog(p){return isNil(p)?'':letters(Left(p))+' '+rprog(Right(p));}
function prog(){
return lprog(Left(Left(Right(T))))
+' [ '+letters(Right(Left(T)))+' ] '
+rprog(Right(Left(Right(T))));
}
onkeydown=e=>{
e.preventDefault();
//console.log(e.key);
switch(e.key){
case'Shift':
case'Enter':
case'Control':
case'Alt':
case'Tab':
case'CapsLock':return;
default:
T=Pair(Pair(Left(T), list(e.key)), Right(T));
edit();
console.log(prog());
break;
}
};
The gray word 0, is ‘dup’ and pressing the sequence of keys ‘a:0,a.’ is expected to execute a dup on the stack of T. From a full restart this should add a Nil onto T so that the left part of T is a stack of two Nils.
While I know from past consequences that starting with ‘a:0,a.’ is not as good a test as starting from something like ‘abc,d;f:a:0,a.’ because the latter tests the discrimination of the finding, I’ll keep it simple for now and just start with ‘a:0,a.’.
The consequence expected did not occur and since everything up to and including the compilation has checked out thus far, I’ll start by examining the program from there.
First, a description of the contingency I saw to myself when writing the program. The occasion is post compilation. The front of the list of T is the compiled program and the top of the stack of T is the color of the word just compiled and the second from the top is the spelling of the word.
Since the top is already known to be a color and since the color of blue is a list of length 4 then it can be detected as distinct from gray, red, and green by the fact that if you go down its right part three times (?) it will not be Nil. This is another boundary problem. It conforms to the clopen convention since what is being tested is if the top of the stack is not in ‘[gray, blue)’. Since it is already known that the top is not in and ‘[0, ..)’ this means it is in the intersection of their complements which is just ‘[blue]’ i.e. the top is blue. Note, the order with which these clopen intervals are contemplated is that of the alphabet.
It seems like that the problem is going down the right part three times. I’ll write out the steps in detail using a LISP like notation for the binary trees. Gray is ‘()’, i.e. the empty list Nil, green is ‘(())’, i.e. the list with one item Nil, red is ‘(() ())’, i.e. the list with two items both Nil, and blue is ‘(() () ())’, i.e. the list with three items each Nil. And, already, I can see that I need two rather than three ‘right();’ -s since if ‘(() () ())’ is atop the stack then ‘part(); nip()’ changes it into ‘(() ())’ and another ‘part(); nip();’ turns it into ‘(())’ which, after a ‘not()’, makes it Nil where as starting with any of the other colors the same procedure will make the top not Nil.
I’ll make that change and run the test again.
Ok, that definitely got things going, and now I am stuck in an endless loop. This is a key characteristic of crash based programming. Sadly an endless loop is just a very slow crash e.g. it won’t go on forever because eventually there will be no electricty to supply the machine either because the machien falls apart or the electricity is disconnected for some other reason.
Oh! I forgot to note that because the test I ran begins with ‘:a’ then I also checked the boundary of detecting a blue atop the stack since red is right before blue alphabetically.
My computer is very slow. This is kinda on purpose. But, I do not want to wait very long to restart the browser window and script that run my javascript. This happens often enough when working on such things that a general method presented itself to me. I make a global counter and have it count down to zero every time any while loop goes through an iteration. I’ll call it ‘limit’ and set it to 100. When the value of ‘limit’ hits zero this means the program crashed.
I’ll have the value of ‘limit’ printed in the console after everything else. I find and replace each ‘while(isTopNil())’ with ‘while(isTopNil() && –limit)’ to achieve the desired consequence.
Because of how I’ve unraveled everything, there are only three occurrences of ‘while(isTopNil())’ to even replace. These are also the only places where something is going wrong.
Since adding the parts involving ‘limit’ may have introduced their own bugs, I’ll now debug this new program based on what the previous version actually did i.e. I expect the value of limit to go to 0 and for everything to stop.
Something completely unexpected happened. The console printed out ‘uncaught reference error’ in the definition of ‘edit’. I apparently wrote ‘iNilTop’ instead of ‘isTopNil’ inside a conditional. But, this is unexpected in that I thought such an error would have appeared before setting up the code involving limit. It didn’t, and this is probably because before the code that prints the error could be executed, the while loops where spinning out of control. This is obviously metaphor, but I have no interest in chasing down how javascript is executed by the browser.
I restarted everything from scratch and input ‘a:0,a.’ and was further surprsied by the output. The value of limit is negative two and not zero as I expected. It occurred to me immediately that I had not forseen the consequences of the conditional execution of ‘&&’. To make sure that ‘–limit’ is executed on every step of each while loop, I have to put it first and then the other conditions after that i.e. substitute ‘while(isTopNil() && –limit)’ with ‘while(–limit && isTopNil)’.
Having done that I’ll now run the same test input from scratch. I do not know what to expect.
Ha, now the value of limit is negative four. I’m going to avoid tracing down what is going on by going straight to a strict inequality ‘0<–limit’.
Ok, did that, and now the value of ‘limit’ is negative two. I can interpret this to mean that execution went through two while loops after the value of limit was zero.
Now, debugging can turn to an examination of the definition of ‘green’.
I’ll have to pick this up after I eat some lunch.
The hardest function is probably ‘id’ so ‘next’ should be easier. Starting from
function next(){
// get next word
push();
dup();
fore();
pop();
part(); drop(); // left
part(); nip(); // right
if(isNilTop()){ // no next word
drop();
}else{ // do color
part();
if(isNilTop()){ // gray
drop();
part(); nip(); if(isNilTop()){ drop(); dup(); }else{
part(); nip(); if(isNilTop()){ drop(); graypop(); }else{
part(); nip(); if(isNilTop()){ drop(); graypush(); }else{
part(); nip(); if(isNilTop()){ drop(); swap(); }else{
part(); nip(); if(isNilTop()){ drop(); nil(); }else{
part(); nip(); if(isNilTop()){ drop(); pair(); }else{
part(); nip(); if(isNilTop()){ drop(); part(); }else{
drop();}}}}}}}
}else{
part(); nip(); // right
if(iNilTop()){ // green
drop();
green();
}else{
drop(); drop();
}}
next();
}}
I got to:
// get first word
push();
dup();
fore();
pop();
part(); drop(); // left
part(); nip(); // right
dup();
not();
// while next word isn't nil
while(isNilTop()){
drop();
// get color of word
part();
// if color is gray
if(isNilTop()){ // gray
drop();
part(); nip(); if(isNilTop()){ drop(); dup(); }else{
part(); nip(); if(isNilTop()){ drop(); graypop(); }else{
part(); nip(); if(isNilTop()){ drop(); graypush(); }else{
part(); nip(); if(isNilTop()){ drop(); swap(); }else{
part(); nip(); if(isNilTop()){ drop(); nil(); }else{
part(); nip(); if(isNilTop()){ drop(); pair(); }else{
part(); nip(); if(isNilTop()){ drop(); part(); }else{
drop();}}}}}}}
// if color is not gray
}else{
// if color is green
part(); nip(); // right
if(iNilTop()){ // green
drop();
green();
// if color is not green
}else{
drop();
drop();
}}
// get next word
push();
dup();
fore();
pop();
part(); drop(); // left
part(); nip(); // right
dup();
not();
}
drop();
Now to put it all together and read it anew.
var Nil = new Object();
function isNil(p){return Object.is(p, Nil);}
function Pair(l,r){return new Array(l,r);}
function Left(p){if(isNil(p)) return Nil; else return p[0];}
function Right(p){if(isNil(p)) return Nil; else return p[1];}
var T=Nil;
function isTopNil(){return isNil(Right(Left(T)));}
function drop(){T=Pair(Left(Left(T)), Right(T));}
function dup(){T=Pair(Pair(Left(T), Right(Left(T))), Right(T));}
function pop(){T=Pair(Left(Left(T)), Pair(Right(Left(T)), Right(T)));}
function push(){T=Pair(Pair(Left(T), Left(Right(T))), Right(Right(T)));}
function swap(){T=Pair(Pair(Pair(Left(Left(Left(T))),
Right(Left(T))), Right(Left(Right(T)))), Right(T));}
function nil(){T=Pair(Pair(Left(T), Nil), Right(T));}
function pair(){T=Pair(Pair(Left(Left(Left(T))),
Pair(Right(Left(Left(T))), Right(Left(T)))), Right(T));}
function part(){T=Pair(Pair(Pair(Left(Left(T)),
Left(Right(Left(T)))), Right(Right(Left(T)))), Right(T));}
function nip(){pop(); drop(); push();}
function over(){pop(); dup(); push(); swap();}
function dig(){over(); pop(); nip();}
function bury(){dig(); dig(); push(); push();}
function unbury(){bury(); bury();}
function select(){
if(isNilTop()){
drop();
nip();
}else{
drop(); drop();
}}
function not(){
nil();
nil(); nil(); pair();
unbury();
select();
}
function or(){
dup();
select();
}
function and(){
not();
swap();
not();
or();
not();
}
function back(){
part();
swap(); part();
bury();
pair(); pair();
}
function fore(){
part(); part();
bury();
pair();
swap(); pair();
}
function green(){
nil(); nil(); nil(); pair(); pair(); pair(); // redify
push();
dup();
fore();
pop();
// find
nil(); not(); // prime the pump
while(isNilTop()){
// setup this iteration
drop();
back();
// match?
over(); over();
part(); drop(); // left
part(); nip(); // right
// id
// make a new empty list
nil();
// put the trees at the front of it
pair(); pair();
// prime the pump
nil();
// while the list is not empty
// and the top two items are not nil
while(isNilTop()){
drop();
// get the top two trees from the list
push();
part(); part();
pop();
// are they both nil?
over(); over(); and();
// if they're both nil
if(isNilTop()){
drop();
// drop them: they're identical
drop(); drop();
// is the list nil?
push();
dup();
pop();
// if they're not both nil
}else{
drop();
// is one of them nil?
over(); over(); or();
// if one is nil
if(isNilTop()){
drop();
// then the other one isn't
// they're not identical
// put them back on the list
push();
pair(); pair();
// the list is not nil! It's over.
dup();
pop();
// if they're both not nil
}else{
drop();
// break into their left and right parts
part(); unbury(); part(); unbury();
// put them on the list
push();
pair(); pair(); pair(); pair();
pop();
nil();
}}}
drop();
push();
// end of id
// no more?
over();
part(); drop(); // left
or(); // match or no more?
not();
}
drop();
nip();
// end of find
pop();
}
function edit(){
dup();
// is letter
part(); nip(); // right
part(); nip(); // right
part(); nip(); // right
part(); nip(); // right
not();
if(isNilTop()){
drop();
pair();
}else{
// compile word
over(); over();
pair();
push();
part(); bury();
swap(); pair();
swap(); pair();
// blue?
part(); nip(); // right
part(); nip(); // right
part(); nip(); // right
not();
if(isNilTop()){ // blue
drop();
green();
// next
// get first word
push();
dup();
fore();
pop();
part(); drop(); // left
part(); nip(); // right
dup();
not();
// while next word isn't nil
while(isNilTop()){
drop();
// get color of word
part();
// if color is gray
if(isNilTop()){ // gray
drop();
part(); nip(); if(isNilTop()){ drop(); dup(); }else{
part(); nip(); if(isNilTop()){ drop(); drop();}else{
part(); nip(); if(isNilTop()){ drop(); push(); swap(); pop(); pop(); }else{
part(); nip(); if(isNilTop()){ drop(); push(); push(); swap(); pop(); }else{
part(); nip(); if(isNilTop()){ drop(); swap(); }else{
part(); nip(); if(isNilTop()){ drop(); nil(); }else{
part(); nip(); if(isNilTop()){ drop(); pair(); }else{
part(); nip(); if(isNilTop()){ drop(); part(); }else{
drop();}}}}}}}}
// if color is not gray
}else{
// if color is green
part(); nip(); // right
if(iNilTop()){ // green
drop();
green();
// if color is not green
}else{
drop();
drop();
}}
// get next word
push();
dup();
fore();
pop();
part(); drop(); // left
part(); nip(); // right
dup();
not();
}
drop();
// end of next
}else{
drop(); drop();
}
nil();
}}
Here’s an updated version of ‘id’ that uses a list instead of a stack to do the tree traversal.
function id(){
// make a new empty list
nil();
// put the trees at the front of it
pair(); pair();
// prime the pump
nil();
// while the list is not empty
// and the top two items are not nil
while(isNilTop()){
drop();
// get the top two trees from the list
push();
part(); part();
pop();
// are they both nil?
over(); over(); and();
// if they're both nil
if(isNilTop()){
drop();
// drop them: they're identical
drop(); drop();
// is the list nil?
push();
dup();
pop();
// if they're not both nil
}else{
drop();
// is one of them nil?
over(); over(); or();
// if one is nil
if(isNilTop()){
drop();
// then the other one isn't
// they're not identical
// put them back on the list
push();
pair(); pair();
// the list is not nil! It's over.
dup();
pop();
// if they're both not nil
}else{
drop();
// break into their left and right parts
part(); unbury(); part(); unbury();
// put them on the list
push();
pair(); pair(); pair(); pair();
pop();
nil();
}}}
drop();
push();
}
Since, ‘id’ only appears in ‘find’ then I’ll go ahead and put that code there where it goes:
var Nil = new Object();
function isNil(p){return Object.is(p, Nil);}
function Pair(l,r){return new Array(l,r);}
function Left(p){if(isNil(p)) return Nil; else return p[0];}
function Right(p){if(isNil(p)) return Nil; else return p[1];}
var T=Nil;
function isTopNil(){return isNil(Right(Left(T)));}
function drop(){T=Pair(Left(Left(T)), Right(T));}
function dup(){T=Pair(Pair(Left(T), Right(Left(T))), Right(T));}
function pop(){T=Pair(Left(Left(T)), Pair(Right(Left(T)), Right(T)));}
function push(){T=Pair(Pair(Left(T), Left(Right(T))), Right(Right(T)));}
function swap(){T=Pair(Pair(Pair(Left(Left(Left(T))),
Right(Left(T))), Right(Left(Right(T)))), Right(T));}
function nil(){T=Pair(Pair(Left(T), Nil), Right(T));}
function pair(){T=Pair(Pair(Left(Left(Left(T))),
Pair(Right(Left(Left(T))), Right(Left(T)))), Right(T));}
function part(){T=Pair(Pair(Pair(Left(Left(T)),
Left(Right(Left(T)))), Right(Right(Left(T)))), Right(T));}
function nip(){pop(); drop(); push();}
function over(){pop(); dup(); push(); swap();}
function dig(){over(); pop(); nip();}
function bury(){dig(); dig(); push(); push();}
function unbury(){bury(); bury();}
function select(){
if(isNilTop()){
drop();
nip();
}else{
drop(); drop();
}}
function not(){
nil();
nil(); nil(); pair();
unbury();
select();
}
function or(){
dup();
select();
}
function and(){
not();
swap();
not();
or();
not();
}
function back(){
part();
swap(); part();
bury();
pair(); pair();
}
function fore(){
part(); part();
bury();
pair();
swap(); pair();
}
function green(){
nil(); nil(); nil(); pair(); pair(); pair(); // redify
push();
dup();
fore();
pop();
// find
nil(); not(); // prime the pump
while(isNilTop()){
// setup this iteration
drop();
back();
// match?
over(); over();
part(); drop(); // left
part(); nip(); // right
// id
// make a new empty list
nil();
// put the trees at the front of it
pair(); pair();
// prime the pump
nil();
// while the list is not empty
// and the top two items are not nil
while(isNilTop()){
drop();
// get the top two trees from the list
push();
part(); part();
pop();
// are they both nil?
over(); over(); and();
// if they're both nil
if(isNilTop()){
drop();
// drop them: they're identical
drop(); drop();
// is the list nil?
push();
dup();
pop();
// if they're not both nil
}else{
drop();
// is one of them nil?
over(); over(); or();
// if one is nil
if(isNilTop()){
drop();
// then the other one isn't
// they're not identical
// put them back on the list
push();
pair(); pair();
// the list is not nil! It's over.
dup();
pop();
// if they're both not nil
}else{
drop();
// break into their left and right parts
part(); unbury(); part(); unbury();
// put them on the list
push();
pair(); pair(); pair(); pair();
pop();
nil();
}}}
drop();
push();
// end of id
// no more?
over();
part(); drop(); // left
or(); // match or no more?
not();
}
drop();
nip();
// end of find
pop();
}
function graypop(){push(); swap(); pop(); pop();}
function graypush(){push(); push(); swap(); pop();}
function next(){
// get next word
push();
dup();
fore();
pop();
part(); drop(); // left
part(); nip(); // right
if(isNilTop()){ // no next word
drop();
}else{ // do color
part();
if(isNilTop()){ // gray
drop();
part(); nip(); if(isNilTop()){ drop(); dup(); }else{
part(); nip(); if(isNilTop()){ drop(); graypop(); }else{
part(); nip(); if(isNilTop()){ drop(); graypush(); }else{
part(); nip(); if(isNilTop()){ drop(); swap(); }else{
part(); nip(); if(isNilTop()){ drop(); nil(); }else{
part(); nip(); if(isNilTop()){ drop(); pair(); }else{
part(); nip(); if(isNilTop()){ drop(); part(); }else{
drop();}}}}}}}
}else{
part(); nip(); // right
if(iNilTop()){ // green
drop();
green();
}else{
drop(); drop();
}}
next();
}}
function edit(){
dup();
// is letter
part(); nip(); // right
part(); nip(); // right
part(); nip(); // right
part(); nip(); // right
not();
if(isNilTop()){
drop();
pair();
}else{
// compile word
over(); over();
pair();
push();
part(); bury();
swap(); pair();
swap(); pair();
// blue?
part(); nip(); // right
part(); nip(); // right
part(); nip(); // right
not();
if(isNilTop()){ // blue
drop();
green();
next();
}else{
drop(); drop();
}
nil();
}}
Next to do is to turn ‘next’ into its iterative form.
This continues work on my programming environment.
There are a few bugs in the last implementation of the core logic to my programming environment e.g. the definition I gave to ‘nor’ actually made it into ‘or’, the definition of ‘nfi’ accidentally repeated the definition of ‘fi’, the method of introducing the truth functions was wrong (based on the ‘nor’ mistake), and there was a problem with the compiler (as far as I can recall).
Here are some of the inspiring programming languages that went into my programming environment:
Ultimately, the inspiration for the present design of my programming environment is my further work on Quine’s predicate functor logic.
The next step in work is input and output. Input and output is actually not the kind of problem it seems to be when you’re taught it. It is closer to a big problem lurking in the way I’ve implemented everything. Remember, the goal is simplicity, and one feature of that is that any implementation must automatically mirror the way it would be implemented in my programming environment.
The looming problems are of two similar kinds: conditional execution and recursive definition. The function defining ‘id’ replaces the top two items on the stack with Nil if they are identical and something not Nil if they are not identical. Identity is indiscernibility with respect to a predicate of ordered pairing i.e. ordered pairs are identical if and only if their left parts are identical and their right parts are identical.
The recursion part interests me the most. There are only three defined functions that execute themselves: id, find, and next. The idea is to get all this down to a single while loop. The method is what is commonly called ‘a virtual machine’. I don’t like that phrase, but it is the one that people are familiar with. I already started making the virtual machine in #2025-1007-1413.
It ends up that the heart of the virtual machine is just the function ‘gray’.
var Nil = new Object();
function isNil(p){return Object.is(p, Nil);}
function Pair(l,r){return new Array(l,r);}
function Left(p){if(isNil(p)) return Nil; else return p[0];}
function Right(p){if(isNil(p)) return Nil; else return p[1];}
var T=Nil;
function isTopNil(){return isNil(Right(Left(T)));}
function drop(){T=Pair(Left(Left(T)), Right(T));}
function dup(){T=Pair(Pair(Left(T), Right(Left(T))), Right(T));}
function pop(){T=Pair(Left(Left(T)), Pair(Right(Left(T)), Right(T)));}
function push(){T=Pair(Pair(Left(T), Left(Right(T))), Right(Right(T)));}
function swap(){T=Pair(Pair(Pair(Left(Left(Left(T))),
Right(Left(T))), Right(Left(Right(T)))), Right(T));}
function nil(){T=Pair(Pair(Left(T), Nil), Right(T));}
function pair(){T=Pair(Pair(Left(Left(Left(T))),
Pair(Right(Left(Left(T))), Right(Left(T)))), Right(T));}
function part(){T=Pair(Pair(Pair(Left(Left(T)),
Left(Right(Left(T)))), Right(Right(Left(T)))), Right(T));}
function graypop(){push(); swap(); pop(); pop();}
function graypush(){push(); push(); swap(); pop();}
function gray(){
part(); nip(); if(isNilTop()){ drop(); dup(); }else{
part(); nip(); if(isNilTop()){ drop(); graypop(); }else{
part(); nip(); if(isNilTop()){ drop(); graypush(); }else{
part(); nip(); if(isNilTop()){ drop(); swap(); }else{
part(); nip(); if(isNilTop()){ drop(); nil(); }else{
part(); nip(); if(isNilTop()){ drop(); pair(); }else{
part(); nip(); if(isNilTop()){ drop(); part(); }else{ drop();}}}}}}}}
Now, using defined functions must be as simple as replacing the defined function with its definition i.e. there can be no mention of the function in its definition as we are trying to make that operation possible rather than assume it is there already. This is called ‘unrolling’ and is much like what people do when they learn to program in assembly.
The function gray is only used in one place: inside the definition of ‘hue’. So rather than introduce ‘gray’ as a separately defined function we will just put the code directly where it goes in ‘hue’. I’ve put some of the other functions in this clump of code to help me see the next step more clearly to myself.
var Nil = new Object();
function isNil(p){return Object.is(p, Nil);}
function Pair(l,r){return new Array(l,r);}
function Left(p){if(isNil(p)) return Nil; else return p[0];}
function Right(p){if(isNil(p)) return Nil; else return p[1];}
var T=Nil;
function isTopNil(){return isNil(Right(Left(T)));}
function drop(){T=Pair(Left(Left(T)), Right(T));}
function dup(){T=Pair(Pair(Left(T), Right(Left(T))), Right(T));}
function pop(){T=Pair(Left(Left(T)), Pair(Right(Left(T)), Right(T)));}
function push(){T=Pair(Pair(Left(T), Left(Right(T))), Right(Right(T)));}
function swap(){T=Pair(Pair(Pair(Left(Left(Left(T))),
Right(Left(T))), Right(Left(Right(T)))), Right(T));}
function nil(){T=Pair(Pair(Left(T), Nil), Right(T));}
function pair(){T=Pair(Pair(Left(Left(Left(T))),
Pair(Right(Left(Left(T))), Right(Left(T)))), Right(T));}
function part(){T=Pair(Pair(Pair(Left(Left(T)),
Left(Right(Left(T)))), Right(Right(Left(T)))), Right(T));}
function graypop(){push(); swap(); pop(); pop();}
function graypush(){push(); push(); swap(); pop();}
function zero(){nil();}
function succ(){nil(); cons();} // natural sucessor
function one(){zero(); succ();}
function two(){one(); succ();}
function redify(){two(); pair();}
function green(){redify(); pup(); fore(); pop(); find(); pop();}
function hue(){
part();
if(isNilTop()){ // do gray word
drop();
part(); nip(); if(isNilTop()){ drop(); dup(); }else{
part(); nip(); if(isNilTop()){ drop(); graypop(); }else{
part(); nip(); if(isNilTop()){ drop(); graypush(); }else{
part(); nip(); if(isNilTop()){ drop(); swap(); }else{
part(); nip(); if(isNilTop()){ drop(); nil(); }else{
part(); nip(); if(isNilTop()){ drop(); pair(); }else{
part(); nip(); if(isNilTop()){ drop(); part(); }else{ drop();}}}}}}}}
}else{
part(); nip(); if(isNilTop()){ drop(); green(); }else{ drop(); drop();}}}
Green is something to worry about later. Now, ‘hue’ only occurs in ‘next’ which is great! So we can put its definition there.
var Nil = new Object();
function isNil(p){return Object.is(p, Nil);}
function Pair(l,r){return new Array(l,r);}
function Left(p){if(isNil(p)) return Nil; else return p[0];}
function Right(p){if(isNil(p)) return Nil; else return p[1];}
var T=Nil;
function isTopNil(){return isNil(Right(Left(T)));}
function drop(){T=Pair(Left(Left(T)), Right(T));}
function dup(){T=Pair(Pair(Left(T), Right(Left(T))), Right(T));}
function pop(){T=Pair(Left(Left(T)), Pair(Right(Left(T)), Right(T)));}
function push(){T=Pair(Pair(Left(T), Left(Right(T))), Right(Right(T)));}
function swap(){T=Pair(Pair(Pair(Left(Left(Left(T))),
Right(Left(T))), Right(Left(Right(T)))), Right(T));}
function nil(){T=Pair(Pair(Left(T), Nil), Right(T));}
function pair(){T=Pair(Pair(Left(Left(Left(T))),
Pair(Right(Left(Left(T))), Right(Left(T)))), Right(T));}
function part(){T=Pair(Pair(Pair(Left(Left(T)),
Left(Right(Left(T)))), Right(Right(Left(T)))), Right(T));}
function nip(){pop(); drop(); push();}
function over(){pop(); dup(); push(); swap();}
function dig(){over(); pop(); nip();}
function bury(){dig(); dig(); push(); push();}
function unbury(){bury(); bury();}
function id(){
if(isNilTop()){
drop();
}else{
swap(); drop();
part(); unbury();
part(); unbury();
id();
bury();
id();
and();
}}
function back(){
part();
swap(); part();
bury();
pair(); pair();
}
function fore(){
part(); part();
bury();
pair();
swap(); pair();
}
function find(){
over(); over();
part(); drop(); // left
part(); nip(); // right
id();
if(isNilTop()){
drop(); nip();
}else{
drop();
back();
find();
}}
function green(){
nil(); nil(); nil(); pair(); pair(); pair(); // redify
push();
dup();
fore();
pop();
find();
pop();
}
function graypop(){push(); swap(); pop(); pop();}
function graypush(){push(); push(); swap(); pop();}
function next(){
// get next word
push();
dup();
fore();
pop();
part(); drop(); // left
part(); nip(); // right
if(isNilTop()){ // no next word
drop();
}else{ // do color
part();
if(isNilTop()){ // do gray word
drop();
part(); nip(); if(isNilTop()){ drop(); dup(); }else{
part(); nip(); if(isNilTop()){ drop(); graypop(); }else{
part(); nip(); if(isNilTop()){ drop(); graypush(); }else{
part(); nip(); if(isNilTop()){ drop(); swap(); }else{
part(); nip(); if(isNilTop()){ drop(); nil(); }else{
part(); nip(); if(isNilTop()){ drop(); pair(); }else{
part(); nip(); if(isNilTop()){ drop(); part(); }else{ drop();}}}}}}}}
}else{ // do green word
part(); nip();
if(isNilTop()){
drop();
green();
}else{
drop(); drop();
}}
next();
}}
function isABC(){
part(); nip(); // right
part(); nip(); // right
part(); nip(); // right
part(); nip(); // right
not();
}
function isBlue(){
part(); nip(); // right
part(); nip(); // right
part(); nip(); // right
not();
}
function compile(){
over(); over();
pair();
push();
part(); bury();
swap(); pair();
swap(); pair();
}
function edit(){
dup();
isABC();
if(isNilTop()){
drop();
pair();
}else{
compile();
isBlue();
if(isNilTop()){
drop();
green();
next();
}else{
drop(); drop();
}
nil();
}}
A lot of little edits there to bring things together. Also a showcase of an extremely important practice: replacing definitions with comments. Factoring a routine into subroutines, which is what programming well used to entail, was done first by way of effective commenting (well, that and basic operations that controlled what was executed when).
A hint about how this all works: gray words are what other programming languages would call literals or machine code.
There is another big hint that isn’t here: every recursion can be turned into an iteration. This is of foundational importance to the practices of mathematics, but it also has the practical effect of giving hope that there is a way to replace recursively defined functions with while loops of appropriate operations.
Now some further substitution of definitions:
var Nil = new Object();
function isNil(p){return Object.is(p, Nil);}
function Pair(l,r){return new Array(l,r);}
function Left(p){if(isNil(p)) return Nil; else return p[0];}
function Right(p){if(isNil(p)) return Nil; else return p[1];}
var T=Nil;
function isTopNil(){return isNil(Right(Left(T)));}
function drop(){T=Pair(Left(Left(T)), Right(T));}
function dup(){T=Pair(Pair(Left(T), Right(Left(T))), Right(T));}
function pop(){T=Pair(Left(Left(T)), Pair(Right(Left(T)), Right(T)));}
function push(){T=Pair(Pair(Left(T), Left(Right(T))), Right(Right(T)));}
function swap(){T=Pair(Pair(Pair(Left(Left(Left(T))),
Right(Left(T))), Right(Left(Right(T)))), Right(T));}
function nil(){T=Pair(Pair(Left(T), Nil), Right(T));}
function pair(){T=Pair(Pair(Left(Left(Left(T))),
Pair(Right(Left(Left(T))), Right(Left(T)))), Right(T));}
function part(){T=Pair(Pair(Pair(Left(Left(T)),
Left(Right(Left(T)))), Right(Right(Left(T)))), Right(T));}
function nip(){pop(); drop(); push();}
function over(){pop(); dup(); push(); swap();}
function dig(){over(); pop(); nip();}
function bury(){dig(); dig(); push(); push();}
function unbury(){bury(); bury();}
function id(){
if(isNilTop()){
drop();
}else{
swap(); drop();
part(); unbury();
part(); unbury();
id();
bury();
id();
and();
}}
function back(){
part();
swap(); part();
bury();
pair(); pair();
}
function fore(){
part(); part();
bury();
pair();
swap(); pair();
}
function find(){
over(); over();
part(); drop(); // left
part(); nip(); // right
id();
if(isNilTop()){
drop(); nip();
}else{
drop();
back();
find();
}}
function green(){
nil(); nil(); nil(); pair(); pair(); pair(); // redify
push();
dup();
fore();
pop();
find();
pop();
}
function graypop(){push(); swap(); pop(); pop();}
function graypush(){push(); push(); swap(); pop();}
function next(){
// get next word
push();
dup();
fore();
pop();
part(); drop(); // left
part(); nip(); // right
if(isNilTop()){ // no next word
drop();
}else{ // do color
part();
if(isNilTop()){ // do gray word
drop();
part(); nip(); if(isNilTop()){ drop(); dup(); }else{
part(); nip(); if(isNilTop()){ drop(); graypop(); }else{
part(); nip(); if(isNilTop()){ drop(); graypush(); }else{
part(); nip(); if(isNilTop()){ drop(); swap(); }else{
part(); nip(); if(isNilTop()){ drop(); nil(); }else{
part(); nip(); if(isNilTop()){ drop(); pair(); }else{
part(); nip(); if(isNilTop()){ drop(); part(); }else{ drop();}}}}}}}
}else{ // do green word
part(); nip();
if(isNilTop()){
drop();
green();
}else{
drop(); drop();
}}
next();
}}
function edit(){
dup();
// is letter
part(); nip(); // right
part(); nip(); // right
part(); nip(); // right
part(); nip(); // right
not();
if(isNilTop()){
drop();
pair();
}else{
// compile word
over(); over();
pair();
push();
part(); bury();
swap(); pair();
swap(); pair();
// is blue
part(); nip(); // right
part(); nip(); // right
part(); nip(); // right
not();
if(isNilTop()){
drop();
green();
next();
}else{
drop(); drop();
}
nil();
}}
Let’s work on ‘find’ for now. The aim is to get the function
function find(){
over(); over();
part(); drop(); // left
part(); nip(); // right
id();
if(isNilTop()){
drop(); nip();
}else{
drop();
back();
find();
}}
into a form that doesn’t use ‘find’ in its own definition (again, this is because we are trying to ‘push out’ the implicit recursions permitted by the native language). Doing this also presented me with a missing condition!
function find(){
// match?
over(); over();
part(); drop(); // left
part(); nip(); // right
id();
// no more?
over();
part(); drop(); // left
or(); // match or no more?
if(isNilTop()){
drop(); nip();
}else{
drop();
back();
find();
}
Ok, so now how to flip this around so that it’s a ‘while’. First, an overt convention: we find backwards from the current word, that means the current word is not eligible to be “found”. Here is a quick and dirty way to flip things around:
function find(){
nil(); // prime the pump
while(isNilTop()){
// setup this iteration
drop();
back();
// match?
over(); over();
part(); drop(); // left
part(); nip(); // right
id();
// no more?
over();
part(); drop(); // left
or(); // match or no more?
not();
}
drop();
nip();
While this could be shortened with a ‘do while’ I’m going to stick with ‘while’. It seems it would be regrettable to lose the independently defined function of ‘find’, but the problem to solve is starting everything up from the minimally mutaliting foundation. In any effective programming environment it will be found that the primitive operations have a double life if they are to be part of the hierarchy of langauges constructed in such an environment.
So now the state of affairs is as follows:
var Nil = new Object();
function isNil(p){return Object.is(p, Nil);}
function Pair(l,r){return new Array(l,r);}
function Left(p){if(isNil(p)) return Nil; else return p[0];}
function Right(p){if(isNil(p)) return Nil; else return p[1];}
var T=Nil;
function isTopNil(){return isNil(Right(Left(T)));}
function drop(){T=Pair(Left(Left(T)), Right(T));}
function dup(){T=Pair(Pair(Left(T), Right(Left(T))), Right(T));}
function pop(){T=Pair(Left(Left(T)), Pair(Right(Left(T)), Right(T)));}
function push(){T=Pair(Pair(Left(T), Left(Right(T))), Right(Right(T)));}
function swap(){T=Pair(Pair(Pair(Left(Left(Left(T))),
Right(Left(T))), Right(Left(Right(T)))), Right(T));}
function nil(){T=Pair(Pair(Left(T), Nil), Right(T));}
function pair(){T=Pair(Pair(Left(Left(Left(T))),
Pair(Right(Left(Left(T))), Right(Left(T)))), Right(T));}
function part(){T=Pair(Pair(Pair(Left(Left(T)),
Left(Right(Left(T)))), Right(Right(Left(T)))), Right(T));}
function nip(){pop(); drop(); push();}
function over(){pop(); dup(); push(); swap();}
function dig(){over(); pop(); nip();}
function bury(){dig(); dig(); push(); push();}
function unbury(){bury(); bury();}
function select(){
if(isNilTop()){
drop();
nip();
}else{
drop(); drop();
}}
function not(){
nil();
nil(); nil(); pair();
unbury();
select();
}
function or(){
dup();
select();
}
function and(){
not();
swap();
not();
or();
not();
}
function id(){
if(isNilTop()){
drop();
}else{
swap(); drop();
part(); unbury();
part(); unbury();
id();
bury();
id();
and();
}}
function back(){
part();
swap(); part();
bury();
pair(); pair();
}
function fore(){
part(); part();
bury();
pair();
swap(); pair();
}
function green(){
nil(); nil(); nil(); pair(); pair(); pair(); // redify
push();
dup();
fore();
pop();
// find
nil(); // prime the pump
while(isNilTop()){
// setup this iteration
drop();
back();
// match?
over(); over();
part(); drop(); // left
part(); nip(); // right
id();
// no more?
over();
part(); drop(); // left
or(); // match or no more?
not();
}
drop();
nip();
pop();
}
function graypop(){push(); swap(); pop(); pop();}
function graypush(){push(); push(); swap(); pop();}
function next(){
// get next word
push();
dup();
fore();
pop();
part(); drop(); // left
part(); nip(); // right
if(isNilTop()){ // no next word
drop();
}else{ // do color
part();
if(isNilTop()){ // gray
drop();
part(); nip(); if(isNilTop()){ drop(); dup(); }else{
part(); nip(); if(isNilTop()){ drop(); graypop(); }else{
part(); nip(); if(isNilTop()){ drop(); graypush(); }else{
part(); nip(); if(isNilTop()){ drop(); swap(); }else{
part(); nip(); if(isNilTop()){ drop(); nil(); }else{
part(); nip(); if(isNilTop()){ drop(); pair(); }else{
part(); nip(); if(isNilTop()){ drop(); part(); }else{
drop();}}}}}}}
}else{
part(); nip(); // right
if(iNilTop()){ // green
drop();
green();
}else{
drop(); drop();
}}
next();
}}
function edit(){
dup();
// is letter
part(); nip(); // right
part(); nip(); // right
part(); nip(); // right
part(); nip(); // right
not();
if(isNilTop()){
drop();
pair();
}else{
// compile word
over(); over();
pair();
push();
part(); bury();
swap(); pair();
swap(); pair();
// blue?
part(); nip(); // right
part(); nip(); // right
part(); nip(); // right
not();
if(isNilTop()){ // blue
drop();
green();
next();
}else{
drop(); drop();
}
nil();
}}
Now the next function to fix up into an interation is ‘id’. Starting from
function id(){
if(isNilTop()){
drop();
}else{
swap(); drop();
part(); unbury();
part(); unbury();
id();
bury();
id();
and();
}}
Identity is established here by a method called ‘tree recursion’. This checks the right branchs first, then the left branchs, and then if they’re both correspondingly identitcal the whole thing is. We can stop checking a pair if both are nil (in which case they are identical) or if one of them is nil and the other isn’t (in which case they are not identical), but if they’re both not nil then we can’t stop checking. So the condition for ‘going on’ is ‘they’re both not nil’.
Now, we must confront an important problem when dealing with trees: iterating with a stack. While it’s funny, we’re going to put a stack on the stack (so we can stack while we stack).
function id(){
nil();
swap(); pair();
swap(); pair();
pop();
nil();
while(isNilTop()){ // the stack is not nil
drop();
push();
part(); swap(); part(); bury(); // pop top two subtrees
pop(); // put stack out of the way
over(); over(); and(); // are they both nil?
if(isNiltop()){ // yes
drop();
drop();
drop();
push();
dup();
pop();
not();
}else{
drop();
// is one nil and the other not?
over(); over(); or();
if(isNilTop()){ // yes
drop();
push();
swap(); pair();
swap(); pair();
dup();
pop();
}else{// they're both not nil
drop();
part();
unbury();
part();
unbury();
push();
swap(); pair();
swap(); pair();
swap(); pair();
swap(); pair();
pop();
nil();
}
}
drop();
push();
}
That is more or less the solution, but I’m tired and will have to take a rest before I can see through it all.
Rather than introduce the two constructions of composition and contingency as functions of functions in the javascript code that simulates my programming environment, I’m just going to make everything as boring and pedestrian as possible.
var Nil = new Object();
function isNil(p){return Object.is(p, Nil);}
function Pair(l,r){return new Array(l,r);}
function Left(p){if(isNil(p)) return Nil; else return p[0];}
function Right(p){if(isNil(p)) return Nil; else return p[1];}
var T=Nil;
function isTopNil(){return isNil(Right(Left(T)));}
function drop(){T=Pair(Left(Left(T)), Right(T));}
function dup(){T=Pair(Pair(Left(T), Right(Left(T))), Right(T));}
function pop(){T=Pair(Left(Left(T)), Pair(Right(Left(T)), Right(T)));}
function push(){T=Pair(Pair(Left(T), Left(Right(T))), Right(Right(T)));}
function swap(){
T=Pair(Pair(Pair(Left(Left(Left(T))),
Right(Left(T))), Right(Left(Right(T)))), Right(T));}
function nil(){T=Pair(Pair(Left(T), Nil), Right(T));}
function pair(){
T=Pair(Pair(Left(Left(Left(T))),
Pair(Right(Left(Left(T))), Right(Left(T)))), Right(T));}
function part(){
T=Pair(Pair(Pair(Left(Left(T)),
Left(Right(Left(T)))), Right(Right(Left(T)))), Right(T));}
function drop2(){drop(); drop();}
function pop2(){pop(); pop();}
function push2(){push(); push();}
function pair2(){pair(); pair();}
function part2(){part(); part();}
function pup(){push(); dup();}
function dush(){drop(); push();}
function nip(){pop(); dush();}
function dip(){drop(); nip();}
function over(){pop(); dush(); swap();}
function over2(){over(); over();}
function dig(){over(); pop(); nip();}
function dig2(){dig(); dig();}
function bury(){dig2(); push2();}
function unbury(){bury(); bury();}
function party(){part(); unbury();}
function party2(){party(); party();}
function l(){part(); drop();}
function r(){part(); nip();}
function r2(){r(); r();}
function r3(){r(); r2();}
function r4(){r2(); r2();}
function lr(){l(); r();}
function cons(){swap(); pair();}
function cons2(){cons(); cons();}
function snoc(){part(); swap();}
function snoc2(){snoc(); snoc();}
function select(){if(isNilTop()) dip(); else drop2();}
function nor(){dup(); select();} // joint denial
function not(){dup(); nor();} // negation
function or(){nor(); not();} // alternation
function IF(){not(); or();} // converse conditional
function nif(){IF(); not();} // complementary converse conditional
function nfi(){nfi(); not();} // complementary conditional
function fi(){nfi(); not();} // conditional
function nand(){not(); fi();} // alternative denial
function and(){nand(); not();} // conjunction
function xor(){over2(); nand(); bury(); or(); and();} // exclusive alternation
function iff(){xor(); not();} // biconditional
function zero(){nil();}
function succ(){nil(); cons();} // natural sucessor
function one(){zero(); succ();}
function two(){one(); succ();}
function three(){two(); succ();}
function id(){if(isNilTop()){ drop(); }else{
swap(); drop(); party2(); id(); bury(); id(); and(); }}
function back(){snoc(); part(); bury(); pair2();}
function fore(){part2(); bury(); pair(); cons();}
function find(){over2(); lr(); id();
if(isNilTop()){ dip(); }else{ drop(); back(); find(); }}
function graypop(){push(); swap(); pop2();}
function graypush(){push2(); swap(); pop();}
function gray(){drop2(); r();
if(isNilTop()){ drop(); dup(); }else{ r();
if(isNilTop()){ drop(); graypop(); }else{ r();
if(isNilTop()){ drop(); graypush(); }else{ r();
if(isNilTop()){ drop(); swap(); }else{ r();
if(isNilTop()){ drop(); nil(); }else{ r();
if(isNilTop()){ drop(); pair(); }else{ r();
if(isNilTop()){ drop(); part(); }else{ drop();}}}}}}}}
function redify(){two(); pair();}
function green(){redify(); pup(); fore(); pop(); find(); pop();}
function hue(){part(); if(isNilTop()){ drop(); gray(); }else{ r();
if(isNilTop()){ drop(); green(); }else{ drop2(); }}}
function word(){pup(); fore(); pop(); lr();}
function next(){word(); if(isNilTop()){ drop(); }else{ hue(); next(); }}
function isABC(){r4(); not();}
function isBlue(){r3(); not();}
function compile(){over2(); pair(); push(); part(); bury(); cons2();}
function edit(){dup(); isABC(); if(isNilTop()){ drop(); pair(); }else{
compile(); isBlue(); if(isNilTop()){ drop(); green(); }else{ drop2() } nil(); }}
As crazy as it sounds I’m going to start writing about my programming environment from the beginning all over again. I’m trying to tell a new story like it’s an old story. The only thing I can do is at least make this new story old for me.
Ah, but, before I do that, I have to straighten out some mistakes I saw that I made in the last note with the javascript code for my programming environment.
Programs are tapes of words that can be shifted back and fore. You can edit programs by moving them back or fore and splicing in new words or deleting old ones. Some people like to do it live and others like to do it in post.
I didn’t implement the compiler correctly in the last note. There are also some little simplifications that need to be made to clean everything up a bit.
Oops, future John here again. Past John had to go to a spooky play and didn’t get to finish this note.
(you can skip this first paragraph if you want)
Yesterday I wrote more than I ever have before on the exact details of my programming environment. Since I’ve never done that outside of a purely technical and overly precise way, it has taken me a while to find a good way of explaining how everything works that’s accessible. When I explain things to myself, I can reference everything that I already know. Because of my unique history, there is little that I share with a lot of people. There’s more that I share with other people than the little that it takes to explain how my programming environment works, but none of that majority of similar stuff is sufficiently close to a technical and precise explanation. I’ve probably already lost some people. That’s okay. This is how I figure out how not to lose people. I have to start by losing people before I can bring people along with me. I wasn’t born a ringleader.
It occurred to me last night that a way of getting a self sustaining interpreter for my language, and for any language, is to start by implementing it in some other language and then push out its purportedly autonomous interpreter from there.
So, there are two ways to combine operations to form compound operations: compositionally and contingently. Composing one operation with another returns an operation that does the one and then the other. Conditionalizing one operation with another returns an operation that does the one if the top of the stack is Nil and does the other one otherwise. Here are the definitions of composition and conditionalization in javascript:
function comp(first, second){
return function(){
first();
second();
};
}
function cond(ifnil, ifnot){
return function(){
if(isNil(RL(T)))
ifnil();
else
ifnot();
}
}
So, all together, the native functions, followed by the functions for the basic operations (drop, dup, pop, push, swap, nil, pair, part) and the basic constructions of new functions from old ones (comp, cond) are
function Pair(l,r){return [l,r];}
var Nil = new Object();
function isNil(p){return p==Nil;}
function Left(p){return isNil(p) ? Nil : p[0];}
function Right(p){return isNil(p) ? Nil : p[1];}
function P(l,r){return Pair(l,r);}
function L(p){return Left(p);}
function R(p){return Right(p);}
function LL(p){return L(L(p));}
function LR(p){return L(R(p));}
function RL(p){return R(L(p));}
function RR(p){return R(R(p));}
function LLL(p){return L(LL(p));}
function LRL(p){return L(RL(p));}
function RLL(p){return R(LL(p));}
function RRL(p){return R(RL(p));}
function LLLL(p){return L(LLL(p));}
function RLLL(p){return R(LLL(p));}
var T=Nil;
function drop(){T=P(LL(T),R(T));}
function dup(){T=P(P(L(T),RL(T)),R(T));}
function pop(){T=P(LL(T),P(RL(T),R(T)));}
function push(){T=P(P(L(T),LR(T)),RR(T));}
function swap(){T=P(P(P(LLL(T),RL(T)),RLL(T)),R(T));}
function nil(){T=P(P(L(T),Nil),R(T));}
function pair(){T=P(P(LLL(T),P(RLL(T),RL(T))),R(T));}
function part(){T=P(P(P(LL(T),LRL(T)),RRL(T)),R(T));}
function comp(first, second){
return function(){
first();
second();
};
}
function cond(ifnil, ifnot){
return function(){
if(isNil(RL(T)))
ifnil();
else
ifnot();
}
}
With comp and cond it is possible to easily define mux i.e. it is no longer a basic operation.
var mux = cond(comp(drop, comp(pop, comp(drop, push))),
comp(drop, drop));
Two of these parts of mux occur often enough to be given their own names:
var drop2 = comp(drop, drop);
var nip = comp(pop, comp(drop, push));
so that then ‘mux’ is defined simply by
mux = cond(comp(drop, nip), drop2);
Ok. With that example out of the way, here is everything reprogrammed with these new constructions of comp and cond in mind:
function Pair(l,r){return [l,r];}
var Nil = new Object();
function isNil(p){return p==Nil;}
function Left(p){return isNil(p) ? Nil : p[0];}
function Right(p){return isNil(p) ? Nil : p[1];}
function P(l,r){return Pair(l,r);}
function L(p){return Left(p);}
function R(p){return Right(p);}
function LL(p){return L(L(p));}
function LR(p){return L(R(p));}
function RL(p){return R(L(p));}
function RR(p){return R(R(p));}
function LLL(p){return L(LL(p));}
function LRL(p){return L(RL(p));}
function RLL(p){return R(LL(p));}
function RRL(p){return R(RL(p));}
function LLLL(p){return L(LLL(p));}
function RLLL(p){return R(LLL(p));}
var T=Nil;
function drop(){T=P(LL(T),R(T));}
function dup(){T=P(P(L(T),RL(T)),R(T));}
function pop(){T=P(LL(T),P(RL(T),R(T)));}
function push(){T=P(P(L(T),LR(T)),RR(T));}
function swap(){T=P(P(P(LLL(T),RL(T)),RLL(T)),R(T));}
function nil(){T=P(P(L(T),Nil),R(T));}
function pair(){T=P(P(LLL(T),P(RLL(T),RL(T))),R(T));}
function part(){T=P(P(P(LL(T),LRL(T)),RRL(T)),R(T));}
function comp(first, second){
return function(){
first();
second();
};
}
function cond(ifnil, ifnot){
return function(){
if(isNil(RL(T)))
ifnil();
else
ifnot();
}
}
var drop2 = comp(drop, drop);
var push2 = comp(push, push);
var nip = comp(pop, comp(drop, push));
var over = comp(pop, comp(dup, comp(push, swap)));
var over2 = comp(over, over);
var dig = comp(over, comp(pop, nip));
var dig2 = comp(dig, dig);
var bury = comp(dig2, push2);
var unbury = comp(bury, bury);
var left = comp(part, drop);
var right = comp(part, nip);
var mux = cond(comp(drop, nip), drop2); // multiplexer
var nor = comp(dup, mux); // joint denial
var not = comp(dup, nor); // negation
var or = comp(nor, not); // alternation
var IF = comp(not, or); // converse conditional
var nif = comp(IF, not); // complementary converse conditional
var nfi = comp(swap, nif); // complementary conditional
var fi = comp(nfi, not); // conditional
var nand = comp(not, fi); // alternative denial
var and = comp(nand, not); // conjunction
var xor = comp(over2, comp(nand, comp(bury, comp(or, and)))); // exclusive alternation
var iff = comp(xor, not); // biconditional
Instead of needing to say ‘function’ everywhere and have all those empty parenthesies, just a few comps and conds are needed. Sadly, because of the direct way that I’ve taken this, the definition of an actually recursive function involves an unfortunate notation particular to javascript. It is demonstrated in the definition of identity, which goes as follows (don’t be scared):
var partunbury = comp(part, unbury);
var partunbury2 = comp(partunbury, partunbury);
var id = function f(){
return cond(drop, comp(swap, cond(drop, comp(partunbury2, comp(f, comp(bury, comp(f, and)))))))();
}
First, let me just describe in plane English how to check if two binary trees are identical. They are identical if and only if their left parts are identical and their right parts are identical.
Since all of our binary trees eventually end in leafs of Nil, it is easy enough to deal with the simple cases where one or the other pair is already Nil. Both the pairs are ontop of the stack. If the top tree is Nil, then it matches the second from top tree if and only if that tree is Nil also. It just so happens that Nil is what we’ve been using for true. So if the top tree is Nil, we just need to drop it and then if the second from top tree is Nil we are left with true, which is what we want. But, if the second from top tree is not Nil then we are also left with what we need: a non-Nil tree is false and it is false that a non-Nil tree is identical to a Nil tree.
This basic case is setted by the first part of the definition ‘cond(drop’.
Now, if the top tree is non Nil, then either the second from top of the tree is Nil or non-Nil. We swap the top and second from top and check everything the same way that we did in the previous basic case i.e. these two basic cases are basically swaps of each other.
All together the basic cases are settled by ‘cond(drop, comp(swap, cond(drop, …’.
The next part of the definition just breaks each of the trees into their left parts and their right parts and then checks if those corresponding parts are identical. This is where the specifics of defining recursive functions in javascript come in. I would like to write the definition as follows:
var id = cond(drop, comp(swap, cond(drop, comp(partunbury2, comp(id, comp(bury, comp(id, and)))))))
This doesn’t work because the moment javascript hits the first ‘id’ on the right hand side of the equality it replaces it with whatever ‘id’ has already been defined as, and it hasn’t been defined yet because that’s what we’re doing at that exact moment. So, it effectively replaces each occurrence of ‘id’ on the right hand side with ‘undefined’ which is not a function. The console will tell you as much if you try to execute ‘id()’.
So we have to tell javascript that we’re making a recursive function (a function that repeats part of itself), and it’s going to tentatively be called ‘f’ and we’ll be able to reference ‘f’ appropriately.
Another way the definition of ‘id’ could have been given is as follows (and this should have been the one I started with):
function id(){
cond(drop, comp(swap, cond(drop, comp(partunbury2, comp(id, comp(bury, comp(id, and)))))))();
}
Either way, such definitions are trivial in the language being constructed (just make the appropriate green word and it goes back to the nearest matching red and goes forth from there).
The functions being defined now are to build up the interpreter, compiler, and executer of my programming environment. Programs are stacks. The last word coded is pushed atop the stack. We manipulate where we are when executing a program by spliting it into a stack and a list. This is accomplished by two operations: back and fore.
A program mid execution is an ordered pair whose left part is the stack of the program up to the word that is currently being executed. It’s right part is a list of everything that it will do next assuming the current word being executed doesn’t change that. Back and fore manage the movement back and forth through the program.
var back = comp(part, comp(pop, comp(part, comp(push, comp(pair, pair)))));
var fore = comp(part, comp(part, comp(pop, comp(pair, comp(push, pair)))));
The next ‘big’ function is called ‘find’. It goes back through a program and looks for the first matching red word that it can find by going deeper into the stack half of the currently executing program.
(As a side note, it just occurred to me that I can go ahead and use the latest definition of ‘id’ that I made and avoid having to go through the stuff about recursion because people are unlikely to notice that there is even a problem to be solved with that notation, and it is a bit closer to how my programming environment works. Sorry for putting you through that recursive talk.)
Find works on more than just red words. It expects a pair whose left part is a stack and whose right part is a list and it will search down the stack, going ‘back’ each step, until it finds a tree that matches the one at the second from top of T.
var lr = comp(left, right); // notice this is ***backwards*** from 'LR(x)'
function find(){
comp(over2, comp(lr, comp(eq, cond(comp(drop, nip), comp(drop, comp(back, find)))))))();
}
A few things to note. The first is very important and the second not so much. The first is that this is the second time we’ve seen ‘comp(drop, nip)’. This happens a lot and the more it happens the more it needs its own name. I call it ‘dip’. You can (and should) call it what you want. Maybe you’ll call it ‘drip’.
var dip = comp(drop, nip);
It’s worth it to go back and update everything with ‘dip’ in mind. The reason this is true is interesting. There are only a hand full of secondary words that keep turning up again and again. If there were always new ones then I’d say don’t bother going back over everything replacing anything like ‘comp(drop, nip)’ with ‘dip’ or ‘comp(drop, comp(nip, ..))’ with ‘comp(dip, ..)’. Anyway, this is something that is only uncovered by solving problems like we are right now.
Here’s everything written again (since I already know that ‘dip’ comes up, it only saves us a very tiny bit, but it’s never a bad time to look over all the code again and see if there’s something there that shouldn’t be because teh best thing you can do when programming is to eliminate eliminate eliminate! There can’t be problems in parts of the program that no longer exist, and as long as the program gets shorter and shorter there are fewer places where any problems whatsoever can occur).
function Pair(l,r){return [l,r];}
var Nil = new Object();
function isNil(p){return p==Nil;}
function Left(p){return isNil(p) ? Nil : p[0];}
function Right(p){return isNil(p) ? Nil : p[1];}
function P(l,r){return Pair(l,r);}
function L(p){return Left(p);}
function R(p){return Right(p);}
function LL(p){return L(L(p));}
function LR(p){return L(R(p));}
function RL(p){return R(L(p));}
function RR(p){return R(R(p));}
function LLL(p){return L(LL(p));}
function LRL(p){return L(RL(p));}
function RLL(p){return R(LL(p));}
function RRL(p){return R(RL(p));}
function LLLL(p){return L(LLL(p));}
function RLLL(p){return R(LLL(p));}
var T=Nil;
function drop(){T=P(LL(T),R(T));}
function dup(){T=P(P(L(T),RL(T)),R(T));}
function pop(){T=P(LL(T),P(RL(T),R(T)));}
function push(){T=P(P(L(T),LR(T)),RR(T));}
function swap(){T=P(P(P(LLL(T),RL(T)),RLL(T)),R(T));}
function nil(){T=P(P(L(T),Nil),R(T));}
function pair(){T=P(P(LLL(T),P(RLL(T),RL(T))),R(T));}
function part(){T=P(P(P(LL(T),LRL(T)),RRL(T)),R(T));}
function comp(first, second){
return function(){
first();
second();
};
}
function cond(ifnil, ifnot){
return function(){
if(isNil(RL(T)))
ifnil();
else
ifnot();
}
}
var drop2 = comp(drop, drop);
var push2 = comp(push, push);
var pair2 = comp(pair, pair);
var part2 = comp(part, part);
var nip = comp(pop, comp(drop, push));
var dip = comp(drop, nip);
var over = comp(pop, comp(dup, comp(push, swap)));
var over2 = comp(over, over);
var dig = comp(over, comp(pop, nip));
var dig2 = comp(dig, dig);
var bury = comp(dig2, push2);
var unbury = comp(bury, bury);
var left = comp(part, drop);
var right = comp(part, nip);
var lr = comp(left, right); // notice this is ***backwards*** from 'LR(x)'
var mux = cond(dip, drop2); // multiplexer
var nor = comp(dup, mux); // joint denial
var not = comp(dup, nor); // negation
var or = comp(nor, not); // alternation
var IF = comp(not, or); // converse conditional
var nif = comp(IF, not); // complementary converse conditional
var nfi = comp(swap, nif); // complementary conditional
var fi = comp(nfi, not); // conditional
var nand = comp(not, fi); // alternative denial
var and = comp(nand, not); // conjunction
var xor = comp(over2, comp(nand, comp(bury, comp(or, and)))); // exclusive alternation
var iff = comp(xor, not); // biconditional
var partunbury = comp(part, unbury);
var partunbury2 = comp(partunbury, partunbury);
function id(){cond(drop, comp(swap, cond(drop,
comp(partunbury2, comp(id, comp(bury, comp(id, and)))))))();}
var back = comp(part, comp(pop, comp(part, comp(push, pair2))));
var fore = comp(part2, comp(pop, comp(pair, comp(push, pair))));
function find(){
comp(over2, comp(lr, comp(eq, cond(dip, comp(drop, comp(back, find))))))();
}
The second important thing was to notice that the definition of ‘lr’ is backwards from the upper case definition of ‘LR’ in that the one executes left then right, and the other takes the left of the right of its argument. This is unavoidable. It is the mathematical equivalent of knowing which right turn you have to take to make a left at the stop sign.
There are only two pieces missing from this clockwork programming environment: how to deal with words of a given color and how to edit a program.
The first color words to deal with are the gray ones. Now, unlike in the last note, there is a bit of artistry that must be introduced to keep a tight ship.
The assumption is that the top of the stack of the tree spells out the gray word. This means that what is to be done is seen by looking through the length of the top as if it was a list. There’s no easy way to creep up on this, so I’ll just say before hand that the way I do it right now is not the way that it will be done later.
var gray = cond(drop2, comp(right,
cond(comp(drop, dup), comp(right,
cond(comp(drop, pop), comp(right,
cond(comp(drop, push), comp(right,
cond(comp(drop, swap), comp(right,
cond(comp(drop, nil), comp(right,
cond(comp(drop, pair), comp(right,
cond(comp(drop, part), drop)))))))))))))));
First, remember, this isn’t how it will stay. But, there are only two things that may need to change later: the occurences of ‘pop’ and ‘push’. Note that before executing the quoted operation a drop must be executed so that it doesn’t get in the way of dropping, dupping, popping, pushing, swaping, nil-ing, pairing, or parting the appropriate items!
Now the next color to deal with is green. The assumption is that the top of the tree spells out a green word. Assuming the color of red is marked by a Nil, and that a word is an ordered pair whose right part is its color and whose left part spells it out, all that needs to be done is to put a nil atop the stack and pair it with the spelling of the green word so that we can go find it.
We have what we want to find, but the question is where we’re going to go looking for it. The assumption is that the top of the list of the tree is where the green word just came from i.e. that the first item of the list is a running program, that is a pair whose left part is everything before the green word and the right part is everything after the green word.
We push the program to the top of the stack, duplicate it, move it fore so that when we return to it we’re ready to get on where the green word left off, pop that baby back up to the front of the list, and then find our way through the duplicate we just made. It’s that simple.
var green = comp(nil, comp(pair, comp(push, comp(dup, comp(fore, comp(pop, find))))));
Now, there needs to be a way of geting the next word and doing the right thing if its gray or green (if it’s red or blue then there’s nothing that needs to be done). Let’s assume the word is on top of the stack. It just occurred to me that if we make nil stand for gray and the pair whose left and right part is nil stand for green then this dispatch based on color can be short and simple. Let’s make that the new convention. Here’s the function first and then we’ll go back and make all the edits with this new convention in mind.
var hue = comp(part, cond(comp(drop, gray), comp(right, cond(comp(drop, green), drop2))));
Then, ‘green’ has to be redefined, we can write a helper to simplify this if further changes occur:
function Pair(l,r){return [l,r];}
var Nil = new Object();
function isNil(p){return p==Nil;}
function Left(p){return isNil(p) ? Nil : p[0];}
function Right(p){return isNil(p) ? Nil : p[1];}
function P(l,r){return Pair(l,r);}
function L(p){return Left(p);}
function R(p){return Right(p);}
function LL(p){return L(L(p));}
function LR(p){return L(R(p));}
function RL(p){return R(L(p));}
function RR(p){return R(R(p));}
function LLL(p){return L(LL(p));}
function LRL(p){return L(RL(p));}
function RLL(p){return R(LL(p));}
function RRL(p){return R(RL(p));}
function LLLL(p){return L(LLL(p));}
function RLLL(p){return R(LLL(p));}
var T=Nil;
function drop(){T=P(LL(T),R(T));}
function dup(){T=P(P(L(T),RL(T)),R(T));}
function pop(){T=P(LL(T),P(RL(T),R(T)));}
function push(){T=P(P(L(T),LR(T)),RR(T));}
function swap(){T=P(P(P(LLL(T),RL(T)),RLL(T)),R(T));}
function nil(){T=P(P(L(T),Nil),R(T));}
function pair(){T=P(P(LLL(T),P(RLL(T),RL(T))),R(T));}
function part(){T=P(P(P(LL(T),LRL(T)),RRL(T)),R(T));}
function comp(first, second){
return function(){
first();
second();
};
}
function cond(ifnil, ifnot){
return function(){
if(isNil(RL(T)))
ifnil();
else
ifnot();
}
}
var drop2 = comp(drop, drop);
var push2 = comp(push, push);
var pair2 = comp(pair, pair);
var part2 = comp(part, part);
var nip = comp(pop, comp(drop, push));
var dip = comp(drop, nip);
var over = comp(pop, comp(dup, comp(push, swap)));
var over2 = comp(over, over);
var dig = comp(over, comp(pop, nip));
var dig2 = comp(dig, dig);
var bury = comp(dig2, push2);
var unbury = comp(bury, bury);
var left = comp(part, drop);
var right = comp(part, nip);
var lr = comp(left, right); // notice this is ***backwards*** from 'LR(x)'
var mux = cond(dip, drop2); // multiplexer
var nor = comp(dup, mux); // joint denial
var not = comp(dup, nor); // negation
var or = comp(nor, not); // alternation
var IF = comp(not, or); // converse conditional
var nif = comp(IF, not); // complementary converse conditional
var nfi = comp(swap, nif); // complementary conditional
var fi = comp(nfi, not); // conditional
var nand = comp(not, fi); // alternative denial
var and = comp(nand, not); // conjunction
var xor = comp(over2, comp(nand, comp(bury, comp(or, and)))); // exclusive alternation
var iff = comp(xor, not); // biconditional
var zero = nil;
var succ = comp(nil, comp(swap, pair));
var one = comp(zero, succ);
var two = comp(one, succ);
var three = comp(two, succ);
var partunbury = comp(part, unbury);
var partunbury2 = comp(partunbury, partunbury);
function id(){cond(drop, comp(swap, cond(drop,
comp(partunbury2, comp(id, comp(bury, comp(id, and)))))))();}
var back = comp(part, comp(pop, comp(part, comp(push, pair2))));
var fore = comp(part2, comp(pop, comp(pair, comp(push, pair))));
function find(){
comp(over2, comp(lr, comp(eq, cond(dip, comp(drop, comp(back, find))))))();
}
var gray = cond(drop2, comp(right,
cond(comp(drop, dup), comp(right,
cond(comp(drop, pop), comp(right,
cond(comp(drop, push), comp(right,
cond(comp(drop, swap), comp(right,
cond(comp(drop, nil), comp(right,
cond(comp(drop, pair), comp(right,
cond(comp(drop, part), drop)))))))))))))));
var redify = comp(two, pair);
var green = comp(redify, comp(push, comp(dup, comp(fore, comp(pop, comp(find, pop))))));
var hue = comp(part, cond(comp(drop, gray), comp(right, cond(comp(drop, green), drop2))));
Another change got made to green because the last version didn’t put the place found back onto the front of the list.
The next operation is ‘next’. It expects the currently executing program to be at the front of the list. Gets the next word and moves the program fore one step so that it’s ready for the next next. Then it does whatever is next. But, the question is this: what if there is no next word? Well, it will just stop.
var fetch = comp(push, comp(dup, comp(fore, comp(pop, lr))));
function next(){ comp(fetch, cond(drop, comp(hue, next)))();}
Now, all that is left is editing, but before getting to that, there is a problem that I’ve ignored. Gray words don’t work right. If you think through a push or a pop there will be a problem. The problem is that there is no place we’re keeping the currently running program that won’t get mixed up by a push or a pop from executing a gray word. I already mentioned this before, the way to avoid this is to redefine what push and pop do when they are read as gray words so that they don’t move the currently executing instructions from the front of the list where fetch, next, AND green expect them to be (that ‘pop’ at the end of the definition of ‘green’ makes it so everything gets going when ‘next’ is executed).
var graypop = comp(push, comp(swap, pop2));
var graypush = comp(push2, comp(swap, pop));
var gray = var gray = cond(drop2, comp(right,
cond(comp(drop, dup), comp(right,
cond(comp(drop, graypop), comp(right,
cond(comp(drop, graypush), comp(right,
cond(comp(drop, swap), comp(right,
cond(comp(drop, nil), comp(right,
cond(comp(drop, pair), comp(right,
cond(comp(drop, part), drop)))))))))))))));
So, all together we have this:
function Pair(l,r){return [l,r];}
var Nil = new Object();
function isNil(p){return p==Nil;}
function Left(p){return isNil(p) ? Nil : p[0];}
function Right(p){return isNil(p) ? Nil : p[1];}
function P(l,r){return Pair(l,r);}
function L(p){return Left(p);}
function R(p){return Right(p);}
function LL(p){return L(L(p));}
function LR(p){return L(R(p));}
function RL(p){return R(L(p));}
function RR(p){return R(R(p));}
function LLL(p){return L(LL(p));}
function LRL(p){return L(RL(p));}
function RLL(p){return R(LL(p));}
function RRL(p){return R(RL(p));}
function LLLL(p){return L(LLL(p));}
function RLLL(p){return R(LLL(p));}
var T=Nil;
function drop(){T=P(LL(T),R(T));}
function dup(){T=P(P(L(T),RL(T)),R(T));}
function pop(){T=P(LL(T),P(RL(T),R(T)));}
function push(){T=P(P(L(T),LR(T)),RR(T));}
function swap(){T=P(P(P(LLL(T),RL(T)),RLL(T)),R(T));}
function nil(){T=P(P(L(T),Nil),R(T));}
function pair(){T=P(P(LLL(T),P(RLL(T),RL(T))),R(T));}
function part(){T=P(P(P(LL(T),LRL(T)),RRL(T)),R(T));}
function comp(first, second){
return function(){
first();
second();
};
}
function cond(ifnil, ifnot){
return function(){
if(isNil(RL(T)))
ifnil();
else
ifnot();
}
}
var drop2 = comp(drop, drop);
var push2 = comp(push, push);
var pair2 = comp(pair, pair);
var part2 = comp(part, part);
var nip = comp(pop, comp(drop, push));
var dip = comp(drop, nip);
var over = comp(pop, comp(dup, comp(push, swap)));
var over2 = comp(over, over);
var dig = comp(over, comp(pop, nip));
var dig2 = comp(dig, dig);
var bury = comp(dig2, push2);
var unbury = comp(bury, bury);
var left = comp(part, drop);
var right = comp(part, nip);
var lr = comp(left, right); // notice this is ***backwards*** from 'LR(x)'
var mux = cond(dip, drop2); // multiplexer
var nor = comp(dup, mux); // joint denial
var not = comp(dup, nor); // negation
var or = comp(nor, not); // alternation
var IF = comp(not, or); // converse conditional
var nif = comp(IF, not); // complementary converse conditional
var nfi = comp(swap, nif); // complementary conditional
var fi = comp(nfi, not); // conditional
var nand = comp(not, fi); // alternative denial
var and = comp(nand, not); // conjunction
var xor = comp(over2, comp(nand, comp(bury, comp(or, and)))); // exclusive alternation
var iff = comp(xor, not); // biconditional
var zero = nil;
var succ = comp(nil, comp(swap, pair));
var one = comp(zero, succ);
var two = comp(one, succ);
var three = comp(two, succ);
var partunbury = comp(part, unbury);
var partunbury2 = comp(partunbury, partunbury);
function id(){cond(drop, comp(swap, cond(drop,
comp(partunbury2, comp(id, comp(bury, comp(id, and)))))))();}
var back = comp(part, comp(pop, comp(part, comp(push, pair2))));
var fore = comp(part2, comp(pop, comp(pair, comp(push, pair))));
function find(){
comp(over2, comp(lr, comp(eq, cond(dip, comp(drop, comp(back, find))))))();
}
var graypop = comp(push, comp(swap, pop2));
var graypush = comp(push2, comp(swap, pop));
var gray = cond(drop2, comp(right,
cond(comp(drop, dup), comp(right,
cond(comp(drop, graypop), comp(right,
cond(comp(drop, graypush), comp(right,
cond(comp(drop, swap), comp(right,
cond(comp(drop, nil), comp(right,
cond(comp(drop, pair), comp(right,
cond(comp(drop, part), drop)))))))))))))));
var redify = comp(two, pair);
var green = comp(redify, comp(push, comp(dup, comp(fore, comp(pop, comp(find, pop))))));
var hue = comp(part, cond(comp(drop, gray), comp(right, cond(comp(drop, green), drop2))));
var fetch = comp(push, comp(dup, comp(fore, comp(pop, lr))));
function next(){ comp(fetch, cond(drop, comp(hue, next)))();}
All that’s left now is to get the editor working. The editor is where blue words get everything started. It’s also where programs are constructed word by word. A word is constructed by spelling it out and then by coloring it. If it was the other way around then blue words wouldn’t work as simply as they can. That is, you first spell out the blue word and then after spelling it you color it blue and that starts everything off running.
So the editor expects the top of the stack to be the next letter to add to the spelling of the word under construction or the color of the word that is then done being constructed.
If the top of the stack is not a color then it is a letter and gets paired with spelling that has been made thus far. If the top of the stack is a color then it also gets paired with the spelling form a complete word whose spelling is its left part and whose color is its right part. Then the word has to be put onto the end of the program under construction. If the word is colored blue then we treat it like a green word.
var rr = comp(right, right);
var r3 = comp(rr, r);
var r4 = comp(r3, r);
var isLetter = comp(r4, not);
var isBlue = comp(r3, not);
var cons = comp(swap, pair);
var compile = comp(over2, comp(pair, comp(push, comp(cons, pop))));
var edit = comp(dup, comp(isLetter, cond(comp(drop, pair),
comp(compile, comp(isBlue, comp(cond(comp(drop, green), drop2), nil))))));
Sadly, I don’t have time to test everything, but here it all is after looking over everything I’ve done here and making a few little changes e.g. ‘cons’ was not defined where it otherwise would be e.g. after ‘pair’.
function Pair(l,r){return [l,r];}
var Nil = new Object();
function isNil(p){return p==Nil;}
function Left(p){return isNil(p) ? Nil : p[0];}
function Right(p){return isNil(p) ? Nil : p[1];}
function P(l,r){return Pair(l,r);}
function L(p){return Left(p);}
function R(p){return Right(p);}
function LL(p){return L(L(p));}
function LR(p){return L(R(p));}
function RL(p){return R(L(p));}
function RR(p){return R(R(p));}
function LLL(p){return L(LL(p));}
function LRL(p){return L(RL(p));}
function RLL(p){return R(LL(p));}
function RRL(p){return R(RL(p));}
function LLLL(p){return L(LLL(p));}
function RLLL(p){return R(LLL(p));}
var T=Nil;
function drop(){T=P(LL(T),R(T));}
function dup(){T=P(P(L(T),RL(T)),R(T));}
function pop(){T=P(LL(T),P(RL(T),R(T)));}
function push(){T=P(P(L(T),LR(T)),RR(T));}
function swap(){T=P(P(P(LLL(T),RL(T)),RLL(T)),R(T));}
function nil(){T=P(P(L(T),Nil),R(T));}
function pair(){T=P(P(LLL(T),P(RLL(T),RL(T))),R(T));}
function part(){T=P(P(P(LL(T),LRL(T)),RRL(T)),R(T));}
function comp(first, second){return function(){
first(); second(); };}
function cond(ifnil, ifnot){return function(){
if(isNil(RL(T))) ifnil(); else ifnot();}}
var drop2 = comp(drop, drop);
var pop2 = comp(pop, pop);
var push2 = comp(push, push);
var pair2 = comp(pair, pair);
var part2 = comp(part, part);
var nip = comp(pop, comp(drop, push));
var dip = comp(drop, nip);
var over = comp(pop, comp(dup, comp(push, swap)));
var over2 = comp(over, over);
var dig = comp(over, comp(pop, nip));
var dig2 = comp(dig, dig);
var bury = comp(dig2, push2);
var unbury = comp(bury, bury);
var left = comp(part, drop);
var right = comp(part, nip);
var rr = comp(right, right);
var r3 = comp(rr, r);
var r4 = comp(r3, r);
var lr = comp(left, right); // notice this is ***backwards*** from 'LR(x)'
var cons = comp(swap, pair);
var mux = cond(dip, drop2); // multiplexer
var nor = comp(dup, mux); // joint denial
var not = comp(dup, nor); // negation
var or = comp(nor, not); // alternation
var IF = comp(not, or); // converse conditional
var nif = comp(IF, not); // complementary converse conditional
var nfi = comp(swap, nif); // complementary conditional
var fi = comp(nfi, not); // conditional
var nand = comp(not, fi); // alternative denial
var and = comp(nand, not); // conjunction
var xor = comp(over2, comp(nand, comp(bury, comp(or, and)))); // exclusive alternation
var iff = comp(xor, not); // biconditional
var zero = nil;
var succ = comp(nil, comp(swap, pair));
var one = comp(zero, succ);
var two = comp(one, succ);
var three = comp(two, succ);
var party = comp(part, unbury);
var party2 = comp(party, party);
function id(){cond(drop, comp(swap, cond(drop,
comp(party2, comp(id, comp(bury, comp(id, and)))))))();}
var back = comp(part, comp(pop, comp(part, comp(push, pair2))));
var fore = comp(part2, comp(pop, comp(pair, comp(push, pair))));
function find(){
comp(over2, comp(lr, comp(eq, cond(dip, comp(drop, comp(back, find))))))();
}
var graypop = comp(push, comp(swap, pop2));
var graypush = comp(push2, comp(swap, pop));
var gray = cond(drop2, comp(right,
cond(comp(drop, dup), comp(right,
cond(comp(drop, graypop), comp(right,
cond(comp(drop, graypush), comp(right,
cond(comp(drop, swap), comp(right,
cond(comp(drop, nil), comp(right,
cond(comp(drop, pair), comp(right,
cond(comp(drop, part), drop)))))))))))))));
var redify = comp(two, pair);
var green = comp(redify, comp(push, comp(dup, comp(fore, comp(pop, comp(find, pop))))));
var hue = comp(part, cond(comp(drop, gray), comp(right, cond(comp(drop, green), drop2))));
var fetch = comp(push, comp(dup, comp(fore, comp(pop, lr))));
function next(){ comp(fetch, cond(drop, comp(hue, next)))();}
var isLetter = comp(r4, not);
var isBlue = comp(r3, not);
var compile = comp(over2, comp(pair, comp(push, comp(cons, pop))));
var edit = comp(dup, comp(isLetter, cond(comp(drop, pair),
comp(compile, comp(isBlue, comp(cond(comp(drop, green), drop2), nil))))));
That’s it. That’s the logic. There’s some io stuff that I didn’t do and there may be some big gap I don’t see right now, but one day at a time.
This continues work on my programming environment. I’m not going to link it back to other stuff because that’s boring. I might do it later if what I end up writing here matters at all.
I was writing a story about about how my programming environment is made of a forest of programmable trees. They purportedly listen for mands (a technical term from Skinner’s “Verbal Behavior”) and reinforce the speaker/programmger accordingly.
I made the mistake of taking ‘mux’ as a primitive as if it made it possible to write an executer from it. The executer reads a binary tree and takes a corresponding action based on what it reads. This lets the tree command itself. This is how you get automatic rather than manual programming. It’s the difference between cruse control and manual.
In the code for the forest, the function that manages mands is called ‘act’. There is a special list called ‘mand’ that holds a list of mands to be executed cyclically. Technically, ‘act’ is composed of two functions ‘fetch’ and ‘next’. The function ‘fetch’ pops the top item of the list of mands and appends it to the bottom of the list. The function ‘next’ checks input/output conditions and if they are appropriate plans the next action.
The three functions ‘act’, ‘next’, and ‘fetch’ can be put into one that is more organized and may yield even simpler functions if later factored into parts. The key step is ‘next’. In one of the earlier notes on my programming environment, I showed that a tree needs to not only be able to execute a command/instruction, but it also has to schedule, in some way, the execution of the next command. It has to choose what to do next (to put it in democratic terms).
Wrapping everything up into ‘next’ encapsulates the thinking that goes into setting up the ‘fetch-execute’ or ‘instruction’ cycle familiar to hardware designers.
I’ll go ahead and setup the input/output stuff too just to really get the ball rolling. Also, I’ll simplify the logic of following instructions to avoid some endless loops. The list of mands will just get shorter each step rather than having the top of the list append to the end of the list after the relevant action has been taken. The reason to avoid this right now is because the operation of appending is primitive recursive and as a supplementary operation it is not helpful to include any kind of recursion except in the controlling environment. I have not made this clear here, but hopefully I’ll find a good way of making it clear. It is most important when it comes to stepping away from a single tree and programming with a whole forest of trees that speak and listen to each other.
This is quick code. This is not quality code. Some say that you should never write quick code. I know of no other way to start writing quality code.
var mands=Nil; // instructions
var tx=Nil; // transmitting
var rx=Nil; // receiving
function next(){ // follow next instruction
let mand=L(mands); // next instruction
mands=R(mands); // ready the next mand
function check(){ // instructions are coded as the length of a list
if(isNil(mand))return true; // reached the end of mand?
mand=R(mand); // drop the top of the list
return false; // haven't reached the end of mand
}
switch(true){ // decode instruction
case check(): drop(); break; // a length zero list encodes drop
case check(): dup(); break; // a length one list encodes dup
case check(): pop(); break; // and so on
case check(): push(); break;
case check(): swap(); break;
case check(): nil(); break;
case check(): pair(); break;
case check(): part(); break;
case check(): mux(); break;
default: break;
}
if(rx||tx)return; // if receiving or transmitting wait
next(); // plan to do the next thing
}
As the comments say each command is encoded as the length of a list. So the list of commands, called ‘mands’, is a list of lists each length of which corresponds to a basic operation on the tree. It is customary to grab the first command and then immediately set up the list of commands so that it is ready for the next command to be grabbed the next time around. There are various reasons for delaying this set up, but I’ve not done that here.
The function ‘check’ is just a local (and trashy) way to check if the length of the current mand is zero or not. If it’s not then it shortens the list by removing its first item and then that shorter list is checked to see if it is length zero. Eventually, since we’ve made a prior agreement, this checking process will reach the end of the length of mand and when it does we do the corresponding basic operation e.g. if the mand is initially of length zero then drop is executed, and if the mand is initially of length one then the first check is false and the next check is true so that dup is executed. If instead of lengths of lists I was using numbers then ‘0’ would encode drop, ‘1’ would encode dup, and so on.
This encoding and decoding is the crux of stored programming and is made possible by a fact first uncovered by Godel in his incompleteness proofs. The general method is called ‘the arithmetization of syntax’ but it’s really a fancy name for an old method that is familiar to anyone who reads a newspaper: quoting. Codes are just obscure ways of quoting something that we said e.g. I said ‘drop’ and quoted it as a length zero list. When we quote people we tend to do so in a way that makes it relatively easy to reconstruct what they might have said on the spot, but since I’ve decided that everything is a binary tree (aka ordered pair) then my only way of making quotations is with ordered pairs rather than letters of the alphabet.
After the relevant basic operation is done, ‘next’ checks to see if the tree is transmitting or receiving. None of the basic operations I’ve introduced thus far change whether the tree is transmitting or receiving (speaking or listening). I’ll change ‘transmitting’ and ‘receiving’ as ‘rx’ and ‘tx’ into ‘speaking’ and ‘listening’ because those are better names for what is going on. I’ll also give below the supplementary code that sets up the basic operations on the tree that are included in the switch statement of ‘next’.
function Pair(l,r){return [l,r];}
var Nil = new Object();
function isNil(p){return p==Nil;}
function Left(p){return isNil(p) ? Nil : p[0];}
function Right(p){return isNil(p) ? Nil : p[1];}
function P(l,r){return Pair(l,r);}
function L(p){return Left(p);}
function R(p){return Right(p);}
function LL(p){return L(L(p));}
function LR(p){return L(R(p));}
function RL(p){return R(L(p));}
function RR(p){return R(R(p));}
function LLL(p){return L(LL(p));}
function LRL(p){return L(RL(p));}
function RLL(p){return R(LL(p));}
function RRL(p){return R(RL(p));}
function LLLL(p){return L(LLL(p));}
function RLLL(p){return R(LLL(p));}
var T=Nil;
function drop(){T=P(LL(T),R(T));}
function dup(){T=P(P(L(T),RL(T)),R(T));}
function pop(){T=P(LL(T),P(RL(T),R(T)));}
function push(){T=P(P(L(T),LR(T)),RR(T));}
function swap(){T=P(P(P(LLL(T),RL(T)),RLL(T)),R(T));}
function nil(){T=P(P(L(T),Nil),R(T));}
function pair(){T=P(P(LLL(T),P(RLL(T),RL(T))),R(T));}
function part(){T=P(P(P(LL(T),LRL(T)),RRL(T)),R(T));}
function mux(){T=P(P(LLLL(T), RL(T)==Nil ? RLL(T) : RLLL(T)), R(T));}
var mands=Nil; // instructions
var speaking=Nil; // transmitting
var listening=Nil; // receiving
function next(){ // do next instruction
let mand=L(mands); // get next instruction
mands=R(mands); // set up getting next instruction
function check(){ // instructions are quoted as the length of a list
if(isNil(mand))return true; // reached the end of mand?
mand=R(mand); // drop the first item of the list
return false; // haven't reached the end of mand
}
switch(true){ // disquote instruction
case check(): drop(); break; // a length zero list quotes drop
case check(): dup(); break; // a length one list quotes dup
case check(): pop(); break; // and so on
case check(): push(); break;
case check(): swap(); break;
case check(): nil(); break;
case check(): pair(); break;
case check(): part(); break;
case check(): mux(); break;
default: break;
}
if(speaking||listening)return; // if receiving or transmitting wait
next(); // plan to do the next instruction
}
In general, it is better to speak of quoting and disquoting than it is coding and decoding. Codes suggest a cryptic mysticism that quotations don’t. Also, quotatoin is more cryptic than most people suspect. And yet, we find many instances of the phrase “You misquoted me!” or “You’re quote was out of context!”. I’d rather people see codes and quotes as similar than as different even in casual conversation.
Programmers familiar with programming, e.g. in javascript as above, will notice that there is something very wrong with the definition of ‘next’. It ends with a call to itself! Naively this is ok. But, if you plan on running this code for any length of time then you’ll get the classic “stack overflow” error. This stack is much like the list of the tree. It is part of how the ‘flow of control’ is managed. When a stack overflow error is thrown, it means that the machine has no more room to for keeping track of what it has to do next. Specifically, after it finishes doing what this function is telling it to do, it won’t know what function to do next. Everything has to be written down because it has no clue what it’s doing or what it’s going to do next.
One of the solutions that people are familiar with is also needlessly complicated. It’s called “tail recursion” and involves what is called a “tail call”. It’s an obscure way of saying that if the last thing a function tells a machine to do is to ‘call’ another function, then it’s just as well to ‘jump’ directly to that function rather than complete ‘the call’ by coming back to the previous function and doing nothing.
This has got me off on a rant. I’m going to rant. We used to just call things routines and subroutines. These words are much stronger than ‘function’. Now we are forced to speak about the differences between functions and routines. Not every routine is a function. Whether a subroutine is a function or not is sometimes obviously true and most of the time a technical thornbush. Now we have methods, procedures, functions, routines, and programs. The word ‘program’ is the best. Better than routine.
We can give a program a name. A program is just a special kind of quotation. Programs go hand in hand with schedules. Scheduling a program does not guarentee that the consequences of executing the program are going to occur. They are planned to occur, but that doesn’t promise us any more than when a phycisist says there will be an eclipse at this time on this day.
Oh, we also have macros, macros in addition to functions. Goodness help me. Goodness help us!
I have a general way that I like to set up scheduling operations.
var scheduled=[];
var going=false;
function schedule(it){
scheduled.push(it);
if(going)return;
going=true;
while(scheduled.length) scheduled.shift()();
going=false;
}
For now the scheduler is just implemented in javascript. It expects you to schedule the name of a function without any variables and it runs through everything that’s been scheduled and executes it. Usually I separate out the planning phase and the ‘going’ phase, but, in this case, the ‘plan’ function would just push the name of a function onto the end of the list named ‘scheduled’ and then call another function, I usally call it ‘go’, which has the single while loop that runs everything.
Scheduling is something that I first take as exterior to the language I’m designing. I’ve also been a bit misleading in my implementation of ‘next’, but first I’ll go ahead and revise it with the scheduler in mind:
function next(){ // do next instruction
let mand=L(mands); // get next instruction
mands=R(mands); // set up getting next instruction
function check(){ // instructions are coded as the length of a list
if(isNil(mand))return true; // reached the end of mand?
mand=R(mand); // drop the first item of the list
return false; // haven't reached the end of mand
}
switch(true){ // disquote instruction
case check(): drop(); break; // a length zero list quotes drop
case check(): dup(); break; // a length one list quotes dup
case check(): pop(); break; // and so on
case check(): push(); break;
case check(): swap(); break;
case check(): nil(); break;
case check(): pair(); break;
case check(): part(); break;
case check(): mux(); break;
default: break;
}
if(speaking||listening)return; // if receiving or transmitting wait
schedule(next);
}
Since I’ve allowed myself to use javascript without first implementing everything in my own language (something which isn’t hard, but it has to start somewhere, usually on a host system, and while the hardware may mirror the design of the programming environment, it may not, and there is nothing to do but submit to hegemony for now, but hopefully not forever).
Those familiar with LISP will recognize much of what is going on thus far. The bulk of ‘next’ is what a LISPer would call ‘apply’. The other component in a LISP interpreter is ‘eval’. Apply applies primitive operations to arguments, or, if the operation being applied is not primitive, it looks for its definition and sends it over to eval. Eval evaluates commands before picking what to apply. Eval is where macros come in. Macros are just an elaborate way for a machine to pick what to do based on what it sees before it goes and does it.
In my programming environment there are single words, and each of them is of one color: red, green, blue, and gray. Gray words are primitive words. Hence, the part of ‘next’ that has all those ‘case check()’ statements is really just the way of dealing with gray words.
I usually factor working with gray words out into its own function, just as LISP people factor out application from evaluation. Here is one way I might factor out executing gray words:
function gray(word){ // gray words are primitive words
function check(){ // instructions are coded as the length of a list
if(isNil(word))return true; // reached the end of mand?
word=R(word); // drop the first item of the list
return false; // haven't reached the end of mand
}
switch(true){ // disquote instruction
case check(): drop(); break; // a length zero list quotes drop
case check(): dup(); break; // a length one list quotes dup
case check(): pop(); break; // and so on
case check(): push(); break;
case check(): swap(); break;
case check(): nil(); break;
case check(): pair(); break;
case check(): part(); break;
case check(): mux(); break;
default: break;
}
}
So then, the definition of ‘next’ can be written just
function next(){ // do next instruction
let mand=L(mands); // get next instruction
mands=R(mands); // set up getting next instruction
gray(mand);
if(speaking||listening)return; // if receiving or transmitting wait
schedule(next);
}
Here it is now clear that each mand in the list of mands must be a gray word. Setting up the appropriate gray words is what other programming langauges call ‘compiling’. But, everything doesn’t need to be gray words. That would make it impossible to define new words.
A note on colors. There are a few ways to deal with them. One way is to separate out how a word is spelled from the color it is spelled with. Directly, words have a color and a spelling.
I’ll get right to it. Programs are a list of words. When you see a gray word you do what it says to do. When you see a green word you mark your place in the program and search back word by word for the nearest red word that is spelled the same as the green one. If you find such a matching red word then you start going forward and follow on with the program from there. If you don’t find one then usually everything is designed to crash so that you know you’ve done something very wrong. An alternative is you can have the machine silently return back to where you left off i.e. the place you marked right after the green word whose matching red you failed to find.
Blue words are special. They are like green words, but they make things happen the moment you’re done typing them. There are a few different ways to deal with blue words. Each way is something you have to be careful about because blue words do things NOW and not later.
In my programming environment there is no difference between programming and editing and executing. It’s all just programming. The moment you finish typing a blue word it gets treated like you just ran into a green word spelled the same way. That means, everything stops and you go looking backwards word by word for the first red word spelled the same as the blue word you just finished typing.
Blue words only do anything the moment after you type them. Otherwise, they are treated just like when you run into a red word: nothing happens.
So green and blue words are alike except that if you run into a blue word while running a program nothing happens and when you finish typing a blue word it gets treated as if you just ran into a green word spelled the same as the blue word.
When green words mark where they were at so that you can return to them aftwards this all happens on the list of the tree. Yes, the tree is still here, doing all these things. Also, red words don’t work exactly the way that I said. Neither do programs.
A program is really like a list of red words followed by a list of green, blue, and gray words. Let us agree that the right part of an ordered pair is where we’re going to put the color of a word and the left part is going to be where we spell out.
Sadly, that’s all I have time for today. Just as things were getting juicy.
Hey, we need to talk about behavior. Why don’t you take a seat next to me right here. I know it feels a bit like being sent to the principle’s office. No one wants to talk about behavior who isn’t a bit afraid of what that might entail. In the past I’ve scared some people off just by talking about behavior the way that I do. Is there really something so frightening about the way I talk about behavior as opposed to the ways that others talk about it? Or, is it just a difficult topic to deal with?
I take the responsibility for the discomfort I’ve caused people when speaking of behavior and, specifically, when telling them about the control of human behavior. That phrase, “the control of human behavior”, is certainly weighty. It’s like something you’d hear in a monologue by a James Bond villian. “And now Mr. Bond once I flip this switch and activate my brain ray satellite all the people of Earth will be under my control.”
There’s no good way to start on the control of human behavior. This is probably a good thing. It probably helps to make people uncomfortable when discussing something so profoundly consequential. At the same time it makes the whole topic seem much larger and overbearing than it actually is. It’s no more overwhelming than talking about physics, but it is immediately personal in a way physics may never be.
One way of approaching the science of behavior is through the history of psychology. Something about starting way off makes people feel in control of subject. It has to do with how foolish the ancient theories are compared to the common sense of today. But, once the march through history arrives at our front door, it is far more unsettling than it has any right to be.
The history of pschology is a no go. It would take too long to get started on the stuff that matters most and the urgent demands of daily events leaves little room for dottling. Though, I wouldn’t want to sound self important. As if the only way to save the world is with the science of behavior. This is certainly my opinion, belief, or conviction, but it’s no more my aim “to control the world for good” than it is to follow the Bond villian and “controlt he world for bad”.
Rather than continue to fumble for a place to start talking about behavior, I’ll go straight to the single paragraph summary of the first chapter of B. F. Skinner’s “Beyond Freedom and Dignity” (BFD) as it is better than anything I might have to say:
“Almost all our major problems involve human behavior, and they cannot be solved by physical and biological technology alone. What is needed is a technology of behavior, but we have been slow to develop the science from which such a technology might be drawn. One difficulty is that almost all of what is called behavioral science continues to trace behavior to states of mind, feelings, traits of chacter, human nature, and so on. Physics and biology once followed similar practices and advanced only when they discarded them. The behavioral sciences have been slow to change partly because the explanatory entities often seem to be directly observed and partly because other kinds of explanations have been hard to find. the environment is obviously importgant, but its role has remained obscure. It does not push or pull, it selects, and this function is difficult to discover and analyze. The role of natural selection in evolution was formulated only a little more than a hundred years ago, and the selective role of the environment in shaping and maintaining the behavior of the individual is only beginning to be recognized and studied. As the interaction between organism and evnironment has come to be understood, however, effects once assigned to states of mind, feelings, and traits are beginning to be trace to accessible conditions, adn a technology of behavior may therefore become available. It will not solve our problems, however, until it replaces traditional prescientific views, and these are strongly entrenched. Freedom and dignity illustrate the difficulty. They are the possessions of the autonomous man of traditional theory, and they are essential to practices in which a person is held responsible for his conduct and given credit for his achievements. A scientific analysis shifts both the responsibility and the achievement to the environment. It also raises questions concerning “values.” Who will use a technology and to what ends? Until these issues are resolved, a technology of behavior will continue to be rejected, and with it possibly the only way to solve our problems.” pg. 24-25 BFD
If you made it through that paragraph then you are among the few. Heck, if you’re reading anything I’m writing here, then you are among the very few. I once thought of building a culture entirely based around that single word “few”. When you say it aloud it sounds like “phew”, a sigh of relief, and often the first and only way to impose a calming kind of self control.
But, having gotten through that paragraph and having become one of the few people that reads what I write here, what is there to do about it? Or, said another way, what is our problem? There must be something wrong if we are all here now reading this. It is not as if I’m writing this because I enjoy hearing myself talk. I do, but that’s not why I’m writing this.
Something about the way the world works is not working. If, like me, you have some history with the sciences, then there appears to be promise in any practices brought under scientific control. Most people fear this. Control of any kind is to be obliterated. Controllers are to be dismantled. They are seen as an over threat to our freedom and dignity as people who are otherwise in control of their life and work. I can do what I want to do, can’t I?
So much has been said about behavior that is prescientific that it is hard to get to where the science starts. All that I can say is that it helps to start with some other science, e.g. I started with physics and math, and carefully extend their experimental and theoretical practices to the subject of behavior. Why behavior? Why not mind? I have no short answer to that.
If you go to a psychologist they are going to talk with you about what you think and feel, and, if they are good, may only indirectly approach your behavior as a seperate subject. Even then, the approach to behavior is going to be remote from the places where it actually occurs (unless you are exceptionally rich and the psychologist comes to you, even then, what can be seen by the psychologists eyes can easily be shown to be inadequate to dealing with behavior scientifically).
Your reports on what you have done, the occasions on which you did them, and the consequences are sketchy at best. We do not yet live in a world where the experimental practices characteristic of a science, specifically a science of behavior, are present in everyday life. To even suggest something like that is to flame fires of revolt as an escape from controlling interference over everyday life. Records of your behavior in everyday life are almost never used to help you e.g. your search history is sold to the highest bidder because it can be used to present you with advertisements that are designed to take your money without you feeling as if it has been stolen.
As long as we proceed in a prescientific way, the cultures of the world are to be stuck in a diabolical equilibrium. Controllers continue to give people what they want, and, if they can, to control what you want so that they can give you what they have in exchange for what they want. This battle of wants is an example of Skinner’s “traditional prescientific views, and these are strongly entrenched”.
To start by telling someone that there common sense explanations for behavior are bankrupt is to end the conversation before it has even begun. This is where it helps to start with some other science, like physics or chemistry or biology, and creep up on a science of behavior. There are problems with that method too and they are somewhat new problems. The sciences of physics, chemistry, and biology are presently stuck on experimental and theoretical practices which test or evluate hypotheses. Modern students of science are taught “the scientific method” and it is a step by step guide to how to deal with hypotheses. It is as if science starts with an idea and the rest follows from there.
To even contend something other than that science starts with ideas is foreign to most people. To tell them that science starts in the laboratory not in the library is to condemn those dedicated to scientific theories. It isn’t, but that is the feeling that comes from suggesting such an alternative. Maybe you’re more sophistocated than the average student of science or more enlightened than the professional scientist. Maybe you’ve seen some shit, and know your way around this world of uncertainty. Maybe you’ve been snowed in by a few blizzards of data and lived to tell the tale.
I’m going to stop here because this note doesn’t seem to be working. I don’t want to delete it or pretend that it didn’t happen. It’s important to me, for the sake of accuracy if not completeness, to leave the mistakes on the page for all to see. Before I can know something I have to start by not knowing it already. If I knew it already then I’d be a fool. Even the things that I seem to know already often yield upon further reflection.
Maybe I’ll switch gears and reflect on why this entry isn’t working. I wanted to talk about the science and technology of behavior. I have tried many times to start such a talk. Each time I end up getting stuck in the weeds well before getting to any kind of proper introduction. This time I thought that if I started with something that someone else wrote, in this case Skinner, then I might see how to go on from there. It’s possible that is still a helpful way to begin, but that the conditions today are not ripe for effective writing.
In the past I kept a cumulative record of my keyboard. I could look at each key press and see the shape of the curve of responses. I had to drop that because I spent more time thinking about how to program it than responding to it as an experimental analysis of behavior. That is, I worried about setting up the aparatus rather than uncovering what it revealed about my own behavior. This happens so easily with such things. It occurs to me that I may just be manufacturing excuses.
Ugh, this has definitely devolved into something going nowhere.
So, the problem is to start talking about the science of behavior. If that was the real problem then stating it would give me some hint as to how to solve it. As stated I see no hint. So what is the real problem?
The real problem has to be something closer to the conditions underwhich I’m writing right now. Whatever words are coming are not the ones that I had planned on when I started out writing about the science of behavior. I’m stuck in a place that I seem to get stuck often. Writing about writing rather than writing about what it is I started writing about.
I could stop writing right now. I could get up and walk around and perhaps do something else. After doing that something else for a while I might be ready to come back and write about the things I set out to write about. That also might not happen. It might be better for me to sit here and write my way through this obstacle.
Is this an obstacle? Is this verbal behavior obstructing the other verbal behavior that is more appropriate to the topic at hand? Am I still steadfastly discovering what it is that I have to say by sitting here and writing like this, or am I simply responding by spinning my wheels?
There are still fragments of what I wanted to say that are strong enough that I can almost feel them on my finger tips. Some decaf coffee might help. A few sips of the good stuff can get things moving in the right direction. A little bit of momentum can carry me through the toughest of spots.
Ah, it occurred to me that what I have set out to write on is too large as a single problem. Breaking it into pieces is what I’m forgeting to do. Most of the time I can take on the whole thing all at once details and all. That’s one of my strenths and also translates into a weakness when it comes to writing to those who are not already familiar with the whole that I have in mind. This is certainly an example of that error at work.
It’s not enough to break down “write on the science and technology of behavior” into “write on an introduction to the science and technology of behavior”. Both are, in my lights, clearly not what I am aiming to write about, nor are they what I am likely to write about effectively. This writing here is evidence of the present insurmountability. A different line of attack is needed to get from here to a better vantage point.
I haven’t floundered like this in a long while. Usually I’m away from the keyboard when I get ‘lost in thought’ like this. It’s kinda cool that I’m able to make a record of my thinking as it’s happening. Later, if I read this later that is, I should be able to see more clearly where I went wrong and what I got stuck on. There’s always something I’m stuck on, behind, or something like that. When I write in circles like this it comes down to something simple and obvious but which I only see with the power of hindsight.
It’s possible to leave biology and enter into behavior by focusing on how cells produce forces.
This is future John. That’s where past John left it, and I think he did all that he could. Good job past John. Sorry I’ll never be able to complement you myself. Here’s to future future John doing better than past John!
This continues my notes on volume 1 of Russel and Whitehead’s “Principia Mathematica” from here #2025-0927-1559.
The end of Quine’s “Set Theory and its Logic” is a brilliant panorama of set theories including Russell’s and Quine’s own New Foundations. As with almost everything Quine wrote, there is something subtle yet profoundly new to his presentation of Zermelo-Frankel set theory. By passing through stages from Russell’s theory he lands firmly in Zermelo-Frankel’s set theories as a result of a theory of cumulative types. This is accomplished in a way that is abhored by modern type theorists: by basic predicate logic. General variables replaced the adornments of so called ‘typed variables’ and all that goes as type theory is treated deftly by predicates appropriate to the type theory of interest.
Why type theorists reject the methods of predicate logic is clear enough from anyone sufficiently familiar with the methods of predicate logic: there is no room for myth or mysticism in the extensional contexts of a theory with a predicate logic. It is much like the brash requirement that an objector “put up or shut up”. It is not so confrontational, but this is the sense that I get from reading type theorists. There is a similr kind of fear of predicate logic in the methods of category theorists. Ask one to give you the predicate logic of their category theory and they scatter or loosely mention something about Lawvere. How mathematicians and logicians have come to accept such weak methods given what was learned through Frege is beyond me. It is so far beyond me that I almost take it as a personal attack on the profession.
Unlike Quine’s method of analyzing set theories, including Russell’s, most modern type theorists are working in a secondary theory. The primary theory is something like Quine’s protosyntax, or what is now most often called bit theory or binary tree theory. The difference between bit theory and binary tree theory is how you order their items.
I aim to deal with the origins of Russell’s theories of types in this note. They result from his own work and from Frege’s analysis of judgement.
Russell’s theory of judgement and perception together with Poincare’s vicious-circle principle are the access points proffered to the reader as proto type theorist.
The vicious circle principle is an example of a geometrical metaphor of misdirection. The geometric metaphors, especially those which mention circles, cyclical motion, or spirals are among the spatial metaphors that misdirect an effective analysis of behavior and its origins. Instead of saying that, e.g., that history occurs in cycles as if what has happened before is happening again, it is much more accurate to speak of schedules.
“An analysis of the paradoxes to be avoided shows that they all result from a certain kind of vicious circle. The vicious circles in question arise from supposing that a collection of objects may contain members which can only be defined by means of the collection as a whole.” pg. 38 PM1
Thanks to hindsight, we now know that Russell’s paradox is not a paradox, and that the vicious circle principle is not a principle. Other than in remote places in Quine’s work, I have not seen sources that give appropriate attention to these gems of hindsight so I’ll give them some time here.
Russell’s paradox is most often presented as follows: is the class of classes not classed as themselves classed as itself? The argument goes that if such a class is classed as itself then it is one of the classes not classed as themselves, a contradiction. From whence, if such a class is not classed as itself then it is one of the classes not classed as themselves and hence is classed as itself, a contradiction. To emphasize the eroneous hint of paradox, the conclusion of this argument is often presented as “If it is classed as itself then it’s not and if it’s not it is.” as if it is in some fluxuating state of selfclasshood and nonselfclasshood. This way of dealing with the history of classes obscures not only the logic but also the theory.
First, a key distinction which was not available to Russell is that between the copula ‘is’ of predication and the predicate ‘in’ of classification (or what is more often known as membership, belongingness, etc.). I emphasize a predicate of classification rather than one of membership or belongingness because I wish to make it clear where among my metaphysical trichotomy of predicating predicates, classing classes, and attributing attributes I am speaking. It is crossing these methods that is as dangerous as crossing the streams in Ghostbusters.
The copula ‘is’ is a grammatical connective, it is also called a grammatical atom. The grammatical construction of predication pairs a predicate with variables (better to call them pronouns and to later distinguish between free and bound pronouns than to get stuck speaking of names). We mark predication with copuli. One place predication pairs a one place predicate with a single pronoun e.g. ‘it is red’ predicates the one place predicate ‘red’ of the pronoun ‘it’ and ‘it’ is free in the predication ‘it is red’.
The predicate ‘in’ or, better ‘classed’, is a two place predicate. It is predicated of two pronouns. Two place predications are usually marked with ‘is’ and a new grammatical atom ‘to’ e.g. ‘x is classed to y’ predicates ‘classed’ of ‘x’ with ‘y’ and ‘x’ and ‘y’ are free in the predication ‘x is classed to y’.
Russell’s paradox is now cast as bad grammar i.e. sometimes mistaking a predicate for a copula (as an inert grammatical atom that happens to aid in a grammatical analysis of a predication into its purportedly grammatical parts) and sometimes mistaking a copula for a predicate. When I use the two place predicate ‘classed’ it is harder to make this mistake but there is still a good enough explanation as to why such mistakes occurred in the first place.
So to put Russell’s paradox in my metaphysical trichotomy: it is crossing the boundary between predicating predicates and classing classes without passing through immigration. On the predicate side there is no paradox: there is only a validitiy of predicate logic. On the class side there is no paradox: there is only a conclusion of class theories.
Predicate abstraction is needed in order to vividly reveal the metaphysical mistake. Predicate abstraction is the logical import of the relative clause of English e.g. ‘who loved Dick’, ‘which is uptown from here’, or ‘that was late’. If you look up the definition of ‘relative clause’ you’ll get something like “A relative clause is a clause that modifies a noun or noun phrase and uses some grammatical device to indicate that one of the arguments in the relative clause refers to the noun or noun phrase.” from https://en.wikipedia.org/wiki/Relative_clause.
In logic, relative clauses are predicates. This is not obscure since, in English grammar, a complete simple sentence which contains a relative clause is analyzed into its subject and predicate such that the predicate is identified as the relative clause.
There is a slight hitch here since in English the copula, e.g. ‘is’, can often be found as part of what comes out as the predicate of a sentence. Such nonuniformities are carried over to the uniform regimentation characteristic of logical (and grammatical) methods by dropping the copula from English predicates e.g. The predicate of the sentence “It is red” is the clause “is red” in English and is “red” in a logical theory with the one place predicate ‘red’.
In english, the pronouns ‘who’, ‘which’, and ‘that’ in the relative clauses “who loved Dick”, “which is uptown from here”, and “that was late” are often shuffled into some special grammatical category e.g. relative pronouns. The pronouns of logic (historically known by the misleading name ‘variables’) are of a singular sort. They can later be classified as bound or free depending on whether they occur within a predicate abstract or not, but I’ll get to that in a second.
Quine’s method of regimentation of relative clauses into predicate abstracts embraces a verbal practice familiar to mathematicians and marked in English sentences by the phrase “such that” as in “x such that x is even and x is greater than twenty”. Examples: “who loved Dick” is regimented as “x such that x loved Dick”, “which is uptown from here” is regimented as “x such that x is uptown from here”, and “that was late” is regimented as “x such that x was late”. Any pronoun other than ‘x’ will do in each of these cases, though when substituting in and out of predicate abstracts for the sake of generating validities from validities more care must be taken to preserve validity across instances either by restricting grammatical constructions or by restricting substitution.
Having regimented the English relative clause into the strict ‘such that’ clause of matheamtical English a final step of regimentation is introduced to finally paraphrase the relative clause of English as a predicate abstract of logic. It eases the grammatical analysis of the scope of a predicate abstract.
Quine chooses the somewhat radical route of paraphrasing “who loves Dick” as “{x: x loves Dick}” which can be read aloud in Mathematical English as “x such that x loves Dick” though, here, it is obvious that something of the scope that can be analyzed by pairing off left and right curly brackets is lost to some other context in both the English and its mathematical variant. This notation is radical because historically curly brackets have been associated with what some called “class builder notation” e.g. “{x: x loves Dick}” is read as “the class of x such that x loves Dick”. Further emphasizing this association, Quine goes so far as to use the epsilon of class-membership that is most familiar to modern students rather as the copula of predication. Interestingly, he draws the line at letting “=” be short for the grammatical connective of coextension!
My personal preference has evolved into a further plying of mathematical English into that which is easily seen to be as formal as any effective grammar of logic. The relative clause “who loves Dick” is paraphrased as the predicate abstract “(x such that x loves Dick)” where the parenthesis set up the scope of the pronoun ‘x’ bound by the context “(x such that … )”. This is largely inspired by the familiar method of function abstraction in theories of lambda calculi and programming languages descending from McCarthy’s LISP.
All at once, the predicate abstract ‘(x such that x is red)’ abstracts the predciation ‘x is red’ by binding the free occurrence of ‘x’ in ‘x is red’ with the binder “(x such that …)” so that ‘x’ is bound in the predicate abstract ‘(x such that x is red)’.
Schematically, with the predicate letter ‘F’, “(x such that x is F)” abstracts the predication “x is F” which is sometimes written as just “Fx” but which I shall not write as such here because I am emphasizing the difference between the copula ‘is’ and the two place predicate ‘classed’.
One last uniform method of regimentation is helpful when dealing with predicating and abstracting many place predicates. The two place predication “x is classed to y” of the two place predicate ‘classed’ with the pronouns ‘x’ and ‘y’ is paraphrased as “x with y is classed”. This is terrible English, be it mathematical or not. A better method has yet to evolve. Continuing, the three placed predication “x stole y from z” of the three place predicate ‘stole’ with the pronouns ‘x’, ‘y’, and ‘z’ is paraphrased as “x, y, with z is stole”. This suggests that “x, y, with z is stole” is short for the somewhat more acceptable “x with y with z is stole”.
The general problem of ‘with’ is subject to a similar problem as that with the copula ‘is’. Here, the use of ‘with’ along with the commas ‘,’ is just a use of grammatical atoms to clearly indicate which of the many grammatical constructions of predication are participating in predication examined (often for the sake of simplifying grammatical analysis).
Now that predicate abstracts are firmly planted as part of predicate logic, that is as the logical import of the English relative clause (thanks to Quine), there is just a bit more to do to get back to Russell. Quine stopped at presenting predicate abstraction as a mere huerustic. He said so in the final chapter to “Methods of Logic 4th edition” on “Axiomatic Set Theory”. This was misleading of him though, as he had gone much further than a mere heuristic when invoking the same notation under the name of “the virtual theory of classes” in “Set Theory and its Logic” and the earlier work “O Sentido da nova logica” section 51 of 1942.
I must also make it clear that Quine is not unqiue in his recognition of the importance of something like predicate abstraction. He is responsible for clearly and exactly linking it up with the relative clause of English. He showed how predicate abstraction is burdened with the whole business of bondage and freedom. Others mixed it with quantification. We can be sure that Quine made this link in a special way by looking at how it interlocks with his method of predicate functors and the elimination of pronouns.
There is further uniqueness that went unmentioned by Quine in the ways that predicate abstracts permit a smooth flowing exposition of so called ‘free logics’ where free pronouns come to play the part of dummy singular terms.
This all is not as obscure as it might seem since the entirety of Russell’s type theoretic distinctions are derived from mishandling names. Schofinkel (1924) and Curry’s combinators, Tarski’s cylindrical algebra (1952), Bernay’s modification of Tarski (1959), and Halmos’s polyadic algebras (1956) deal in their own way with predicate abstraction but only indirectly because they are aimed just at eliminating variables and only do so within a given theory whereas Quine’s methods of predicate functors eliminate variables from logical methods entirely.
A bit more history. The epsilon as copula is Peano’s and it is also Peano who, notationally, did most of what goes as predicate abstraction, as Peano used a flipped epsilon as short for ‘such that’ though Peano was closer to Russell and closer still to the theories of function application and combinator application in letting the right way epsilon go as ‘classed’ which is the ‘application’ of class theories. Peano and Russell both adopted the methods of class and relation abstraction. Russell used a hat over variables and let the alphabetical order of the variables indicate the arrangement of items that may be so related, and peano used the backwards epsilon, though I think only with one variable, I’m not quite sure about that.
This is all a horrible way of saying that Quine’s methods are not unfounded or unanticipated. They are presented as distinct methods though and this is often an essential ingrediant in any effect method.
Back to predicate abstracts and their part in the grammatical constructions of logic. In a truth functional and quantificational logic, predicate abstraction is an abbreviation. The variable bound by a predicate abstract comes to be bound by a quantifier. Technically, the truth functional and quantification logic must be a truth functionl and quantificational theory of identity. This comes from the fact that Wang’s premise of identity theory is the machinery which permits the quantificational abbreviation of predicate abstraction to yield the premises of concretion that I am now to explain.
Truth functional and quantificational logic is peculiar on theories of identity (and equivalently adequate logical methods such as predicate functor logic, equivalent adequacy can be established by supplementing the one with the other and proving the grammatical category of the other equivalent to some grammatical category of the one, and then vice versa). A given logical theory has only finitely many items in its lexicon and from these a predicate of indiscernability can be constructed such that it is always coextensive with any predicate of identity that may be already among the lexical predicates. Such a predicate of lexical indiscernabiltiy can be investigated by way of the methods of schematic theories. Schematic theories are contrasted from their instances by predicate letters which are reserved as schematically lexical. This means that unlike the other predicate letters in a schema, lexical letters (as they may now be called) may only be substituted by predicates of a contemplated lexicon when generating instances of schema in which they occur. Sometimes a special alphabet is reserved for such work, but that is rarely needed in practice. Marks can be attached to the familiar predicate letters rendering them newly lexical.
As a consequence of all that I have just said, the quantificational abbreviation of predicate abstraction can be introduced into any such logical theory since Wang’s schematic premise of identity is a validity of logical theories. The only reason it is not a validity of logic is because the methods of logic proper are independent of lexicon (and theory).
I now come to explain where concretion is among these preliminary remarks. The schematic premises of concretion are the theoretical import of a profound principle called ‘predicational completeness’. It was identified by Quine and explained as follows. Each sentence in which a given name occurs is equivalent to a one place predication of a one place predicate with that name.
In simple cases this seems silly e.g. ‘It is red’ is already in a form where one place predication of the one place predicate ‘red’ with ‘it’ is equivalent to, identical in fact, ‘It is red’. Where things go strange is when there are multiple subjects e.g. ‘He stole it from her’ is equivalent to ‘He is who stole it from her’ which, paraphrased as a one place predicate of a one place predicate abstract with ‘he’, is equivalent to ‘he is (x such that x stole it from her)’. Predicate abstracts bestow a language with this precious property of predicational completeness. Upon reflection we may see that much of what goes as philosophy is the result of this almost pedestrian analysis of the relative clause into predicate abstraction.
Independent of the coextension of identity with lexical indiscernability, predicate abstraction can be introduced as basic grammar and put to work in any theory with the aid of the premises of concretion (each instance of the schematic premises of concretion). Concretion is the principle which generalizes the equivalence between “he stole it from her” and “he is who stole it from her”.
The remaining schematic premises follow on with each of the many grammatical constructions of predication.
I have presented the premises without presenting the grammatical category of predicate abstracts. I shall do that now. An n place predicate abstract pairs a zero place predicate (sentence) with n pronouns and is usually marked by the grammatical atom ‘such that’, and perhaps even with parentheses.
Such ‘tagged’ trees are frequently useful as they aid in any grammatical analysis. When tagged as in the example, that is when tagged by the name of the gramamtical category into which they are constructed, such trees are called “abstract syntax trees”. The word ‘syntax’ comes from an unhappy time before the distinction between what were once called ‘syntax and semantics’ was untangled. Technically, they were put back together again, like Humpty Dumpty after his fall.
Having introduced the grammatical category of predicate abstracts, each instance of the schematic premise of concretion is part of any theory in such a logic. In almost all cases, these premises are redundant. It is also an appealing part of logical methods generally sice what are sadly called ‘extra-logical axioms’ (as if there are ‘nonlogical premises’ and hence ‘logical premises’), like those premises which are instances of a the schematic premise of concretion, are dispensible i.e. there is nothing gained from adding a grammatical category of predicate abstracts to the grammar of truth functional and quantificational logic, especially when a further restriction is placed on the premises of any theory which includes such predicate abstracts.
The explanation for the redundancy just mentioned is to show that the premises of concretion are conclusions from Wang’s premises of identity. In fact, concretion is equivalent to an instance of Wang’s schematic premise of identity which substitutes its predicate letter for another instance of Wang’s schematic premise of identity. (Technically, if Wang’s schematic premise is presented as a universal closure then some care must be taken when explicating the definition of quantificational validity e.g. to segregate the vacuous validities of a vacuous universe from the others as in Quine’s methods of “free variable logics”).
The lexical status of ‘classed’ permits the contextual elimination of any instance of the schematic singular description built up from only the two place predicate ‘classed’.
Singular desriptions prompt contemplation of their descriptional premises i.e. speaking of ‘the such and such’ tends to proceed from tacit premises that there is at least one such and such and that there is at most one such and such. It is the first of these descriptional premises (called “the descriptional premise of existence” and the other is called “the descriptional premise of uniqueness” though they can be combined into one like the schema “some item is (x such that Fx and each item is (y such that Fy if and only if x is identical to y))”).
an inconsistency outright. So, by an augmentation of Quine’s main method that accomodates interchange of equivalents, the descriptional premise of existence of Russell’s one place predicate of nonselfclassification is an inconsistency. Therefore, its negation is a conclusion in any theory whose lexicon includes the two place predicate ‘classed’. Careful attention must be paid to the wording of my conclusion.
I’m somewhat surprise at how good it has felt to write this way. In the past I couldn’t let a single sentence sit upon the page unless I saw its interlocking connections with the basic principles that I picked to govern my behavior. I’d get stuck talking about behavior rather than about all those mental things, e.g. beleiving, that becloud any effective methods of dealing with the world.
It all spoke towards what I might call an insecurity in my own philosophical foundations. There were doubts which I have eliminated. Since they are doubts about the nature of logic, there was practically nothing I could speak of that wouldn’t bring them up. Doubt can be crippling. It can spiral. You can doubt you’re doubting (contrary to anything Descrates may have said). There’s so much uncertainty. So little credibility. Yet, “Act with vigor without absolute certainty” was always looming since I’d read Russell’s words.
As my grasp of what I call “the metaphysical trichotomy” strengthens, I have found myself rarely paralyzed by the precision that was once necessary to get my stories straight. The metaphysical trichotomy is a division of metaphysical methods into
Among these three, the two which have occurred most throughout history are the classing of classes and the attributing of attributes. This is surprising since Aristotle wrote of subjects and predicates as if they were some common sense starting point of a rigorous analysis. Though, anything that is said of Aristotle is perhaps subject to ammendation as much as anything that is said of any fragmentary source of historical record.
Perhaps what has survived of Aristotle is nothing like what he said to his inner circle. The difference between the exoteric and the esoteric is largely a measure of the size of the community. As much as Aristotle’s private thoughts from his esoteric works bleed into his public presentations in his exoteric works, we are stuck with what has survived when peicing him together. He is no more special than anyone else from history in this regard and the challenge is ever present.
The predicates of Aristotle were, until quite recently, the predicates of the modern world. There really has been so little progress in philosophy that most of what Aristotle had to say is still relevant to today. The same can be said of almost all the well known philosopher’s contributions to their subject.
Aristotles predicates are attributes. They are one place relations. Whether they are relations in extension or intension is somewhat obscure, but, being a philosopher, we can err on the side of intension. I for one have no grasp of ‘intension’ except as another way of saying ‘not extensional’.
Presumably the division of complete sentences into their subjects and predicates are somehow responsible for the mirror world of attributes and particulars that pervade speaking of the intensional.
The word ‘attributes’ and ‘attribution’ are a catch all for intensional stuff. Intensionalists will not like such a broad restriction. Further work can be done to set up better methods of dealing with intensional context. This is easy enough to admit. But, there’s not much more to be gained from the trichotomy so early in an introduction of it.
A skip back to Aristotle to clear up the correspondance between his talk of substance and my talk of attributes. Aristotle’s substances are what attributes are attributed of. Paralle, following the trichotomy, names or singular terms are what predicates are predicated of, and items are what classes are classed of. Though it is more familiar these days to speak of pure class theories or pure set theories, there is nothing to be gained from mathematical methods which do not link up with the rest of the world.
Properly, I take items only and solely as what one place predicates denote. There is reason to suspect a difference between items so defined and items as that which are classed by classes.
Attributes are elsewhere called properties, intensions, concepts, qualities, aspects, natures, characters, features, types, exemplifiables, essences, meanings, and any other stuff that is somehow only mentioned in nonextensional conversations and contexts.
There are parallels along the trichotomy throughout the history of philosophy that I submit are so helpful they can scarcely be ignored. An example is the difference between atomic and compound particulars parallel to Aristotle’s difference between substance and matter.
Because I am reading over a lot of articles at once I’m taking notes that do not flow, but I would rather record what is occurring as it occurs then try to reconstruct such occurrences from memory.
Properties are universals, though there are some who may say that they are particulars and there such properties are called tropes, and substances are purportedly particulars. Properties are said to be exemplified or to be instantiated.
I didn’t get out as much as I may have hoped, but this is a starting point. Here is a list of the pages I happened to have open before I dropped this line of inquiry:
There is so much that I now must do that I can’t sort through it all and organize it into a plan without writing about it here. Most of the time when I make plans I don’t write them out. I see them to myself and can go from there. It’s a pleasing thing to close your eyes and see yourself doing the things that you need to in order to get things done. Sometimes it’s a frustrating thing. You see how easy it was in your head and then when you go to do it you yearn for that ease of imagination.
I’ve not spoken much about my imagination because it was so far from the language that I was permitted to use while I was solving the problems of logic that presented themselves to me over the past decade. It’s hard to look back and accept that I spent a whole decade working on something that I only just recently began to understand. The moment I understood what I was woriking on, the solutions were practically there.
This is not unfamiliar. When I showed that all compactly supported real valued functions have a finitely n refinable extension there was a lot of time spent just understand what the hell I was actually trying to do.
Among the things that I’ve figured out to myself, the countable union of countable sets is countable is the most interesting. This is a rudimentary problem in any regular real analysis course, but for me it was part of a class run by the Moore Method. Now it’s called “The discovery method” or something else which is a bit more politically neutral.
R. L. Moore rubbed some people the wrong way, rightfully in some cases, with the intensly confrontational style of his teaching. He would construct a synthetic mathematical environment in the classroom and get people in front of an evolving mathematical audience in order to construct new mathematicians.
Before I was in my first Moore method class I was just a physicist. After a Moore method set theory course I was a mathematician. You’re given a list of definitions and theorems. That’s all. There are some mild threats about having to take a test on anything that you get wrong in front of the rest of the class, or on anything that the class lets slip by in a purported proof you present to them.
Later, as I began to learn more about the experimental analysis of behavior, it would become clear just how lucky I was to have a Moore method class. That I had multiple courses led this way was a boon. It is also clearer to me how difficult it would be for anyone without my accidental chain of mutually reinforcing teachers to surmount the peculiar obstacles set up by that book length list of definitions and theorems.
The effectiveness of the Moore or Discovery method in mathematics, and in any other largely logical subject for that matter, is in its accidental arrangement of an effective program of instruction. Such a program can be made much more effective than it was, but the logic of mathematics is such that there’s very little difference between what you can prove and what you can know (and when you are likey to know it).
As much as the premises and conclusions of a logical theory are beyond any temporal ordering (even though those like Russel and Poincare presented their principle as a vicious circle or cycle that forces sets to evolve like in Godel’s constructive hierarchy) there is no way to teach a person but in this world of spacetime chunks, or so common sense would have us say.
Already, in an indirect way, which is the only way I really know how to go sometimes, I’ve laid out some of the aims of any plan I may have had the mind to construct here. Mathematics is a collection of reinforcing practices of a verbal community. Logic is even more decidedly verbal and grammar, upon which a repertoire of logic is based, is perhaps the most strictly verbal subject there is outside the science of verbal behavior itself (of which grammar is merely a small derivative of the far more profound autoclitic behavior).
To collect everything mentioned thus far among the aims of any effective plan that may come from these contemplations: logic, math, science of verbal behavior, and an effective technology of teaching.
Not each of these is represented entirely in what I’ve written thus far. That’s okay. There isn’t a sure fire plan for making plans. Some people, like George Polya or Ray Dalio present their principles for solving problems, but those are their ways and not the ways of a science of problem solving. Yes, there is such a science and it is not some cognitive myth or mysticism of the modeling mind or the mind as processor.
I may have dropped the strictures of a technical vocabulary under the control of an experimental analysis of behavior, but the conclusiosn from such a science still hold me in check. That is one of the crazy good things about science. You don’t have to be wedded to it in order to benefit from everything that it gives the world.
This is also one of its weaknesses. Some people, those who know very little about how science works, can slip into some pretty strange beliefs if they aren’t shown the links from what they can do today with their machines and the laboratory conditions those technologies appeared from. Take the Newcomen atmospheric engine. It is presented as an invention of necessity. There was water in a mine and someone with money wanted that water out so they could mine more money.
At little dig into history reveals that there was a tapestry of past discovery that propped up the practices and behaviors that gave the world a Newcomen and his engine. When we turn a key and turn on an engine today, we aren’t reminded of that tapestry any more than Newcomen was when he fired up his boiler.
This is the primary reason that people continue to say “Those who don’t learn from history are doomed to repeat it.” If you’re the kind of person that thinks human behavior is beyond scientific control then it is easy to tell me I’m wrong about that. It was certianly hoped by B. F. Skinner, among all the other experimental analysts of behavior, that there wouldn’t be so much work wasted on bringing the public up to speed with the kind of control available from such a science.
On one hand it is simple for people to see that something is at work when the protogambler finds themselves satisfied sitting at the slot machine for hours on end. Or when you walk into a bingo hall and find the most fastideous dabbers. Still, those unfamiliar with the sciences, especially the laboratory control of human and nonhuman organisms, see something more at work. They see something of themselves in what others are doing. It can’t possibly be that there is something uncertain about my contact with the world. I know myself as well, if not better, than anyone else. What I experience is true, or at least it must count as knowledge. If I can’t trust my eyes, my nose, or my skin, then where do I begin? I can’t possibly begin outside myself, although there may very well be some all knowing entity of which I amonly a part and which exists outside of all of us.
The history of psychology is largely the history of philosophy. Especially the further you go back. You’ll find Plato writing of a Socrates who spoke of memory like a glob of wax. Some people have a firm block, others have a soft one, some have a big ball, others a small one. Whatever they have, the world, we are told, imprints itself upon our wax and there are only so many places ripe for the stamping. They overlap, they break down, and, generally, the globs are fallable impressions of the world which we hold within us and which limit what we may ever know about the imprinting world.
What goes as common sense is really the consequences of a verbal community. The way we talk about the world and how it works is something that we are taught by the people we listen and speak to. Sure, the world is still out there whether we talk about it, at least the consequences come whether we speak of them or not. But, when people speak of common sense being common or uncommon, it is a question of the kind of control that one verbal community has with respect to another.
Damn, I have to shift gears to woodoworking right now. I’ll leave this as it is, look back at it later today and collect together some of the points to aim at in my evolving “post logic” plans.
Finally, months after starting and stopping, I am back to woodwork. A while ago I made plans with my dad to make a clock out of wood. That means cutting the gears and assembling everything only and solely with wood. The first step was to clean up some junk in the garage to make room for a workbench. This turned into far more of a challenge than I expected.
My dad gets attached to things and puts them into places which almost turn into sacred spots. While things started smoothly enough, we soon got into a row over where to put what. I thought some of it should go in the trash and he thought it should go on a high shelf. A few days of this and it was destine to end as it did. It fizzled out and without fanfair ended.
Doing my best not to make the same mistake twice, I went ahead and just started building things on my own. But, lest I be accused of a kind of intellectual dishonesty, I’d never say that “I can just build things.” Without the tools that have been provided by my father, and without the rich history of teachers I’ve had throughout elementary and graduate school, I would not have quickly found my way.
Can you imagine a world without hammers? One where you had to make up the idea of hammers before you even started to think of nails? Or saws? Can you think of a world without saws? This reminds me of a parody of instructional videos from the fifties and sixties. A little animated character pops onto a black and white screen. His name is something like “Springy the Spring” and he’s here to tell you about a world without springs. You are made to see that such a world is akin to hell and that your ignorance of springs is a disservice to your species. Apparently, in this springless world, the springs became conscious and went about chastising humans who pretended like the springs were never there to begin with even though they otherwise knew that the springs had built spacefairing vessles and set up shop on the far side of the moon.
It’s not as hard to imagine a world without hammers as it might first seem. Our world hasn’t had hammers for most of its history. If an account of the cosmos left out hammers then it would still be appropriately complete. How do we get from a hammerless world to a world with hammers and saws and all sorts of tools?
Some people like to say that our brains got us here. We fell from trees because our heads were too heavy. We had such a headache when we hit the ground that we stopped walking around on our hands and feet. With our hands on our balooning heads, struggling to hold them and their precious cargo up, we walked on two legs. After we took an aspirin, our hands were free and what else were our brains going to do but put them to work whacking things with rocks. Eventually we whacked some rocks with rocks and built some edges and cut some grass and wove some rope and lever armed our way into hammering our enemies into houses.
Some like to go further. It isn’t just our brains. There are other organisms that have brains, and they don’t sit around making hammers. What makes us different, us being humans, is our will. We have a spirit of discovery, invention, and adventure. Sure, it has something to do with our brains. Brains are a vital organ after all. But, it’s something beyond brains that gets us into these elaborate worlds of tools upon tools.
What difference does it make if we make hammers because of our brains or our will or because some moon god slept too late and got caught by the sun god and dropped a ray of inspiration upon us in his anger?
The question would be empty if we didn’t have some measure of control over the things that we make. If we wish to find out how to make better tools we might want to know how we came to make the tools that we already have. Is that enough? Is it enough to look upon the things that we have done in order to see the ways in which we may do them better? Or is there more to it? Is there always more to it? Are we at the whim of a world which doesn’t give a damn about us but only every so often flips a coin to reveal a boon?
Somehow I’ve gotten far off track. I meant to explain my woodworking and got stuck questioning the origins of technology. This is familiar. I’ve always got something on my mind, even if I insist upon eliminating all talk of minds.
Let me start with what is in front of me. Tage Frid wrote a series of three books that are the “Feynman Lectures on Physics” of woodworking. I can’t stay away from such monolithic efforts. They speak to my depths. The world is not some spasm. It’s sensible. It sets itself out in a way that brings joy if contemplated just right. I look for that right way, and books like Frid’s help me catch a glimpse of what’s likely to work and what’s likely to fail.
My relationship with books is simple. I love them and they love me back. In my dreams. Wherever my inclination to have and to hold books came from I’ve never lost it and much of my life is organized around it. I like to read the books that I get. I don’t read all of them because that would take too long. I do read some of them many times over and these are often the ones that have taught me the most e.g. ‘Methods of Logic 4th edition’ by Quine.
Tage Frid’s books purport to teach woodowrking. The collection is called “Tage Frid Teaches Woodowkring”. Each page is full of pictures and diagrams which show step by step methods of construction and deconstruction. They hold still the things that videos do not and have a narrow focus as any written piece must have to get its point across. These books are as much a source of inspiration as they are education. Seeing someone working wood gives me the gist of what must be done to have a given effect but there are always some snags when I do it myself and it’s harder to have a video handy than it is a book when in the workshop. Maybe that is just a weakness of the design of my current workshop which is little more than a suped-up saw horse holding a table which is also a big vice.
I’m really interested in the design of worshops. Woodworking is, for me, part of any effectively programmable environment. That means that it extends our control, and countercontrol, over the world around us. This is most profound in the tools for making tools. These being the things that interest me the most across the board from woodworking to digital circuit design. The tools that make tools are made by tools and the methods of making tools are probably the closest humanity has come to really examining systematic methods of discovery and invention.
As much as I deplore whatever it is that props up Youtube as a company, I have no doubt that it has given me access to some great instructional videos on woodworking. None is better, to me, than Paul Seller’s delightful videos that can be found here https://www.youtube.com/@Paul.Sellers. His even more delightful website can be found here https://paulsellers.com/ and all sorts of things can be explored from there.
Paul has a calming effect on me, as any good woodworker does. This all started, I must say, from my early childhood. My dad was always working on soemthing at his bench. I was always watching either Norm Abram’s “The New Yankee Workshop” https://www.newyankee.com/ or Bob Ross’s “The Joy of Painting” https://www.bobross.com/. There were also a lot of cooking shows, e.g. Julia Child’s “The French Chef”, and craft shows. Whatever I made when I was young, nothing was ever as good as my imagination. I still have that problem, especially when I discoverd in the past two years that I have some skill for drawing.
In addition to all the math, logic, and science that I write about here, I’m going to start writing more about my sensitivities as an artist.
Science and art are both two sides of the same coin that is sensitivity. Scientists and artists are taught by communities which care a great deal for the deferred consequences of their practices. When you are told to cut here and not there you can see the difference in the final product. Whether you take the advice or not, you find yourself learning how to carefully select one response rather than an other through a complex process of setting up self controls. Artists and scientists know and teach the power of building a world that works better than it once did. Most of the time all it takes is controlling yourself as you would others. This isn’t as sinister as it sounds. You’d not pay a piano teacher that didn’t successfully control your piano playing.
They are the ones who are most likely to leave it better than they found it.
I came here to tell you about my programming language, but because I’m impatient and don’t like explaining myself I might not. There isn’t a day that goes by when I’m not fighting myself to get things done. Most people are this way. Probably all people are this way. I’d not want to meet someone who gets everything done that they planned on doing. Or who gets everything done whether they planned it or not. What would a person like that even be like?
There was a character in the Batman cartoons I watched when I was young who looked at his watch all the time and one day lost his papers while out to lunch. I’ll call him watch man. That’s not his name. He’s got a proper Batman villian name, but I don’t remember what it is now.
The wind blows away his papers and he runs after them and a bird poops on his lunch all while he’s somehow late for the most important meeting of his life. He misses the meeting, and misses out on that future promotion. He misses out on an entire life of luxury, at least to his mind. This drives him mad. He’s down, out, and constantly looking at his watch. It’s all that keeps him tickity-boo (I heard that once on a British show and it still sounds like a well wound clock clicking along like a cosmic timekeeper).
Watch guy, or man, or whatever, didn’t notice that in addition to loosing his papers, his lunch, his job, and almost losing his watch, he lost his mind. Somewhere between the bird taking a shit and that shit landing on his lunch he went one hundred percent wackadoo. This is how all good Batman villians start up.
Causality screwed Dr. Clock and he was happy to screw his courage to the sticking place. It was super sticky. He stuck it to the world and vowed to beat those thugs in their office suits with their smarmy smiles. Those thin lips that tightened like an untuned piano string. They just loved to take it out on the little guy. The guy who couldn’t make it to the boardroom on time. How was this tiny man going to make it in life if he couldn’t make it to work?
I think Mr. Cuckcoo put someone between the largest pair of minute and hour hands he could find. Wait, maybe this was a Superman villian. Maybe he put Louis Lane on the Metropolis equivalent of Big Ben. Either way it was comically huge. There has to be a book on “Knots for Villians” because they all seem to know how to tie the same unbreakable-until-it-isn’t knot.
Anyway, where was I. Oh, yeah! I was going to talk about my programming language. It has nothing to do with Tick-Tock McClock. Honestly. I just had to get that out there.
Here’s how my programming language works. First, I don’t have any cool way of explaining it. It’s not like Star Trek where a robot comes out and dances until you lose track of all the plot’s loose ends and the credits roll and a beer comercial reminds you how thirsty you are for a cold one and that there’s only one beer cold enough for you and it’s Ocho Equis a beer from a company founded in Mexico and owned by an international Dutch brewing conglomerate you have to Google because its name is deliberately designed to be hard to remember.
My programming environment is really an environment though. When I think about it I go to a forest in my mind. I don’t believe in minds, but never mind. That’s something for another mind. I mean another time.
There’s a forest in the dark place of my mind. I’ve mentioned this dark place elsewhere, but not here. It’s a really not scary place. It’s dark yes, but in a “anything could happen beyond that shadow” kind of way. Really warm and full of opportunity and surprises.
So I go to my dark forest and come upon a tree. It’s white, obviously. I think its a white oak. Is there such a thing as a white oak? Let me look that up. One second. Ok. It’s called a white oak but its bark is actually “light gray”. Lets agree that it is early white with like darker gray lines between the bits of bark. Also, it’s not glowing, but it’s white enough that in the pitch blackness of my dark place it almost looks like it’s glowing. Oh, and it’s nighttime. That doesn’t really matter, but it is.
You can talk to the tree. But before we start talking to it, I forget to describe its shape. Later you’ll find out that this one tree is part of a forest and that the trees are talking to each other. I may have already mentioned the forest thing, but I made the specific tree I walked up on seem a lot bigger than it was. It’s not really a ‘tree’ tree yet. It’s more like a sapling. It’s more like a stick in the ground. Like a nub really. I’m tempted to dig around and see if it isn’t still sticking out of its nut.
The stick was a lie. It’s actually just a leaf sorta sticking out of the ground. I’m not sure if the leaf is silver or what. I am sure of one thing, and it is only something you can see to yourself in your own dark place: the leaf has two parts and each of them is identical to itself. When you look at it, there’s no mistaking it as a whole leaf, but at the same time you know it has two parts, a left part and a right part, and that when you focus on them, or try to, you somehow see the whole leaf all over again. It’s not an optical illusion either. Something about it just makes sense.
If you happen to look around at any of the other trees in the forest, maybe catch one of them out of the corner of your eye, you’d see that all the leaves on all the trees are like this. All the leaves on all the trees are somehow made of two parts, a left part and a right part, and are somehow identical to each of their parts. Don’t ask me to draw this for you because I can only see it to myself and wouldn’t know how to make you see it to yourself. I don’t know if it’s even safe to look at it for too long.
Ok. So we’ve got a weird leaf sticking out of the ground. Now there’s another even weirder thing: the trees are listening. It’s straight out of “The Wizard of OZ” but, rather than the trees being alive, these trees are just really good listeners. If you can get past the weird leaves then you can get on board with the weird listening trees.
There’s a whisper. You turn in the direction it’s coming from and see nothing. When you turn back to the leaf you find a tiny but proper sapling. From the ground, two branches grow out away from each other. In a way just as weird as how the leaf is identical to its left and right parts, whenever you look at the tree you always see the whole thing as having a unique left part and a unique right part. It’s like when old video games use two dimensional sprites to fake a three dimensional object e.g. like the original DOOM’s pickup sprites.
If you notice a difference between what I was writing before and what I’m writing now it’s because I wrote this part a day later. Rather than break everything up into multiple notes, I really want to just give this one a good enough ending to be somewhat complete.
Something about that whisper changed the tree. When you turn around, instead of seeing a strange leaf popping out of the ground, you see a ‘V’ shape. It’s milky white, mercurial, and, if you touch it, warm.
At the end of the right branch of the ‘V’ is a leaf, and at the end of the left branch is another ‘V’ shape. Now that you notice it, there’s a faint smell of oak in the air and maybe even a low hum, but you can’t tell if you’re hearing it or feeling it within you.
So, up the left branch from the ground you run into a ‘V’ that splits into its own left and right part. The right part ends in a leaf. The left part ends in another ‘V’ both branches of which end in leaves.
All together it looks something like this
\/
\/
\/
but everything isn’t as straight as it looks there. All those left parts stretch out from the ground like a skelatal spine. There right parts are like leafy ribs.
You blink and catch a haste of whispers in the wind. By the time your eyes open, your little tree has grown and changed. Maybe it even groaned? It looks like this
\/
\/ \/
\ \/ /
\/ /
\/
and you can still see a strange substances swirling on or beneath the surface of the tree. Maybe it even glitters a bit. It’s too strange to say for sure. What can clearly be said, is that there is a left spine and a right spine.
You happen to be right in the path of the latest whisper and clearly here ‘dup’ and watch as the tree shutters, shifts, and slips into this shape:
\/
\ \/ \/
\/ \/
\ \/ /
\/ /
\/
Stunned, you compose yourself, and shout ‘DUP’. Sensing your urgency, the tree appears to change faster than it ever has before:
\/
\ \/
\/
\ \/ \/
\/ \/
\ \/ /
\/ /
\/
As anyone would, you shout ‘DUP’ again and again, and watch as the left spine grows to what you suppose must be a sky. If there was a moon here, it certainy would have reached beyond it by now.
The next logical thing is to grab onto to left spine and yell ‘DUP’ a few times to see if it can carry you up. It has to doesn’t it? If it supports all those duplicates you sent up earlier, then of course it will support your weight. You jump up and reach for the the second limb from the ground. Your hands pass through it like its made of water. It’s warm. Your hands are dry. You’re butt is on the ground. Peculiar.
A thought flashes before you “Did that tree just drop me?” and the tree wooshed. What? “DROP” you yell. Woosh. Without opening your mouth or saying a thing aloud, you yell “drop” to yourself and the tree moves again. As fast as you can, you repeat it to yourself. You’re so focused you shut your eyes and when you open them again you see the tree like this:
\/
\ \/
\/
\ \/
\/
\ \/ \/
\/ \/
\ \/ /
\/ /
\/
Well, now you look bored. Oh, yeah, I’ve been with you this whole time. It was kinda awkward watching you yell “DUP” alot and then you just closed your eyes for half an hour. I’m tired too. Let’s get something to eat and come back to this later. Deal?
This continues my notes on the first volume of Russell (and Whitehead’s) “Principia Mathematica” from #2025-0922-1511.
While many of the past notes have probably contained more quotes from Russell’s “An Inquiry into Meaning and Truth” (IMT), this note deals squarely with volume one of “Principia Mathematica” (PM1).
My logical analyses of ‘grammatical’ and ‘denotes’ in the previous note from this series are put on the shelf for a moment. They do more to show my response to Russell’s theory of types than to show Russell’s theory of types.
This has more to do with the inscrutibility of Russell’s explanations than with Russell’s work in PM1. Like Frege’s ‘Begriffsschrift’, PM1’s commentaries (or ‘summaries’ as they are called there) are to be entirely ignored when it comes to accurately interpretign the text.
I know that some people, most people probably, dislike this method that ignores the commentary. It is simple enough to justify it. I no more know what I’m saying here than Russel knew when he said things there. We discover what we he to say. This never ends. This also explains why books like IMT and ‘Human Knowledge: it’s scope and limits’ were written by Russell. This also explains why early Russell and late Russell don’t always agree.
We don’t express what we mean, we discover what we have to say.
With that little diatribe out of the way (I can not promise that more may not come), I move into a more technical interpretation of Chapter II ‘The Theory of Logical Types’. This is preliminary to a detailed examination of section C on the general theory of classes, and, prior to that, those parts of section B that deal with the axiom of reducibility and its application in the parts on identity.
“The theory of lgoical types, .., recommended itself to us in the first instance by its ability to solve certain contradictions. .. But the theory in question is not wholly dependent upon this indirect recommendation: it has also a certain consonance with common sense which makes it inherently credible.” pg. 39 PM1
Whatever recommendation there is from common sense is out the window once it is emphasized that such common sense is after the fact. The admission that the common sense is after the fact does not make whatever it is that is being called common sense common sense. This is an example of a subtle trick where history is cast as developmentalistic or progressive e.g. ‘the arc of the universe bends toward justice’.
Beyond this literary dodge, there are the consequences of the theory which Russell proposed. The axiom of reducibility renders his work on logical types impotent. The theory of logical types, and the axiom of reducibility which appears from it, are presented as if they properly tame intensional contexts into extensional ones where the extensional ones merely abbreviate some intensional context with which we are directly (mentally) connected as perceiving, judging, thinking, or reasoning beings.
But, what is said is said of the commentary and not of the work itself. The work which includes the axiom of reducibility is, just as the notation of Frege’s ‘Begriffsschrift’, impeccable. Its contributions to science are a consequence of this impeccability, not the ‘peccable’ commentary on the impeccable.
I’ll do my best to render Russell’s commentary impeccably. I’ll avoid predicate functor logic for now. It eliminates all talk of variables outright, and yet almost all of what Russell has to say is about variables.
“See the last section of the present Chapter. Cf. also H. Poincare, “Les mathematiques et la logique,” Revue de Metaphysique et de Morale, Mai 1906, pg. 307” pg. 39 pm1
“More generally, given any set of objects such that, if we suppose the set to have a total, it will contain members which presuppose this total, then such a set cannot have a total. By saying that a set has ‘no total’, we mean, primarly, that no significant statement can be made about ‘all its members’. Propositions, as teh above illustration shows, must be a set having no total. The same is true, as we shall shortly see, of propositional functions, even when these are restricted to such as can significantly have as argument a given object a. In such cases, it is necessary to break up our set into smaller sets, each of which is capable of a total. This is what the theory of types aims at effecting.” pg. 39 PM1
What Russell says as ‘suppose the set to have a total’ is said by Bourbaki as ‘collectivizing relation’, though the link is not so tight except that they are vaguely about yet another way of talking that appeared later from von Neumann: that between classes and sets where sets are defined as classes which are members of some class and proper classes are defined as class which are the members of no class.
The quote demonstrates a problem that I have with Russell’s writing generally. He allows himself to be wildly imprecise. So imprecise in fact that whatever accuracy he has aimed at is entirely obscured except for those who read him forwards and backwards. That is, you must read on and return from where you began with what he later has to say freshly in your recent history.
I have to linger here because this gets at a general method of obscurity that is common among those who write as if the reader is not changed by what they read i.e. as if the reader can intervene on their own behalf and choose not to be changed by their exposure to such writings.
Bluntly, the writer controls the reader. For those who reject the accuracy of this description, I can only hope that they will accept the following: the reader has far less control over the writer than the writer does the reader.
Russell writes as if he has not already reached his destination. As if he is uncovering new discoveries upon the page. That is, as if he is discovering what he has to say. But he is not. He has already told us this multiple times. The writer who plays at impartiality or objectivity is a surrepticious controller. They are also a controller who does not deal well with control.
The difference between what Russell writes in PM1 and what history preceeded the printing on the page is that between the notes written here and the memos which have yet to appear here. These notes record occurrences which purportedly lack the purpose of ‘expressing ideas’ or ‘communicating thoughts’. They are, from a literary point of view, a mess. To the surrepticiously controlling writer, these notes are noise. Somewhere within these notes is a signal, a meaning, or a point. While I admit that I may, or may have already, written something like ‘my point is’ I can, at least here, make it clear that such sentences are inaccurate.
I can also now submit the past few points as evidence the relevance of my earlier warning that I can not promise more diatribes shall not come.
Onto a clearing up of ‘sets having or not having a total’.
The notation used here in my interpretation is mostly that of Quine’s “Methods of Logic 4th Edition”. The fourth edition is profoundly different from the prior editions, and it must be carefully studied by those who might otherwise reject it as idiosyncratic.
Talk of sets obscures the verbal practices which are at work when speaking of classification. Classifying classes supplants talk of ‘belongingness’ or ‘membership’ or ‘this set is in that set’. The key is classification. This is where the constraint on extensionality comes from. A classification which is beyond intersubjective agreement between qualified witness is bankrupt. If two scientists on different sides of the world can’t agree on their classifications then their correspondance is corrupted beyond repair. A new classification must be established.
An extentionally classificatory predicate such as that of Zermelo and Frankel’s set theory is typically what mathematicians who analyze their activity into the shuffling of sets hither and thither (I say it that way so it sounds as mystical as it seems to me) is oddly mixed with an irrelevant theory of identity. This continues to cause great confusion on the role of identity in logic as opposed to its role in theories of identity (what Bourbaki calls ‘equalitarian theories’, or what is analyzed by ‘univalence’ by homotopy type theorists).
A brief side note on the differences between classification and categorification. In English, and not mathematical English, a category is “a specifically defined division in a system of classification: a class” and reaches back through Aristotle’s Organon. A world with different kinds of things may have among its things kinds. Kinds, sorts, traits, essences, natures, and all such stuff looms. That’s all I have to say about that for now.
As a reminder for those who are not familiar with schematic methods in logic, the ‘F’ in ‘F extensionally classifies’ is a predicate letter, not a predicate. The predicates ‘is a member of’ or ‘belongs to’ or ‘is classed as’ are not to be confused with the predicate letter ‘F’. Each of those prediates purports to, when substituted for the predicate letter ‘F’ in ‘F extensionally classifies’, be a part of a truth e.g. each item is (x, y, and z such that each item is (w such that w is classed as x if and only if w is classed as y) and x is classed as z, only if y is classed as z, i.e. is classed as extensionally classifies. I know that ‘is classed as extensionally calssifies’ is not proper English and would, if spoken to an English speaker, would turn heads. Yet, it is precisely on this strange matter that Russell’s summaries turn.
An instance of that schema is one where ‘G’ is replaced by ‘u such that not Fuu’ i.e. ‘some item is (x such that x is F total and each item is (y such that (u such that not Fuu)y if and only if Fyx)’.
‘not Fuu if and only if Fuu’ is a truth functional inconsistency, therefore ‘some item is (x such that x is F total and each item is (y such that not Fyy, if and only if Fyx)’ is an inconsistency.
So, ‘u is F total’ and the validity imply that ‘some item is (y such that not Fyy, or exclusively Fyu)’.
Said another way ‘some item is (x such that x is F total)’ implies ‘some item is (y such that not Fyy, or exclusively Fyx)’. So that if there is any ‘F’ totality (call it a set if you want to) then there is something, call it ‘u’, such that not …
I’m a closeted storyteller. When I was young I used to tell everyone stories. Sometimes my younger sister would ask me to tell her a bedtime story. We’d get under the covers with a flashlight and I’d make up a story about a princess who could fly and talk to dragons. Somehow I made it so that everyone was part of the story. If my sister wanted the princess to explore a cloud city then that became part of the dream world.
Not many of those early stories stuck with me. Everything gets forgotten eventually.
As I grew up, fewer people listened to my stories. They told me that they didn’t have time to listen. That I was being weird for weaving fantasy and fiction into the real problems of the world. The real world was full of chores, homework, and the stuff that we all have to do. It’s that last one. The stuff that we have to do that always got me into trouble. I didn’t see what it was that we had to do. What we were doing was all that we had to do. This moment was the one to live in. The next one hasn’t happened. The old ones have made their contribution and are gone. We’re here now and we’re living in a story world. At least I was.
It’s really hard for me to write about losing touch with my storyteller. Letting go of the world that is right in front of me is now something I struggle to do. It’s strange to look back and see a boy that ran with the wind and smiled even when he fell down.
That storyteller never died. He just got gobbled up by a world that he didn’t understand. He grew up. He got a life where a new kind of story could be told, at least for a while. Science, math, logic, and all those great things are just stories. They’re the best stories as far as I’m concerned. Logic is the best story I’ve ever heard. It’s one that I keep telling myself over and over again. There’s always something new to it. But, there has always been something missing from it. Well, it’s not logic that was missing something. It was me.
When people repeatedly tell you that you’re telling stories at the wrong time or in the wrong way then you start to think that it’s your stories that are wrong. Eventually you even start to think that the storyteller within you is wrong. Maybe even you are the one who is wrong. Maybe the people are right. No. That can’t be right.
The people around you are the ones who are wrong. Well, they’re not wrong, so much as they are unprepared for the strengths of a scintillating storyteller. My powers of distraction are still profound to this day. That’s all thanks to my fabled roots.
I still have doubts though.
I had let the world tell me my stories were wrong without thinking it over for myself first. It’s not as if I knew how to think for myself when I was young anyway.
By the time I was taught to think for myself I had lost something of the storytelling spark. If there is one thing I could tell my younger self it’s that it’s never too late to be what you could have been.
Actually, you don’t ever really lose your sparks, you just lose sight of them. They’re off in the distance or tucked away somewhere you haven’t been for a long while. Were they ever there to begin with? None of this stops you from being and becoming who you could have been. It’s magical. At least, it is if you let it be.
It feels embarassing for me to write like this, but I know that this is how I have to write. I don’t meant that this is the way to write. That’s a different problem entirely. E. B. White and the Olgilvey folks can tell you how to do that. Anyone can read them and work on writing well. What I’m talking about is how to write like I write. I mean like the way I write when I’m not worried about who might be looking over my shoulder, or who might say my stories are weird, or that now is not the time for storytelling because there are so many much more important things to do than “ruin the signal to noise ratio”.
Let me go back to those earlier days. Those are the days where the world crept in before I knew to fight it off with with whatever I had on hand. All my teachers knew that I was a talker. If you were one of my teachers then I’m so sorry. My mouth never shut. I had to talk about what we were working on. I had to ask questions “So is it like this? blah blah blah…” as if I was the only other person in the room. I was the only one who was seeing this new part of the world for the first time. No one else saw it how I saw it. What an ego.
Even now I see how silly I was. It makes me smile.
The principle of the school knew me by name. I was outside their office almost every day, sometimes multiple times a day. I wasn’t even there to ‘see the principle’. My poor teachers sent me away so they could have a moment of silence. There was always a book in my hand. I’d sit with my back against the wall to the principles office, right next to the door, with my little legs out blocking the hallway. Even when I was supposed to be out of the way I couldn’t help but get in the way. That’s how it goes for storytellers. I know that now. I don’t know how I could have known that then.
It’s confusing to me that the people of the world haven’t figured out how to deal well with each other. Given how old the human species is, and how many children were born and raised throughout history, I’d expect at least some of the basics to be firmly figured out. It’s as if we only just started having grandparents and grandchildren. As if we’re figuring it all out for the first time. As if all the other times are different. Or as if this time it’s different. It’s outrageous. I’m not crazy for expecting that humanity should have made some progress since we slipped and felled our nearest ancestors into extinction. Here’s to the next few millennia I guess.
But, forget all that. Let’s get back to me, me, and me. That’s all a storyteller cares about. Are you, the listener, attending to me? Are you listening? Are you with me? Are you laughing, smiling, frowning, crying or cackling? Are you responding in any way at all to me? If not, why not? Why not!
This continues my explanation of my programming environment from #2025-0923-2200.
function Pair(l,r){return [l,r];}var Nil = new Object();function isNil(p){return p==Nil;}x ? y : z. When it’s executed, if x is true it returns y, otherwise it returns z. It’s helpful because you can put it places where if(x) y; else z; doesn’t work. Find more details elsewhere.function Right(p){return isNil(p) ? Nil : p[1];} function P(l,r){return [l,r];}
function L(p){return Left(p);}
function R(p){return Right(p);}
function LL(p){return L(L(p));}
function LR(p){return L(R(p));}
function RL(p){return R(L(p));}
function RR(p){return R(R(p));}
function LLL(p){return L(LL(p));}
function LRL(p){return L(RL(p));}
function RLL(p){return R(LL(p));}
function RRL(p){return R(RL(p));}
function LLLL(p){return L(LLL(p));}
function RLLL(p){return R(LLL(p));}
var T=Nil;function drop(){T=P(LL(T),R(T));}function dup(){T=P(P(L(T),RL(T)),R(T));}function pop(){T=P(LL(T),P(RL(T),R(T)));}function push(){T=P(P(L(T),LR(T)),RR(T));}function swap(){T=P(P(P(LLL(T),RL(T)),RLL(T)),R(T));}function nil(){T=P(P(L(T),Nil),R(T));}function pair(){T=P(P(LLL(T),P(RLL(T),RL(T))),R(T));}function part(){T=P(P(P(LL(T),LRL(T)),RRL(T)),R(T));}function mux(){T=P(P(LLLL(T), RL(T)==Nil ? RLL(T) : RLLL(T)), R(T));} function Pair(l,r){return [l,r];}
var Nil = new Object();
function isNil(p){return p==Nil;}
function Left(p){return isNil(p) ? Nil : p[0];}
function Right(p){return isNil(p) ? Nil : p[1];}
function P(l,r){return Pair(l,r);}
function L(p){return Left(p);}
function R(p){return Right(p);}
function LL(p){return L(L(p));}
function LR(p){return L(R(p));}
function RL(p){return R(L(p));}
function RR(p){return R(R(p));}
function LLL(p){return L(LL(p));}
function LRL(p){return L(RL(p));}
function RLL(p){return R(LL(p));}
function RRL(p){return R(RL(p));}
function LLLL(p){return L(LLL(p));}
function RLLL(p){return R(LLL(p));}
var T=Nil;
function drop(){T=P(LL(T),R(T));}
function dup(){T=P(P(L(T),RL(T)),R(T));}
function pop(){T=P(LL(T),P(RL(T),R(T)));}
function push(){T=P(P(L(T),LR(T)),RR(T));}
function swap(){T=P(P(P(LLL(T),RL(T)),RLL(T)),R(T));}
function nil(){T=P(P(L(T),Nil),R(T));}
function pair(){T=P(P(LLL(T),P(RLL(T),RL(T))),R(T));}
function part(){T=P(P(P(LL(T),LRL(T)),RRL(T)),R(T));}
function mux(){T=P(P(LLLL(T), RL(T)==Nil ? RLL(T) : RLLL(T)), R(T));}
The rest of the operations are defined from these basic ones.
This is not entirely true. At the end, when we must leave our pure world of pairs, we must find a way to link back up with reality. Such operations as deal with input and output are not too complicated e.g. some versions of Charles Moore’s Forth work with just two such functions called ‘key?’ (which returns true if there is a key press from the programmer waiting to be dealt with and if there is it puts the name of the key pressed somewhere convenient, otherwise it returns false) and ‘emit’ (which sends a character to the host system). Another minimal input and output method is that of Paul Grahams bel https://www.paulgraham.com/bel.html with one operation that sends a bit and another that gets a bit if there’s one waiting at the door.
Some of first defined operations are helpful recombic ones: nip and over.
The op called ‘nip’ drops the second from top fruit from the tree.
function nip(){pop(); drop(); push();}
The op called ‘over’ duplicates the second from top of the pile and puts it ontop of the pile.
function over(){pop(); dup(); push(); swap();}
Next is an op named ‘dig’: it moves the second from top of the pile to the beginning of the list.
function dig(){over(); pop(); nip();}
Next, ‘bury’: it buries the top of the pile under the two fruit below it.
function bury(){dig(); dig(); push(); push();}
Next, ‘unbury’: it undoes ‘bury’.
function unbury(){bury(); bury();}
Who knew that two buries makes an unbury? What this means is that if you bury three times in a row then you get right back to where you started. There’s a name for that.
Next ‘nop’: short for ‘no op’ as in ‘an operation that appears to do nothing at all’. It’s not really used except when you want to mention that you’re wasting time doing nothing.
function nop(){bury(); unbury();}
function nip(){pop(); drop(); push();}
function over(){pop(); dup(); push(); swap();}
function dig(){over(); pop(); nip();}
function bury(){dig(); dig(); push(); push();}
function unbury(){bury(); bury();}
function nop(){bury(); unbury();}
Next are the defined nonrecombic operations e.g. left, right, enlist, and enpile.
Just like ‘pair’, the op named ‘left’ does what Left does but to the top of the tree i.e. it replaces the top of the tree with its left part.
function left(){part(); drop();}
The op called ‘right’ does the similar.
function right(){part(); nip();}
Since everything in the tree is made from Nil, then if you go down the right parts of the tree you’ll eventually hit Nil. This means that the end of the list is Nil. Similarly, if you go down the left parts of the tree you’ll eventually hit Nil, and this means that Nil is the end of the pile.
From this happy accident, we make a convention: lists end down their right parts with Nil and piles (aka stacks) end down their left parts with Nil.
The definitions of enlist and enpile follow this convention and put the top into a new list or a new pile.
function enlist(){nil(); pair();}
function enpile(){nil(); swap(); pair();}
The sequence of two operations, ‘swap(); pair();’, at the end of the definition of enpile is sometimes given a special name: cons. This comes down from the programming language LISP.
It might be better to put the definitions of enlist and enpile later because they are really only interesting when you start prepending and appending things, which is not so easy when all you have is mux.
Next is the all important truth-value operations. They are all derived from ‘nor’. Nil is truth i.e. the truth-value of true. Everything other than Nil is falsehood i.e. the truth-value of false. So, the op named ‘nor’ is Nil if the top nor the second from top is Nil.
function nor(){dup(); mux();}
function nor(){dup(); mux();} // joint denial
function not(){dup(); nor();} // negation
function or(){nor(); not();} // alternation
function IF(){not(); or();} // converse conditional
function nif(){IF(); not();} // complementary converse conditional
function nfi(){swap(); nif();} // complementary conditional
function fi(){nfi(); not();} // conditional
function nand(){not(); fi();} // alternative denial
function and(){nand(); not();} // alternation
function xor(){over(); over(); nand(); bury(); or(); and();} // exclusive alternation
function iff(){xor(); not();} // biconditional
Technically, those operations are better described as some of the bit operations, where Nil is treated as the bit named ‘1’ and everything else is treated as the bit named ‘0’. Even better, they are some of the remainders of polynomials divided by two. For more on that read R. L. Goodstein’s 1962 “Fundamental Concepts of Mathematics”.
That’s it for the simplest of the defined operations. I could define the rest of the bit operations, e.g. half adder, but that isn’t really important right now.
I have reached the point where I must either double down on sticking with mux or allowing for another primitive. What I’ll do is go ahead with the shim and then go through why it seems natural at first but breaks all the rules.
I just spent about twenty minutes thinking this all over and it occurred to me that mux was not what I planned it to be. That’s not a problem though because an operation like mux helps anyway.
Now that all of the basic operations and the basic definitions have been assembled the problem to solve is this: how do we make these operations run themselves?
So far, the operation of defining operations has been offloaded onto the native environment i.e. javascript. Every time we want to define a new operation from old ones we have to write function followed by the defined operations name and then a list of the basic operations it is made from.
There’s another even more important operation that we’ve offloaded onto the native environment: the operation of choosing what operation to execute next!
I thought I defined mux so that it could let us define these operations, but I was mistaken. In my attempt to simplify everything, I over simplified. No matter. An appropriate operation is likely to occur to me and if it doesn’t then there is always the slightly obscure way that already works. I’m pretty sure the obscure way can be replaced by a more accessible way.
In order for the tree to run itself it has to be able to look at part of itself and decide what to do next based on what it sees. This requires, at least, that one of the basic operations of the tree is that it does one thing when it sees one thing and does another thing when it sees another thing.
Mux does one thing when it sees one thing and another thing when it sees another thing, but neither of the things it does leads to doing a next thing.
So in order for the tree to run itself it has to do things based on what it sees and then plan on doing something after that. I find a nice way of making this clear and showing how it can be done with basic operations.
function Pair(l,r){return [l,r];}
var Nil = new Object();
function isNil(p){return p==Nil;}
function Left(p){return isNil(p) ? Nil : p[0];}
function Right(p){return isNil(p) ? Nil : p[1];}
function P(l,r){return Pair(l,r);}
function L(p){return Left(p);}
function R(p){return Right(p);}
function LL(p){return L(L(p));}
function LR(p){return L(R(p));}
function RL(p){return R(L(p));}
function RR(p){return R(R(p));}
function LLL(p){return L(LL(p));}
function LRL(p){return L(RL(p));}
function RLL(p){return R(LL(p));}
function RRL(p){return R(RL(p));}
function LLLL(p){return L(LLL(p));}
function RLLL(p){return R(LLL(p));}
var T=Nil;
function drop(){T=P(LL(T),R(T));}
function dup(){T=P(P(L(T),RL(T)),R(T));}
function pop(){T=P(LL(T),P(RL(T),R(T)));}
function push(){T=P(P(L(T),LR(T)),RR(T));}
function swap(){T=P(P(P(LLL(T),RL(T)),RLL(T)),R(T));}
function nil(){T=P(P(L(T),Nil),R(T));}
function pair(){T=P(P(LLL(T),P(RLL(T),RL(T))),R(T));}
function part(){T=P(P(P(LL(T),LRL(T)),RRL(T)),R(T));}
function mux(){T=P(P(LLLL(T), RL(T)==Nil ? RLL(T) : RLLL(T)), R(T));}
function nip(){pop(); drop(); push();}
function over(){pop(); dup(); push(); swap();}
function dig(){over(); pop(); nip();}
function bury(){dig(); dig(); push(); push();}
function unbury(){bury(); bury();}
function nop(){bury(); unbury();}
function left(){part(); drop();}
function right(){part(); nip();}
function nor(){dup(); mux();} // joint denial
function not(){dup(); nor();} // negation
function or(){nor(); not();} // alternation
function IF(){not(); or();} // converse conditional
function nif(){IF(); not();} // complementary converse conditional
function nfi(){swap(); nif();} // complementary conditional
function fi(){nfi(); not();} // conditional
function nand(){not(); fi();} // alternative denial
function and(){nand(); not();} // alternation
function xor(){over(); over(); nand(); bury(); or(); and();} // exclusive alternation
function iff(){xor(); not();} // biconditional
A new attempt to write on the logic of my programming environment.
The basic items are ordered pairs. In my current implementation I tame an array into a few basic functions for allocating fresh ordered pairs and collecting any of the garbage ones every so often.
The way I do that is complicated and not important in the beginning. It’s easier to build everything ontop of a programming language that takes care of its own garbage.
There are many programming langauges. I’m picking javascript because you can play with it in any modern browser that lets you mess with their console.
var Nil;
function Pair(l,r){return [l,r];}
function Left(p){if(p==Nil)return Nil; else return p[0];}
function Right(p){if(p==Nil)return Nil; else return p[1];}
function P(l,r){return Pair(l,r);}
function L(p){return Left(p);}
function R(p){return Right(p);}
function LL(p){return L(L(p));}
function LR(p){return L(R(p));}
function RL(p){return R(L(p));}
function RR(p){return R(R(p));}
function LLL(p){return L(LL(p));}
function LRL(p){return L(RL(p));}
function RLL(p){return R(LL(p));}
function RRL(p){reutrn R(RL(p));}
Actually, everything introduced thus far is really all just helper functions for defining the tiny set of basic functions from which everything is actually defined. I have been able to get the basic functions down to only three in number, but they are very hard to understand without first getting familiar with a larger number of basic functions. No two of the larger number of basic functions can be defined in terms of the others (the same is true of the smaller number).
var T=Nil;
The Left part of the tree is called ‘the stack’ or ‘the pile’, and the Right part of the tree is called ‘the list’. The Right part of the pile is called ‘the top’. So executing ‘L(T)’ returns the pile and executing ‘R(T)’ returns the list. Executing ‘RL(T)’ returns the top and executing ‘LR(T)’ returns the item that begins the list.
function drop(){T=P(LL(T),R(T));}
function push(){T=P(P(L(T),LR(T)),RR(T));}
function pop(){T=P(LL(T),P(RL(T),R(T)));}
function dup(){T=P(P(L(T),RL(T)),R(T));}
function swap(){T=P(P(P(LLL(T),RL(T)),RLL(T)),R(T));}
Any number of copies of any item in the tree can be put anywhere in the tree with just drop, push, pop, dup, and swap. Together they are called a recombically complete set of operations because they can recombine the tree into any other tree that has exactly the same items as it does.
The smallest recombically complete set of operations I’ve uncovered has only three ops. If you find a smaller set then tell someone about it (hopefully it gets back to me one way or another). It would be a really big deal if you can find fewer. I’ll reveal my smallest set later on. As a hint, two of them are drop and push, and the third is almost like the operation named ‘over’ that is introduced next.
The next two operations are called ‘pair’ and ‘part’, note that ‘pair’ has a lowercase ‘p’. The new op named ‘pair’ does what Pair does, but now only with the top two items on the tree. There are two ways that it could be defined. The first way pairs the top with the second from the top, and the other pairs the second from top with the top. They are the same except for a swap.
function pair(){T=P(P(LLL(T),P(RLL(T),RL(T))),R(T));}
function part(){T=P(P(P(LL(T),LRL(T)),RRL(T)),R(T));}
function nil(){T=P(P(L(T),Nil),R(T));}
function mux(){T=P(P(L(LLL(T)), RL(T)==Nil ? RLL(T) : R(LLL(T))), R(T));}
The definition of mux looks a lot worse than it is. It does showcase the number of parentheses that are saved by the little abbreviations made earlier.
Together pair, part, nil, and mux are the nonrecombic operations. Everything that they do can be done by just one basic op. That single op can be absorbed into any recombically complete collection of ops. The trick is to just do everything that pair, part, nil, and mux do and then sort through it all to get at only the things you want using the recombically complete set.
function P(l,r){return Pair(l,r);}
function L(p){return Left(p);}
function R(p){return Right(p);}
function LL(p){return L(L(p));}
function LR(p){return L(R(p));}
function RL(p){return R(L(p));}
function RR(p){return R(R(p));}
function LLL(p){return L(LL(p));}
function LRL(p){return L(RL(p));}
function RLL(p){return R(LL(p));}
function RRL(p){reutrn R(RL(p));}
function drop(){T=P(LL(T),R(T));}
function push(){T=P(P(L(T),LR(T)),RR(T));}
function pop(){T=P(LL(T),P(RL(T),R(T)));}
function dup(){T=P(P(L(T),RL(T)),R(T));}
function swap(){T=P(P(P(LLL(T),RL(T)),RLL(T)),R(T));}
function pair(){T=P(P(LLL(T),P(RLL(T),RL(T))),R(T));}
function part(){T=P(P(P(LL(T),LRL(T)),RRL(T)),R(T));}
function nil(){T=P(P(L(T),Nil),R(T));}
function mux(){T=P(P(L(LLL(T)), RL(T)==Nil ? RLL(T) : R(LLL(T))), R(T));}
As strange as these basic operations may seem, they are actually designed from the basic building blocks of logic. Specifically, they are designed from predicate functor logic which is the same as truth-functional and quantificational logic but without any variables. This is significantly different from, e.g., Moses Schönfinkel and Haskel Curry’s combinators (which are sadly coupled with the unhappy phrase ‘combinatory logic’ which is actually not logic but a theory of function application). Elsewhere in these notes I’ve developed predicate functor logic from multiple starting points, and shall not go into it any further here.
They are part of my work on going through programs I wrote and subjecting them to a logical analysis with an aim to porting them to a predicate functor logic programming language based on Quine’s main method.
All the way back in #2025-0707-1949, I started to explain what I there called ‘my stack based programming language’. It’s a bit strange to call it ‘stack based’ when the basic items are ordered pairs. While the raw material are ordered pairs, the methods are like those of stack based langauges descendant from Charles Moore’s Forth.
The ordered pairs are dealt with uniformly by way of K. P. Lee’s 1980 ‘A Linear Algorithm for Copying Binary Trees Using Bounded Workspace’. I mentioned it in note already mentioned, and again when it came up among a list of papers related to garbage collection in LISP. That note is #2025-0904-1549.
Lee’s algorithm is delightful and appears to be almost as perfect as one can hope when it comes to copying binary trees.
It turns into a garbage collection algorithm in my programming language because everything there part of a single binary tree that need only be copied every once and a while. If the orderd pairs are stored in an array then you can only use half of the array at any one time, though there are some access points to better methods which are not of interest to me now.
This continues notes from #2025-0921-1542 on the first volume of Whitehead and Russell’s ‘Principia Mathematica’.
The last note was big and deep. My aim here is to summarize the conclusions.
“Thus classifcatory propositions, or such as assign predicates, would seem to be really propositoins asserting similarity. If so, the simplest propositions are relational.” pg. 40
“Everything has the arrangement it has, and is incapable of re-arrangement.” pg. 43
“When we observe that the order of words can be changed, and that we can say ‘Caesar killed Brutus’ just as easily as ‘Brutus killed Caesar,’ we are apt to think that the words are definite things which are capable of different arrangements. This is a mistake: the words are abstractions, and the verbal utterances can only have whichever order they do have.” pg. 43
“In a language like English, in which the order of the words is essential to the meaning of the sentence, we can put the matter of non-symmetrical relatoins as follows: Given a set of words which is capable of forming a sentence, it often happens that it is capable of forming two or more sentences of which one is true while the others are false, these sentences differing as tot he order of the words. Thus the meaning of a sentence, at any rate in some cases, is determined by the series of words, not by the class. IN such cases, the meaning of the wentence is not obtainable as an aggregate of the meanings of the several words. When a person knows who Brutus was, who Caesar was, and what killing is, he still does not know who killed whome when he hears the sentence ‘Brutus killed Caesar’; to know this he requires syntax as well as vocabulary, since the form of the sentence as a whole contributes to the meaning.” pg. 45
“The importance of atomic forms and their contradictories is that– as we shall see– all propositions, or at least all non-psychological propositions, justified by observation without inference, are of these forms. That is to say, if due care is taken, all teh sentences which embody empiral physical data will assert or deny propositions of atomic form. All other physical sentences can theoretically be either proved or disproved (as the case may be), or rendered probably or improbable, by sentences of these forms; and we ought not to include as a datum anything capable of logical proof or disproof by means of other data. But this is merely by way of anticipation.
In a sentence of atomic form, expressed in a strictly logical language, there are a finite number of proper names (any finite number from one upwards), and there is one word which is not a proper name. .. We can distinguish proper names from other words by the fact that a proper name can occur in every form of atomic sentence, whereas a word which is not a proper name can only occur in an atomic sentence which has the appropriate number of proper names. .. Sometimes for the sake of uniformity, predicates are called monadic relation words.” pg. 53-54
“I come now to the parts of speech, other than conjunctions, that cannot occur in atomic forms. Such are ‘a’, ‘the’, ‘all’, ‘some’, ‘many’, ‘none’. To these, I think ‘not’ should be added; but this is analogous to conjunctions.” pg. 54
“Propositions containing ‘all’ or ‘none’ can be disproved by empirical data, but not proved except in logic and mathematics.” pg. 55
“Until we have examined everything, we cannot be sure but that something unexamined is human but immortal. Since we cannot examine everything, we cannot know general propositions empirically.” pg. 55
“No proposition containing the (in the singular) can be strictly proved by empirical evidence. .. To prove that Scott was the author, we should have to survey the universe and find that everything in it either did not write ‘Waverly’ or was Scott. This is beyond our powers.” pg. 55
“Empirical evidence can prove propositions containing ‘a’ or ‘some’, and can disprove propositions containing ‘the’, ‘all’, or ‘none’. It cannot prove propositions containing ‘a’ or ‘some’, and cannot prove propositions containing ‘the’, ‘all’, or ‘none’. If emprical evidence is to lead us to disbelieve propositions about ‘some’ or to believe propositions about ‘all’, it must be in virtue of some principle of inference other than strict deduction– unless, indeed, there should be propositions containing the word ‘all’ among our basic propositions.” pg. 55-56
This continues notes from #2025-0917-1724 on the first volume of Whitehead and Russell’s ‘Principia Mathematica’.
Though what follows may still appear to be far from the contents of volume one of Russell and Whitehead’s “Principia Mathematica”, Russell’s “Principles of Mathematics” and “An Inquiry into Meaning and Truth”, I can assure the wary reader that these are solid steps in the direction of all these works and all that are related to them in their logical and scientific interests.
Most importantly, to the notes here at least, the theories of grammar and denotation which follow, and the premises presented of them, are my only way of firmly grasping what Russell has to say about the hierarchy of languages that not only compose his method of types, but also the propositional methods which are so characteristic of his phenominalism e.g. where he takes sentences as instances of propositions and sentential utterances as instances of sentences, I shall take instances of verbal responses and verbal schema (which are purportedly a special kind of autoclitic).
Such abbreviations have to be added to the growing list of waypoints through predicate functor logic.
Just two lists of premises of logical theories of the grammar of predicate functor logics are needed: the constructive and the lexical. The constructive premises are the same for all theories of the grammar of predicate functor logics. The lexical premises are different for each such theory depending on the lexical predicates of each such theory.
Following Russell and Quine, rather than Carnap, the lexical premises of each such theory share at least lexical predicates (what I shall call ‘the singularly designative one place predicates of lexicality’ or ‘the predicates of lexical designation’ or ‘the lexical designators’) which are part of each such theory i.e. ‘designates-grammatical’, ‘designates-drops’, ‘designates-hems’, and ‘designates-huhs’.
This is an important point which Russell urged in his earlier and later writings. It is also a point which is easier to appreciate after the relevant premises have been given as I shall now.
The constructional premises establish that each of the predicates ‘drops’, ‘hems’, and ‘huhs’ permit unambiguous methods of grammatical analysis and synthesis i.e. the grammatical methods of grade school students. The fields of ‘drops’, ‘hems’, and ‘huhs’ are that of ‘grammatical’ i.e. constructed compounds and their components are grammatical. Each constructed compound is unique in that no distinct constructions have corresponding components that are indistinct. This is the synthetic part of grammar e.g. from two grammatical sentences we compose their conjunction and it inherets its grammaticality from them through the construction.
The constructional premises may lack the features required to guarentee a unique grammatical analysis.
Oh! This may be more interesting than I thought. With just the constructive premises we may define ‘lexical’ as ‘x such that x grammatical and not some (y and z such that x drops y, x hems y, or x huhs y with z)’
Originally, I had collected in this point a number of quotes from the chapter ‘The Object-Language’ of Russell’s ‘An Inquiry into Meaning and Truth’ (IMT) that I thought were about the predicate ‘lexical’. After assembling the quotes and chasing down a reference to ‘atomic forms’ in the chpater ‘Sentences, Syntax, and Parts of Speech’ it occurred to me that the chapter ‘The Object-Language’ is actually about the denotative premises that are to follow, along with the quotes that once went here, and that ‘Sentences, Syntax, and Parts of Speech’ is more about the grammatical premises that have been given.
“Sentences can be interrogative, optative, exclamatory, or imperative; they may also be indicative. Throughout most of the remainder of our discussions, we may confine ourselves to indicative sentences, since these alone are true or false. In addition to be true or false, indicative sentences have two other properties which are of interest to us, and which they share with other sentences. The first of these is that they are composed of words, and have a meaning derivative from that of the words that they contain; the second is that they have a certain kind of unity, in vritue of which they are capable of properties not possessed by their constitutent words.” pg. 34
“A single sentence, for our purposes, must be one which says something that cannot be said in two separate simpler sentences.” pg. 35
“Consider next such a sentence as ‘I should be sorry if you fell ill.’ This cannot be divided into ‘I shall be sorry’ and ‘you will fall ill’; it has the kind of unity that we are demanding of a sentence. But it has a complexity which some sentences do not have; neglecting tense, it states a relation between ‘I am sorry’ and ‘you are ill’. We may interpret it as asserting that at any time when the second of these sentences is true, the first is also true. Such sentences may be called ‘molecular’ in relation to their constituent sentences, which, in the same relation may be called ‘atomic’.” pg. 35
“Whether any sentences are ‘atomic’ in a nonrelative sense, may, for the present, be left an open question; but whenever we find a sentence to be molecular, we shall do well, while we are considering what makes a unity of sentences, to transfer our attention, in the first place, to its atoms. Roughly, an atomic sentence is one containing only one verb; but this would only be accurate in a strictly logical language.” pg. 35
“When we say “I said ‘A’” what we really mean is “I made a noise closely similar to the noise I am now about to make: ‘A’” This, however, is a digression.” pg. 37
“We will rever to the supposition that I say first ‘A’ and then ‘B’. We will call the particular occurrence which was my first utterance ‘Tom’ and that which was my second utterance ‘Harry’. Then we can say ‘Tom preceded Harry.’. This was what we really meant to say when we said “The sound ‘A’ preceded the sound ‘B’”; and now, at last, we seem to have reached an atomic sentence which does not merely classify.” pg. 37
“It might be objected that, when I say “Tom preceded Harry,” this implies “Tom occurred” and “Harry occurred” just as when I said “The sound ‘A’ preceded the sound ‘B’,” that implied “‘A’ occurred” and “‘B’ occurred”. This, I think, would be a logical error. When I say that an unspecified member of a class occurred, my statement is significant provided I know what class is meant; but in the case of a true proper name, the name is meaningless unless it names something, and if it names something, that something must occur. This may seem reminiscent of the ontological argument, but it is really only part of the definition of ‘name’. A proper name names something of which there are not a plurality of instances, and names it by a convention ad hhoc, not by a description composed of words with previously assigned meanings. Unless, therefore, the name names something, it is an empty noise, not a word. And when I say “Tom preceded Harry,” where “Tom” and “Harry” are names of particular noises, we do not presuppose “Tom occurred” and “Harry occurred,” which are both strictly meaningless.” pg. 37
“In practice, proper names are not given to single brief occurrences, because most of them are not sufficiently interesting. When we have occasion to mention them, we do so by means of descriptions such as ‘the death of Caesar’ or ‘the birth of Christ’. To speak for the momen in terms of physics, we given proper names to certain continuous stretches of spacetime, such as Socrates, France, or the moon. In former days, it would have been said that we given a proper name to a substance or collection of substances, but now we have to find a different phrase to express the object of a proper name.” pg. 38
“A proper name, in practice, always embraces many occurrences, but not as a class-name does: the separate occurrences are parts of what the name means, not instance of it. Consider, say, ‘Ceasar died’. ‘Death’ is a generic word for a number of occurrences having certain resemblances to each other, but not necessarily any spatiotemporal interconnection; each of these is a death. ‘Caesar’, on the contrary, stands for a series of occurrences, collectively, not severally. When we say ‘Caesar died,” we say that one of the series of occurrences which was Caesar was a member of the class of deaths; this occurrence is called “Caesar’s death”.” pg. 38
“A proper name may be assigned to any continuous portion of spacetime, or to a quality, as I shall argue in Chapter VI. Two parts of one man’s life may have different names; for instance, Abram and Abraham, or Octavianus and Augustus. “The universe” may be regarded as a proper name for the whole of spacetime. We can give a proper name to very small portions of spacetime, provided they are large enough to be noticed. If I say ‘A’ once at 6 P.M. on a given date, we can given a proper name to this noise, or, to be still more particular, to the auditory sensation that some one person present has in hearing me. But even when we have arrived at this degree of minuteness, we cannot say that we have named something destitute of structure. It may therefore be assumed, at least for the present, that every proper name is the name of a structure, not of something destitute of parts. But this is an empirical fact, not a logical necessity.” pg. 39
“If we are to avoid entanglement in questions that are not linguistic, we must distinguish sentences, not by the complexity which they may happen to have, but by that implied in their form. “Alexander preceded Caesar” is complex owing to the complexity of Alexander and Caesar; but ‘x preceded y’ does not, by its form, imply that x and y are complex. In fact, since alexander died before Caesar was born, every constituent of Alexander preceded every constituent of Caesar. We may thus accept ‘x precedes y’ as an atomic form of proposition, even if we cannot actually mention an x and a y which give an atomic proposition. We shall say, then, that a form of proposition is atomic if the fact that a proposition is of this form does not logically imply that it is a structure composed of subordinate propositions. And we shall add that it is not logically necessary that a proper name should name a structure which has parts.” pg. 39
Recast, ‘lexical’ for ‘x such that x grammatical and not x compound’.
Russell uses the terms ‘atomic’ and ‘molecular’ where I use ‘lexical’ and ‘compound’.
With this definition of lexical it may seem as if the premises of lexicality could be dropped entirely, except perhaps to say that ‘some item is (x such that x lexical)’.
This theory of grammar is anticipated by the more familiar methods of grammar involving calculi which contain construction rules that do not have antecedents i.e. the items of the lexicon are degenerate constructions.
While there is still more to say about the chapter on ‘Sentences, Syntax, and Parts of Speech’ I prefer to move onto the premises of the theory of denotation.
Just as with the theory of grammar, the premises are divded into two groups: the constructional and the lexical.
Before leaving the grammatical premises, I must note that rather than the predicate functor ‘functive’ the predicate functor ‘one-to-one’ may be needed to guarentee a unique gramamtical analysis and synthesis.
I give the premises of denotation in a truth functional and quantificational logic because I have yet to put them in their predicate functor form, and because it is easier for most humans to read at first sight than those who are not already familiar with predicate functors.
Since it was already hinted at repeatedly in the chapter ‘syntax, sentences, and parts of speech’, I can say that the constructive premises are, for Russell, a step up from the lexical premises.
First I shall give the lexical premises which are from the grammatical theory where ‘lexical’ was not first contemplated as being defined as ‘x such that x grammatical and not x complex’ i.e. with the predicates ‘designates-drops’, ‘designates-hems’, ‘designates-huhs’, and ‘designates-grammatical’.
It just occurred to me that ‘grammatical’, like ‘lexical’, can probably be defined if the premises of the simplified theory of grammaticality are altered.
‘one-hot k .. F’ denotes something if and only if that something is denoted by one and only one of the predicates .. and ‘F’.
I went ahead and changed the predicate functor from ‘functive’ to ‘one-to-one’ because I submit that the previous premises did not guarentee unambiguous grammatical analyses and syntheses. I also had to add the one-hot premise for the same reason.
The one-hot premise makes it so that the grammatical constructions each generate uniquely identifiable compounds i.e. a compound is either a drop, a hem, or a huh.
No longer are the fields of the predicates of grammatical constructions ‘drops’, ‘hems’, and ‘huhs’ set to be coextensive with the predicate ‘grammatical’. I have simply required that each component of one construction is also a component of the other constructions, so that, e.g., if some item is (y such that y drops x) then some item is (y such that y hems x) and some item is (y and z such that y huhs x with z).
I have not yet introduced an abbreviation of predicate functor logic which corresponds to the now familiar definition of ‘domain’ and ‘codomain’, ‘range’, ‘image’, or the presumably more general ‘source’ and ‘cosourse’ or ‘target’ (though ‘source’ is technically coextensive with ‘item’). For predicates generally, I speak only of projections.
From these premises, ‘grammatical’ is short for ‘field drops’ or ‘field hems’ or ‘field huhs’. They are each coextensive with the others. And, specifically, ‘grammatical’ is equivalent to ‘proj 1 1 0 drops’, ‘proj 1 1 0 hems’, ‘proj 1 2 0 huhs’, or ‘proj 2 2 0 huhs’.
Now ‘lexical’ is short for ‘and grammatical not compound’.
This shows that, yes, ‘grammatical’ can be define in a theory of grammar with only three predicates ‘drops’, ‘hems’, and ‘huhs’ which correspond to the grammatical constructions in any theory with a predicate functor logic.
Note, the predicate functor abbreviations I’ve used here can be given as abbreviations in a truth functional and quantificational logic as in note #2025-0908-1633.
The problem posed by eliminating the predicates of designation is whether such .. wait.
Whether this analysis solves the problems presented by Russell in the following quotations and how far they go in seperating what he calls the object words from the logical words will have to wait for some other day, perhaps tomorrow even.
“the hierarchy must extend upwards indefinitely, but not downwards, since, if it did, language could never get started. There must therefore, be a language of lowest type. I shall define one such language, not the only possible one (My hierarchy of languages is not identical with Carnap’s or Tarski’s.). I shall call this sometimes the ‘object language’, sometimes the ‘primary language.’ My purpose, in the present chapter, is to define and describe this basic language. the languages which followin the hierarchy I shall call secondary, tertiary, and so on; it is to be understood that each language contains all its predecessors.
The primary language, we shall find, can be defined both logically and psychologically; but before attempting formal definitions it will be well to make a preliminary informal exploration.” pg.76
“we can now partially define the primary or object-language as a language consisting wholly of ‘object-words’ (There must be syntax, but it need not be rendered explict by the use of syntactical words, such as ‘is’.), where ‘object-words’ are defined, logically, as words having meaning in isolation, and psychologically, as words which have been learnt without its being necessary to have previously elarnt any other words. These two definitions are not strictly equivalent, and where they conflict the logical definition is to be preferred. They would become equivalent if we were allowed to suppose an indefinite extension of our perceptive faculties. We could not, in fact, recognize a chiliagon by merely looking at it, but we can easily imagine beings capable of this feat. On teh other hand, ti is clearly impossible that any being’s knowledge of langauge should begin wtih an understanding of ‘or,’ alothough the maning of this word is not learnt from a formal definition. Thus in addition to the class of actual object-words, there is a class of possible object-words. For many purposes the class of actual and possible object-words is more important than the class of actual object-words.” pg. 79-80
“In the learning of an object-word, there are four things to be considered: the understanding of the heard word in the presence of the object, the understanding of it in the absence of the object, the speaking of the word in the presence of the object, and the speaking of it in the absence of the object. Roughly speaking, this is the order in which a child acquires these four capacities.” pg. 80-81
“The meaning of an object-word can only be learnt by hearing it frequently pronounced in the presence of the object. The association between word and object is just like any other habitual association, e.g., that between sight and touch. When the association has been established, the object suggests the word, and the word suggests the object, just as an object seen suggests sensations fo touch, and an object touched in the dark suggests sensations fo sight. Association and habit are not specially connected wtih language; they are characteristics of psychology and physiology generally. ..
As soon as the association between an object-word and what it means has been established, the word is ‘understood’ in the absence of the object, that is to say, it ‘suggests’ the object in exactly the same sense in which sight and touch suggest one another” pg. 81-82
“Generally, when you hear an object-word which you understand, your beahviour is, up to a point, that which teh object itself would have caused. this may occur without any ‘mental’ intermediary, by the ordinary rules of conditioned reflexes, since the word has become associated with the object.” pg. 82-83
“We learn, by direct association with what the word means, not only proper names of the people we know, class-names such as ‘man’ and ‘dog,’ names of sensible qualities such as ‘yellow,’ ‘hard,’ ‘sweet,’ and names of actions such as ‘walk,’ ‘run,’ ‘eat,’ ‘drink,’ but also such words as ‘up’ and ‘down’, ‘in’ and ‘out,’ ‘before’ and ‘after,’ and even ‘quick’ and ‘slow’. But we do not learn in this way either complicated words such as ‘dodecahedron’ or logical words such as ‘not,’ ‘or,’ ‘the,’ ‘all,’ ‘some.’ Logical words, as we have seen, presuppose language; in fact, they presuppose what, in an earlier chapter we spoke of as ‘atomic forms.” pg. 83
“One of the things that have seemed puzzling about language is that, in ordinary speech, sentences are true or false, but single words are neither. in the object-language this distinction does not exist. Every single word of this language is capable of standing alone, and, when it stands alone, means that it is applicable to the present datum of perception.” 92
“Let us now re-state the whole matter of the object-language.
An object-word is a clas of similar noises or utterances such that, from habit, they have become associated with a class of mutually similar occurrences frequently experienced at teh saem tiem as one fo the noises or utterances in question. That is to say, let A1, A2, A3, .. be a set of similar occurrences, and let a1, a2, a3, .. be a set of similar noises or utterances; and suppose that when A1 occurred you heard the noise a1, when A2 occurred you heard the noise a2, and so on. After this happened a great many times, you notice an occurrence An which is like A1, A2, A3, .. and it cuases you, by association, to utter or imagine a noise an which is like a1, a2, a3, .. . If, now, A is a class of mutually similar occurrences of which A1, A2, A3, .. An are members, and a is a class of mutually similar noises or utterances of which a1, a2, a3, .. an are members, we may say that a is a word which is the name of the class A, or ‘means’ the class A. This is more or less vague, since there may be several classes which satisfy teh above conditions for A and a. A child learning the object-language applies Mill’s Canons of Induction, and gradually corrects his mistakes.” pg. 92-93
“At the lowest level of speech, the distinction between sentences and single words does not exist. At thsi level, single words are used to indicate teh sensible presence of what they designate. It is throught his form of speech that object-words acquire their meaning, and in this form of speech each word is an assertion. Anything going beyond assertions as to what is sensibly present, and even some assertions which do not do so, can only be effected by means of sentences; but if sentences contain object-words, what they assert depends upon the meaning of the ojbect words. There are sentences containing no object-wrods; they are those of logic and matheamtics. But all empirical statements contain object-words, or dictionary words defined in terms of them. Thus the meaning of object-words is fundamental in the theory of empirical knowledge, since it is thruogh them that language is connected with nonlinguistic occurrences in the way that makes it capable of expressing empirical truth or falsehood.” pg. 32-33
Went to ‘Beautiful: The Carole King Musical’ and loved each one of the songs.
Link to Cormac McCarthy’s tips on writing about science: https://gwern.net/doc/science/2019-savage.pdf
I know that somehwere inside me is a writer. He got covered up a long time ago. When I was young I used to tell stories to my friends. They’d hang on every one of my words. We’d all be on the adventure together. Whether it was on the jungle gym or the bus ride home we had the best of times exploring my imagination. I feared that power.
It came back to me for a bit during college. It was always there. Something like that never really goes away. Even if you get stuck on logic and science like me, there’s always a piece of you that keeps the power of story telling alive.
A poem I wrote for a college credit gave me a glimpse at what it would be like to have stuck with telling stories instead of doing science. It’s only recently that I figured out I can do both. Or, better, I can do it all. Or, best, I can do whatever I set out to do. The limit of what I can do is what I set for myself.
Most of the time I stick to the ‘personal score card’ method of Warren Buffett. My score card has been a bit too personal as of late. It’s probably a good thing to compare and contrast with how others keep score just in case you might find something you like. Not that you have to worry about the scores they’ve got relative to yours.
I know that I can write better.
It took me this long to figure out that I can write the dashes which turn up in the hyperlinking anchors as part of the subsection title (or I could just smoosh all the numbers together, but I find that hard to read at a glance or while I’m scrolling quikckly back and forth in time). I’m contemplating not even making each of these notes a subsection and just leaving them as their own sections under a top most section titled ‘NOTES’ which may or may not contain some short sentences on how notes work, though I’ve already explaiend that in the hint #how-to-do-it.
It also occurred to me that I can put the pound at the front of the entry as well e.g. ‘#2025-0917-1724’ instead of ‘2025-0917-1724’. This would make it even easier to copy and paste hyperlink anchors. All of this has reignited by intense desire to present the page as a javascript snippet that you must manually copy and paste into your browser to effectively use. Ah, they joys of obscurity.
This continues notes from #2025-0915-1331 on the first volume of Whitehead and Russell’s ‘Principia Mathematica’.
In the last note I presented, at the end, the premises of a logical theory of the grammar of logical theories which did not depend upon concatenation, e.g. as in Quine’s theory of protosyntax, nor trees, e.g. as in Feferman’s theory of finitary inductively presented logics.
I was quick to see that uniqueness of existential joint denials had not been guarenteed by this initial list of premisses. It was also apparent that a single predicate in the lexicon could potentially be predicated in more than one way.
The predicates of predication, e.g. ‘w, x, y, z such that w predicates x with y and with z’ and ‘u, v, w, x, y, and z such that u predicates v with w, x, y, and with z’, are an example of a peculiarity of most grammars of logic which is often overlooked because of the general utility of concatenation or tupling: there is not a single gramamtical construction called ‘predication’, rather there are grammatical constructions of predication for each of the many places of each predicate in the lexicon e.g. ‘is red’ is traditionally predicated of just one pronoun (a better name than ‘variable’) and ‘fathered’ is traditionally predicate of just two pronouns (perhaps identically spelled out as when we write ‘each item is (x such that not x fathered x)’).
It’s at this moment that I just jumped ahead to predicate functor logics like those I’ve contemplated lately, rather than go through the depths of shoehorning some theory of sequences and satisfaction into a theory like the one above.
I am glad that I worked out some of the corrections along this direction because it indirectly forced me to finally give a definition of the counting quantifiers in a truth functional and quantificational logic (TQL) which then prompted me to render the appropriate abbreviation for the same in predicate functor logic.
This continues notes from #2025-0913-1705 on Whitehead and Russell’s ‘Principia Mathematica’.
Quine more than anyone else emphasized that validitiy is the synthesis of grammar and denotation. In his “Principles of Philosophy Second Edition” (POL) he gives a preformal description of grammar and truth (truth is degenerated denotation as shown in Quine’s “From Stimulus to Science” (FSS)). What Quine calls “Tarski’s definition of truth” in POL is properly the presentation of a theory of truth. In FSS Quine takes greater care to emphasize that a consistent theory of denotation involves a hierarchy of theories of denotation. Neither in FSS nor in POL does Quine take the same kind of care he gives to the hierarchy of theories of denotation with his theory of grammar.
Russell and Whitehead’s type theory imposes restrictions upon what propositions are to be taken as meaningful. The meaning of meaning is mysterious, and, thankfully, it comes out as something amenable to paraphrase into grammar.
For Russel, and almost anyone else, meaning is far more than anything that can be caught by grammar e.g. synonymy is more than can be read from a grammatical analysis.
What Quine shows in POL is that grammar is as much ‘about the world’ as denotation. This is in harmony with Russell’s repeated reminder in ‘An Inquiry into Meaning and Truth’ (IMT) that language is part of the same world that the nonlinguistic world is i.e. unlike the division of the world into, what some philosophers have mandated as seperate but equal, the mental and the material, the division of the world into linguistic and nonlinguistic is not as seperate or equal as it may seem.
For Russell, experience is immediately self evident (at least in theory) and language is merely an expression of experience (as far as that goes). Hence, he distances himself from the logical positivists that permit their protocol sentences to be the raw material from which, following Mach, the mental and material worlds are logically constructed e.g. as in Carnap’s ‘Aufbau’.
Russell’s propositional methods in ‘Principia Mathematica Volume One’ (PM1) are under a similar control as to those he presents in IMT except perhaps that IMT is a much more careful consideration of all that propositional methods have to give all the sciences and not just mathematics in particular.
My aim here is to give a more detailed explination of the parallels between theories of grammar and theories of denotation which bring Quine’s methods closer to Russell’s in PM and IMT.
Beyond Quine and Russell’s methods are the calculi and automata that set up grammars in modern works in theories of mathematical logic and computation. And yet, these are not so far from Quine’s methods at the end of his ‘Mathematical Logic’ (ML) where a theory of protosyntax provides, a necessarily incomplete, explication of the formal methods from which grammar has benefited.
My methods are the slightest generalization of Quine’s protosyntax with an eye to accomodating the propositionalists who cannot let go of their nonextensional methods. Instead of building concatenation conspicuously into grammar as Quine does with his protosyntax and instead of building trees into grammar as abstract syntax trees do, I prefer to deal with grammar in the same way that Tarski deals with truth, and closer to how Quine describes grammar in POL and Russell describes it, indirectly, in PM (in the course of his explanation of his theory of types) and, similarly, in IMT.
Although grammar is mentioned by Russell, it is so strongly distinguished from meaning that he misses those parts of his analysis of meaning which actually turn into analyses of grammar. Hopefully I can bring out those points more conspicuously.
Thus, I would perhaps call my methods those of a logical theory of logical grammars. The grammars are logical in that they are constructed to accomodate logical practices as opposed to those of English, German, French, and other purportedly less formal languages.
The theory is logical in that its compounds are composed from a list of basic predicates (as integral words or phrases such as ‘fathered’, ‘red’, ‘is next to’), predicated of an appropriate string of varibles, and assembled by the grammatical construction of iterated existential joint denial (where the existential joint denial of grammatical compounds with a variable not occuring in them is the joint denial of them).
The difference between my methods and those of Quine at the end of ML in the chapter on protosyntax is that rather than define the predicates of grammar from the predicates of a covert theory of concatenation I leave them undefined.
In Tarski’s analysis of truth, the predicate being analyzed is ‘is true’. It is found that there can be no consistent theory of a predicate ‘is true’ when the disquotational principle is regimented with respect to the predicate ‘is true’. The principle of diquotation is exemplified by the classic example: ‘snow is white’ is true if and only if snow is white. It is precisely on the matter of diquotation that Russell, and almost all others, have made their mistakes.
The mistake is the result of mistaking the grammar of logical schema with the grammar of logical sentences e.g. predicate letters in a schema of quantificational and truth functional logic are not variables. Beyond mistaking letters of a schema for variables of a sentence, there is a partcularly tricky situation which confused both Russell and Wittgenstein and which is brought up in IMT: logical sentences about logical schema.
To bring up the problem of theories of logical schema is premature, but points in the direction that my work on grammar shall go as I work over PM and IMT.
A logical theory of grammar gives the premises which connect grammatical lexicon, construction, and categories: lexicon are the raw material from which different categories are constructed.
It shall be seen from this particular example of a theory of a grammar of a logical theory that Quine’s arguments that the predicate ‘is grammatical’ does not have a transcendent definition is further justified by the problems posed by, and solved by, specifying this theory.
As with Tarski’s theory of truth, the lexical elements are the trickiest. In the case of Tarski’s analysis, disquotation comes through as a template upon which relevant premises are presented. In the case of my analysis of grammar, there is nothing quite like diquotation available because, and this is important, the subject of the theory is purely linguistic.
Russell rightly makes much noise about the difference in how to treat the linguistic parts of our world and their links with the nonlinguistic parts of the world. What Russell says is a difference between a word, or sentence, and its quotation, is seen by Quine as a difference between a word, or a sentence, and spelling out words, or sentences.
Note that the definition of the counting quantifiers is made with respect to the predicate ‘is identical to’ and that, in standard logical langauges, we can take any theory which does or does not have a predicate of identity as coextensive with or defined by (respectively) indiscernability with respect to that language’s lexical predicates.
I shall begin by giving the premises as they occurred to me and then contemplate them and their consequences with an aim at unearthing any missing premises. My past has primed me to begin with a fragmenetary collection of premises, no matter how weak or poorly considered they may later end up showing themselves as, rather than trying to capture each key premise from the haze of preformal contemplations up to that point.
Already as I wrote them it appeared to me that there were insufficient premises to proved for the unique decomposition of an existential joint denial into its binding pronoun (i.e. variable) and grammatical subcomponents.
It also occurred to me that there is no guarentee that a lexical predicate is not predicable in more than one way e.g. schematically, ‘Fx’ and ‘Fxy’ are permitted by my premises and yet they are not common in logical practice.
This continues notes from #2025-0911-1654 on Whitehead and Russell’s ‘Principia Mathematica’.
Part I of PM1 is on mathematical logic which is to extend its symbolism into the propositions of mathematics.
Propositions are presupposed by Russel as that which the symbolisms are about when the symbolic constructions are sensible.
Point (2) of page 2 distinguishes symbolism from language. This distinction is no longer appropriate (why?).
“aids the intuition in regions too abstract for the imagination readily to present to the mind the true relation between ideas employed.” pg.2
“And thus the mind is finally led to construct trains of reasoning in regions of thought in which the imagination would be entirely unable to sustain itself without symbolic help. ORdinary langauge yields no help. ITs grammatical structure does not represent uniquely the relations between the ideas involved.” pg. 2
“(a) Most mathematical investigation is concerned not with the analysis of the complete process of reasoning, but with the presentation of such and abstract of the proof as is sufficient to convince teh properly instructed mind.” pg. 3
My method, in contrast to Russell’s, is to step back to grammar prior to logic: grammar being the method by which we distinguish logic as a language from languages such as English and German. Grammar does not tell us whether an item is true or false, only whether it is a grammatical compound of a logical language or not.
Grammar is no starting or stopping point e.g. autoclitic behavior can be examined prior to grammatical behavior and grammatical behavior can be constructed from it (thanks to Skinner’s work in ‘Verbal Behavior’).
This is far more significant than it may seem. While Russell in, e.g., ‘An Inquiry into Meaning and Truth’, confronts the psychological origins of knowledge, he leaves them quite far out from his purportedly direct knowledge of the world through his personal experience. It shall come as no surprise that I trace the origins of his particular blend of explication, experience, and experiment to his carelessness with quotation marks. But, more of that when it comes.
The notation of PM is that of Peano’s ‘Formulario Mathematico’ along with additions carefully crafted by Russell and Whitehead from the deferred consequences of thier use.
“In ordinary mathematics, a variable generally stands for an undetermined number or quantity.” pg.4
“In mathematical logic, any symbol whose meaning is not determinate is called a variable, and the various determinations of which its meaning is susceptable are called the values of the variables.” pg. 4
In the traditional view, symbols symbolize or mean something other than whatever it is that a symbol presents itself as e.g. be it an assembly of marks on the page or those strange elements of Jung’s archetypes that haunt images and our purported unconscious experiences with them.
The link between a symbol and what it symbolizes is presented as a kind of puzzle, like that of more modern works on naming objects which emphasize a cosmic blessing of conspicuous contact between an utterance and an object.
Presumably the meaning of a symbol can be different from what the symbol symbolizes: this much is made clear by Russell’s definition of ‘variable’ as ‘any symbol whose meaning is not determinate’.
Variables are purportedly special symbols whose meanings are on the boundary of obscurity. The meanings of a variable can, purportedly in some cases, be determined from further scrutiny (perhaps by perfection of some perceptive apparatus along with some evolution of a reasoning organ), and as far as they can be so determined, such meanings are especially called ‘values of the variables’.
The phrase ‘values of variables’ and related articulations are legion in any realm of philosophy which leverages notation as a tool for thought. Quine is even known more for his principle ‘To be is to be the value of a variable’ than he is for his more profound ‘To be is to be denoted by a one place predicate.’ which follows from his work on predicate functor logic.
The rest of the variables are ushered in by the grammatical construction of accentuation (which presumably involves a grammatical partical called an accent).
Though this grammatical paraphrase of Russell’s mystical definition of ‘variable’ appears weak it is spartan. It is strong enough to carry through the elaborate arguments involved in talk of bondage, freedom, quantification, substitution, etc.
Talk of variables as more than bits of grammar has enhanced the hazy mist of not determinant, or indeterminate, meanings which carry with them a spirit of variability or variation. This plays itself out in PM (and almost everywhere else) as a problem of variables having sometime one meaning, sometimes another, and sometimes, as I have said twice before, a mystical haze or mass of indeterminate meaning that has some structural resemblance to a whole which is somehow distinct from the mereological sum of its propositional parts i.e. an indeterminate whole which is somehow distinct from the mereological sum of its determinate meanings.
When compared to such talk of meanings my grammatical methods are hardly loaded with the heaps of history that we inherit through the etymology of the mentalistic phrases of our native language. Perception, judgement, reason, inference, and other mysterious operations of the thinking mind are profound when compared to the largely incomplete methods of modern sciences of behavior. Yet, there is now, more than ever, an alternative, provided largely by Skinner’s works, to the experiential literature which floods the sciences of philosophy and psychology.
I’m hopeful that further careful reading of Russell’s work and those earlier and later works which pass through it shall reveal to me an effective way of freeing those like Russell and Carnap from their phenomenalism.
How does talk of the meaning of variables come out in a predicate functor world i.e. when variables are properly eliminated from logic as they are Quine’s predicate functor methods what is left of the problems posed by variables either with respect to Russell’s definition or the grammatical definition?
The lexicon of a predicate functor language is populated entirely of predicates as integral words or phrases such as ‘red’ and ‘round’ and ‘atop’. The question as to whether there are intensional or extensional items named by, or, strictly, designated by these lexical elements is a question about the theory of the world in which such a language occurs. The items of the world are purportedly those which are denoted by the one place predicates of the predicate functor language. So, the question as to whether there exist items designated by the predicates of the lexicon of the language is a question about a special part of the theory of the world: the theory of denotation.
Those familiar with PM will notice that something like a hierarchy of langauges is forced upon anyone examining truth or denotation as a result of Tarski’s analysis of truth and that this is anticipated by Russell’s type theory in PM and further emphasized in “An Inquiry into Meaning and Truth” (IMT).
In both PM and IMT there is recourse to experiences and expressions which obscures an almost neutral analysis of the conclusions which link a theory of grammar, a theory of denotation, and a theory of the world.
Presumably, the theory of the world must, in addition to being consistent, contain the theory of its language’s grammar and its langauge’s truth. Something like such a theory is analyzed here as part of the proper paraphrasing of the premises and conclusions of PM.
Turning now to values of variables. Talk of values of variables is difficult for me to pin down precisely. In the sentences of English, relative clauses do the work of binding pronouns, e.g. ‘it such that it is red and it is round’. This is distinguished from the dangling pronouns of an otherwise complete sentence, e.g. ‘It is red and it is round.’
In the latter sentence with the dangling pronouns, the grammarian asks “Who or what is it?” or, better, “What does ‘it’ refer to in the sentence ‘It is red and it is round.’?” If the immediately preceeding sentence is ‘He dropped the block.’ then may answer “‘it’ refers to ‘the block’” or, perhaps, “‘it’ refers to the block”? How about “‘it’ refers back to ‘the block’” or “‘it’ refers to the same thing as ‘the block’”? Even if ‘the block’ refers to no thing?
There is much more to be said about ‘range of values of a variable’ and its connection to ‘significance’.
I shall return to the preface and the introduction and the beginning of Chapter 1 “Preliminary Explanations of Ideas and Notation”, but now wish to leap ahead, still in chapter one, to the primitive propositions that play such a central role in PM and as access points to the peculiarities of the phenomenology promoted by Russell’s ‘ideas and notation’.
Translation of “primitive propositions employed in the calculus of propositions.” pg. 13
The paraphrase reveals some of the predicates found in the language of a logical theory of the grammar and truth (better, denotation, but that shall be left for later) of a logical language.
The predicate ‘x, y, and z such that x conditionalizes y with z’ is a predicate in a logical theory of grammar where the grammatical construction of conditoinalization is either defined, e.g. with respect to negation and alternation or just alternative denial or perhaps just joint denial, or is part of the lexicon. In either case, it can be taken by those who are wedded to propositions as being “about them” even though they play no proper part in my paraphrase.
Properly, Tarski’s analysis of truth yields a heirarchy of theories of grammatical and truth predicates, though, until now, Quine has obscured the purely grammatical elements with his and Tarski’s principle of disquotation.
Though what I have said of my own work is perhaps obscure with respect to what has been quoted from PM, I can assure the intreped reader that each of the technical distinctions I’ve brought up appear in one form or another in PM (some explicitly).
(1) Anything implied by a true premise is true \pp
this is the rule which justifies inference.
(2) \thm \dott p \or p \dot \imp \dot p \pp
i.e. if p or p is true, then p is true.
(3) \thm \dott q \dot \imp \dot p \or q \pp
i.e. if q is true, then p or q is true.
(4) \thm \dott p \or q \dot \imp \dot q \or p \pp
i.e. if p or q is true, then q or p is true.
(5) \thm \dott p \or ( q \or r ) \dot \imp \dot q \or (p \or r) \pp
i.e. if either p is true or “q or r” is true, then either q is true or “p or r” is true.
(6) \thm \dottt q \imp r \dot \imp \dott p \or q \dot \imp \dot p \or r \pp
if q implies r, then “p or q” implies “p or r”.
Note Russell’s use of quotations where I strongly recommend commas, or, if familiar methods of carrying over grouping into punctuation are to be entirely dropped, full parenthetical methods.
Presumably, what Russell here calls ‘primitive propositions’ can be paraphrased as I have done and shown to be conclusions in a consistent theory of truth (or denotation) like that from Tarski’s analysis of Quine’s extension of it to denotation (where truth becomes degenerate denotation).
By such paraphrasing it can perhaps be shown that something like Russell’s hierarchy of langauges (as in the type theory of PM or its further elaboration in “An Inquiry into Meaning and Truth”) fits smoothly into Tarski’s analysis.
This begins my notes on Whitehead and Russell’s ‘Principia Mathematica’.
As much as my early history as a mathematician is involved with Bertrand Russell’s writings, I never expected to be taking notes on his and Whitehead’s ‘Principia Mathematica’. Happily I got to read through it as an undergraduate, but my outlook has changed significantly since then. It is the most recent work on predicate functor logic that prompts me to begin taking notes here on Volume One.
Most of what goes as theories of relations can be carried over to pure logic. Russell thought, perhaps until the end of his life, that what he had worked on in Principia Mathematica was logic. At best it as a logical theory of sets, better, a logical theory of classing classes.
It is much easier for me to know read Russell after having traced the schematic methods of Quine through the symbolic methods of Carnap. Symbolic methods are plauged by careless designation: symbols symbolize the things they perpurt to designate.
Whereas predicate functors are purely grammatical constructions, e.g. the conjunction of a pair of sentences or predicates is marked by a grammatical particle ‘and’ without ‘and’ designating anything whatsoever, the propositional functions of ‘Principia Mathematica’ and the connectives of anything like that found in Carnap’s ‘Introduction to Symbolic Logic and its Applications’ are wrapped up with reference and meaning, e.g. a connective designates its propositional function.
“For what follows we must come squarely to terms with ‘denote’. Since it is often used interchangeably with ‘designate’, and a singular term normally designates one and only one object, readers are apt to think of denotations as relating a predicate likewise to a single object, namely the class of all the things it is true of, or a property shared by them. in my use of ‘denote’, as in John Stuart Mill’s (‘A System of Logic’ 2 volumes, London 1868, Chapter 2 Section 5), a predicate denotes rather each separate thing of which it is true. The class or property is not involved.
For years, to obviate confusion, I avoided ‘denote’ altogether in favor of ‘true of’; but that evasion would be impracticable in these pages, where denotation is becoming the center of action. Unlike Mill, I still withold the word from singular terms; they are well served by ‘designates’.
A word of caution is in order regarding ‘predicate’ too. Some logicians take a predicate as a way of building a sentence aroudn a singular term or, more concretely, as what Peirce called a rheme (volume 2 paragraph 95), a sentence wtih blanks in it, these being distinctively marked in the case of a many-place predicate. This version covers, implicitly, the potential output of predicate abstraction or predicate functors. But a predicate in my sense is always an integral wrod, phrase, or clause, grammatically a noun, adjective, or verb. some are generated from others by grammatical constructions, notably the relative clause or, formally, predicate abstraction, or predicate functors.
Denotation, explained thus far only for one place predicates, is needed also for predicates of two ro more places. Two place predicates such as transitive verbs, may be said to denote ordered pairs, and n place predicates denote sequences of length n” pg. 60-61 “From Stimulus to Science” by Quine
Reading Carnap finally filled in the gaps for me (though I do not agree with his outlook, he is able to explain himself far better than any other logicians following Frege and Russell).
“The theoretical treatment of any domain fo objects consists in setting up sentences concerning teh objecs of the domain (sentences ascribing certain properties and relations to the objects in question), and in establishing rules according to which other sentences can be derived from those given. .. To form sentences concerning the individuals of a given domain there must first of all be available in the language two kinds of signs: 1. names for the individuals of the domain– we call these individual constants; 2. designations for the properties and relations predicated of the individuals– we call these predicates.
.. It is only when we move away from pure logic (i.e. from consideration of the skeleton language to be constructed in what follows) that we speak of the interpretation of the separate individual constants and predicates.
.. ‘P’ might designate teh property Spherical. (I prefer this mode of expression to the more elaborate turn of phrase ‘the property of being spherical’. Similarly, I write ‘the property Prime Number’, ‘the property ODD’, etc. Again, I use ‘the class Spherical’ in place of ‘the class of spherical individuals’; and analogously, ‘the class Blue’, etc. And again, I say ‘the relation Greater’ rather than ‘the relation that obtains between x and y when x is greater than y’; and similarly ‘the relation Similar’, ‘the relation Father’, etc.) Now suppose that, in addition to designating the property Spherical by ‘P’, we take ‘a’ to designate teh sun and ‘b’ to designate the moon. Then in our symbolic language we write the sentence ‘P(a)’ for ‘the sun is sphereical’. Similarly, ‘P(b)’ is the translation into our symbolic language of the English sentence ‘the moon is spherical’.
..
Regarding Terminology. 1. In orderinary word language there is no word which comprehends both properties and relations. Since such a word would serve a useful purpose, let us agree in what follows that the word ‘attribute’ shall have thsi sense. Thus a one-place attribute is a property, and a two-place (or a many-place) attribute is a relation. 2. Similarly, it is useful to havea comprehensive term for the designations of one- and many-place attributes. For this let us follow Hilbert and use the word ‘predicate’. (Heretofore, this word has been confined mostly to properties or to designations of them, and has not included many-place attriutes or predicates.) Thus a one place predicate is a sign for a one place attribute (i.e. for a property); and in general an n place predicate is a sign for an n place attribute. 3. Let us always distinguish clearly between signs and what id esignated. Failure to observ e this distinction has in the past occasioned much confusion in logic and in philosophy generally. In speaking about an expression, let us always put the expression in quotation marks or use some special designation for it, e.g. a German letter as in 21a. We make but one exception to this practice: we omit quotation marks in case the expression stands on a line either alone or with a designating number or letter; see e.g. our enumeration of formulas in T8-2. Suppose, e.g., ‘Pa’ is taken as a symbolic translation of ‘a is old’; then we say: “P (but not: ‘P’) is a one-place attribute, viz. the property Old; this attribute is designated by a one-place predicate ‘P’”. Similarly, we say: “the two-place relation R exists between such and such persons”, “the two-place predicate ‘R’ occurs in such and such a sentence”. And simiarly: ‘the individual a …’, “the name ‘a’ …”.
2b. Sentential constants. It is often burdensome to work with sentences that are entirely written out like ‘Pa’ or ‘Rac’, esxpecially if they are even longer or are repeated frequently in the same connection. We therefore use on occasion the letters ‘A’, ‘B’, ‘C’ as abbreviations for any sentences whatever of the symbolic language. These letters are called sentential constants (or: propositional constants). E.g. in a certain case ‘A’ might be taken as an abbreviation for ‘Pa’; as soon as ‘P’ and ‘a’ are interpreted, ‘A’ is also interpreted. In our use of a sentential constant we will for the most part leave open what particular sentence it stands for as an abbreviation.” pg. 4-6 ISA
More examples from Carnap shall be forthcoming, but first, if the distinctions being raised, e.g. that between Quine and Carnap, are not clear or do not clearly present themselves as consequential, then a much more detailed explanation of the problems and solutions has already been given in the following entry on Goldfarb’s “Deductive Logic” where a use and mention mistake is made in the very section on use and mention: #2025-0820-1850.
Note, Carnap gets away with his symbolic methods in a way that Goldfarb does not: as long as Carnap does not claim anything more than symbolic logic as his subject then he is fine to carelessly deploy designations and names with impunity, but symbolic logic is not to be confused with schematic logic or pure logic (Carnap does make this confusion, but it is forgivable because it is a mistake in name and not a mixture of methods: his methods are strictly symbolic and do not rise to the schematic).
Logic does not presuppose attributes, classes, or predicates. Logic may be said to be about predicates in as much as it is about validity as truth under each supplemented lexical substitution. It is a use and mention error to take schema as sentences i.e. to take an assembly of marks deployed in the development of logic as referring or meaning, or even ranking among the compounds of some grammatical construction from lexical elements.
It is von Neumann 1927 “Zur Hilbertschen Beweistheorie” through Quine that grants us the utility of schematic methods in logic without anything like the trappings of signs, symbols, or formula of theories of logic as contemplated by Russell and Carnap and their descendant.
“10a. Predicate variables. According to our treatment of the universal quantifier and the existential quantifier, a sentence of the form ‘(x)(..x..)’ is true if and only if the sentential formula ‘..x..’ holds for every individual; and a sentence of the form ‘(E x)(..x)’ is true if and only if the formula ‘..x..’ holds for at least one individual.
Now, it is easy to see taht the sentence ‘(x)Px only if Pa’ (i.e. ‘(x)(Px) only if Pa’) is true in every possible case, no matter what the facts are regarding the individual a and the property P. ONly two cases need to be distinguished. Case (1): the individual a has property P. In this case, ‘Pa’ is true; hence (by Truth-table I(4)) teh whole sentence is true. Case (2): a fails to have property P. In this case, the sentence ‘(x)Px’ is false because it asserts that all individuals have property P; hence (again by the truth-table) the whole sentence is true. The sentence in quenstion is thus necessarily true, regardless of the facts. We may also see this immediately from the word-language version of ‘(x)Px only if Pa’. “If all individuals are P, then a is P”. Indeed, the sentence ‘(x)Px only if Pa’ can be included among sentences that are L-true in our technical sense, provided we extend in a suitable way the rules governing value-assignments. Let us make that extension now.
Let us agree that free variables and descriptive signs count as value-bearing signs. (Thus, in ‘(x)Px only if Pa’ only ‘P’ and ‘a’ are value bearing.) As values of individual signs, let us take all individuals of the domain in question; and as values of one-place predicates, let us take all classes of these individuals (i.e. all subclasses of the domain in question).
Let us agree to regard a one-place atomic formula as true at a given value assignment if and only if the individual (assigned as the value fo the individual sign) belongs to teh class (assigned as the value of the predicate). Further, we agree to regard a universal sentence (say ‘(x)Px’) as true at a given value-assignment provided the operand of this sentence (here ‘Px’) is true at each value-assignment to ‘x’, in view of the assignment already given to the remaining value-bearing signs (here, only ‘P’).
In view of the above, it is readily seen that the sentence ‘(x)Px only if Pa’ is true at every value-assignment to the value-bearing signs ‘P’ and ‘a’, and hence is L-true. (The argument is essentially the same as that given at the beginning of this section; we repeat it here because the formation must now be phrased in terms of value-assignments. Case (1): the value-assignment to ‘P’ and ‘a’ is such taht the individual assigned to ‘a’ does in fact belong to the class assigned to ‘P’. At this value-assignment, ‘Pa’ is true; and hence the whole sentence ‘(x)Px only if Pa’ is true. Case (2): the value-assignment to ‘P’ and ‘a’ is such that the individual assigned to ‘a’ does not belong to the class assigned to ‘P’. At this value-assignemnt, ‘(x)Px’ is false since ‘Px’ is not true at every value-assignment to ‘x’ (in particular ‘Px’ is not true if we assign to ‘x’ the individual presently assigned to ‘a’); hence at thsi value-assignemnt the whole sentence ‘(x)Px only if Pa’ is again true. consequently the sentence is true at every value-assignment.) Similarly, the open formula ‘(x)Px only if Py’ is L-true; for the possible value-assignment to the free variable ‘y’ are identical with those of ‘a’.
It is further evident that any other sentence with teh same form as ‘(x)Px only if Pa’, but with a different predicate in place of ‘P’, is true just as ‘(x)Px only if Pa’ is. E.g. ‘(x)Qx only if Qa’ is true. Now we saw earlier that sentential variables are useful because they facilitate the creation of open L-true formulas from which L=-true sentences can be obtained by arbitrary substitutions. Here, analogously, it is useful to introduce predicate variables. Let us agree to use ‘F’, ‘G’, ‘H’, ‘K’ (and other letters, as occasion demands) for predicate variables, and to count as expressions substitutable for these variables either predicate constants or other predicate variables. In making value-assignments for a sentential formula, we assign classes of individuals to one-place predicate variables and also to one-place predicate constants. Thus e.g. the open formula ‘(x)Fx only if Fa’ is L-true, since in fact the possible value-assignments to ‘F’ are the same as those originally possible for ‘P’. From this L-true formula our earlier L-true sentences can then be obtained by substituting ‘P’ for ‘F’, or else ‘Q’ for ‘F’. The open formula ‘(x)Fx only if Fy’ with both ‘F’ and ‘y’ as free variables is also L-true, and is in fact the most general formula of the form considered here; it has as substitution instances the previous L-true formulas of this section. ‘(x)Fx only if Fy’ is a purely logical formula, devoid of descriptive constants.
10b. Intensions and extensions. Our practice has been to define L-concepts on the basis of value-assignments. Now let us take up several questions regarding the sorts of values we have used in such assignment. Why do we take the values of sentential variables to be truth-values and not propositions? Of course, it is simpler to work with just two truth-values than with indefinitely many propositions. But teh question is: Is this simplification justifiable? A similar question occurs in connection with one-place predicate variables: Is it justifiable to take as values of these predicate variables just classes of individuals, rather than properties?
In order to resolve these questions we introduce here the semantic concepts of intensions and extensions. (A reader concerned chiefly to master the technique of the symbolic language, and having less interest in semantic and philosophic matters, may omit this section.)
A one-place predicate designates a property. (E.g. ‘Book’ designates the property of beign a book; ‘Blue’ designates the colour blue, a property of certain things.) We shall call this property the intension of the predicate. By the extension of a predicate we shall understand the class of individuals having the property designated by the predicate. (E.g. the extension of ‘Book’ is the class of books; and the extension of ‘Blue’ is teh class of blue things.) Analogously, we consider the intension of a two-place predicate to be the two-place relation designated by the predicate, and the extension of a two-place predicate to be the class of ordered pairs of individuals for which the predicate holds (i.e. the class of ordered pairs that satisfy the relation designated by the predicate). (E.G. the intension of the predicate ‘Fa’ is the relation of fatherhood, and the extension of this predicate is the class of paris comprising a father and one of his children.) In general, for any natural number n, n greater than or equal to two, we take the intension of an n-place predicate to be the n-place relation designated by that predicate, and its extension to be the class of ordered n-tuples for which the predicate holds.
While not customary, it is useful to make analogous distinctions for individual constants (or, more general, for closed individual expressions). Suppose the father of Peter Brown is also may of Lexington. Then the two phrases ‘the father of Peter Brown’ and ‘the mayor of Lexington’ (more precisely, the individual expressions in our symbolic language that correspond to these two phrases; such expressionare introduced later (35) as ‘descriptions’) refer to the same individual. Of these two phrases therefore, we say that they have the same extension, viz. this particular individual. On the other hand, it is evident that the two phrases have different senses. By the intension of an individual expression we understand its sense. This is a concept similar to property or relation, but of a different type for which there is no established designation; we agree to use for it the term ‘individual concept’. We will become aquanted later with still other such concepts, among them functions like e.g. the arithmetic sum-function designated ‘+’. By teh intension of such a function sig (or: functor) we understand the function designated by the sign; by its extension we understand the value-distribution of the function (a notion to be explicated later).
Next, suppose that the symbolic language contains variables whose substitutable expressions include the constants and closed compound expressions of some fixed kind. Following out the distinction between the intension and the extension of a constant, it is possible here to set up a similar distinction between the value-intensions and the value-extensions of a variable. Expressions substitutable for a variable have both intensions and extensions; we count all such intensions among the value-intensions of the variable, and similarly all such extensions among the value-extensions of the variable. When we think of the ‘values’ of a variable, we usually have its value-intensions in mind. However, in examining the L-truth of logical formulas constructed in a loanguage with so simple a structure as the symbolic languages treated in this book, ti is quite sufficient to consider the values of a variable as its value-extensions. E.g. the values (regarded as value-intensions) of the sentential variables ‘p’, ‘q’, etc. , are propositoins; as we have see however the tautological character of (say) the formula ‘p or not p’ can be ascertained without considering numerous (in some circumstances, infinitely numerous) propositions, but simply the two truth-values which are the value-extensions of the variable ‘p’.
So far as teh truth-value of a sentential compound is concerned, it is sufficient to consider just the value-extensions (the truth-values) of constituent sentential variables because the truth-value of this compound is uniquely determined by teh truth-values of its components; i.e. the sentential connectives used in such compounds are themselves extensional. Again, the truth-value of an atomic sentence obviously depends only on the extension of its predicate and teh extensions of its individual constants; hence, an atomic sentence is also extensional. Continuing, the truth-value of a universal sentence depends only on the extension of the property determined by teh operand of the quantifier (i.e. on whether this property attaches to all individuals, or not); thus a universal sentence is also extensional. The same remark applies to an existential sentence. INdeed, each of our symbolic langauges– the present language A, and the languages B and C to be introduced later– is an extensional language in the following sense: a sentence in any one of these languages does not change tis truth value if any expression in the sentence is replaced by another with the same extension. Consequently, it suffices for the evaluation of any formula to consider simply the possible extensions of the formula’s descriptive constants and the value-extensions of the formula’s variables.
A symbolic language which, in contrast to the one treated here, also contains symbols for the so-called logical modalities– i.e. such concepts as necessity, possibility, impossibility, contingency and the like– is not extensional. (For suppose it is not raining here now. Then the sentence ‘it is raining’ is false, and so has the same extension (or truth-value) as the L-false sentence ‘it is raining and it is not raining’. Now let this last sentence be a component in a larger modal sentence; when ‘it is raining and it is not raining’ is replaced by ‘ti is raining’, the truth-value of the whole modal sentence does not always remain unchanged. E.g. the modal sentence ‘it is impossible that it is raining and it is not raining’ is true, whereas the sentence ‘it is impossible that it is raining’ (produced therefrom by the indicated replacement) is false– for while it is note the case that it is raining here now, this case is nevertheless logically possible. Thus symbolic languages with modality symbols are generally not extensional.) In such non-extensional symbolic languages, one must consider intensions as well as extensions as values of descriptive conastants and variables.
Most systems of symbolic logic employ an extensional language, for the reason taht such a language has raidcally simpler structure and hence simpler constitutive rules. However, it cannot justly be said that this procedure compels a neglect of the logical modalities: these can be expressed in another way, viz. in the metalanguage, with the aid of L-concepts. Instead of saying a certain proposition (or state of affairs) is necessary– or impossible, or possible, or contingent–, we say that a corresponding sentence (i.e. one that designates the proposition in question) is L-true– or L-false, or not L-false, or L-indeterminate, respectively. E.g. let ‘A’ designate the proposition (the possible state of affairs) that it is raining here now; ‘A or not A’ then designates the proposition that it is raining or it is nto raining. Within a modal language containing words one would say ‘it is necessary that it is raining or it is not raining’; in a symbolic modal language having the symbol ‘N’ for ‘necessary’, the sentence would appear as ‘N(A or not A)’. By contrast, this sentence cannot be stated in our object language because this language is extensional; however we can formulate the corresponding sentence “the sentence ‘A or not A’ is L-true” in our metalanguage.
Intensions and extensions of the chief types of expressions:
- Expressions
- Sentence
- Individual constant
- One-place predicate
- n-place predicate (n greater than one)
- Functor
- Intensions
- Proposition
- Individual concept
- Property
- n-place relation
- Function
- Extensions
- Truth-value
- Individual
- Class of individuals
- Class of ordered n-tuples of individuals
- Value-distribution.” pg. 38-42 ISA
Symbolic methods resort to ‘expressions’ because a formula is said to express (symbolically) a proposition or a state of affairs. So, although I may speak of sentences, predicates, and functors, they are not the expressions that Carnap (or Russell) are talking about. Carnap’s use of functor in ISA corresponds to functional intensions and value-distributions.
“The word ‘functor’, grammatical in import but logical in habitat, becomes useful at this point. A functor is a sign that attaches to one or more expressions of given grammatical kind or kinds to produce an expression of a given grammatical kind. The negation sign is a functor that attaches to a statement to produce a statement, and to a term to produce a term. The alternation sign, conditional sign, and biconditional sign are functors that join statements in pairs to produce statements and join terms in pairs to produce terms. ‘some’ is a functor that attaches toa term to form a statement; so is ‘each’. Finally ‘includes’, ‘properly includes’, and ‘coextensive’ are functors that join terms to form statements. Functors of this last kind are called copulas; in particular ‘properly includes’ means ‘are’.” pg. 129 Methods of Logic 4th Edition by Quine
“The grammarian’s question is, then, what strings of phonemes belong to the language? What strings, that is, ever get uttered or could get uttered in the community as normal speech? The grammarian’s job is to demarcate, formally, the class of all such strings of phonemes. Formally? THis means staying within a purely mathematical theory of finite strings of phonemes. More explicitly, it means saying nothing that could not be said by means of a technical vocabulary in which, besides the usual logical particles and any desired auxiliary apparatus from pure mathematics, there are only the names of the ponemes and a symbol signifying the concatenation of phonemes.
A mere listing of strings would already be formal, but it would not suffice, since the desired strings, though finite in lenght, are infinite in number. So the grammarian has recourse to recursion: he specifies a lexicon, or list of words, together with various grammatical constructions, or steps that lead to compound expressions from constituent ones. His job is to devise his lexicon and his constructions in such a way as to demarcate teh desired class: the class of all the strings of phonemes that could be uttered in normal speec. The strings of phonemes obtainable from the lexicon by continued use of the construction should all be capable of occurring in normal speech; and, conversely, every string capable of occurring in normal speech should be obtainable from the lexicon by the constructions (or should at least be a fragment of a string which as a whole is obtainable from the lexicon by constructions).
When we analyze a complex expression according to the constructions involved, we get something of the form of an inverted tree, like a genealogy. The complex expression is at the top. Below it, at the next level, are the ‘immediate constituents’– one or two or more– from which the complex expression was got by one application of some one construction. Below each of these constituents are its immediate constituents; and so on down. Each branch of teh tree terminates downward in a word.
..
As an aid to specifying the constructions, the lexicon is classified into gramatical categories. For we want to be able to specify a construction by saying what operation is to be performed upon any expression of such and such category; or perhaps what operation is to be performed upon any pair of expresions, one of this cateogry and one of that. Since the compound expressions obtained by constructions are to be available as constituents under further constructions, we must also say what category each construction issues in.
Thus a construction is specified in this vein: take any expressions, belonging respectively to such and such categories, and combine them in such and such a distinctive way; the result will belong to such and such a category. Commonly the distinctive way of combining the constituents will be marked by the insertion of a distinctive particle; examples are ‘or’, ‘plus’, ‘and’, ‘but’. Also there are constructions that operate on single constituents, rather than combining two or more; one such is negation, which consists in prefacing the constituent with the partical ‘not’.
The constructions serve to add complex members to the categories, which had begun with word lists. A construction may even start a new category, which had no simple members; for instance, the class of sentences. Teh constructions, once specified, apply over and over, swelling the several categories ad infinitum.
The categories are what we used to call the parts of speech, though they need no preserve the traditional lines of cleavage. One of our categories might be that of singular terms. Another might be that of copulas. Another might be that of intransitive verbs. Another might be that of adjectives. One of our constructions might be that of applying ‘not’ to a copula to get a complex copula. Another might be that of prefixing a copula to an adjective to get a complex intransitive verb: ‘is mortal’, ‘is not mortal’. Another might be that of joining a singular term to an intransitive verb to get a sentence: “Tom is mortal”, “Tom is not mortal”. Another might be that of joining two sentences by an ‘or’ to get a sentence: ‘Tom is mortal or Tom is not mortal’. What grammar tells us thus indirectly through its lexicon, categories, and constructions is not that this last sentence is true, but just that it is English.” pg. 16-18 Philosophy of Logic Second Edition Quine
So, taking Quine’s definition of functor from Methods of Logic 4th edition (he does give a different definition of ‘functor’ in the philosophy) with the grammatical distinctions in Philosophy of Logic second edition: functors are the grammatical particles of a grammatical construction on gramamtical categories.
It is important to distinguish grammatical particles from functors: not all gramatical particles need be functors. I leave this open because though the logical methods I am allied to are those where the distinction between particle and functor is immaterial, I anticipate the odd grammar that may appear where some technical matter turns on the distinction between functor and particle. For more see pages 26 thru 30 of Quine’s Philosophy of Logic Second Edition.
All of these preliminary quotations from Quine and Carnap have been carefully presented so that upon dipping into Principia Mathematica I can make clear what are the problems and their solutions. They are all principle concerned with the difference between logic and mathematics (and that specialization of mathematics called ‘mathematical logic’).
“The mathematical treatment of the principles of mathematics, which is the subject of the present work, has arisen from the conjunction of two different studies, both in the main very modern. On the one hand we have the work of analysts and geometers, in the way of formulating and systematising their axioms,a nd the work of Cantor and others on such matters as the theory of aggregates. On the other hand we have symbolic logic, which, after a necessary period of growth, has now, thanks to Peano and his followers, acquired the technical adaptability and the logical comprehensiveness that are essential to a mathematical instrument for dealing with what have hiterto been the beginnings of mathematics. From the combination fo these two studies two results emerge, namely (1) that what were formerly taken, tacitly or explicitly, as axioms, are either unnecessary or demonstrable; (2) that the same methods by which supposed axioms are demonstrated will given valuable results in regions, such as infinite number, which had formerly been regarded as inaccessible to human knowledge. Hence the scope of matheamtics is enlarged both by the addition of new subjects and by a backward extension into provinces hitherto abandoned to philosophy.” Preface, v, Principia Mathematica Volume 1.
My claim is that it is the projection of strictly logical methods onto the methods of mathematics that produces the two results enumerated in the first paragraph of the preface (not the mathematization of the logical methods of mathematics). It is perhaps best explained by the analogy of seperation of church and state.
An technology of teaching is effective when it constructs the terminal repertoire specified.
A terminal repertoire is effective when it makes a person who responds appropriately to the world.
A response is appropriate when it contributes to the survival of the culture which aided in its construction.
Logical and scientific responses have contributed most to the survival of the cultures which aided in their construction.
This continues notes from #2025-0826-1635 on 2023 “The Earth and Its Peoples 8th ed.” (EP) by Richard Bulliet, Pamela Crossley, Daniel Headrick, Steven Hirsch, and Lyman Johnson.
The fragments of notes from chapter 2 will come after I have finished notes on chapter one exhaustively. I shall start at number 19 of the notes on chapter one, leaving point 20 until after it and so on. Here it will begin anew as point 1.
Between 50-46k years ago humans migrated out of eastern and southern Africa throughout south Asia. Around 46k years ago they populated western and eastern Asia and Australia. Some 46-10k years ago they were in northeastern Asia and north and south America.
1-2: Technology and Culture in the Ice Age
“Learned patterns fo action and expression constitute culture. .. Culture includes both material objects, such as dwellings, clothing, tools,a nd crafts, and nonmaterial values, beliefs, and languages. .. Among humans, instincts are less important thant he cultural traditions that each generation learns from its elders.” pg. 13
1-2a: Food Gathering and Stone Tools
“Because the tools that survive are mostly made of stone, the extensive period of history from the appearance of the first fabricated stone tools around 2.6M years ago until the appearance of metal tools around 6k years ago has been called the Stone Age.” pg. 13
The name is misleading: “Early humans also made useful objects out of bone, skin, wood, plant fibers, and other materials less likely than stone to survive the ravages of Time.”
“Early scholars recognized two phases of the Stone Age: the Paleolithic (pay-lee-oh-LITH-ik) (Old Stone Age), down to 8000 BCE, and the Neolithic (NEE-OH-LITH-IK) (New Stone Age), which is associated with the rise of agriculture. Modern scientists have developed more complex schemes with many subdivisions.” pg. 13
“early humans depended heavily on vegetable foods such as leaves, seeds, and grasses, but during the Ice Age the consumption of highly nutritious animal flesh increased. .. increased meat eating and toolmaking– appear to be closely linked.” pg. 13
“Specimens of crude early tools found in the Great Rift Valley of eastern Africa reveal that Homo habilis made tools by chipping flakes off the edges of volcanic stones.” pg. 13
“Lacking the skill to hunt and kill large animals, *Homo habilis probably obtained animal protein by scavenging meat from kills made by animal predators or resulting from accidents. .. used large stone ‘choppers’ for cracking open bones to get at the nutritious marrow .. such tools are found far from the volcanic outcrops where they were quarried.. people carried them long distances for use at kill sittes and camps.” pg 13.
“Members of Homo erectus were also scavengers .. they made more tools for butchering large animals, including a hand ax formed by removing chips from both sides of a stone to produce a sharp outer edge. .. suitable for skinning and butchering animals, for scraping skins clean for use as clothing and mats, for sharpening wooden tools, and for digging up edible roots .. can also be hurled accurately for nearly 100 feet (30 meters), .. projectile to fell animals .. hunted elephants by driving into swamps .. trapped and died.” pg. 13
“Homo sapiens .. array of finely made tools, they tracked and killed large animals. Sharp stone flakes chipped from carefully prepared rock cores were used in combination with other materials. Attaching a stone point to a wooden shaft made a spear. Embedding seveer sharp stone flakes in a bone handle produced a sawing tool. .. may of caused ecological crises. between 10k and 13k years ago .. giant mastodons and mamoths .. disappeared from Africa, Southeast Asia, and northern Europe. In North america around 11k years ago, three-fourths of the large mammals became extinct .. giant camels, ground sloths, stag-moose, giant cats, mastodons, and mammoths .. In australia .. similar events .. the Ice age .. distinguish .. climate change and human predation.” pg. 13-14
“anthropologists do not believe that early humans depended primarly on meat for their food .. present-day foragers (hunter and food-gathering peoples) in Africa .. nourishment from wild vegetable foods, .. meat .. feasts, .. . The same .. true for Stone Age .. tools .. left few traces because they were .. perishable .. they .. used skins and mats woven from leaves for collecting fruits, berries, adn wild seeds .. dug edible roots out of the ground with wooden sticks.” pg. 14-15
Collation of timeline from ‘The Little Book of Big History’ by Ian Crofton and Jeremy Black 2017
Although it can be time consuming, when I collect and collate past notes into present ones, as I shall do here, I often type them out rather than copy and paste. It helps me to catch old errors and every so often introduces new ones. Not all new errors are helpful, but sometimes they are. Presumably, they are helpful enough that their tolerance and that of the difficulty of typing rather than copypasting is worth it in the end.
Fragments of notes and two selections from Durant’s ‘The Pleasures of Philosophy’.
Arguing is said to express reasoning, arguments to express reasons, and logic to study correct, appropriate, effective, faultless, proper, good, right, flawless, strict, errorless, infallible, impeccable, or accurate reasoning.
Here are two selections from Durant’s 1929 ‘The Mansions of Philosophy’ (reprinted in 1952 as ‘The Pleasures of Philosophy’, the edition from which I quote)
“Logic is a poor hors d’ourvre for the feast of philosophy; it dulls a thousand appetites for every one it whets. We suspect logic because we have learned that most reasoning is desire dressed in a little rationality; we pretend to be constructing edifices of impartial thought, when actually we are selecting only such facts and agreements as will give dignity to some personal or patriotic wish. We suspect logic because middle age has taught us that life is larger, surer, profounder than our syllogisms; logic is static, puffed up with ‘invariable truths,’ while life is fluent and changeful, and surprises all formulas. ‘The number of things that reason at first refused to recognized, and yet had in the end to admit, is considerable.’ Perhaps in our youth we memorized all the rules of perfect thinking, only to find that the pursuit of knowledge, the recognition of truth, and the wisdom of life, fell incalculably outside this elegantly ordered realm. How gladly we would leave to the end this lgoic that can make even philosophy dry and spiritless, rather than set it here as a barrier to problems less basic, possibly, but much more directly vital to our lives! And yet we must not; we cannot ride forth on our quest of truth without determining in advance what we are looking for, by what road we propose to seek it, and how we shall know it if we come upon it. Any other order would not be logical!” pg. 15-16
“Where, then, shall be the place of reason in this ridiculously plebeian logic of ours, that confirms the prejudices of the commonest man in the street? Its function here, as elsewhere, is to coordinate– sensations into ideas, ideas into knowledge, knowledge into wisdom, purposes into personality, individuals into society, societies into peace. The role of reason in the conquest of truth is secondary but vital: it must weave the chaos and contradictions of many sens into unified and harmonious conclusions which it shall hold subject to verification or rejection by subsequent sensation. It is not half so certain as sensation; for ‘in transcending what is given by actual perception we without doubt make use of an inference’; and every inferential step awa from immediate sensation lowers the probability of our truth. But this, too, is a gamble that life must make; we must attempt the reconcilliation of discordant senses and partial views; if we are to extend our understanding and our mastery. Just as Kohler’s chimpanzees reasoned best when they took in the entire situation, so for ourselves reasoned truth, like philosophy and wisdom, like morality and beauty, is total perspective, the harmonious union of the part with the whole. Through sensation we stand firmly with our feet on the earth; through reason we lift the mind’s eye beyond the present scope of sense, and conceive new truths which some day the senses may verify. Sensation is the test of truth, but reason is its discoverer.” pg. 20
Happened to catch an episode of Jude Judy’s ‘Justice on Trial’ and enjoyed it quite a lot: I would enjoy seeing more mock trial shows.
A delightful conversation with the wonderfully talented Polyducks https://polyducks.co.uk prompted me to dig back into the foundations of the science of behavior.
One of the key paths that needs smoothing out is that from social behavior through verbal behavior to scientific and logical behavior. I shall collect notes from work done by Skinner and others along this path and attempt to link everything together with a aim towards the construction of a logical theory.
“The use of concepts and laws derived from an experimental analysis in the interpretation of daily life is also a source of misunderstanding. An analogy from another science may be helpful. Geophysics interprets the present condition of the accessible parts of the earth in terms of presumed conditions in the mantle and core. It appeals quite freely to the physical laws derived from laboratory analyses of matter under various pressures and temperatures, even though it is merely an assumption that comparable states actually prevail in the interior of the earth. In the same way familiar facts about verbal behavior are interpreted with principles derived from the laboratory study of contingencies of reinforcement, even though the contingencies maintained by the verbal environment cannot be precisely ascertained. In both these examples principles derived from research conducted under the favorable conditions of the laboratory are used to give a plausable account of facts which are not at the moment under experimental control. Neither account can at the present time be proved, but both are to be preferred to treatments which lack the same kind of experimental support.” pg. 100 Skinner ‘Contingencies of Reinforcement’
This distinguishes interpretation from speculation, though the boundary is not as clear as that between, e.g., preformal and formal mathematics. What Skinner emphasizes is that the vast majority of verbal behavior is not strictly logical or scientific and yet much of what passes as scientific and logic behavior is not as strict as vast majority of scientists and logicians purport.
As a side note: almost all (all?) of Skinner’s writings are available through the B. F. Skinner foundation https://www.bfskinner.org/.
“A natural datum in a science of behavior is the probability that a given bit of behavior will occur at a given time. An experimental analysis deals with that probability in terms of frequency or rate of responding. Like probability, rate of responding would be a meaningless concept if it were not possible to specify topography of response in such a way that separate instances of an operant can be counted. The specification is usually made with the help of a part of the apparatus– the ‘operandum’– which senses occurrences of a response. In practice, responses so defined show a considerable uniformity as the organism moves about in a framework set by its own anatomy and the immediate environment.” pg. 83.
An experiment in writing without restraints.
It is sad but true that I have not yet written substantially about behavior in these notes. So much of the last decade of my life and work has been dedicated to behavior and yet there is nothing here that would make that clear. Sure, I mention B. F. Skinner more often than most, but here I have mentioned Quine more than Skinner. To those that know me ‘in real life’ this is a surprise: they wait for Skinner to enter every conversation (and, unsurprisingly, they now bring Skinner up before I do).
I am not a disciple of Skinner. I may call myself more of a radical behaviorist than just a regular old behaviorist, but that does not, to me at least, then imply that I am devoted to all that Skinner has said or all that has been attributed to what Skinner may or may not have said. This must be written explicitly because there is so much about Skinner that can either be misunderstood or when understood is very unlike what I have got from reading him.
It is not uncommon for me to read something that so many others have read and to discover something about it that is at odds with prevailing interpretations. This happens when I read Skinner and when I read Quine. It happens because I have read almost everything that they have written and reread it all again. There are always gaps in my reading: I am not a reading machine even though I may dream of one day being one.
The Skinner we have from his writings is not the same Skinner we have from the rest of his work. This is true of anyone who has done work which the public has been forced to recognize because of its consequences to culture. Little has had a more profound effect on the way we live and work than the science which sometimes goes by the peculiar name ‘the experimental analysis of behavior’.
The phrase ‘the experimental analysis of behavior’ is a mouthful. Sometimes it is interchangeable with ‘the science of behavior’ and sometimes it is not. When I speak with others about behavior, especially when I speak to them about human behavior, I often have to say ‘the experimental analysis of behavior’ rather than ‘the science of behavior’ because most people are quick to separate humans from the rest of the behaving organisms the world has come to know.
Another reason to emphasize the experimental analysis of behavior is to point away from ‘the analysis of mind’ or ‘psychological analysis’, e.g., as in Freud’s early science. Skinner said on more than one occasion that Freud was one of the first behaviorists, and this is as much a word game as it is a concrete observation: Freud linked present behavior to past consequences.
As an aside, I’m writing more casually than I have up to this point because I can not stand the stark entries I’ve made as of late. They are practically impenetrable to most people and I do not like that. I’d rather write in a way that is at least accessible to someone who, unlike me, is less sensitive to precision: inaccuracy is inexcusable but imprecision is often acceptable.
How to start up the science of behavior is as difficult a problem as how to start up the science of logic. There are no obvious entryways into the bits that matter most. Everything is either obvious or obscure.
There is the temptation, one which has been around for centuries, to flirt with philosophy or with some other science which is in vogue e.g. cognitive science is still popular enough to command the attention of most casual readers. A kind of watered down cognitive science is still popular among the influencers who profess a sort of mystic mastery over the minds of the masses.
It is popularity gives the public a feeling of familiarity: if they don’t already know something about it they feel as if there is something there to be known. It is surprising how often I have spoken either about behavior or logic and the listener has said ‘I do not know what you are talking about’. Yet, if I was to tell a fantasy story about a unicorn’s search for the lost love of their youth, there is suddenly something to know.
It is not as if I suck at communicating, though I in no way profess to be good at writing or speaking. I lack the rhetorical flare that comes from that false sense of certainty behind any hard won piece of proper literature (be it an essay or an encyclopedic entry).
I am uncertain in almost all that I do and this is the ultimate weakness in an effective writer. Writers are tyrrants. They take complete control of the reader and are bound by no laws, no reasons, no limits other than what they can get away with, and a good writer can get away with the world.
A good story well told beats any sort of scientific explanation or experiment every time. It doesn’t just beat it, it extinguishes it. Science lives or dies by the stories that people tell: no experiment in the world can undo the damage done by a good story. Good stories are their own death, and so often that death is delayed again and again. People can’t drop a good story even if it harms their friends and family. The world is powerless, and the writer knows all of this, and somehow they still publish.
I’m glad to say that this note has now become less about the science of behavior and more about the pains caused by my subpar writing. I’m uncertain and fallible. Raw strength sniffs out whatever whiff of weakness there is to be found in unfamiliar words and sentences. Those who would preach to the choir are at least assured of the interests of their audience. Pick love, money, power, or any other of the more complex derivatives of food, sex, and water and you’ll find a choir waiting for you to sing for your supper.
How am I to tell a good story when I dispise them so? Can science go on without stories? Others have told me that there is a compromise e.g. imprecision is unavoidable. If you had to sit around all day and speak down to ever nanometer you’d never reach any part of the world beyond the tip of your nose. Are accuracy and imprecision enough to spawn stories? Why not just look at what others have written?
Writers of history have done the most to help me understand the kinds of stories that can be told which are both accurate and imprecise. Imprecision is the order of the day in history. But, historians are plagued by the same old problem: who cares? So what if so-and-so won the war some two centuries ago: why are you wasting my time with this?
To say “Well, some people like history and some people don’t.” is to condemn those who do and those who don’t. They are born that way, or they were brought up that way, and there’s nothing you can do about their birth or their upbringing: they are in the past and inaccessible, much less manipulable. Science is where such misguided excuses end: the more we know about how the world works the more control we have over how the world works out.
This has been another strange experiment in a less restricted kind of writing.
Sequence of links to papers that are relevant to garbarge collection of binary trees.
Most people learn logic for the same reason they learn any other foreign language: they love someone (or something) and wish to know them (or it) in their native tongue.
Those in love with ideas, items, and words are ready for logic. Those who love this or that philosopher may only be as prepared to learn logic as they are to learn their native language, be it Latin, Greek, Gemran, French, Spanish, English, Chinese etc.
The language of philosophy’s most well known child, science, is logic. Contemporaries would point to something beyond logic, be it a deviation or the divinations of statisticians and probability theorists (the modern day Oracle of Delphi).
At least the probability theorist has the courage to put ‘theorist’ in their name. Those that would dodge the dismissive of ‘theory’ are doomed to be found out and live in constant fear that they shall be caught by their assumptions.
The days where philosophy gave us a consistent way of life are gone, but the delights of philosophic practices are as alive as ever. Those that condemn human behavior for its limited fallibility have condemned their life no matter how consistent it may seem.
There is no better reason for anyone to read what I’ve written here than there is for them to go to a pizza parlor just because I recommended it. It is easy enough for me to admit that I hide in obscurities because I am destine to lose to anyone who professes the profound. The richness of my methods are mundane when compared to the new. Mine are old. Not so old as the stories we have told since we found stories to tell, but much older than the new and much less ornate in their presentation.
I am told that this is a failure on my part. Though I object, perhaps by saying this is a feature, I am at a loss to defend myself against the profound, the ambitious, or the earnest. The performance is lost on me, my life, and my work.
At best I meddle in the world of words, never quite reaching the items and ideas that dazzle and confound the best and worst of the world (though to call me a nominalist would be a dire insult).
What logic brings to me is my love. If your love is science, and if through it you have found a kind of sublime satisfaction, then logic is the language in which you may know your love better than you know your self.
Once upon a time, logic was correct reasoning. Humans were special because they reasoned in ways that no other creature could. The division of logic into its purportedly formal and informal parts was a grand mistake: at best they are poor names for poles between which the practices of paraphrasing are more or less effective (but do not be fooled into taking effectiveness as a criterion of logic).
Arguments express inferences, or so we are told: what evidence there is for either the argument or the inference is in need of closer examination.
An argument was then correct when its conclusions can be inferred from its premises: it is seen by some as evidence for the inference which is part of an act of reason.
When an argument is said (erroneously) to express an inference, the premises and conclusions of the argument are said to express what the inference is about. The things inferences are about are called propositions: they are also the things that are said (erroneously) to be true or false.
When it is said that a sentence of an argument is true, it is true only because it expresses a true proposition (again this is the classical error).
This has been a poorly written polemic.
My work is largely inaccessible because of how bad I am at writing. Almost all of the writing I’ve done throughout my life is aimed at a kind of self discovery: it doesn’t take into account what other people are interested in or in what they tend to think.
While there’s room for the kind of writing that leads to discovery, it can only go so far. It might help one or two people who are already on a similar discovery expedition.
When I work to know a thing my view of the world narrows upon it. I am blind to what else the world has to offer and may bash myself upon an obstacle heralded no more covertly than the rising of the sun.
This blindness seems characteristic of any intense investigation. In the past, metaphors made up for such mistakes. A reader may question whether the metaphor carries the day, but may never make it back to the skirmish from which it was won. My fear, and it is so very frequent, is that without logic I shall always be blind to the largest of obstacles: doubt.
To know logic has been my thing. Beyond logic is theory and my theory of the world is bound to B. F. Skinner’s radical behaviorism.
The basic things of the world are contingencies. They are consequences from responses on occasions. No more is known of them than that.
Theories project logic upon the world and work with the rest of our repertoires to deal with what troubles us most.
The premises of a theory activate the network of validities uncovered by logic, and through them reach conclusions which check the rest of the world. Conclusions go extinct before premises, but no premise is left unthreatened by contravened conclusions.
I am failing to write freely.
The wikipedia page on binary relations https://en.wikipedia.org/wiki/Binary_relation includes a subsection on ‘Types of Binary Relations’. These are properly taken as types of two place, or two zero place, or one one place, predicates.
The sum of the number of left and right places of a predicate is its degree. The organization of predicates of any degree with respect to the types of binary relations proceeds in the same way as when the methods of integral and differential calculus are extended to their multivariable generalizations.
Thankfully, my names for the different kinds of funtional abbreviations of logic are consistent with those in the wikipedia article.
Upon reading the wikipedia article, it occurred to me to distinguish between injectiveness, perhpas just injective, and injectivity. Maybe better, the difference between injective and injection.
An injection is total and injective i.e. functional and injective.
The difference between functive and functional just occured to me. This would be carried on as injective and injectional, surjective and surjectional, and bijective and bijectional.
Another edit that must be made to my organization of predicates is to change ‘graph abbreviations’ to ‘multigraph abbreviations’. They were always multigraph abbreviations and it is only when the multigraph abbreviations are restricted to predicates of degree two that they become graph abbreviations.
Connectivity may also be better explained as the inclusion of the appropriately degreed nonidentity of the predicate with the alternation of the predicate and its inverse.
A great list of great lists is found at http://www.greaterbooks.com/.
My interest is in the verbal control of human behavior e.g. the textual and vocal responses that are shaped by verbal behavior which was emitted hundreds if not thousands of years ago.
Books are the tip of a verbal iceberg which is forever submerged in the murky waters of unrecorded history.
Completely unrelated, a link to Carnap’s 1950 ‘Empiricism, Semantics, and Ontology’: https://www.phil.cmu.edu/projects/carnap/editorial/latex_pdf/1956-ESO.pdf
In #2025-0830-1425 the world, camera, and image plane coordinate frames were constructed. All that is left, to finish setting up the camera obscura problem (of predicting where points are projected from the world through the pinhole onto the back wall), is to set up the image sensor or image receptor coordinate frame.
In the dark room version of the camera obscura presented first in #2025-0828-1649 the back wall became what is now called the image sensor. I’ve suggested multiple times now that it is better named ‘image receptor’ because ‘sensing’ is a far more elaborate repertoire than ‘receiving’ and receptors are the basis of our contact with the world at the ends of our peripheral nervous system.
Whether it’s called the image sensor or image receptor, there are only a few key measurements that complete the description of the intrinsic or internal parameters of the camera.
The image sensor is a rectangular part of the image plane. Each of its sides are parallel to one of the basis vectors of the image plane coordinate frame and perpendicular to the other i.e. its upper and lower sides are parallel to the floor and its left and right sides are perpendicular to the floor.
It has a definite width and height. Both are measured in (meter) stick units of the image plane coordinate frame.
In the idealized camera obscura, the principle point (the origin of the image plane coordinate plane) is at the intersection of the diagonals of the image sensor. In reality, finding the location of the principle point relative to the image sensor coordinate frame is part of a general problem called ‘camera calibration’.
The origin of the image sensor coordinate frame is the upper left hand corner of the image sensor (when looking through the pinhole to the image plane). Therefore, the origin of the image sensor is designated by ‘(-half of the width of the image sensor, -half of the height of the image sensor)’ in the image plane coordinate frame.
The image sensor is made of rectangular pixels and the units of the image sensor coordinate frame are pixels. So, when talking about lengths it can either be given in the units of the image sensor coordinate frame or in the units of the image plane coordinate frame.
Since the pixels are rectangular and not square, three horizontal pixels may be longer, shorter, or the same length as three vertical pixels. Thus, lengths are most often given in (meter) stick units because those are the same units as the world, camera, and image plane coordinate frames.
Unlike the basis vectors of the world, camera, and image plane, the basis vectors of the image sensor coordinate frame are possibly different lengths even though they are both given in pixel units. Thus, we differentiate between horizontal and vertical pixel units.
A horizontal pixel unit is the width (in the world, camera, or image plane units) of a pixel, and a vertical pixel unit is the height of a pixel. The first basis vector points from the upper left to the upper right of a single pixel and the second basis vector points from the upper left to the lower left of a single pixel.
Since pixels are rectangles and not points, the pixel at the origin of the image sensor coordinate frame intersects the origin at its upper left hand corner. While it may seem obvious that the upper left corner of the image sensor is the upper left corner of a pixel, this is not always the case, and there are other ways of picking the origin that may be helpful e.g. some times it helps to pick the center of the upper left corner pixel as the origin.
Converting between pixel units and (meter) stick units involves two densities: horizontal and vertical pixel density. Horizontal pixel density is the number of horizontal pixles per meter and vertical pixel density is the number of vertical pixels per meter.
Thus, the principle point is designated by ‘(the product of the horizontal pixel density with half the width of the image sensor, the product of the vertical pixel density with half the height of the image sensor)’ in the image sensor coordinate frame.
Now, to solve the problem of projecting a point in the world onto the image sensor, we need to go from the world coordinate frame to the camera coordinate frame to the image plane coordinate plane to the image sensor coordinate frame!
Continuing the timeline from #2025-0831-1507 of ‘The Little Book of Big History’ by Ian Crofton and Jeremy Black 2017.
A very important ascii map
180 150W 120W 90W 60W 30W 000 30E 60E 90E 120E 150E 180
+90N-+-----+-----+-----+-----+----+-----+-----+-----+-----+-----+-----+
| . _..::__: ,-"-"._ |7 , _,.__ |
| _.___ _ _<_>`!(._`.`-. / _._ `_ ,_/ ' '-._.---.-.__|
|.{ " " `-==,',._\{ \ / {) / _ ">_,-' ` mt-2_|
+ \_.:--. `._ )`^-. "' , [_/( __,/-' +
|'"' \ " _L oD_,--' ) /. (| |
| | ,' _)_.\\._<> 6 _,' / ' |
| `. / [_/_'` `"( <'} ) |
+30N \\ .-. ) / `-'"..' `:._ _) ' +
| ` \ ( `( / `:\ > \ ,-^. /' ' |
| `._, "" | \`' \| ?_) {\ |
| `=.---. `._._ ,' "` |' ,- '. |
+000 | `-._ | / `:`<_|h--._ +
| ( > . | , `=.__.`-'\ |
| `. / | |{| ,-.,\ .|
| | ,' \ / `' ," \ |
+30S | / |_' | __ / +
| | | '-' `-' \.|
| |/ " / |
| \. ' |
+60S +
| ,/ ______._.--._ _..---.---------._ |
| ,-----"-..?----_/ ) _,-'" " ( |
|.._( `-----' `-|
+90S-+-----+-----+-----+-----+----+-----+-----+-----+-----+-----+-----+
Map 1998 Matthew Thomas. Freely usable as long as this line is included
There is also a program that displays maps in the terminal:
Corrected the schematic abbreviation by the predicate functor of functionality in #2025-0830-2137 whose compositional component was missing ‘j’.
Continuing the timeline from #2025-0829-1654 starting on page 99 of ‘The Little Book of Big History’ by Ian Crofton and Jeremy Black 2017.
A reminder that the abbreviations of logic (and their schematic descendants) do little more than set the stage for validities (e.g. ‘coex and F F F’) and that validities do little more than set the stage for theories.
It occurred to me that philosophy as a science can be defined as that which seperates theory from logic e.g. set theory from its logic. This would explain the back and forth between philosophers, logicians, and theoraticians (and experimentalists).
Having completed in #2025-0830-1409 the schematic graph abbreviations of logic I turned forward to the order abbreviations. But, before that, I looked back at what I had done thus far and found further simplifying improvements.
Going back to the schematic compositional abbreviations of logc set out in #2025-0828-2145 a simplification can be introduced by moving from ‘resultant’ to an intermediate predicate functor ‘composition’. I’m unsettled on the phrase ‘composition’, but accept that there is room for a definition of its sort in the methods of logic as I give them (that is with a stack based outlook).
The definition of resultants does not change. It is still
(I had to correct this when looking back at it: it had ‘each’ where it needed ‘some’) It is very unfamiliar for those not already steeped in stack based methods of programming e.g. be they in a high level language like Forth or a low level language like x86 assembly.
To help explain how resultants work in a stack based world for those not already familiar with them, I’ll skip to unraveling the resultant:
(I still have to do this for most of the other schematic abbreviations).
So, ‘res j k m F G’ for ‘some j 0 (j bury j+k) and (k drop) F (m prop) G’ i.e. ‘(..u..w: some ..v is such that ..u..vF..x..y and ..u..v..wG..y :..x..y)’.
now, the result of ‘STEP’ with ‘STEP’ is two steps of execution of a stack machine i.e. the following are equivalent
the last of which follows from “1 2 3 STEP ‘plus’ ‘times’” and “1 2 3 6 STEP ‘times’”.
Thus the definition of resultant of a pair of predicates enforces a method of program composition at a very early stage without ever mentioning such complex theories.
The drawback is but momentary: the definition of resultant doesn’t appear to have the features familiar to those who were erroneously trained as mathematicians or computer scientists to deal with the composition of two place relations rather than predicates and rather than predicates regimented on the expedience of stack notation.
The intermediate definition of composition is introduced now (it was already hidden in the iterate abbreviations):
I still refuse to give any abbreviation what so ever for truth functional and quantificational schema like
or, better, their predicate abstract logic analogue
which is closest to Godel 1940 and Peano 1911 which are to be greatly preferred, as Quine strongly suggests on pages 24 and 25 of ‘Set Theory and its Logic Revised Edition’, to modern methods (which were the methods by which I was originally instructed when I was taught a strictly deductive mathematical logic).
With the predicate functor ‘compose’ in hand, the abbreviation of the predicate functor of iteration is much shorter (and more familiar):
The predicate functor ‘compose’ also comes in handy when giving better abbreviations for the predicate functor of functionality:
Note this new abbreviation brings ‘func’ into harmony with the other three place predicate functors ‘inject’, ‘surject’, ‘biject’, and ‘corr’. This is a grand improvement over the abbreviations in #2025-0828-2145.
Oh, it seems there are also abbreviations for ‘cw’ and ‘ccw’ that I have yet to include:
There are perhaps others of similar familial resemblence which I have accidentally omitted: this is a large part of why I distinguish between notes and memos! Memos set a high bar.
In preperation for the schematic order abbreviations of logic:
Having set up the world and camera coordinate frames in #2025-0829-1227 there remain two coordinate frames to set up (those of the image plane and the image sensor) before solving the problem of calculating the point projected through a pinhole from the world onto the sensor, better the receptor, can be solved.
The third basis vector of the camera coordinate frame points into the dark room or box of our camera obscura. All the work put into setting up the world and camera coordinate frames is collected under the phrase ‘extrinsic parameters’ since they deal with stuff that is outside the dark room.
Even though I didn’t overtly say that the points of the world to be projected through the pinhole are to be outside the room, I’ve tacitly made descriptions of the projected points as being exterior when I spoke of the trunk of a tree being projected upside down onto the back wall of the dark room. Now that I’ve overtly required that the points to be projected be outside the room, that can be added to the list of things that fall under the premises of the forward model of pinhole cameras.
The remaining parameters to be explained are called ‘intrinsic parameters’ becuase they deal with the interior of the camera obscura.
Just as the first measurement in the world coordinate frame was the location of the origin of the camera coordinate frame with respect to the origin of the world coordinate frame, so too is the first measurement in the camera coordinate frame the origin of the image plane coordinate frame the first measurement in the camera coordinate frame.
The line along the third basis vector of the camera coordinate frame (the one whose tail is at the pinhole and that points into the dark room and is perpendicular to the window with the pinhole in it) is called the optical axis.
The optical axis intersects with the back wall of the dark room. The back wall of the dark room is called the image plane because it is the two dimensional surface on which the points projected by the pinhole are imaged.
The point where the optical axis intersects with the image plane is called ‘the principle point’ (of the camera).
In a standard camera (or in a standard room) the back wall is perpendicular to the wall with the pinhole in it so that the optical axis is perpendicular not only to the image plane of the back wall but also to the window with the pinhole in it. Complex cameras can include elaborate mirrors (and even more elaborate lenses) that bounce projected points that would otherwise hit the back wall towards some other surface: nothing like that is contemplated here.
The distance from the pinhole to the principle point is called the ‘effective focal length’ of the camera, or just the focal length.
Since the third basis vector of the camera coordinate frame points right at the principle point from the pinhole, then the coordinate of the principle point in the camera coordinate frame is ‘(0, 0, the focal length)’.
The principle point is the origin of the image plane coordinate frame.
The image plane coordinate frame has only two basis vectors: only two measurements are needed to locate any point on the back wall with respect to the principle point i.e. with respect to the origin of the image plane coordinate frame.
Just as some basic conventions constrained the way the camera coordinate frame was set up, there are basic conventions when setting up the image plane coordinate frame that keep calculations nice and tidy.
The first and second basis vectors of the image plane coordinate frame are parallel to the first and second basis vectors of the camera coordinate frame. So, there are no more than two ways each of them can point i.e. they can point in the same direction as the bases vectors of the camera coordinate frame or in the opposite direction of the bases vectors of the camera cordiante frame.
Because of how image sensors work, and because of how images are drawn on computer screens (a pixel at a time), the bases vectors of the image plane coordinate frame are made to point in the opposite directions of those of the camera coordinate frame.
So, since the second basis vector of the camera coordinate frame points up, then the second of the image plane coordinate frame points down, and since, when looking through the pinhole to the back wall, the first basis vector of the camera coordinate plane points left, then the first of the image plane coordinate frame points right (when peeping through the pinhole to the back wall).
To recap:
With the schematic compositional and functional abbreviations completed in #2025-0828-2145 I move on to the schematic graph abbreviations of logic.
But, first, it occurred to me that I forgot to put two fundamental abbreviations in the sameness section:
Though they may seem silly in their brief insignificance, they complete a wide range of validities that are often misinterpreted as algebraic identites.
Notes on ‘The Little Book of Big History’ by Ian Crofton and Jeremy Black 2017
Notes on ‘Geography’ by Crash Course.
When organizing my reading, writing, and thinking on my history of the world, I have found geography to be extremely helpful e.g. its divisions of the planet into the litho-, hydro-, bio-, and atomo- spheres. It is funny, in a good way, that the division of the world, e.g. by Aristotle, into earth, water, air, and fire lives a ghostly life through this broad organizational scheme of geography.
I do not mean to imply that the division of the world into its four spheres is the one that I shall settle on. For me, the world is still to be organized around its contingencies where a contingency is a consequence from a response on an occasion e.g. a door opens from a push on a lever, a heart beats faster from a kiss on a cheek, and mountains rise from shifts on tectonic plates.
I’m disinclined to fixate on the ‘space’, ‘place’, etc. principles of geography because they mismanage the social contingencies which are better handled on the side of the science of behavior (though, perhaps this conclusion will be contravened by contemplations recorded here).
“Physical geograph is all about recognizing the characteristics of teh environment and the processes that create, modify, and destroy those environments” (1:35 ‘What is physical geography? Crash Course Geography #4’)
Ecosystems are composed of the interconnected systems of the atmo, hydro, litho, and bio spheres.
What causes these systemst to interact? The sun and its insolation (incoming solar radiation).
Topography: shape of the land.
Geomorphology: the origin and evolution of the shape of the Earth’s surface e.g. weathering, glaciers, streams.
Pedology: study of soil types and how they form.
Hydrology: how water is moved managed and distributed above and below the Earth’s surface.
Climatology: study of atmospheric weather patterns over time
Oceanography: study of the past present and future features of the oceans
Meteorology: study of atmospheric processes and phenomena
Biogeography: distribution of plants and animals (fauna and flora) in an area
Ecologists study the biosphere: what physically happens as different species interact across the landscape
When geographers study the biosphere they focus on how ecological processes are distributed across space, and how species move and change over time. They focus on spatial patterns in the landscapes.
13.7B years ago the universe began!
4.5B years ago in the Milkyway Galaxy a solar nebula collapsed under its own gravity: most of its mass went to the center as what we now call the sun, and the rest, a very tiny amount, spun out and formed the rest of the solar system.
That tiny amount spun out the Earth and other planets: they rotate counterclockwise because they are all gravitationally condensed parts of that tiny bit of the collapsed solar nebula that made our sun.
As a consequence of the solar nebula collapse, the earth rotates around the sun and rotates around a central axis which passes through the north and south poles. A full rotation occurs once about every 24 hours.
At the equator the Earth spins at 1.6k kilometers per hour (km/h) which is about 13 times faster than a Cheetah can run at top speed.
In St. Petersburg, at 60 degrees North latitude, the spin speed is half that at the equator i.e. about 800 km/h.
At the poles, it is 0 km/h.
The sun moves in the sky because of the relative motion of an observer on the surface of the Earth. From the moon it is easy to see the Earth rotating because the moon rotates on its axis at the exact same rate it orbits Earth i.e. on Earth we always see the same side of the moon.
The Circle of Illumination: the dividing line that seperates the half of Earth that’s currently receiving light and solar energy from the other half that’s in total darkness.
The path of the Earth around the sun is called an orbit: a revolution (traveling from one point on the orbit and back to it) takes about 365 days.
Plane of the Ecliptic: the imaginary plane that contains Earth’s orbit
The Earth is tilted 23.5 degrees from a line perpendicular to the plane of the ecliptic through the center of the Earth (probably happend when something hit Earth billions of years ago).
Insolation: incoming solar radiation. It’s how the different Earth systems.
In note #2025-0828-1649 we set up the world coordinate frame with a stick we found next to the tree whose upside down projection first presented the problem of modeling a pinhole camera. With the counterclockwise bases vectors we were able to measure where points are in the world coordinate frame.
The task is still to calculate the point projected through the pinhole in the window from the tree onto the back wall of the dark room.
The first point measured from the origin of the world is the location of the pinhole i.e. the world coordinates of the camera.
The world coordinates of the camera are also the world coordinates of the origin of the camera coordinate frame! So, all together there are two origins: the origin of the world coordinate frame and the origin of the camera coordinate frame.
The origin of the camera coordinate frame also comes with three basis vectors. Unlike with the bases vectors of the world coordinate frame, the bases vectors of the camera coordinate frame come with some important constraints i.e. we can’t just pick up a stick and set up a counterclockwise coordinate frame by turning it at right angles around the origin willy nilly.
The most important basis vector of the camera coordinate frame is the third one: it points perpendicular to the window with the pinhole in it and it points into the dark room. The line through the third basis vector of the camera coordinate frame is called ‘the optical axis’.
The second most important basis vector is, surprise, the second one. It’s often called ‘the up vector’ because when you’re using a camera without twisting or turning it in interesting, and perhaps even artistic ways, it is perpendicular to the optical axis and perpendicular to the ground. I’ve also seen it called ‘the top vector’ because it is perpendicular to the top plate of a camera.
The least important basis vector is the first and it is set up by requiring that the camera coordinate frame be a counterclockwise one i.e. it is perpendicular to the third basis vector (the optical one) and the top or up vector and if you go from the first to the second and then the third basis vector it makes a counterclockwise path along the cricle through their tips when you look through that circle at the origin.
To recap:
With the schematic sameness abbreviations of logic written out in #2025-0826-1442, I move onto the schematic compositional abbreviations.
Now for the functional abbreviations.
Work on a memo on how to program your own virtual worlds from scratch.
Light bounces off a tree and passes through a tiny pinhole in a very dark room and is projected onto the back wall. Surprisingly, the image of the tree on the wall is upside down.
If someone covers up the pinhole and carefully measures, e.g. in meters, where a point on the tree is relative to the hole in the window, can you mark on the back wall exactly where that point will be projected when the pinhole is uncovered?
Setting up and solving this problem produces the forward imaging model of a pinhole camera: how to go from a point in the world to a point on an image sensor, e.g. the back wall of the dark room.
When you try to go from a point on the image sensor, e.g. the back wall, to a point in the world you get the backward imaging model of a pinhole camera. Interestingly, at least two camera obscura (the fancy word for ‘pinhole camera’) are needed to calculate the location of a point in the world!
Technically, the forward imaging model transforms the world coordinates of a point into the image sensor coordinates of its projected point.
I’ll do my best to explain everything as concretely as possible and first walk through each of the steps that lead from the simple camera obscura experiment to the simple geometry and algebra that goes into drawing virtual worlds on a screen and controlling them with your mouse and keyboard.
This little project is one of the better ways to see how the methods of linear algebra solve all sorts of practical problems using only a handfull of vector and matrix methods. But, I’m not going to focus on the linear algebra at all, because it isn’t actually needed to navigate the steps that make up the forward imaging model of a pinhole camera.
In fact, the forward imaging model is not even linear at first! It can be turned into a linear problem by the beautifully general methods of homogenous coordinates. No need to worry about that now though.
For those who aren’t scared off by seeing technical terms they do not yet know feel free to read any sentence that starts with ‘technically’. Otherwise, feel free to skip over any such sentence.
The first step in setting up the pinhole camera problem is to pick a point in the world and call it the origin. More than one origin turns up as we go along, so this one will be called ‘the world origin’ or ‘the origin of the world’, or, technically, the origin of the world coordinate frame.
The world coordinate frame is a fancy way of mentioning how we’re going to measure things from the origin (of the world!) e.g. how we’re going to figure out where a point on the tree is at in the world with respect to the origin.
Pick up a stick and put one end at the origin and the other end of that stick now becomes the end point of our first basis vector. Vectors are no more complicated than that: one point is their tip and the other point is their tail.
What makes the origin and the other end of the stick into a basis vector is that we’re going to use the length of the stick as our basic or basis unit. Everything else we do will be in basic stick units.
If we happen to use a meter stick, then other people with a meter stick can more easily recreate all the measurements we’ll need to make. There is nothing more special to a meter than that.
Technically, the meter is special in that it is convenient when making physical calculations. Even more technically, there are stacks of papers on how we arrived at the standard units of measurement used throughout the sciences and you can read more by searching for them.
The line that goes through the first basis vector of the world coordinate frame is called ‘the x-axis’, or, more precisely, ‘the world x-axis’, or, technically, ‘the x-axis of the world coordinate frame’.
The word ‘coordinate’ in the phrase ‘the world coordinate frame’ refers to the triple of numbers that we’ll use to write down our measurements once we’ve set up our world coordinate frame e.g. ‘(2,3,4)’ designates a triple whose first coordinate tells us to go two stick units in the direction of the first basis vector (and along the x-axis) as the first step in finding the point with coordinate (2,3,4) in the world coordinate frame.
By keeping the tail of our stick at the origin of the world and rotating it at a right angle from the first basis vector, we get the point at the tip of the second basis vector. The line that goes through the second basis vector is called the ‘y-axis of the world coordinate frame’.
Especially the series “Antikythera Fragments - Ancient Tool Technology”
But, if you just want to see how geometric constructions work, then its best to look for YouTube videos on drafting techniques or those on geometric constructions outright e.g.
To get the third basis vector we move the stick so that it is perpendicular to the first and second basis vector. Sadly, there are two ways that you can do this! One is called the clockwise way and the other is called the counterclockwise way.
To figure out whether the third basis vector makes a clockwise or counterclockwise coordinate frame you have to draw a circle through the three points at the tips of each basis vector and then look at the origin through the circle and see if when you go from the first to the second to the third basis vector you go clockwise or counterclockwise around the circle.
We want a counterclockwise frame. It is most often known as a ‘right handed frame’ becuase if you put out your thumb, pointer finger, and middle finger on your right hand as if you were making them into the first, second, and third basis vector then you make a counterclockwise path around the points at their tips.
The three triples ‘(1,0,0)’, ‘(0,1,0)’, and ‘(0,0,1)’ designate the tips of the first, second, and third basis vector. Any point in the world can be reached, in only one way, by going out along the first axis, then the second, then the third and writing out the number of steps along each basis vector in a triple e.g. ‘(3,-4,5)’ designates the point at three units out from the origin in the direction of the first basis vector, then negative four units out from the origin in the direction of the second basis vector (i.e. four units out from the origin in the opposite direction of the second basis vector), and five units out in the direction of the third basis vector.
These triples of numbers are all that you get when you’re calculating where the point designated by such triples is projected on the back wall through the pinhole. They can be more than just integers e.g. ‘(3.5, -4.32, 1.6)’.
Paul Graham is reading the 1721 edition of “The Works of the Right Hounourable Joseph Addison, Esq” a four volume work on the cofounder of The Spectator and whose books were banned by the Catholic Church until 1966.
I originally linked to the 1761 editions when I wrote up a tweet giving the full name of ‘Addision’ as ‘Joseph Addison’ because I didn’t know of him. While I had a weak response to ‘The Spectator’ I’m pretty sure it was just that there is a modern magizine called ‘The Spectator’.
Notes on Chapter 1 ‘Nature, Humanity, and History, to 3500 BCE’ from 2023 “The Earth and Its Peoples 8th ed.” (EP) by Richard Bulliet, Pamela Crossley, Daniel Headrick, Steven Hirsch, and Lyman Johnson.
“A creature with a body much like that of modern humans but with a face that had heavy brown ridges and a low forehead, like the faces of apes.” EP pg. 5
In 1871 Charles Darwin (1809–1882) published ‘The Descent of Man and Selection in Relation to Sex’ which extended his 1859 book ‘On the Origin of Species by Means of Natural Selection, or the Preservation of Favoured Races in the Struggle for Life’ to the evolution of humans and predicted that humans first appeared in Africa because of their similarities to African apes.
In 1891 in the Indonesian island of Java Eugene Duboi’s team found the 700,000 and 1,490,000 year old skullcap of ‘Java man’ (Homo erectus erectus).
In 1924 “Raymond Dar found the skull of a creature he named Australopithecus africanus (aw-strah-loh-PITH-uh-kuhs ah-frih-KAH-nuhs) (African southern ape), which he argued was transitional between apes and early humans” (EP pg. 6).
In 1929 near Peking (old Beijing) China, W. C. Pei found the skullcap of ‘Peking man’ and it was similar to the one found by Duboi.
“Since 1950, Louis and Mary Leaky and their son, Richard, along with many others have discovered a wealth of early human fossils in the exposed sediments of the Great Rift Valley of eastern Africa.” (EP pg. 6)
“Biologists classify australopithescines (aw-strah-loh-PITH-uh-seen) and humans as members of a family of primates known as hominids (HOM-uh-nid) .. The first hominids are now dated about 7M years ago.” (EP pg. 6)
Between 2M and 3M years ago, climate selected genus Homo (to which modern human belong).
In 1960s Olduvai (ol-DOO-vy) Gorge of northern Tanzania, Louis Leakey found fossilized Homo habilis (HOH-moh HAB-uh-luhs) (handy human) whose brain was about 50\% larger than that of australopithecines: “Seeds adn other fossilized remains found in ancient Homo habilis camps indicate that the new species ate a greater variety of more nutritious foods than did australopithecines.” (EP pg. 7)
No later than 80-50k years ago did verbal communities of humans appear. It could be much earlier because the science of behavior has yet to be effectively applied to history outside of what similarities it shares with experimental anthropology.
Notes on chapter 2 ‘The first River Valley Civilizations, 3500-1500 BCE’ of
“Certain political, social, economic, and technological traits are usually seen as indicators of civilization:
- cities as administrative centers,
- a political system based on control of a defined territory rather than kinship connections,
- many people engaged in specialized, nonfood-producing activities,
- status distinctions based largely on accumulations of substantial wealth by some groups,
- monumental buildings,
- a system for keeping permanent records,
- long-distance trade,
- major advances in science and the arts.
The earliest societies exhibiting these traits developed in the floodplains of great rivers: the Tigris (TIE-gris) and Euphrates (you-FRAY-teez) in Iraq, the Indus in Pakistan, the Yellow or Huang He (hwang huh) in China, and the Nile in Egypt.” EP pg. 28-29
With a few simple edits to the abbreviations laid out in #2025-0825-1745 I can more easily reach universal and existential closures while also making way for indiscernibility. The introduction of the abbreviations for indiscernibility are unhappy, but I’d rather get on with it than labor over notation.
‘Some F’ for ‘not Some not F’
That completes the new methods of introducing the Boolean abbreviations. Now for the sameness abbreviations of indiscernibility and (lexical) identity:
‘nid i’ for ‘not id i’
I have all but one or two books after the 24th and none before that one. While I may buy the missing books online, I may also just leave them as little quests.
One thing I know for sure is that this series contains many of the texts that I’ve either read online or have wanted to read e.g. I already dove into Spinoza’s 1677 ‘Ethics’ (written from 1661-1675) and Descarte’s 1701 ‘Rules for the Direction of the Mind’ (written from 1628).
I also got the entire compendium of ‘Harvard Classics’ for $30.
There should be more series like this that are less encyclopedic but more exaustive.
The methods of logic have yet to be adequatly applied to history.
Went to the bookfar again: bought over 30 books for $25.
A grammatically simpler and uniform prefix notation shall be adapted as a result of the notational exploration done in the previous note.
The uniform grammar enforces a prefix notation on composing predicate functors. It eliminates a lot but not all parentheses but they can always be included to guide the eye even when they are otherwise redundant. Each of the predicate functor abbreviations given above are given below with the new notation.
Each predicate functor has a number of places (just like each predicate has a number of left places and a number of right places). For example, huh is a two place predicate functor (just as ‘fathered’ is a 1 1 place predciate since it has one left place and one right place, and ‘pairs’ is a 1 2 place predicate since it has one left place and two right places). Knowing the number of places of a predicate functor makes chaining them together require no parentheses e.g. ‘huh huh F G H’ is equivalent to ‘huh (huh F G) H’ (remember that most parentheses are there to help your eye) and ‘huh huh F G huh H F’ is equivalent to ‘huh (huh F G) (huh H F)’.
I shall continue to keep the infix notation of familiar truth functional (better, denotative functional) and quantificational logic e.g. ‘huh F G’ is equivalent to ‘(..u: some v is such that ..uxvF..y nor ..uxvG..y :x..y)’ just as it was with the old predicate functor notation.
The complementary converse conditional is one of the first examples where proper notational methods make special names superflurous e.g. there is never a need to write ‘not-if F G’ since it is equivalent to ‘not if F G’. This convenience is familiar to any Forth programmer. If you’re a Forth programmer then you can think of predicate functors as defining words like colon and semicolon.
‘not-only-if F G’ and ‘only-if F G’ can both be written as ‘not only if F G’ and ‘only if F G’ by taking ‘only’ as a predicate functor functor, but I shall not do that at this time. Not only does it confuse this prefix notation (which, with appropriate parentheses, is familiar to LISP programmers), it also prompts erroneous talk of ‘higher order predicate functor logics’ which are as bad as talk of ‘higher order truth functional and quantificational logics’.
Went to a bookfair, bought over 60 books for $50.
Anyone reading this note is assumed to have read the slower and gentler explanations in #2025-0823-1417 on the evolution of predicate functor logic from truth-functional and quantification logic thru predicate abstracts and the schematic premise of abstraction and concretion.
Wang’s schematic premise of identity theory ‘Fx if and only if some y is such that Fy and x=y’ implies the schematic biconditionals of concretion
and round out the role of relative clause
Technically, ‘v’ must be replaced by some varaible fresh to the context.
i.e.
Happily, in composing 9.9 I stumbled upon the name ‘cedent’ for what I and most have called ‘sequents’ and it yielded a beautiful way of reading cedents alound as ‘cede’. Happy linguistic accidents are often the best.
This dot notation is experimental and may be replaced by something like
- ‘(1+n)( ~~~ )’ for ‘~~~ (n)( ~~~)’
- ‘(1+n ~~~)’ for ‘~~~ (n ~~~)’
Doing each of the recombic functors above made me notice that I had written the order of predicate functors in the recombic abbreviations backwards! This is something that happens from time to time because there is a bit of confusion when working from the inside out and the outside in with predicate functors and then going over and working with the functional predicates that carry over the recombic functors like the remainder of the quotient of arithmetic operations divided by two do for the truth functional functors (better, denotative functional functors).
That completes the schematic denotative functional, Boolean, quantificational, and recombic abbreviations of logic.
In an attempt to write out the name of the chapter mentioned in the the 19th item of the previous note, I was thrown into reading a number of selections from Russell’s “An Inquiry into Meaning and Truth” (IMT) as a result of having read and made notes in #2025-0817-1523 on “Russell’s Logicism and Theory of Coherence” by Conor Mayo-Wilson in Russell: The Journal of Bertrand Russell Studies. Vol. 31. Summer, 2011. pp. 89-106.
“The third, that of coherence, was discussed and rejected in Chapter X.” pg. 400
“This raises the question: what did Russell understand to be the philosophical importance of logicism? Building on recent work by Andrew Irvine and Martin Godwyn, I argue that Russell thought a systematic reduction of mathematics increases the certainty of known mathematical theorems (even basic arithmetic) facts by showing mathematical knowledge to be coherently organized. The paper outlines Russell’s theory of coherence, and discusses its relevance to logicism and teh certainty attributed to mathematics.” pg. 1
IN recent philosophy we may distinguish four main types of theory as to ‘truth’ or as to its replacement by some concept which is thought preferable. These four theories are:
- The theory which substitues ‘warrented assertability’ for ‘truth’. This theory is advocated by Dr. Dewey and his school.
- The theory which substitutes ‘probability’ for ‘truth’. This theory is advocated by Professor Reichenbach.
- The theory which defines ‘truth’ as ‘coherence’. This theory is advocated by Hegelians and certain lgoical positivits.
- The correspondence theory of truth, according to which the truth of basic propositions depends upon their relation to some occurrence, and the truth of other propositions depends upon their syntactical relations to basic propositions.
For my part I adhere firmly to this last theory. It has, however, two forms between which the decision is not easy. In one form, the basic propositions must be derived from experience, and therefore propositions which cannot be sutably related to experience are neither true nor false. In the other form, the basic propositions need note be related to experience, but only to ‘fact,’ though if they are not related to experience they cannot be known. Thus the two forms of the correspondence theory differ as to the relation of ‘truth’ to ‘knowledge’.
Of the above four theories, I have discussed the third in chapter X;” pg. 362-363 Russell IMT
“But what do I know?
- What I see when I look at them
- What I hear when others read them aloud
- What I see when others quote them in print
- What I see when I compare two copies of the same book.” pg. 180
The experimental analyst of behavior has spent a great deal of time and energy removing such blind spots i.e. in making clear that such naive certainty is extinguished the moment the behavior of organisms is examined in a laboratory environment and is reinforced by vistigial literary environments found in prescientific libraries.
This particular quote from the tenth chapter is also an indication of Russell’s outlook: that which is certain is that with which he is in direct contact.
“Assuming, as I shall do henceforth, that there are basic propositions, it seems to me that, for theory of knowledge, ‘basic propositions’ may be alternatively defined as ‘those propositions about particular occurrences which, after a critical scrutiny, we still believe independently of any extraneous evidence in their favour.’
Let us consider the clauses of this definition, and let us begin at the end. There may be evidence in favoour of a basic proposition, but it is not this evidence alone that causes our belief. You may wake up in the morning and see that it is daylight, and you may see from your watch that it is daylight, and you may see from your watch that it must be daylight. But even if your watch pointed to midnight, you would not doubt that it is daylight. in my scientific system, a number of propositions based on observations support each other, but each is capable of commanding belief on its own account. Moreover mutual support among basic propositions is only possible on the basis of some theory.
There are cases, however-chiefly where memory is concerned– in which our belief, though not inferential, is more or less uncertain. In such cases, a system composed of such beliefs wins more acceptance than any one of them singly. I think Mr. Z. invited me to dinner on Thursday; I look in my diary, and find an entry to that effect. Both my memory adn my diary are fallible, but when tehy agree I think it unlikely that they are both wrong. I will return to this kind fo case later; for the present, i wish to exclude it. It is to be observed, meantime, that a non-inferential belief need not be either certain or indubitable.” 187-188
“We thus arrive at the momentary perception as the least questionable thing in our experience, and as therefore the criterion and touchstone of all other certainties and pseudocertainties.
But for theory fo knolwedge it is not sufficient that we should perceive something; it is necessary that we should express what we perceive in words. now most objectwords are condensed inductions; this is true of teh word ‘dog,’ as we have already had occasion to notice. We must avoid such words, if we wish to be merely recording what we perceive. To do this is very difficult, and requires a special vocabulary. We have seen that this vocabulary includes words such as ‘red,’ and relation-words such as ‘precedes,’ but not names of persons or physical objects or classes of such terms.
We have considered the subject of ‘basic propositions’ or ‘Protokollsatze,’ and tried to show that empirical knowledge is impossible without them. it will be remembered that we defined a ‘basic proposition’ by two characteristics:
- It arises on occasion of a perception, which is the evidence for its truth;
- It has a form such that no two propositions having the form can be mutually inconsistent if derived from different percepts.
A proposition having these two characteristics cannot be disproved, but it would be rash to say that it must be true.
Perhpas no actual proposition quite rigidly fulfils the definition. But pure perceptive propositions remain a limit to which we can approach indefinitely, and the nearer we approach the smaller is the risk of error.
Empirical knowledge requires, however, other premises asserting matters of fact, in addition to pure perceptive propositions. I shall give the name ‘factual premiss’ to any uninferred proposition which asserts something having a date and which I believe after a critical scrutiny. I do not mean that the date is part of theassertion, but merely that soem kind of temporal occurrence is what is involved in teh truth of the assertion.
Factual premisses are not alone sufficient for empirical knowledge, since most of it is inferred. We require, in addition, the premisses necessary for deduction, and those other premisses, whatever they may be, that are necessary for the non-demonstrative inferences upon which science depends.” 189-190 IMT ch. ‘factual premises’
“It must be held, on logical grounds, that no occurrence gives demonstrative grounds in favor of belief in any other occurrence. But the grounds are often such as we cannot fail to accept as giving practical certainty. We saw taht there can be no reason for disbelieving the proposition ‘that is red’ when made in the presence of a red percept; it must, however, be admitted that belief in this proposition is logically possible in teh absence of a red percept. Such grounds as exist for supposing that thsi does not occur are derived from causal laws as to teh occurrence of language. We can, however, in theory, distinguish two cases in relation to a judgement such as ‘taht is red’: one, when it is cause by what it asserts, adn the other, when words or images enter into its causation. In the formercase, it must be true; in the latter, it may be false.” pg. 199-200 ch. ‘factual premises’
“It must be remembered that a factual premiss need not be indubitable, even subjectively; it need only command a certain degree of credence. It can therefore always be reinforced if it is found to harmonize with other factual premisses. What characterizes a factual premiss is not indubitabilitiy, but the fact that it commands a greater or less degree of belief on its own account, independently of its relations to other propositions. We are then led to a combination of self-evidence with coherence: sometimes one factor is very much more important than the other, but in theory coherence always plays some part. The coherence required, however, is not strict logical coherence, for factual premisses can and should be so stated as to be deductively independent of each other. The kind of coherence invovled is a matter which I hope to consider in a later work.” pg. 201-202 ch. ‘factual premises’, Russell ‘Inquiry into Meaning and Truth’
Work on potential memo ‘Stack Notation for Predicate Functor Logic’ set to strengthen #a-stack-notation-for-predicate-functor-logic-2025-0414-1626.
Stack notation for predicate functor logic combines the utility of the Forth and J programming languages with Quine’s elimination of variables from truth-functional and quantificational logic via the additional methods of predicate abstraction and concretion as the logical import of the grammarian’s relative clause.
Instead of a predicate having a certain number of crossreferenced places wherein given names are predicated of it, each predicate has a number of left and a number of right places.
For example, the classic two place predicates ‘father of’, ‘less than’, and ‘member of’ each have one left place and one right place and are therefore called ‘(1,1) place predicates’.
Predications of (i,m) place predicate letters are depicted schematically by prepending i variables and appending m variables to a predicate letter e.g. ‘uFx’, ‘uvGxyz’, and ‘uvwHxyz’.
Thus, a predication such as ‘a pairs b with c’ is rendered as ‘aPAIRSbc’.
What I call ‘free sentences’ are called by Quine, and most others, as ‘open sentences’.
The shift in terminology is justified by noting that what once went as an open sentence, e.g. ‘x is a member of y’, was either taken as a closed sentence where the free variables ‘x’ and ‘y’ were playing the part of names (singular terms) purporting to designate one and only one item throughout a given logical demonstration, or were properly paraphrased as predicates by supplmenting a suppresed prefix of predicate abstraction, e.g. ‘x and y are such that x is a member of y’.
Thus, a free sentence is free to be the component of a predicate abstract or to stand alone as a proper (closed) sentence containing proper names (singular terms).
The schematic biconditionals of concretion can themselves be given as the starting point of a truth functional, predicate abstract, and predicate functor logic, but that shall not be shown here.
Concretion makes each predicate coextensive with each predicate abstract of its predications so that the shapes of predicates and predicate letters in a stack based language can be depicted by a predicate abstract of a predication like ‘aPAIRSbc’ e.g. ‘{xyz: xPAIRSyz}’.
But, the predicate abstract ‘{xyz: xPAIRSyz}’ is not only misleading to those who see the curly brackets as indicating what is most commonly called ‘set builder notation’, but also does not preserve the shape of the predicate shown i.e. ‘aPAIRSbc if and only if {xyz: xPAIRSyz}abc’ suggests that the predicate abstract ‘{xyz: xPAIRSyz}’ has the shape (0,3) whereas its purportedly coextensive predicate ‘PAIRS’ has shape (1,2).
In languages where variables abound, this confusion over shape is obviously eliminated by further and familiar regimentation: ‘xPAIRSyz’ is regimented as ‘PAIRSxyz’ or, with an eye towards the methods of stack based langauges, ‘xyzPAIRS’.
When it comes to eliminating variables and perhaps even introducing them from a variable free predicate functor logic, the methods of stack based languages do pay their way, but to those unfamiliar with overtly stack based langauges, e.g. Charles H. Moore’s Forth programming language or Iverson and Hui’s J programming language, or with the covertly stack based langauges, e.g. John McCarthy and Steve Russell’s LISP programming language, there may appear as if there is little to recommend with such divisions of predicates into their left and right parts.
Interestingly, it was Russell in “An Inquiry into Meaning and Truth” and “Human Knolwedge: its scope and limits” where the problems of analysis, e.g. the logic of a predicate ‘part of’, and individuation are brought down to a division of the world, or at least what can be seen by an eye, into its left and right parts.
Variations on the curly bracket notation abound e.g.
and such grander deviations as
or such pedantic (and wild) deviations as
I do sincerely hope that all this talk fo alternative notations justifies my inclination to avoid notation all togehter for well factored and carefully compounding abbreviations.
Technically, ‘v’ must be replaced by some varaible fresh to the context.
with the same technical stipulation on ‘v’.
Warning! The predicate functor ‘huh’ is an amalgam of what is otherwise presented as three seperate predicate functors: ‘push’, ‘some’, and ‘nor’. Following the changes in shape when they are shown in both parts of the abbreviation can be confusing and can otherwise be worked out, with some effort, from the work just shown.
Technically: given that the reader is careful about the additional constraints on where and when a given variable can be introduced. Such constraints can be formally deduced by going all the way back to the truth functional and quantificational abbreviation of predicate abstracts and the relevant restraints on substutitutions of variables between truth functional and quantificational schema.
Such technicalities are avoided by assuming only what has already been weakly at work with notation such as ‘..x’ i.e. the various parts are to be filled out into the appropriate shapes based on context clues.
While I’ve gone ahead and expanded ‘huh F F’ into the predicate abstract it abbreviates ‘(..u: some v is such that ..uxvF..y nor uxvF..y :x..y)’ so that ‘uh F’ can be written directly as the predicate abstract it abbreviates, I may not always do so, and when I don’t it is appropriate to work it out for yourself.
There is one step in unraveling these abbreviations that I have not overtly shown and so I shall go through the steps of first getting down to the predicate abstract abbreviated by ‘not F’ and then the one abbreviated by ‘push F’.
This unraveling explains to the native speaker of truth functional and quantificational logic (with its abbreviation for predicate abstracts) what it is that the predicate functor logician is talking about when they use the predicate functor of negation as opposed to the truth function of negation.
Thus, ‘push F’ for ‘not uh drop F’ i.e. ‘(..u: ..uxF..y :x..y)’.
The notes in #2025-0821-2007 on the first volume of the “story of civilization” by the Durants sparked the collection of the references to be collated into my history of the world.
The center piece, for better or for worse, is 2023 “The Earth and Its Peoples 8th ed.” by Richard Bulliet, Pamela Crossley, Daniel Headrick, Steven Hirsch, and Lyman Johnson.
A major task is collating the chronology.
I have the remaining books in the Durants’s volumes, and the other two history texts by Susan Wise Bauer. I also have a few texts on the history of philosophy, political philosophy, and histories of economics that must also get folded into my history.
I shall start composing rather than just improvising at the piano. This is largely due to a conversation that I had with a friend.
Continuing work on what I’m now calling “the graph abbreviations” for the memo on predicate logic.
Before I began itemizing my notes and flattening entries down to the hour (if you scroll down far enough you’ll find entries broken into days and then days into entries identified down to the minute), I was writing long paragraphs of notes on the things that I was reading. I stopped doing that because it was unsustainable, and made navigating such entries almost impossible.
I begin again with a new book.
Will and Ariel Durant’s 1936 “The Story of Civilization: Part I Our Oriental Heritage”.
In American English, ‘oriental’ can happily be replaced with ‘asian’.
Since the science of behavior only appeared less than a century ago, there is little to no history of individual behavior which comes close to that which is necessary for a theoretical analysis.
At best, the histories which have been assembled since reports of events began being recorded textually can provide some foundation for a theoretical analysis of cultural scales of change.
The social practices which composed the cultures of the world throughout history are know to us by their often deferred consequences. The past is remote and histories are our weak link to them.
The contingencies which select and maintain social behavior compose a social environment and the shape of the social environment is often mentioned by talk of cultural or social practices e.g. the economic, political, moral, and mental elements mentioned in the Durants’s volumes.
“After fifty thousand years of art men still dispute as to its sources in instinct and in history. What is beauty?– why do we admire it?– why do we endeavor to create it? Since this is no place for psychological discourse we shall answer, briefly and precariously, that beauty is any quality by which an object or a form pleases a beholder. Primarily and originally the object does not please the beholder because it is beautiful, but rather he calls it beautiful because it pleases him. Any object that satisfies desire will seem beautiful: food is beautiful– Thais is not beautiful– to a starving man.” pg. 82
The final sentence, ‘food is beautiful to a starving man’ is the crux of any effective theoretical analysis of cultural practices: as far as we can identify the contingencies of reinforcement (or punishment) that prevailed during a given time, most often by examining the artifacts which helped to maintain social contingencies in the absense of an other organism medaiting their consequences, is as far as we can take whatever history has to offer us.
I hope that as the science of behavior supplants the prescience of mind (and its modern reincarnation in cognitive, information, and systems theories) there will be more that the history of today can give to the people and cultures of tomorrow.
Until then, I have only the assembled histories available to me e.g. the volumes from the Durants and 2023 “The Earth and Its Peoples 8th ed.” by Richard Bulliet, Pamela Crossley, Daniel Headrick, Steven Hirsch, and Lyman Johnson and 2014 “History of the World 6th ed.” by J. M. Roberts and Odd Arne Westad.
It is surprising to many people that science and technology change the control we have over history. The popular story is that history has happened, and that there is some correspondence between the past and its history. Whatever connection there is is established by science in the same way that any other records of events are connected to other records of events.
It is better for me to begin with what textual records there are than to speculate on the scope and limits of such records: such theoretical speculation is for a much later time.
Book Three ‘The Far East’ is broken into two parts: China and Japan.
This particular edition doesn’t have the printed pictures within the text: it does have black and white images at the front of the book. The back and front covers have endsheets that display simple maps of the larger regions dealt with in the text: there is not a one-to-one match between places mentioned in the text and places labeled on the maps.
Since this history is a little less than a century old, there is an additional problem of collating its reports with those made by modern texts, such as the eigth edition of ‘The Earth and its Peoples’.
https://www.etymonline.com/word/civilization
civilization(n.)
1704, in a now-obsolete sense “law which makes a criminal process civil,” from civil + -ization.
The meaning “civilized condition, state of being reclaimed from the rudeness of savage life” is recorded by 1772, probably from French civilisation, serving as an opposite to barbarity and a distinct word from civility, as if from civilize + -ation.
The sense of “a particular human society in a civilized condition, considered as a whole over time,” is from 1857. Related: Civilizational.
Civility, formerly the substantive of both civil and civilize—the latter of which it was not likely to suggest, except by help of its context,—was judiciously relieved of one of its meanings, by civilization. [Fitzedward Hall, “Modern English,” 1873]
https://www.etymonline.com/word/civil
civil(adj.)
late 14c., “relating to civil law or life; pertaining to the internal affairs of a state,” from Old French civil “civil, relating to civil law” (13c.) and directly from Latin civilis “relating to a society, pertaining to public life, relating to the civic order, befitting a citizen,” hence by extension “popular, affable, courteous;” alternative adjectival derivative of civis “townsman” (see city).
Meaning “not barbarous, civilized” is from 1550s. Specifically “relating to the commonwealth as secularly organized” (as opposed to military or ecclesiastical) by 1610s. Meaning “relating to the citizen in his relation to the commonwealth or to fellow citizens” also is from 1610s.
The word civil has about twelve different meanings; it is applied to all manner of objects, which are perfectly disparate. As opposed to criminal, it means all law not criminal. As opposed to ecclesiastical, it means all law not ecclesiastical: as opposed to military, it means all law not military, and so on. [John Austin, “Lectures on Jurisprudence,” 1873] The sense of “polite” was in classical Latin, but English did not pick up this nuance of the word until late 16c., and it has tended to descend in meaning to “meeting minimum standards of courtesy.” “Courteous is thus more commonly said of superiors, civil of inferiors, since it implies or suggests the possibility of incivility or rudeness” [OED].
Civil, literally, applies to one who fulfills the duty of a citizen; It may mean simply not rude, or observant of the external courtesies of intercourse, or quick to do and say gratifying and complimentary things. … Courteous, literally, expresses that style of politeness which belongs to courts: a courteous man is one who is gracefully respectful in his address and manner — one who exhibits a union of dignified complaisance and kindness. The word applies to all sincere kindness and attention. [Century Dictionary, 1895] Civil case (as opposed to criminal) is recorded from 1610s. Civil liberty “natural liberty restrained by law only so far as is necessary for the public good” is by 1640s.
https://www.etymonline.com/word/city
city(n.)
c. 1200, from Old French cite “town, city” (10c., Modern French cité), from earlier citet, from Latin civitatem (nominative civitas; in Late Latin sometimes citatem) originally “citizenship, condition or rights of a citizen, membership in the community,” later “community of citizens, state, commonwealth” (used, for instance of the Gaulish tribes), from civis “townsman,” from PIE root *kei- (1) “to lie,” also forming words for “bed, couch,” and with a secondary sense of “beloved, dear.”
Now “a large and important town,” but originally in early Middle English a walled town, a capital or cathedral town. Distinction from town is early 14c. OED calls it “Not a native designation, but app[arently] at first a somewhat grandiose title, used instead of the OE. burh” (see borough).
Between Latin and English the sense was transferred from the inhabitants to the place. The Latin word for “city” was urbs, but a resident was civis. Civitas seems to have replaced urbs as Rome (the ultimate urbs) lost its prestige. Loss of Latin -v- is regular in French in some situations (compare alleger from alleviare; neige from nivea; jeune from juvenis. A different sound evolution from the Latin word yielded Italian citta, Catalan ciutat, Spanish ciudad, Portuguese cidade.
London is the city from 1550s. As an adjective, “pertaining to a city, urban,” from c. 1300. City hall “chief municipal offices” is first recorded 1670s; to fight city hall is 1913, American English. City slicker “a smart and plausible rogue, of a kind usu. found in cities” [OED] is first recorded 1916 (see slick (adj.)). City limits is from 1825.
The newspaper city-editor, who superintends the collection and publication of local news, is from 1834, American English; hence city desk attested from 1878. Inner city first attested 1968.
https://www.etymonline.com/word/*kei-
*kei-(1)
Proto-Indo-European root meaning “to lie,” also forming words for “bed, couch,” and with a secondary sense of “beloved, dear.”
It might form all or part of: ceilidh; cemetery; city; civic; civil; civilian; civilization; civilize; hide (n.2) measure of land; incivility; incunabula; Siva.
It might also be the source of: Sanskrit Sivah “propitious, gracious;” Greek keisthai “to lie, lie asleep;” Latin cunae “a cradle;” Old Church Slavonic semija “family, domestic servants;” Lithuanian šeima “domestic servants,” Lettish sieva “wife;” Old English hiwan “members of a household.”
Following B. F. Skinner’s hint ‘Etymology is the archeology of thought.’ and digging in and around ‘civilization’ to the PIE root ‘*kei-‘ it is found that the origins of civilization are beloved beds. It is this much wider sense (with all of its metaphorical trappings) that I shall take as ‘civilization’ and I shall leave behind civilization as something opposed to barbarity.
https://www.etymonline.com/word/barbarity
barbarity(n.)
1560s, “want of civilization,” from Latin barbarus (see barbarian (n.)) + -ity. The meaning “savage cruelty, inhuman conduct” is recorded from 1680s.
https://www.etymonline.com/word/barbarian
barbarian(adj.)
mid-14c., “foreign, of another nation or culture,” from Medieval Latin barbarinus (see barbarian (n.)). The meaning “of or pertaining to savages, rude, uncivilized” is from 1590s.
barbarian(n.)
early 15c., in reference to classical history, “a non-Roman or non-Greek,” earlier barbar (late 14c.) “non-Roman or non-Greek person; non-Christian; person speaking a language different from one’s own,” from Medieval Latin barbarinus (source of Old French barbarin “Berber, pagan, Saracen, barbarian”), from Latin barbarus “strange, foreign, barbarous,” from Greek barbaros “foreign, strange; ignorant,” from PIE root *barbar- echoic of unintelligible speech of foreigners (compare Sanskrit barbara- “stammering,” also “non-Aryan,” Latin balbus “stammering,” Czech blblati “to stammer”).
Greek barbaroi (plural noun) meant “all that are not Greek,” but especially the Medes and Persians; originally it was not entirely pejorative, but its sense became moreso after the Persian wars. The Romans (technically themselves barbaroi) took up the word and applied it to tribes or nations which had no Greek or Roman accomplishments.
Also in Middle English (c. 1400) “native of the Barbary coast;” meaning “rude, wild person” is from 1610s. Occasionally in 19c. English distinguished from savage (n.) as being a step closer to civilization. Sometimes, in reference to Renaissance Italy, “a non-Italian.” It also was used to translate the usual Chinese word of contempt for foreigners.
Barbarian applies to whatever pertains to the life of an uncivilized people, without special reference to its moral aspects. Barbarous properly expresses the bad side of barbarian life and character, especially its inhumanity or cruelty: as, a barbarous act. Barbaric expresses the characteristic love of barbarians for adornment, magnificence, noise, etc., but it is not commonly applied to persons: it implies the lack of cultivated taste …. [Century Dictionary, 1889]
So it is that ‘barbarity’ descends from the PIE ‘*barbar-‘ “echoic of unintelligible speech of foreigners”. It is here also that ‘civilization’ clearly ceases to be of any help in separating scientifically significant elements of social environments from irrelevant ones.
“Civilization is social order promoting cultural creation. Four elements constitute it: economic provision, political organization, moral traditions, and the pursuit of knowledge and the arts. It begins when chaos and insecurity end. For when fear is overcome, curiosity and constructiveness are free, and man passes by natural impulse towards the understanding and embellishments of life.” pg. 1
Here we find the developmentalistic metaphors of those like Piaget who speak as if organisms are promoted through distinct stages of progress or perhaps fail to take the step from one stage to the next because of some punitive intervention or lack of innate potential. It is as if one culture creates the next with the first culture coming from some unknown creator.
Cultures are no more created than species are. Social behavior is selected by consequences from variations on past social and nonsocial behavior and, hence, cultural practices as social contingences in a social environment evolve like biological organisms and their behaviors do.
There is no more to natural impulse than what has been selected by consequences from variations on past impulses and nonimpulses (whatever these may be). Impulses and impedence point to the push and pull of forces having conspicuous causes and effects or occasions and responses: the selecting consequences are left unmentioned.
Just as the opposite of reinforcement is not punishment (it is extinction), so too are pairs like ‘encourage’ and ‘discourage’ unfavorable to a theoretical analysis of history.
“Certain factors condition civilization, and may encourage or impeded it. First, geological conditions. Civilization is an interlude between ice ages: at any tiem the current of glaciation may rise again, cover with ice and stone the works of man, and reduce life to some narrow segment of the earth. Or the demon of earthquake, by whose leave we build our cities, may shrug his shoulders and consume us indifferently.” pg.1
Recent work on “Big History” has brought some more attention to the problem of how chemical changes appear from physical ones, how biological changes appear from chemical ones, how behavioral changes appear from biological ones, and how social changes appear from behavioral ones.
The conditions of geology and geography are seen as far more dynamic than they once were: the environments in which social behaviors occur are less a stage on which bits of behaving biology play, and more a cumulative selector.
There are more, and they are to be found in notes that have yet to appear on the science of geology.
Added hint “HOW TO COOK IT”: use a theremometer.
Most people over cook or under cook their meats (and other foodstuffs). Cooking foods to the temperatures given by those who have spent their life cooking is a better place to start than cooking to taste.
Cook to taste after cooking to temperature. You are likely to find that cooking to temperature unlocks flavors and textures that can be startling in their surprise to the tongue and palate.
Today I got Godlfarb’s “Deductive Logic” 2003.
When I got to page 56 I almost threw the book across the room.
It occurs in section 12 “Use and Mention”.
It reveals clearly and exactly that Goldfarb, like Carnap, does not know the difference between a schematic letter and a pronoun or variable.
“Thus in talking about a schema we use a name of that schema, most usually the name obtained by surrounding the schema with quotation marks. What then do we use to talk abotu schemata generally? Naturally, we use expressions such as ‘every schema’ and the like; but for some purposes we must also use variables that range over schemata, which are called syntactic variables. Just as we might use ‘x’ as a numerical variable and say
(4) A number x is odd if and only if x^2 is odd,
we can use ‘X’ and ‘Y’ as syntactic variables and say
(5) A schema X implies a schema Y if and only if the conditional with antecedent X and consequent Y is valid.
Particular instances of (4) are obtained by replacing the variable ‘x’ with the name of a number, for example, ‘3 is odd if and only if 3^2 is odd’. So too particular instances of (5) are obtained by preplacing teh variables ‘X’ and ‘Y’ with names of schemata, for example,
(6) ‘-p’ implies ‘-(p.q)’ if and only if the condition with antecdent ‘-p’ and consequent ‘-(p.q)’ is valid.
Note that the last ten words of (6) constitute a complex name of the schemata ‘-P only if -(p.q)’.” pg. 56-57 Goldfarb “Deductive Logic”.
At best, Goldfarb has weakly presented a substitutional theory of quantification without going out of his way to alert the reader to this fact and to its limitations. I will get to its limitations directly by later quoting a later part of Quine’s “Philosophy of Logic Second Edition” (POL) of 1986.
Also, the following are equivalent:
“This tendency to see set theory as logic has depended early and late on overestimating the kinship between membership and predication. An intermediate notion, attribution of attributes, insinuates itself and heightens the illusion of continuity.
In the innocent ‘Fx’ of the logic of quantification the schematic letter ‘F’ stands in place of a predicate. Or, more explicitly, the combination ‘Fx’ stands in the place of an open sentence in ‘x’; whether the sentence has ‘x’ on one side and an isolated predicate on the other is of no moment. What is important is that in writing ‘F’ and ‘Fx’ we are just schematically simulating sentences and their parts; we are not referring to predicates or other strings of signs, nor are we referring to attributes or sets. Some logicians, however, have taken a contray line, reading ‘F’ as an attribute variable and ‘Fx’ as ‘x has F’. Some, fond of attributes, have done this with their eyes open; others have been seduced into it by a confusion.
The confusion begins as a confusion of sign and object; a confusion between mentioning a sign and using it. Instead of seeing ‘F’ steadfastly as standing in place of an unspecified predicate, our confused logician sees it half the time as naming an unspecified predicate. Thus ‘F’ gains noun status, enabling him to read ‘Fx’ as ‘x has F’ without offend his grammatical ear. Having got this far, he can round out his confusion by calling F an attribute. This attunes his usage to that of the unconfused but prodigal logician who embraces attributes with his eyes open.
The prodigal logician is identifiable with Frege. The confused logician could be Russell, dispite his great contributions.” pg. 66
I make the distinction Quine makes in the second to last paragraph of the quote in 10 by talking about the difference between a schematic letter and a dangling pronoun. Schematic letters are replaced in the mechanical operation of substitution (von Neumann went so far as to say that substitution is executed on a schema in order to further distinguish between the schema and its sentences), but pronouns which occur in a schema (such as ‘x’, ‘y’, or ‘it’) also occur in the substitutional instances of such a schema.
I have gone out of my way when designing my memo on predicate logic to not only avoid pronouns (which is made possible by Quine’s predicate functors) but also to avoid schema. I have gone so far as to avoid notation generally. The price is high, but the pros far outweight the cons: there is no access point for any of the confusion caused over the past century by, e.g., mistaking predicating a predicate for attributing an atribute or classing a class.
It is still my mission to provide a perfectly smooth explanation of predicate logic from the logic of truth functions and quantifications (TQL) or, unhappily, “first-order logic”. This can be accomplished by successive abbreviations through Quine’s method of predicate abstraction and concretion (which is the logical import of the grammarians relative clause).
For those unable to deal with the appearence of ‘x’ in ‘{x}’ it is advised that they introduce a one place predicate ‘X’ to their lexicon (perhaps with appropriate adornments to distinguish it from some other unforunately conflicting predicate) along with the premises 1) ‘some u is such that Xu’ and 2) ‘each u is such that each v is such that, Xu and Xv, only if u=v’. In general, what is ‘learned by ostention’ in a predicate functor community are predicates: names are excommunicated. Said another way, designation and truth are degenerates of denotation.
“The quantifiers ‘some F is such that’ and ‘each F is such that’ are at the heart of the matter. I already deplored this sort of quantification somewhat after the middle of Chapter 2. I think it worth while now to develop my objections in more detail.
Consider first some ordinary quantifications “some x is such that (x walks)’, ‘each x is such that (x walks)’, ‘some x is such that (x is prime)’. The open sentence after the quantifier shows ‘x’ in a position where a name could stand; a name of a walker, for instance, or of a prime number. The quantifications do not mean that names walk or are prime; what are said to walk or to be prime are the things that could be named by names in those positions. To put the predicate letter ‘F’ in a quantifier, then, is to treat predicate positions suddenly as name positions, and hence to treat predicates as names of entities fo some sort. The quantifier ‘some F is such that’ or ‘each F is such that’ says not that some or all predicates are thus and so, but that some or all entitles of the sort named by predicates are thus and so. The logician who grasps this point, and still quantifies ‘F’, may say that these entities are attributes; attributes are for him the values ‘F’, the things over which ‘F’ ranges. The more confused logician, on the other hand, may say that these entities, the values of ‘F’, are predicates. He fails to appreciate the difference between schematically simulating predicates and quantificationally talking about predicates, let alone about attributes.
Even the first logician’s line here is to be deplored. I urged in Chapter 1 that propositions are undesirable; and the same goes for attributes. Attributes are to predicates, or open sentences, as propositions are to closed sentences. Attributes are like propositions in the inadequacy of their individuation. Sets are well individuated by the law of extensionality, which identifies sets whose members are the same; but this law fails for attributes, save as the wrod ‘attribute’ is ill-applied and ‘set’ would serve better. Open sentences that are true of just the same things never determine two sets, but may determine two attributes. What is further required for sameness of attributes is synonymy, in some sense, of the open sentences; and in Chapter 1 we despaired of making satisfactory sense of such synonymy.
Some logicians, for this reason, view the values of ‘F’ as sets. But I deplore the use of predicate letters as quantified variables, even when the values are sets. Predicates have attributes as their ‘intensions’ or meanings (or would if there were attributes), and they have sets as their extensions; but they are names of neither. Variables eligible for quantification therefore do not belong in predicate positions. They belong in name positions.
To put the point another way: even one who admits attributes should not read ‘Fx’ as ‘x has F’, with ‘F’ thus in a name position; rather let him write ‘x has y’, or, if he prefers distinctive variables for attributes, ‘x has #’. Likewise, if someone wants to admit sets as values of quantifiable variables, let him write ‘x in y’; or, if he prefers distinctive variables for sets ‘x in #’. Let him switch explicitly to what I called in Chapter 4 the set-theoretic analogue. The predicate letter ‘F’, like the sentence letter ‘p’, is not a value-taking variable at all, but just a substitution-taking schematic letter.” pg. 66-67
The only thing Quine forgot to mention overtly in these paragraphs (and the next ones I shall further quote to cover just one of the consequent errors which follows from not taking such advice to heart) is that it is only and solely in predicate logic as strictly maintained by Quine in his later work that the definitions of validity in terms of structure, substitution, models, proof, and grammar are equivalent in so weak a theory as protosyntax or what can be made equivalent to a theory of numeric place value notation. One example of the supremacy of this equivalence can be seen in Lindström’s theorem.
For those that would throw out such methods there is nothing more to ask of them but to produce a better alternative. Quine awaited such a day and now so do I.
“If anyone thought attributes a more congenial assumption than sets, he could assume quantification over attributes and then introduce quantification over sets, or a reasonable fascimile, by a certain scheme of contextual definition. Russell took this line. the point of the definition is just to secure the law of extensionality for sets without assuming it for attributes; for this law is the only difference between the two domains. But why would Russell find attributes a more congenial assumption than sets? It was a case rather of not appreciating where elementary logic, in its innocent simulation of predicates, gave way to talk about attributes. The phrase ‘propositional function’, adapted from Frege, cloaked the confusion; Russell used it to refer sometimes to predicates and sometimes to attributes. As a result it was thought by some that Russell had derived set theory, and therewith mathematics generally, from narrowly logical beginnings.
Followers of Hilbert have continued to quantify predicate letters, obtaining what they call a higher-order predicate calculus. The values of these variables are in effect sets; and this way of presenting set theory gives it a deceptive resemblance to logic. One is apt to feel that no abrupt addition to ordinary logic of quantification has been made; just some more quantifiers, governing predicate letters already present. In order to appreciate how deceptive this line can be consider the hypothese ‘some y is such that each x is such that (x in y if and only if Fx)’. It assumes a set {x: Fx} determined by an open sentence in the role of ‘Fx’. This is the central hypothesis of set theory, and the one that has to be restrained in one way or another to avoid the paradoxes. This hypothesis itself falls out of sight in the so-called higher-order predicate calculus. We get ‘some G is such that each x is such that (Gx if and only Fx)’, which evidently follows from the genuinely logical triviality ‘each x is such that (Fx if and only Fx)’ by an elemntary logical inference. There is no actual risk of paradox as long as the ranges of values of ‘x’ and ‘G’ are kept apart, but still a fair bit of set theory has slipped in unheralded.” pg. 67-68
“Instead of reading .. ‘(something x)(is such that Fx)’, we take to read it as ‘(something is)(x such that Fx)’.
Making this shift means recognizing an operator ‘x such that’ of predicate abstraction, in Peano’s notation .., for encapsulating in a self-contained complex predicate all that a sentence affirms of an object. Just substitution ‘x’ for ‘Tom’ in the sentence, if Tom is the object in question, and prefix ‘x such that’ to the resulting open sentence. The complex predicate thus formed, when predicate of anyone, will say of him what the original sentence said of Tom. This is our familiar ‘such that’ idiom, which is matheamtical pidgin English for our indigenous relative clause.”
Then, rather than taking the indefinite singular terms ‘some thing’ or ‘every thing’ I took ‘some item’ and ‘each item’. I then froze those into ‘some item is such that’ and ‘each item is such that’ as canonical quantifiers, from which the problems of cross reference are relieved by the suite of quantifiers ‘some item x is such that’ and ‘each item x is such that’, and then the operation of quantification is factored from the operation of predicate abstraction by switching to ‘some item is’ and ‘each item is’ with the suite of predicate abstractions ‘x such that Fx’, ‘y such that Fy’, and so on. This is all to pave the way for elimination of variables entirely, recognizing ‘some’ as a predicate functor of existential cropping and ‘each’ as its universal analogue and homogenizing and eliminating the variables bound by predicate abstractions via the recombic predicate functors.
“The only propositions of which HObbes’ principle is a sufficient account, are that limited and unimportant class in which both the predicate ant eh subject are proper names. For, as has already been remarked, proper names have strictly no meaning; they are mere marks for individual objects: and when a proper name is predicated of another proper name, all the signification conveyed is, that botht eh names are marks for the same object. But this is precisely what Hobbes produces as a theory of predication in general. His doctrine is full explanation of such predications as these: Hyde was Clarendon, or, Tully is Cicero. It exhausts the meaning of those propositions. But it is sadly inadequate theory of any others. That it should ever have been thought of as such, can be accounted for only by the fact, that Hobbes, in common with other Nominalists, bestowed little or no attention upon the connotation of words: and sought for their meaning exclusively in what they denote: as if all names had been (what none but proper names really are) marks put upon individuals; and as if there were no difference between a proper and general name, except that the first denotes only one individual, adn the last a greater number.”
The word ‘designates’ (as opposed to ‘denotes’) is then reserved for singular (as opposed to general) terms which purport to name one and only one item in the purportedly nonempty universe of discourse.
But for the hindesight of history, there would be an irony to Mill making this distinction between designation and denotation in reaction to Hobbes by reference to propositions and attributes.
“We reflected lately on the connection between existential quantification and alternation. If all the objects are named and finite in number, the quantification is of course dispensable in favor of alternation, and can be viewed as a mere abbreviation. If the objects are infinite in number, on the other hand, this expansion would require an infinitely long alternation. Midway in Chapter 4 we arrived at a view of expressions as finite sequences, in a mathematical sense; and the further step to infinite sequences is in no way audacious. It would, however, be distinctly a departure from all writings on grammar and most writings on logic, including this book, to invoke infinite expressions. Existential quantification over an infinite universe is not dispensable in favor of our alternation notation: finite alternation.
But existential quantification over an infinite universe still admits of an attractive semantical explanation of truth conditions, even though not of elimination, so long as all the objects have names. The quantification is true if adn only if at least one of its instances– got by dropping the quantifier and putting a name for the variable– is true.
All this about existential quantification holds equally, mutatis mutandis, for universal quantification. As long as all objects have names, a universal quantification is true if and only if all its instances are true.
Our standard logical grammar, at its strictest, admitted no names (Chapter 2). Names can be simulated by contextual definitions; this was why we could do without names in our strictest standard grammar. This way of providing for names does not go very well with these semantical explanations of quantification, however, since quantification is needed already in the contextual definition. So we will do well for the space of the present discussion to think in terms rather of standard logical grammar in the broader sense, admitting names.
These truth conditions for quantification, based on substitution of names and variables, have been favored by Ruth Marcus and others. They compare and contrast curiously with the Tarskian one seen in (5) of Chapter 3. They do not share the sort of circularity that was already remarked on midway in Chapter 3: the existential quantifcation is true if some instance is true, and the universal quantification is true if every instance is true. But the great contrast is that (5) of Chapter 3 speaks only of values of variables and makes no appeal to names. For this, (5) pays a price in complexity.
Thus far, no deviation; just different characterizations of the same quantification, so long as everything has a name. But now it must be remarked that this last is a very restrictive condition, even given our provisional new rediness to admit names. In a generous universe there are more things than can be named even with an infiniteude of names. For, let us recall again one of the twin discrepancies noted in Chapter 4 between sets and open sentences. We saw that some of the sets are not determined by any of the sentences. But these ests will lack names; for if a set has a name, say ‘a’, then the set is determined by the sentece ‘x in a’.
A more customary argument to the same purpose makes reference to a classical theorem of set theory, which says taht the irrational numbers cannot all be assigned distinct integers. In contrast, all names can be assigned distinct integers, for instance in Godel’s way. It follows that the irrational numbers can not all be assigned distinct integers.
We saw in Chapter 4 that the substitutional definitions of logical truth came out coextensive with teh model-theoretic definition of logical trut, so long as the object language was rich enough in vocabulary. We now see that an oppostie sort of situation holds for quantification: the substitutional characterization fo quantification is not coextensive witht eh characterization in terms of objects, or values of variables, if we we assume a rich universe. An existential quantification could turn out false when substitutionally construed and true when objectually construed, because of there being objects of the purported kind but only nameless ones. A universal quantification could turn out false when objectually construed and true when substitutionally construed, because of there being objects to the contrary but only nameless ones. And no lvaishness with names can prevent there beign nameless objects in a generous universe. Substitutional quantification is deviant if the universe is rich.
Its strengths. The deviation just now noted consists in leaving certain objects out of account, namely, those without names. But substitutional quantification can deviate also in an opposite way: the substituted expressions can fail to name. The truth condition that we formulated for substitutional quantification spoke expressly fo the substitutino of names, but it would work for any other grammatical category as well as for the category of names. If the category is finite, then the quantifications are again of course eliminable in favor of alternations and conjunctions. We sawa case of precisely this already in teh altter part of Chapter 5, in teh elimination of quantification of the ‘alpha’, ‘beta’, etc., in the positions of a hundred one-place predicates. It is interesting to note now an extension of that eliminable brand of quantification, in the form of an uneliminable brand of substitutional quantification. It builds upon the virtual theory fo classes (Chapter 5), and the names that it draws upon in substitution for its variables ‘alpha’, ‘beta’, ‘etc. are contextually defined abstracts of the virtual theory. If the resulting quantification over simulated classes is seen as cause for rejoicing, then let me remind you in sorrow that these quantifiers are no mere manner of speaking. They are nicely explained, but they are not eliminably defined. Moreover, because vritual class abstracts and corresponding variables come only after ‘in’, the domain of classes afforded is insufficient foundation for any appreciable matheamtics.
When soem new brand of quantification is introduced by definition, and thus eliminable, it of course does nto commit us really to recognizing any objects as values of the variables. It is simulated quantification with a simulated ontological commitment. Our real commitment rests rather with the real quantifiers in the standard language that backs up these false fronts of contextual definition. And note the importance of this stiuplation that the grammar be standard. If modalities or other constructions are admitted in addition to truth functions and quantifiers, they add to the strength or content of the theories in ways that are incommensurable with what mgiht be got by enlaring the universe; incommensurable, that is, except relative to some translation of the whole into standard grammar.
Substitution quantification, on the ohter hand, is neither an eliminable simulation nor a genuine objectual quantification (unless, of course, all things have names). It is not a way, then, of getting along with a null ontology, an empty universe; it is a non-standard idiom, rather, foreign to the langauge in which we talk fo what there is and of values of variables. IF one does still happen to wonder what would be an adequate universe for some theory that comes to him in this non-standard idiom, the thing for him to do is seek one or another reasonable-looking paraphrase of the theory into a standard form in which quantification is objectually construed. Then he can assess the universe of this theory– though the various passable translationas may well call for different universes. An unimaginative way of those translating substitution quantification is to translate it into a metalanguage in which we talk fo strings of signsa nd conctenation and substitution and truth– the sort of language touched on midway in Chapter 3. Identifying the strings with numbers as Godel did, we end up with the positive integers as universe.” pg. 91-94.
note: I discover that 3-6 are wrong in 7 and present the correction in 8 with reflection on what went wrong in 9.
Changed ‘open sentence’ to ‘free sentence’ in hint on how predicate abstracts work.
I submit that the practice of speaking of ‘open sentences’ is incomplete and that what was once called an open sentence is either a poorly paraphrased relative clause, i.e. a predicate, or simply a sentence with dangling pronouns or poorly paraphrased proper nouns.
‘F transitively closes G with respect to H’ for ‘Transitive H, and H includes G, only if H includes F’
Relative transitive closures are conditionals of conjunctions of a transitivity of a first component with the inclusion of the first with a second, with the inclusion of the first with a third.
Premises of transitive closures of one component with respect to an other are relative transitive closures of the one with the other with yet some other.
‘F transitively closes G with respect to F’ implies ‘Transitive F, and F includes G’: ‘F includes F’ is valid and ‘Transitive F, and F includes G, only if F includes F’ with ‘F includes F’ implies ‘Transitive F, and F includes G’.
NO, 6 is wrong, and hence the rest of 3-6 do not work out as I said.
This is a great example of not being governed by the laws of implication: predicates imply validities.
More work on logic memo.
HOW THE ABBREVIATIONS OF LOGIC WORK
1. The truth-functional, better 'denotative-functional', abbreviations are
1. negations are self joint denials,
2. alternations are negations of joint denials,
3. converse conditionals are alternations of their consequent with the negation of their antecedent,
4. complementary converse conditionals are negations of converse conditionals,
5. complementary conditionals are swapped complementary converse conditionals,
6. conditionals are negations of complementary conditionals,
7. alternative denials are conditionals with negated consequents,
8. conjunctions are negations of alternative denials,
9. exclusive alternations are conjunctions of alternative denials with matching alternations,
10. biconditionals are negations of exclusive alternations, and
11. sequents are conditionals of conjunctions with alternations.
2. The Boolean abbreviations are
1. universal closures are negations of existential closures of negations,
2. inclusions are universal closures of conditionals,
3. converse inclusions are swapped inclusions,
4. proper inclusions are conjunctions of inclusions with the negation of their converse, and
5. coextensions are conjunctions of inclusions with their converse.
3. The quantificational abbreviation is
- universal croppings are negations of existential croppings of negations.
4. The recombic abbreviations are
1. dushes are pushes of drops,
2. props are drops of pushes,
3. overs are pushes of hems,
4. oems are hems of overs,
5. dupes are dushes of oems,
6. pops are two drops of oems,
7. nip ns are n pushes of drops of n pops,
8. digs are nips of hems,
9. bury ns are n pushes of n digs,
10. unbury ns are n bury ns,
11. a (i,m)-clockwise turn (CW) is m pops of a bury (i+m) of m pushes, and
12. a (i,m)-counterclockwise turn (CCW) is m pops of a unbury (i+m) of m pushes.
5. (j,k,m)-resultants are j existential croppings of j bury (j+k)s of the conjunction of k drops with m props.
6. First (j\*2)-iterates are their component, and (n+2)th (j\*2)-iterates are (j,j,0)-resultants of their component with j nip (j\*2)s of (n+1)th (j\*2)-iterates of their component.
7. (n+1)th (j\*2)-inverse iterates are (n+1)th (j\*2)-iterates of j bury (j\*2)s.
8. (i+1,m)-indiscernabilities are i, universal croppings of bury 2s of, m, universal croppings of pops of, conjunctions of each biconditional of, drops of with nip 1s of, the 1st thru (i+m)th (i,m)-CWs of their component (Leibniz, Quine).
9. Identities are conjunctions of lexical indiscernabilities (Leibniz, Quine).
10. (i,m)-projections are i, existential croppings of bury 1s of, m, existential croppings of pops.
11. (i,m)-fields are alternations of each (i,m)-projection of the 1st thru (i+m)th (i,m)-CWs of their component.
HOW THE ABBREVIATIONS OF LOGIC WORK
1. The truth-functional, better 'denotative-functional', abbreviations are
1. negations are self joint denials,
2. alternations are negations of joint denials,
3. converse conditionals are alternations of their consequent with the negation of their antecedent,
4. complementary converse conditionals are negations of converse conditionals,
5. complementary conditionals are swapped complementary converse conditionals,
6. conditionals are negations of complementary conditionals,
7. alternative denials are conditionals with negated consequents,
8. conjunctions are negations of alternative denials,
9. exclusive alternations are conjunctions of alternative denials with matching alternations,
10. biconditionals are negations of exclusive alternations, and
11. sequents are conditionals of conjunctions with alternations.
2. The Boolean abbreviations are
1. universal closures are negations of existential closures of negations,
2. inclusions are universal closures of conditionals,
3. converse inclusions are swapped inclusions,
4. proper inclusions are conjunctions of inclusions with the negation of their converse, and
5. coextensions are conjunctions of inclusions with their converse.
3. The quantificational abbreviation is
- universal croppings are negations of existential croppings of negations.
4. The recombic abbreviations are
1. dushes are pushes of drops,
2. props are drops of pushes,
3. overs are pushes of hems,
4. oems are hems of overs,
5. dupes are dushes of oems,
6. pops are two drops of oems,
7. nip ns are n pushes of drops of n pops,
8. digs are nips of hems,
9. bury ns are n pushes of n digs,
10. unbury ns are n bury ns,
11. a (i,m)-clockwise turn (CW) is m pops of a bury (i+m) of m pushes, and
12. a (i,m)-counterclockwise turn (CCW) is m pops of a unbury (i+m) of m pushes.
5. The sameness abbreviations are:
1. (i+1,m)-indiscernabilities are i, universal croppings of bury 2s of, m, universal croppings of pops of, conjunctions of each biconditional of, drops of with nip 1s of, the 1st thru (i+m)th (i,m)-CWs of their component (Leibniz, Quine).
2. Identity is the conjunction of each indiscernability of a lexical atom(Leibniz, Quine).
3. ((i+1)\*2)-identities are conjunctions of each i nip 2s of i nip 1s of the 1st thru ith (i,0)-CWs of their component.
6. The compositional abbreviations are:
1. (j,k,m)-resultants are j existential croppings of j bury (j+k)s of the conjunction of k drops with m props.
2. First (j\*2)-iterates are their component, and (n+2)th (j\*2)-iterates are (j,j,0)-resultants of their component with j nip (j\*2)s of (n+1)th (j\*2)-iterates of their component.
3. (n+1)th (j\*2)-inverse iterates are (n+1)th (j\*2)-iterates of j bury (j\*2)s.
7. The functional abbreviations are
1. i-functionalities are inclusions of ((1+i)\*2)-identities with the conjunction of i drops with i nip i's of their component
2. i-totalities are universal closures of i existential croppings
3. i-partialities are negations of i-totalities
4. (i,j)-injectivities are inclusions of ((1+i)\*2)-identities with the conjunction of i nip js with i nip (i+j)s of their component
5. (i,j)-surjectivities are universal closures of i existential croppings of j (i+j,0)-CWs of their component
6. (i,j)-bijectivities are the conjunction of the (i,j)-injectivity with the (i,j)-surjectivity of its component
7. (i,j)-correlations are the conjunction of the i-functionality with the (i,j)-injectivity of their component
8. The graph abbreviations are
1. i-symmetrics are inclusions of their component with the first (i\*2)-inverse iterate of their component
2. i-nonsymmetrics are negations of i-symmetrics
3. i-asymmetrics are inclusions of their component with the negation fo the first (i\*2)-inverse iterate of their component
4. i-transitivities are inclusions of the second (i\*2)-iterate of their component with their component,
5. i-nontransitivities are negations of i-transitivities,
6. i-intransitivities are inclusions of the second (i\*2)-iterate of their component with the negation of their component,
Time to organize some more of the defined predicate functors.
It is late and I have to get up early tomorrow, but I am reminded of the semisignificant saying of Dr. Larson “A small fish is better than an empty dish.” I can at least be happy that I got out a first draft definition of ((i+1)*2)-identities. I will have to make up a list of translations between the predicate functor, the predicate abstract, and the truth-functional and quantificational sentences for each of these things so that nonnatives can go.
Notes on “Russell’s Logicism and Theory of Coherence” by Conor Mayo-Wilson in Russell: The Journal of Bertrand Russell Studies. Vol. 31. Summer, 2011. pp. 89-106.
“Also, Frege is not mentioned in the essay, which suggests that Quine is thinking primarily of Russell’s logicist program.”
“This idea of contextual definition, or recognition of the sentence as the primary vehicle of meaning, was indispensable to the ensuing developments in teh foundations of mathematics. It was explicit in Frege, and it attained its full flower in Russell’s doctrine of singula rdescriptions as incomplete symbols.” pg. 72
“The two ideals are linked. For, if you define all the concepts by use of some favored subset of them, you thereby show how to translate all theorems into these favored terms. The clearer these terms are, the likelier it is that the truths couched in them will be obviously true, or derivable from bovious truths. If in particular th concepts of mathematics were all reducible to the clear terms of logic, then all the truths of mathematics would go over into truths of logic; and surely the truths of logic are all obvious or at least potentially obvious, i.e., derivable from obvious truths by individually obvious steps.
This particular outcome is in fact denied us, howerever, since matheamtics reduces only to set theory and not to logic proper. Such reduction still enhances clarity, but only because of the interrelations that emerge and not because the end terms of the analysis are clearer than others. As for the end truths, the axiosm of set theory, these have less obviousness and certainty to recommend them than do most of the mathematical theorems that we would derive from them. Moreover, we know from Godel’s work that no consistent axiom system can cover matheamtics even when we renounce self-evidence. Reduction in the foundations of mathematics remains matheamtical and philosophically fascinating, but it does not do what the epistemologist would like of it: it does not reveal the ground of matheamtical knowledge, it does not show how mathematical certainty is possible.” pg. 70 of “Epistomology Naturalized”.
“A philosophical concern that motivated Whitehead and Russell in part, and Frege before them, was as to the nature of mathematical knowledge and the basis of mathematical truth. The conclusion they drew was that mathematics is translatable into pure logic. They counted membership as logic. So mathematical truth is logical truth, they reasoned, and hence all of it must be logically deducible from self-evident logical truths. This is wrong, as transpires in part from Kurt Godels’ paper of 1931 and in part from finidings by Russell himself away back in 1902.
This forlorn hope was not the only point of Russell and Whitehead’s great undertaking. Its other objective, and its enduring value, was simply a deeper understanding of the central concepts of mathematics and their basic laws and interrelations. Their total translatability into just elementary logic and a single familiar two-place predicate, membership, is of itself a philosophic sensation. Modern logic was indispensable to this achievement. An essential factor, of a piece with modern logic, was contextual definition.
Buoyed by their achievement, Russell reflected in 1914 on realizing the dream of empiricist epistemologist: the explicit construction of the external world, or a reasonable facsimilie, from sense impressions, hence from simple ideas. He adumbrated it in Our Knowledge of the External World, and a dozen years later Rudolf Carnap was undertaking to carry it out. Carnp’s effort found expression in Der Logische Aufbau der Welt (1928).” pg. 9-10 of “From Stimulus to Science”
“Carnap’s Aufbau was the culmination of the phenomenalism that evolved through Hobbes, Locke, Berkeley, Hume and had had its roots in Descarte’s doubts adn in the ancient perplexity over knowledge and error. Yet Carnap’s motivation was not this traditional quest for certainty. Rather, his goal was just a systematic integration– what he called a “constitution system”– of our scientific concepts of mind and nature. Whitehead and Russell gave us the construction of mathematicsl from minimal beginnings: Carnap went on from there, accepting logic and mathematics as a finished tool for use in his further constructions. His choice of experiences rather than physical objects as his foundation was, he assures us, just a matter of strategy.
Even so, Carnap’s inspiration was only in part Principia Mathematica. It was more emphatically Russell’s frankly epistemological Our Knowledge of the External World. Thus it is that his constitution system took the form of a “rational reconstruction”– Carnap’s phrase again– of man’s conceptual development. If by 1928 he was seeing this rational reconstruction as merely a good strategy for a constitution system of global science, still there is no mistaking the epistemological predilection that led him to it.” pg. 13-14
“The idea of a self-sufficient sensory language as a foundation for science loses its lustre when we reflect that systematization of our sensory intake is the very business that science itself is engaged in. The memories that link our past experiences with present ones and induce our expectations are themselves mostly memories not of sensory intake but of essentially scientific posits, namely things and events in the physical world. It was perhaps appreciation of this point that led Otto Neurath, Carnap’s colleague in Vienna, to persuade Carnap to give up his methodological phenomenalism in favor of physicalism.” pg. 15
“Studies in the foundations of matheamtics divide symmetrically into two sorts, conceptual and doctrinal. The conceptual studies are concerned with meaning, the doctrinal with truth. The conceptual studies are concerned with clarifying concepts by defining them, some in terms of others. The doctrinal studies are concerned wtih establishing laws and proving them some on the basis of others.” pg. 69-70 “Epistemology Naturalized”
… “The two ideals are linked. For, if you define all the concepts by use of some favored subset of them, you thereby show how to translate all theorems into these favored terms. The clearer these terms are, the likelier it is that the truths couched in them will be obviously true, or derivable from obvious truths.” pg. 70
…“mathematics reduces only to set theory and not to logic proper. Such reduction still enhances clarity, but only because of the interrelations that emerge and not because the end terms of the analysis are clearer than others.” pg. 70
… “Still there remains a helpful thought, regardign epistemology generally, in that duality of structure which was especially conspicuous in teh foundations of mathematics. I refer to the bifurcation into a theory of concepts, or meaning, and a theory of doctrine, or truth; for this applies to teh epistemology of natural knowledge no less than to the foundations of mathematics.” … “On the doctrinal side, I do not see that we are farther along today than where Hume left us. The Humean predicament is the human predicament. But on the conceptual side there has been progress. … It was made by Bentham in his theory of fictions. Bentham’s step was the recognition of contextual definition or what he called paraphrasis. He recognized that to explain a term we do not need to specify an object for it to refer to, nor even specify a synonymous word or phrase; we need only show, by whatever means, how to translate all the whole sentences in which the term is to be used.” pg. 72 … “This idea of contextual definition, or recognition of the sentence as teh primary vehicle of meaning, was indispensable to the ensuing developments in teh foundations of mathematics. It was explicit in Frege, and it attained its full flower in Russell’s doctrin of singular descriptions as incomplete symbols.
Contextual definition was one of two resorts that could be expected to have a liberating effect upon the conceptual side of the epistemology of natural knowledge. The other is resort to the resources of set theory as auxiliary concepts.” pg. 72-73
.. “The two resorts are very unequal in epistemological status. Contextual definition is unassailable. Sentences that have been given meaning as wholes are undeiably meaningful, and the use they make of their component terms is therefore meaningful, regardless of whether any translations are offered for those terms in isolation. … Recourse to sets, on the other hand, is a drastic ontological move, a retreat from the austere ontology of impressions. … these sets, which amount, afterall, to the whole abstract ontology of mathematics.” pg. 73
.. “The issue has not always been clear, however, owing to deceptive hints of continuity between elementary logic and set theory. This is why mathematics was once believed to reduce to logic, that is, to an innocent and unquestional logic, and to inherit these qualities. And this si probably why Russell was content to resort to sets as well as to contextual definition when in Our Knowledge of the External World and elsewhere he addressed himself to the epistemology of natural knowledge, on its conceptual side.
To account for the external world as a logical construct of sense data– such, in Russell’s terms, was the program. It was Carnap, in his Der logische Aufbau der Welt of 1928, who came nearest to executing it.” pg. 73-74
.. “The Cartesian quest for certainty had been the remote motivation of epistemology, both on its conceptual and its doctrinal side; but the quest was seen as a lost cause. To endow the truths of nature with the full authority of immediate experience was as forlorn a hope as hoping to endow the truths of mathematics with the potential obviousness of elementary logic.
“What then could have motivated Carnap’s heroic efforts on the conceptual side of epistemology, when hope of certainty on the doctrinal side was abandoned? There were good reasons still. one was that such constructions could be expected to elicit and clarify the sensory evidence for science, even if the inferential steps between sensory evidence and scientific doctrine must fall short of certainty. The other reason was that such constructions would deepen our understanding of our discourse about the world, even apart from questions of evidence; it would make all cognitive discourse as clear as observation terms and logic and, I must regretfully add, set theory.” pg. 74-75
“Two cardinal tenents of empiricism remained unassailable, however, and so remain to this day. One is that whatever evidence there is for science is sensorty evidence. The other, to which I shall recur, is that all inculcation of meanings of words must rest ultimately on sensory evidence. Hence the continuing attractiveness of the idea of a logischer Aufbau in which the sensorty content of discourse would stand forth explicitly.” pg. 75
.. “But why all this creative reconstruction, all this make believe? The stimulation of his sensory receptors is all the evidence anybody has had to go on, ultimately, in arriving at his picture of the world. Why not just see how this construction really proceeds? Why not settle for psychology? Such a surrender of the epistemological burden to psychology is a move that was disallowed in earlier times as circular reasoning. If the epistemologist’s goal is validation of the grounds of empirical science, he defeats his purpose by using psychology or other empirical science in validation. However, such scruples against circulatrity have little point once we have stopped dreaming of deducing science from observations. If we are out simply to understand the link between observation and science, we are well advised to use any available information, including that provided by the very sciences whose link with observation we are seeking to understand.” pg. 7
“Now that we are permitted to appeal to physical stimulation, the problem dissolves; A is epistemologically prior to B if A is causally nearer than B to the sensory receptors. Or, what is in some ways better, just talk explicitly in terms of causal proximity to sensory receptors and drop the talk of epistemological priority.”pg. 85
“an observation sentence is one on which all speakers of the language give the same verdict when given the same concurrent stimulation. To put the point negatively, an observation sentence is one that is not sensitive to differences in past experience within the speech community.” pg. 86-87
“The old tendency to associate observation sentences with a subjective sensory subject matter is rather an irony when we reflect that observation sentences are also meant to be the intersubjective tribunal of scientific hypotheses. The old tendency was due to the drive to base science on something firmer and prior in the subject’s experience; but we dropped that project.
“Clarification of the notion of observation sentence is a good thing for the notion is fundamental in two connections. These two correspond to the duality I remarked upon early in this lecture: teh duality between concept and doctrine, between knowing what a sentence means and knowing whether it is true. The observation sentence is basic to both enterprises. Its relation to doctrine, to our knowledge of what is true, is very much the traditional one: observation sentences are the repository of evidence for scientific hypotheses. Its relation to meaning is fundamental too, since observation sentences are the ones we are in a position to learn to understand first, both as children and as field linguists. For observation sentences are precisely the ones that we can correlate with observable circumstances of the occasion of utterance or assent, independently of variations in the past histories of individual informants. They afford the only entry to a language.” pg. 88-89
“It is no shock to the preconceptions fo old Vienna to say that epistemology now becomes semantics. For epistemology remains centered as always on evidence, and meaning remains centered as always on verification; and evidence is verification. What is likelier to shock preconceptions is that meaning, once we get beyond observation sentences, ceases in general to have any clear applicability to single sentences; also that epistemology merges with psychology, as well as with linguistics.” pg. 89-90
I may soon have to write for those who are already entrenched in their most cherished methods of argument and debate rather than for those who are already familiar with, e.g., the methods of logic as taught by Quine in “Methods of Logic 4th edtiion”.
The challenge is the same as it has been for about the whole history of philosophy: many things that go without saying don’t.
This is largely because there is a modern canon of philosophy which is under the control of cognitive science and probabilistic and statistical methods.
Said another way, there is a looming influence from information theories that make it appear as if propositions are back in play throughout philosophy, even though there is very little to say of them in the methods of logic or their application.
Another strange practice that turns up in recent readings I’ve done of philosophy: the authors try to understand what the philosopher meant by this or that. As if what was written is just expression of some deeper idea that can somehow be guessed at by the analytic reader who is properly in tune with the state of the writer across their own time period all the way up to ours.
The challenge is met, at least in the beginning, by going straight to the experimental analysis of behavior. But, even here there is a great divide between where modern reading and writing is and the work of those like Skinner and Forester in “Schedules of Reinforcement”.
In modern philosophy, philosophers reason, logicians reason, people reason, and they frequently do so statistically. How am I to square this with the behavior that is actually investigated in experimental laboratories where the schedules of reinforcement are carefully controlled and rates of response are selected by the contingencies connecting various operanda?
“Everyone agrees that data should sometimes change our beliefs.”
“Equipped with probabilism, one can then easily explain how the probabilistic structure of data should incorporated into a rational agent’s beliefs, as the latter also have probabilistic structure.”
What ‘rational agents’?
Typing up handwritten notes from 2025 0723 1609. I hesitate to type these up because the notation used in them is so much more like traditional notation that it may suggest to the reader that there is some enduring value in such methods. I can only emphasize that however familiar are these notational methods, they are superceeded in almost every way by the stack based notation found elsewhere (this same comment applies to the last note 2025 0815 2028).
I write down as many different kinds of notation as I can find in the hopes that something better appears from the work of those who are so obviously better than me. If I recall accurately, these notes are a paraphrase of selections from Quine’s “Methods of Logic 4th edition” (even though I would now see them more clearly as Quine’s paraphrase of Carnap’s “1958 “Introduction to Symbolic Logic and its Applications” (ISLA)).
Note that ‘dup’ and ‘drop’ above are not like those of the stack based methods elsewhere presented. It is impossible for me to sort through all the different kinds of notation that might be used to work with these things, and this is in fact one of the main reasons for presenting how logic works the way that I do: without notation.
This is a note from the John who is typing up these handwritten notes. I’m very happy to have done this because I now have much better names for the predicate functors that correspond to the above and below schematically given predicate functors e.g. ‘symmetrics’ instead of ‘symmetries’, ‘asymmetrics’ instead of ‘asymmetries’, ‘transitives’ instead of ‘transitivities’, and ‘intransitives’ instead of ‘intransitivities’, etc.
Typing up handwritten notes from 2025 0723 1329
##HOW LOGIC WORKS
1. Predicates do or don't denote (are true or false of) where they occur, constructions compound them with connectives, and lexicons itemize atomic ones.
2. Grammars generate categories from recurrent constructions on lexicons.
3. Logic uncovers validities whose supplemented lexical substitutions (SLS) denote everywhere (Quine).
4. Constructions inherit where they denote from their components (Tarski):
5. alethically, i.e.
1. joint denials denote where and only where (waow) each of their components don't,
2. existential closures denote waow there is some where denoted by their component, and
3. components of existential croppings denote the left of waow their compound denotes with some item with the right of the same whereabouts, and
6. recombically, i.e.
4. drops denote the left of the left of waow the compound does with the right of the same whereabouts,
5. pushes denote the left of waow the compound does, with the left of the right of the same whereabouts, with the right of the right of the same whereabouts, and
6. hems denote the left of waow the compound does, with the right of the left of the left of the same whereabouts with the right of the same whereabouts.
## HOW THE ABBREVIATIONS OF LOGIC WORK
1. The truth-functional, better 'denotative-functional', abbreviations are
1. negations are joint denials of their component with itself,
2. alternations are negations of joint denials of their (two) components,
3. converse conditionals are alternations of their left component (consequent) with the negation of their right component (antecedent),
4. complementary converse conditionals are negations of the converse conditional of their components,
5. complementary conditionals are complementary converse conditionals of their right component with their left component,
6. conditionals are negations of complementary conditionals of their components,
7. alternative denials are conditionals of their left component with the negation of their right component,
8. conjunctions are negations of alternative denials of their components,
9. exclusive alternations are conjunctions of alternative denials of their components with alternations of their components,
10. biconditionals are negations of exclusive alternations of their (two) components, and
11. sequents are conditionals of the conjunction of its antecedent components with the alternation of its consequent components.
2. The Boolean abbreviations are
1. universal closures are negations of existential closures of negations of their component
2. inclusions are universal closures of conditionals of their components,
3. converse inclusions are universal closures of converse conditionals of theirs,
4. proper inclusions are conjunctions of inclusions of theirs with the negation of their converse inclusion, and
5. coextensions are conjunctions of inclusions with their converse inclusion.
4. The quantificational abbreviation is
- universal croppings are negations of existential croppings of negations of their component.
5. The recombic abbreviations are
1. overs are pushes of hems of their component,
2. ohems are hems of overs of theirs,
3. dushes are pushes of drops of theirs,
4. props are drops of pushes of theirs,
5. dupes are dushes of ohems of theirs,
6. pops are drops of ohems of theirs,
7. nips are dushes of hems of theirs,
8. digs are nips of hems of theirs,
9. buries are pushes of digs of theirs,
10. bury ns are n pushes of n digs of theirs,
11. unburies are two buries of theirs, and
12. unbury ns are n bury ns of theirs.
6. An (i,j,k)-resultant is j existential croppings of k bury (j+k)s of the conjunction of k drops of the left component with i props of the right component.
## HOW INDISCERNIBILITY WORKS
1. For the predicate letter 'F' write '=F' for '{xy: each {u..v: Fxu..v if and only if Fyu..v, .., and Fu..vx if and only if Fu..vy}'.
2. Write '=(..F,G)', or just '=', for '{xy: (=(..F))xy and (=G)xy}'.
3. Identity and indiscernibility (with respect to the lexicon of predicates of a standard theory) are coextensive.
## HOW LOGIC WORKS
1. Predicates denote or don't (are true or false of) where they occur (Aristotle);
2. grammars generate categories from recurrent connective construction on itemized atomic lexicons (Frege, Hilbert, von Neumann);
3. supplemented lexical substitutions (SLS) of validities denote everywhere (Quine); and
4. denotata are inherited (Tarski) alethically, i.e.
1. joint denials denote where and only where (waow) each of their components don't (Peirce, Sheffer),
2. existential closure denote waow there is some where denoted by their component (Boole, Quine, Hailperin), and
3. existential croppings denote the left part of waow their compound does with some item, with the right part of the same whereabouts (Frege, Russell, Quine), and
5. recombically (Russell, Schönfinkel, Curry, Bernays, Tarski, Quine, Charles H. Moore), i.e.
1. components of drops denote the left part of the left part of waow their comopund does with the right part of the same whereabouts,
2. components of pushes denote the left part of waow their comopund does with the left part of the right part of the same whereabouts, with the right part of the right part of the same whereabouts, and
3. components of hems denote the left part of waow their compound does, with the right part of the left part of the left part of the same whereabouts with the right part of the same whereabouts.
## HOW THE ABBREVIATIONS OF LOGIC WORK
1. The truth-functional, better 'denotative-functional', abbreviations are
1. negations are self joint denials,
2. alternations are negations of joint denials,
3. converse conditionals are alternations of their consequent with the negation of their antecedent,
4. complementary converse conditionals are negations of converse conditionals,
5. complementary conditionals are swapped complementary converse conditionals,
6. conditionals are negations of complementary conditionals,
7. alternative denials are conditionals with negated consequents,
8. conjunctions are negations of alternative denials,
9. exclusive alternations are conjunctions of alternative denials with matching alternations,
10. biconditionals are negations of exclusive alternations, and
11. sequents are conditionals of conjunctions with alternations.
2. The Boolean abbreviations are
1. universal closures are negations of existential closures of negations,
2. inclusions are universal closures of conditionals,
3. converse inclusions are swapped inclusions,
4. proper inclusions are conjunctions of inclusions with the negation of their converse, and
5. coextensions are conjunctions of inclusions with their converse.
3. The quantificational abbreviation is
- universal croppings are negations of existential croppings of negations.
4. The recombic abbreviations are
1. dushes are pushes of drops,
2. props are drops of pushes,
3. overs are pushes of hems,
4. oems are hems of overs,
5. dupes are dushes of oems,
6. pops are two drops of oems,
7. nip ns are n pushes of drops of n pops,
8. digs are nips of hems,
9. bury ns are n pushes of n digs,
10. unbury ns are n bury ns,
11. a (i,m)-clockwise turn (CW) is m pops of a bury (i+m) of m pushes, and
12. a (i,m)-counterclockwise turn (CCW) is m pops of a unbury (i+m) of m pushes.
5. (j,k,m)-resultants are j existential croppings of j bury (j+k)s of the conjunction of k drops with m props.
6. First (j\*2)-iterates are their component, and (n+2)th (j\*2)-iterates are (j,j,0)-resultants of their component with j nip (j\*2)s of (n+1)th (j\*2)-iterates of their component.
7. (n+1)th (j\*2)-inverse iterates are (n+1)th (j\*2)-iterates of j bury (j\*2)s.
8. (i+1,m)-indiscernabilities are i, universal croppings of bury 2s of, m, universal croppings of pops of, conjunctions of each biconditional of, drops of with nip 1s of, the 1st thru (i+m)th (i,m)-CWs of their component (Leibniz, Quine).
9. Identities are conjunctions of lexical indiscernabilities (Leibniz, Quine).
10. (i,m)-projections are i, existential croppings of bury 1s of, m, existential croppings of pops.
11. (i,m)-fields are alternations of each (i,m)-projection of the 1st thru (i+m)th (i,m)-CWs of their component.
## TIE UP
(this has fallen out of sync with all that I'm doing, and I may have to just drop this section because I skip right over it almost every time I have a look at recent notes)
- Predicate (Functor) Logic Memo
- [2025 0608 1515](#2025-0608-1515)
- [2025 0625 2043](#2025-0625-2043)
- [2025 0625 2333](#2025-0625-2333)
- [2025 0627 1410](#2025-0627-1410)
- [2025 0627 1904](#2025-0627-1904)
- Earmuff Experiment
- [2025 0625 1537](#2025-0625-1537)
- [2025 0625 2259](#2025-0625-2259)
- [2025 0629 1518](#2025-0629-1518)
- LISP
- FORTH
- "A first-order axiomatization of the theory of finite trees" by Backofen, Rogers, and Vijay-Shanker 1995
- [2025 0718 1745](#2025-0718-1745)
Some edits and tiny corrections were made in bringing the latest work on the predicate logic memo to the hints section e.g. the bicategoricals of the recombic connectives were missing an initial phrase ‘the components of’, and the wordy sentence about the inheritance of denotata has been replaced with just ‘denotata are inherited’.
No longer is a distinction made about the direction of the inheritance of denotata e.g. Quine emphasizes chasing truth up and down the tree of grammar. Proofs are another way, as Quine also demonstrated, although it was much more remote than may be thought upon reading “Philosophy of Logic 2nd Ed.” (POL) where it is merely mentioned that his main method can be applied outright to sentences rather than to the truth functional and quantificational schema that turn up in his “Methods of Logic 4th ed.” (MOL).
Hailperin’s 1981 “Boole’s Algebra Isn’t Boolean Algebra” is the only other palce other than Quine’s “Methods of Logic 4th edition” (MOL) where I’ve found an accurate report on the actual methods presented in Boole’s 1854 “An Investigation of the Laws of Thought: on Which are Founded the Mathematical Theories of Logic and Probabilities” (aka “Laws of Thought’).
For Quine, and perhaps in part for Hailperin, Boole’s contribution to logic covers the logic of existential closures and denotative functions.
Each item of the grammatical category of denotative functions is equivalent to one in the grammatical category of recurrent joint denial constructions which enforce a uniformity of shape among its non-denotative-functional subcomponents e.g. the joint denial of a pair of predicates with shape (2,2) inherits the shape (2,2).
The existential closure of a denotative functional compound is called by Quine a “Boolean Existence Sentence” schema (on page 116 of MOL Quine was speaking technically of “Boolean Existence Schema” each instance of which can be called a “Boolean Existence Sentence”, though Quine was still wed to ‘statement’ when he wrote MOL and would later drop it for ‘sentence’ to avoid any confusion between the consequences of a verbal response and the verbal response itself).
Denotative functions of existential closures were then called “Boolean sentences”: these have a delightful decision procedure called by Quine “Existential Conditionals”.
The strict correspondence between Boole’s work and Quine’s method of “Existential Conditionals” is sketched out in chapter 20 of MOL “Some Boolean Incidentals” (a sorta catch all for the utilities of Boolean abbreviations).
In Quine’s “Logic Papers” the method of “Existential Conditionals” is obscured by picking up Boole’s equational methods earlier than he does in MOL. There, as in Boole’s “Laws of Thought”, Quine begins by writing “There does not exist” as “0=”.
In MOL Quine almost comes out and says that Booles methods are not those of Boolean algebra, but he merely hints at the past association between an algebra of classes and Boole’s equations, quickly leaving such potential misinterpretations for an abbreviated notation close to that of Boole’s which is equivalent to the method of existential conditionals.
Quine does not take the further step to deal with existential closures of Boolean sentences. It is clear enough upon further reflection that there is strictly nothing added by repeatedly construction existential closures of existential closures, but free wheeling construction of complex existential closures and denotative functions fits with the methods familiar in truth functional and quantificational logic.
Quine also fails to make a distinction between existential closures and repeated existential croppings: when are they equivalent, are they?
In MOL the closest he comes to confronting this problem is when he introduces the predicate functor of existential cropping itself. There he first contemplates the construction of an existential closure of a predicate abstract which is made equivalent to a sequence of quantifiers each variable of which is one bound by the predicate abstract (in the order given by the binding phrase) and these are themselves existential closures of the strict kind contemplated by Quine in the chapters on Boolean logic which are actually only about one place predicates.
The logic of existential closures is, as mentioned in 3, accomodating of predicates of any number of places, but a closer examination of the case Quine excluded, mentioned above in 9, reveals that Quine’s predicate abstracts cover up the evolution of a ‘partial existential closure’ into ‘existential cropping’ proper.
This evolution is primed by predicate functors which are called for if the rules of passage which can be shown with Boolean sentences (and, specifically, Boolean existence sentences) are to carry over to situations described as in 9 e.g. Existential closures of Boolean sentences.
Actually, 13 and 9 do not quite reach the problems hinted at in 12 and 13: it is when Boolean terms (which are here called denotative functional compounds) are mixed with Boolean sentences and Boolean Closures. Some definitions will help make the distinctions clear.
What Quine called Boolean terms are now simply called denotative functions (as shown in 3). Boolean compounds now come to be existential closures or denotative functions of lexical predicates or Boolean compounds. Then, Boolean schema are existential closures or denotative functions of predicate letters or Boolen schema.
By definition the existential closure of a predicate has the shape of a sentence (which, in the stack notation embraced here, means a shape of (0,0)). The constraints on the shape of denotative compounds from the shapes of their components (they must all be the same) makes it so that a Boolean schema which conjoins a predicate letter to an existential closure fixes the shape of the predicate letter to (0,0).
But, how then to preserve the rule of passage which takes the conjunction of predicate letter with an existential closure to the existential closure of the predicate letter with the component of the earlier existential closure?
Quine, as much as Carnap in his 1958 “Introduction to Symbolic Logic and its Applications” (ISLA) gives sentence letters that may end up being purportededly open or closed. The open case is made clear by me when we take a sentence ‘x loves y’ as closed (where ‘x’ and ‘y’ are as much like ‘John’ and ‘Dick’ and are no more like dangling pronouns than ‘John’ and ‘Dick’ are) and ‘x and y such that x loves y’ as the predicate which open sentences are so often taken as (that is, an open sentence is not open at all, it is simply a relative clause whose bound parts have not been properly indicated).
“In the schemata in which we have illustrated quantification thus far, the scope of each occurrence of a quantifier has been a truth function of ‘Fx’, ‘Gx’, etc. But we can be more liberal, allowing some components of the scope to be devoid of ‘x’ and represented simply as ‘p’ and ‘q’.
It makes good sense, e.g., to write ‘some item is (x such that p and Fx)’. This should come out true under just those interpretations of ‘p’ and F’ that make ‘p and Fx’ true of at least one thing x. Those, of course, are just the interpretations that make ‘p’ true and ‘F’ true of something; so teh two schemata ‘each item is (x such that p and Fx)’ and ‘p and each item is (x such that Fx)’ are equivalent.” pg.142 Quine MOL
In a pure predicate functor logic the equivalence of 19, which is given by Quine with variables and quantifiers (the quantifiers themselves been explained by the construction of an existential closure of a one place predicate abstract), is that between the existential cropping of a conjunction of a drop of a predicate on the left with a predicate on the right, and a conjunction of the predicate on the left with the existential cropping of the predicate on the right.
But, note in 20, that existential croppings appear and not existential closures.
Starting instead with a conjunction with a left predicate and an existential closure of a right predicate, the shape of the right predicate dictates an equivalence between the existential closure of the right predicate and, say, (i+m) existential croppings of a recombic combination of the right predciate (whose shape is now obviously (i,m)).
Supposing we leave out the part about the recombic combination of the right predicate (by just supposing we have a lexical predicate of appropraite shape or that we do not look beyond the presently unknown recombic functors as part of our official logical language) we are introduced to exactly one recombic functor by way of maintaining the rule of passage.
Specifically, (i+m) existential croppings of the conjunction of (i+m) drops of the left predicate with the right predicate, is equivalent to the conjunction of the left predicate with (i+m) existential croppings of the right predicate.
Moving from repeated existential croppings to existential closures, there appears to be a ‘drop closure’ such that an existential closure of the conjunction of a drop closure of a left predicate with a right predicate, is equivalent to the conjunction of the left predicate with the existential closure of the right predicate.
This ‘big drop’ is then later seen to be equivalent to appropriately repeated drops.
I must end here, but must also mention that the big problem looming here is that pointed up at the end of Russell’s “An Inquiry into Meaning and Truth” in the chapter on “analysis” where the problem of the predicate ‘is a part of’ appears as it relates to the methods of logic.
As is very clear from the other notes I’ve made here, I find it all but impossible to talk about logic without mentioning ordered pairs, and it may only be in semantic ascent that the problem of ‘is a part of’ occurs. Here, the orderd pair is always as Quine noticed at the end of “Word and Object”: a paradigm of philosophic analysis.
Predicate memo work.
and recombically (Russell, Schönfinkel, Curry, Bernays, Tarski, Quine, Charles H. Moore)
(j,k,m)-resultants are j existential croppings of j bury (j+k)s of the conjunction of k drops with m props.
First (j*2)-iterates are their component, and (n+2)th (j*2)-iterates are (j,j,0)-resultants of their component with j nip (j*2)s of (n+1)th (j*2)-iterates of their component.
(n+1)th (j*2)-inverse iterates are (n+1)th (j*2)-iterates of j bury (j*2)s.
(i+1,m)-indiscernabilities are i, universal croppings of bury 2s of, m, universal croppings of pops of, conjunctions of each biconditional of, drops of with nip 1s of, the 1st, .., thru (i+m)th (i,m)-CWs of their component (Leibniz, Quine).
Identities are conjunctions of lexical indiscernabilities (Leibniz, Quine).
(i,m)-projections are i, existential croppings of bury 1s of, m, existential croppings of pops.
The memo and the hint on predicate (functor) logic may soon diverge. There is a lot that can be said in even less space with unfamiliar three letter words along with terse sentences that have longer explanations in the appropriate memos.
Predicates do or don’t denote (are true or false of) where they occur, grammars recursively construct categories with connectives from lexicon, and validities’s supplemented lexical substitutions (SLS) denote everywhere (Quine).
recombically, i.e.
A (j,k,m)-resultant is j existential croppings of k bury (j+k)s of the conjunction of k drops with m props.
The first iterate of a predicate whose left shape is (j*2) is itself and its (n+1)st iterate is the (j,j,0)-resultant of itself with the j nip (j*2) of its n-th iterate.
A (i,m)-counterclockwise (CCW) turn is m pushes of a bury (i+m) of m pops.
More work on logic memo.
Predicates do or don’t denote (are true or false of) where they occur, constructions compound them with connectives, and lexicons itemize atomic ones.
Grammars generate categories from recurrent constructions on lexicons.
Logic uncovers validities whose supplemented lexical substitutions (SLS) denote everywhere (Quine).
Constructions inherit where they denote from their components (Tarski):
Pushes, pops, dupes, drops, buries, and unburies are recombically complete: any arrangement of any items can be set up from them.
An (i,j,k)-resultant is j existential croppings of k bury (j+k)s of the conjunction of k drops of the left component with i props of the right component.
An (m,n)-resultant of a pair of components where the right part of the shape of the left component and the left part of the shape of the right component are both k is k existential croppings of the conjunction of n props of k pops of the left component with n bury ks of n drops of the right component.
I originally thought that 13 would be simpler to write out than 12, but it is clear that 12 is simpler: this was all but bound to happen because this is precisely the definition which encapsulates the composition of stack operations and notation.
HOW LOGIC WORKS
- Predicates do or don’t denote (are true or false of) where they occur, constructions connect them, and lexicons list atomic ones.
- Grammars generate categories from recurrent constructions on lexicons.
- Logic uncovers validities whose supplemented lexical substitutions (SLS) denote everywhere (Quine).
- Constructions inherit where they denote from their components (Tarski):
- alethically, i.e.
- joint denials denote where and only where (waow) each of their components don’t,
- existential closures denote waow there is some where denoted by their component, and
- components of existential croppings denote the left of waow their compound denotes with some item with the right of the same whereabouts, and
- recombically, i.e.
- drops denote the left of the left of waow the compound does with the right of the same whereabouts,
- pushes denote the left of waow the compound does, with the left of the right of the same whereabouts, with the right of the right of the same whereabouts, and
- hems denote the left of waow the compound does, with the right of the left of the left of the same whereabouts with the right of the same whereabouts.
More work on logic memo.
Predicates do or don’t denote (are true or false of) where they occur, constructions compound them with connectives, and lexicons itemize atomic ones.
Grammars generate categories from recurrent constructions on lexicons.
Logic uncovers validities whose supplemented lexical substitutions (SLS) denote everywhere (Quine).
Constructions inherit where they denote from their components (Tarski):
The wording for 10 of burrying and unburrying has to be fixed to accomodate the standard way of speaking of iterated constructions by predciate functors e.g. ‘bury 6’ is short for ‘6 pushes of 6 digs’ where ‘1 push’ is ‘push of’, ‘2 pushes of’ is ‘push of push of’, and ‘1+n pushes’ is ‘push of n pushes’ and so on.
Furthering 11, ‘4 bury 6’ is short for ‘bury 6 of 3 bury 6’ etc.
Happily, I received my copy of Rudolf Carnap’s 1958 “Introduction to Symbolic Logic and its Applications”.
I shall write up my notes on it here as it occurs to me.
Importantly, the book can be bought from Dover and is hence not nearly as expensive as other technical books of this caliber are.
As with all of Carnap’s technical writings, there are ample divisions of his writing from parts all the way down to subsections.
“Symbolic logic (also called mathematical logic or logistic) is the modern form of logic developed in the last hundred years.” page 1.
It is in this first sentence that there is already a hint at where Carnap goes wrong and where he goes right. He goes wrong in that he does not distinguish between the methods of logic and those of mathematical logic (or what he also calls logistic). He goes write in that he presents much of what logic has contributed to mathematics over the hundred years prior to his publication.
“Such a system is not a theory (i.e. a system of assertions about objects), but a language (i.e. a system of signs and of rules for their use).”
It is interesting that he says this even though he directly references Quine’s “Mathematical Logic” in the sources provided. There, Quine goes out of his way in the last chapter to present his theory of protosyntax and to explain the ways in which it comes to inclusively cover the methods that so often fly under the flag of ‘formal’ or ‘finite’ or, perhaps more restrictively, ‘constructive’.
That is, Quine shows that, in fact, formal languages are as much theories as geometries or algebras.
“We will so construct this symbolic language that into it can be translated the sentences of any given theory about any objects whatever, provided only that some signs of the language have received determinate interpretations such that the signs serve to designate the basic concepts of the theory in question. So long as we remain in the domain of pure logic (i.e. so long as we are concerned with building this language, and not with its application and interpretation respecting a given theory), the signs of our language remain uninterpreted. Strictly speaking, what we construct is not a language but a schema or skeleton of a language: out of this schema we can produce at need a proper language (conceived as an instrument of communication) by interpretation of certain signs.”
The use of ‘symbols’, ‘signs’, and, especially, ‘designates’ is the end of the schematic methods of logic: logic is not about anything in particular and its schematic letters no more symbolize, signal, or designate things than do our native sentences mean their meaning, inform their information, or propose their proposition.
https://www.etymonline.com/word/*deik-
*deik-
Proto-Indo-European root meaning “to show,” also “pronounce solemnly,” “also in derivatives referring to the directing of words or objects” [Watkins].
It might form all or part of: abdicate; abdication; addict; adjudge; apodictic; avenge; benediction; betoken; condition; contradict; contradiction; dedicate; deictic; deixis; dictate; diction; dictionary; dictum; digit; disk; ditto; ditty; edict; Eurydice; index; indicate; indication; indict; indiction; indictive; indite; interdict; judge; judicial; juridical; jurisdiction; malediction; malison; paradigm; policy (n.2) “written insurance agreement;” preach; predicament; predicate; predict; prejudice; revenge; soi-disant; syndic; teach; tetchy; theodicy; toe; token; valediction; vendetta; verdict; veridical; vindicate; vindication; voir dire.
It might also be the source of: Sanskrit dic- “point out, show;” Greek deiknynai “to show, to prove,” dikē “custom, usage;” Latin dicere “speak, tell, say,” digitus “finger,” Old High German zeigon, German zeigen “to show,” Old English teon “to accuse,” tæcan “to teach.”
‘Zeichen’ can be read as ‘sign’, ‘symbol’, ‘token’, ‘mark’, or even ‘character’. It is the last reading ‘character’ that comes out in schematic methods as ‘letters’, and all that the marks mark, signs signal, or symbols symbolize is recurrence of place in quotation upon which the operation of substitution (as given by von Neumann) is to be ‘executed’. The german is ‘ausführen’ as in ‘einen Plan ausführen’ (‘to carry out a plan’) or ‘ein Programm ausführen’ (‘to execute a program’).
“In the present Chapter A, the first of three chapters comprising Part One, we describe a simple symbolic language A containing the following sorts of signs (to be explained later): sentential constants and variables, individual constants and variables, predicate constants and variables of various levels and types, functor constants and variables, sentential connectives, and quantifiers.” page 1.
Predicate letters in truth function and quantification schema are neither constants nor variables. Variables are pronouns (perhaps bound as in a relative clause) and they occur as much in schema as they do in instances of the same. Not even von Neumann made this distinction, and many are prepared to use a new set of letter forms to speak of ‘variable letters’ as dummy variables!
Free variables can even come to play the part of names or constants or, best, singular terms, but in the schematic method, they are at best dummy singular terms, and can perhaps better be called ‘pronoun letters’ if we follow the dummy variable path, or, perhaps better, pronouns which no more dangle than proper nouns do to those inadept at a language.
All of this is to say that to take predicate letters as variables is to misunderstand the power of schematic methods and to mix theory with logic.
“In the course of constructing our symbolic language systems, it frequently happens that a new precisely-defined concept is introduced in place of one which is familiar but insufficiently precise. Such a new concept is called an explication of the old one, and its introduction an explication. (The concept to be explicated is sometimes called the explicandum.) … (FOr a more complete exposition of the methods of explication and the requirements an adequate explicatum must meet, see Carnap (Probability), Chapter I.).”
So it is that Carnap has put yet another book on my reading radar.
“The reduction of the mental to the physical, or indeed of arithmetic to set theory, can be characterized in either of two ways: as explaining or as explaining away. There is no difference, but the first phrasing has a gentler ring. To have rupidiated the life of the mind seems harsher than to have explained it in physical terms.
In some domains the harsher phrasing is better, notably in the various ways of defining natural numbers in set theory. For Frege, twelve was (nearly enough) the class of all dozens; for von Neumann it was the class fo the first twelve natural numbers, from zero through eleven. It cannot be both. We can resolve the dilemma by not identifying twelve with either. We can do without the natural numbers; their work can be done by Frege’s classes or, equally, by von Neumann’s. We proceed to use one or the other set of surrogates, or some thrid, and conveniently callt eh surrogates by the newly vacated numerical names. We have solved the puzzle by speaking of alternative eliminations rather than conflicting explanations. But there is no real difference between the two characterizations.” pg. 86-87 Quine “From Stimulus to Science”.
Going over Carnap’s logical methods in his language C.
To legitimatize this use of connectives in building up compound predicate expressions we introduce the following definitions for predicate expressions of the simplest type. Analogous definitions are understood to hold for each other type of predicate variable.
D28-1.
- (not F)x if and only if not (Fx)
- (F or G)x if and only if (Fx or Gx)
- (F and G)x if and only if (Fx and Gx)
- (F only if G)x if and only if (Fx only if Gx)
- (F if and only if G)x if and only if (Fx if and only if Gx).
As Carnap has set up his language, such “open sentences” as those in D28-1 are read as if they were ammended by a universal closure.
The expedients of D28-1 are employed to great effect by Quine in “Methods of Logic 4th edition” in the chapter named “Boolean Schemata”. Quine adds all that comes with distinguishing between schematisms and symbolisms.
Carnap introduces similar notational expedients in his “Logical Syntax of Language”. I recall reading it and know that it is less than half way through the book, but I do not recall the actual page or chapter.
“The new compounds are as follows: ‘(P or Q)a’ is counted as an abbreviation of the sentences ‘Pa or Qa’, and ‘P or Q’ treated as a predicate expression of the same type as ‘P’, viz. a one-place predicate expression of the first level and of type (0);”
Typing, as far as I am concerned, is a matter of shaping up predicates in the same way that the J programming language shapes up arrays and extends operations with its clever little rank operation.
Note, Carnap’s book appeared before Gerhard Gentzen’s introduction of the upside down ‘A’ as the mark of universal quantification from his works in 1935 ‘Untersuchungen über das logische Schließen’, in English ‘Investigations into Logical Deduction I and II”, had spread. I do not use anything like such notation and have gone ahead and translated the obscure marks into their pidgin English.
“28b. Universality. A property of individuals is called universal provided every individual has this property. Correspondingly, in the terminology of classes: a class of individuals is universal provided every individual belongs to th is class. Generally, a class of any type is said to be universal if each entity of that type belongs to this class. Our symbol for universality is ‘U’; and ‘the class (or property) P is universal’ is rendered ‘U(P)’. Since ‘U(P)’ is clearly synonymous with ‘(x)Px’, the following definition is natural:
D28-2. U(F) if and only if (x)Fx.
Analogous definitions are understood to hold for predicates of any other type, be they one-place or many-place. E.g. ‘(x)(y)Rxy’ can be abbreviated ‘U(R)’.”
The last sentence of the quote in 9 indicates that what is there defined as ‘universal’ is what we now call ‘universal closure’.
Carnap’s methods in “Introduction to Symbolic Logic and its Applications” (ISLA) are the closest thing to the elaborate evolution of predicate functor logic that I’ve had in my written notes for some time. Today I’ll be going through the parts that are relevant to me.
I shall be receiving a print copy of ISLA over the next few days (perhaps even tomorrow): it is much easier for me to read a book than it is for me to read a screen, especially when it comes to cross referencing entries in mathematical or scientific texts.
Carnap introduces “Compound Predicate Expressiosn” in section 28 under the chapter “The Extended Language C”.
Carnap does not distinguish between schematic methods, with its dummy sentence and predicate letters, from symbolic methods, with its designating sentence and predicate names. The presupposition that predicates are names of properties or extensions is, as Quine repeatedly emphasized and which was first made crystal clear through von Neuman’s schematic methods in 1927 “Zur Hilbertschen Beweistheorie”, is resolved by presenting a logical theory of properties or extensions which purport to carry over the logical work of predicates to the items of their respective theories.
There is a historic problem that has persisted since the earliest days of philosophical record: attributing properties to things.
Sadly, attributing a property to a thing has been conflated with predicating a predicate of it. Properties are sometimes even defined as that which can be predicated of an object.
It helps to distinguish between ‘attributing attributes’ and ‘predicating predicates’ and ‘classing classes’.
Historically, all three of these methods have been dealt with as if they were one and the same or as if passage from one to the other goes without saying.
Classes are extensions of predicates, each item classed to a class is so clased because its name predicated to the predicate of which the class is its extension is true.
In the past, it did not make sense, e.g. to Bertrand Russell, that the integral phrases that make up a predicate and a subject of a sentence might carry the load once carried by properties and concepts, better, attributes and classes.
There was the problem of meaning: how could a sentence have ever appeared without the stuff which the sentence purports to be about?
Such stuff was to be attributes in Russell and Whitehead’s “Principia Mathematica” and from there, the classes were constructed through what I am here calling ‘predicates’ (following John Stuart Mill through Quine, as integral words or phrases) and names (or what are often called ‘singular terms’).
For Russell and Whitehead, what I call ‘predicates’ were ‘propositional functions’.
Propositions are the attributes (or properties) of logic, according to those like Russell: they are what are true or false and sentences simply express this propositional property.
It was Frege who pointed the way through predicates (as I take them) from attributes to classes: predicates, as the notation of Frege’s Begriffsschrift, were part of the logic of truth-functions and quantifications and hence subject to the extensionality of such a logic, and thus Russell could round up the attributes (which he thought he knew better than he knew predicates) through propositions, as attributes of attributing attributes of things, and propositional functions, as functions which when applied to things attributes an atribute of it i.e. metaphorically takes a thing and returns an attribute (in this case a proposition).
Looking back it can be hard to see how this path was preferrable e.g. Russel’s own paradox prohibits such free wheeling shifts between attributes, predicates, and classes.
All of this confusion remains to this day, and is present in Carnap’s ISLA.
So much of what Carnap accomplished with language C of ISLA falls squarely under the more restricted methods of predicate functor logic as given, e.g., in my last note.
It is extremely exciting to find someone else who went to all the trouble to collate the many results that bring together the methods of truth-functional, quantificational, predicate abstraction, and predicate functor logic.
What I am beginning to see with greater clarity is that the methods pointed up, e.g., in my last note bring together the methods of relation algebras uncovered by Tarski and the methods of logic uncovered by Carnap, through the careful methods of logic collated by Quine.
So much of Quine follows Carnap that it is tempting to point past Quine to Carnap and leave it there: this would be a huge mistake.
Carnap hastily generalized, as did Church and his followers, the power of lambda abstraction far beyond anything permitted by the methods of logic: such full blown assemblies of set theoretic powers continue to go unntoiced by modern mathematicians.
The method of predicate abstraction as the logical import of the grammarians relative clause clears up the problems caused by taking lambda abstraction as something that may be carelessly carried over beyond a given theory of lambda abstraction and application.
For all the sensitivity that Carnap had to the hierarchy of languages, and for all that he promoted with his existentional thesis, he was too quick to take his notation for a ride.
I have just this day come to see that perhaps the primary importance of the methods of logic, e.g. those started in the last note, is that notation is entirely avoided.
There is a cost and there is a benefit; the cost is that nonnative speakers of the language of predicate logic I’ve introduced are unlikely to follow the later parts without first going over the earlier parts, the benefit is that no single notation is going to mislead me, or any other reader, into seeing metaphors that are not there.
The classic example of a miscarriage of metaphor in the methods of logic is the introduction of quantifiers which purport to bind predicate letters in a schema of predicate logic: this is simply incoherent and reveals that those inclined to bind predicate letters have yet to master the schematic methods so clearly and exactly set out by von Neumann’s 1927 “Zur Hilbertschen Beweistheorie”.
As I reach further out from these new beginnings I hope to take a closer look at Ernst Schröder’s work.
I have read that much of Russell and Whitehead’s work in Principia Mathematica was drawn from Schröder’s work and would like to see that for myself.
It remains a great sadness that the works of those like Hilbert and Schröder are still not yet translated into English. It is quite a strange thing to live in a world which professes its advanced methods without having any of the evidence one would reasonably expect from such purported advancements.
Vaughn Pratt The Origins of the Calculus of Relations suggests Peirce > “Collected Papers of Charles Sanders Peirce edited by Hartshorne and Weiss and published by Harvard University Press in the 1930’”
The problem of closures and croppings has not yet been effectively resolved.
Existential closures denote where and only where (waow) there is somewhere denoted by their component.
Traditionally, closures are defined from croppings as iterated croppings which “close” the shape of their component predicate. Such closed predicates are most often called closed sentences, but in predicate (functor) logic there is no longer a principle role for sentences (be they open or closed) and we are left speaking of predicates alone rather than dealing with the trouble of the grammatical construction called predication or the grammatical category of variables (thus there is also no confusion over the problem of a finite lexicon which contains a finite list of variables and constructs the remaining category of variables by way of, e.g., accentuation).
Components of existential croppings denote the left of with some item with the right of waow the compound does.
Note, the definition of croppings involves the complex singular descriptions ‘the right of’ and ‘the left of’ which are ultimately compounded, in 3, with ‘where’ as in ‘the left of where and only where the compound does with some item with the right of the same whereabouts’.
An alternative to 3 is this: components of existential croppings denote the of waow the compound denotes with some item with the right of the same whereabouts.
Further alternatives may rely on the introduction of ‘predicate’ as that which does or doesn’t denote where it occurs by giving the definition of ‘predicate’ rather as that which does or doesn’t denote a given occasion.
The introduction of occasions is misleading though, as the problem of individuation is bound up with spatiotemporal whereabouts.
HOW LOGIC WORKS
- Predicates do or don’t denote (are true or false of) where they occur, constructions connect them, and lexicons list atomic ones.
- Grammars generate categories from recurrent constructions on lexicons.
- Logic uncovers validities whose supplemented lexical substitutions (SLS) denote everywhere (Quine).
- Constructions inheret where they denote from their components (Tarski):
- recombically, i.e. components of
- drops denote teh left of the left of with teh right of where and only where (waow) their compound does,
- pushes denote the left of with the left of the right of with the right of the right of waow the compound does,
- hems denote the left of with the right of the left of with the right of waow the compound does;
- or alethically, i.e.
- each component of existential joint denials doesn’t denote the left of with some same item with the right of waow the compound does.
I am confronted with the problem of closed quantifications. In the past I have made it clear that the truth functional, or what is better called the denotative functional, connectives combine with a predicate functor of existential closure, as explained by Quine in “Methods of Logic 4th edition” chapter 18 titled “Boolean Schemata”.
The steps in this slower development are closer to each of the problems which have come up in the history of mathematics and its logic e.g. mereolgical problems, collections, assemblies, wholes, parts, individuation.
I finally succeeded in assembling the meat of predicate functor logic independent of truth functional and quantificational logic.
Once upon a time, a woman taught a man to speak: she held up an apple, said ‘red’, he held it up, said ‘red’, she smiled, he held up another, said ‘red’, she frowned, held it up, said ‘green’, etc.
Predicates do or don’t denote (are true or false of) where they occur, constructions connect them, and lexicons list atomic ones.
Grammars generate categories from recurrent constructions on lexicons.
Logic uncovers validities whose supplemented lexical substitutions (SLS) denote everywhere (Quine).
or alethically, i.e.
This will now be added to a hint.
HOW TO MAKE IT LOGICAL
- List the nouns, adjectives, and verbs used to talk about it.
- Define as many from as few as you can.
- List true sentences made from those few.
- Conclude as many from as few as you can.
- Go to 1.
HOW TO MAKE IT LOGICAL
- List the words used to talk about it.
- Define as many from as few as you can.
- List true sentences made from those few.
- Conclude as many from as few as you can.
- Go to 1.
HOW LOGIC WORKS
- Predicates do or don’t denote (are true or false of) where they occur, constructions connect them, and lexicons list atomic ones.
- Grammars generate categories from recurrent constructions on lexicons.
- Logic uncovers validities whose supplemented lexical substitutions (SLS) denote everywhere (Quine).
- Constructions inheret where they denote from their components (Tarski):
- recombically, i.e. components of
- drops denote teh left of the left of with teh right of where and only where (waow) their compound does,
- pushes denote the left of with the left of the right of with the right of the right of waow the compound does,
- hems denote the left of with the right of the left of with the right of waow the compound does;
- or alethically, i.e.
- each component of existential joint denials doesn’t denote the left of with some same item with the right of waow the compound does.
Validity can be defined in purely grammatical terms as transformable into a certain grammatical category with respect to some finite list of transformations.
These are frequently paraphrased as theories of calculi, e.g. syntactic rules of inference, better, of concatenation, e.g. binary strings, or, better still, trees, e.g. binary trees.
Truth functional inconsistency is established by full alternational or conjunctive normal form.
Quantificational inconsistency is established by teh chance construction of a truth-functionally inconsistent conjunction via Quine’s main method (a sequence of transformations which takes a conjunction to endlessly longer ones).
It is important to distinguish between these grammatical definitions and the methods of proof that are so often confused with such grammatical methods in the same way that the predicate of implication is confused with the grammatical particle of conditional compound sentences ‘only if’.
Why analyze validity into grammar and denotation? Why not stick with proof, or with the related definition relative to special grammatical constructions as above?
To show that validity is as transcendent (or immanent) as grammar and denotation are.
The above line of thinking is what led me to ask: does the purported predicate ‘grammatical’ fall prey to Grelling’s paradox?
The answer is no, but perhaps only with respect to one analysis of ‘is grammatical’ as defined from the predicate of lexicality and the predicates of grammatical construction e.g. ‘is lexical’, ‘conjoins’, ‘negates’, ‘existential
This note records my work on solving a problem of substitution: how to deal with notation such as ‘---x---’ and ‘Fx’ where ‘F’ is a schematic predicate letter and ‘x’ is a pronoun (otherwise called a variable).
This is all relative to the methods of logic given by Quine in “Methods of Logic 4th edition” (MOL).
The first place where a notation like ‘---x---’ is used in that book is actually when ‘p’, ‘q’, and ‘r’ are used as dummy closed sentences or schematic closed sentence letters e.g. ‘---p---’.
“But there are useful laws of interchange, the least of which is the first law of interchange: Think of ‘. . . p . . .’ as any schema containing ‘p’, and of ‘. . . q . . .’ as formed from ‘. . . p . . .’ by putting ‘q’ for one or more occurrence of ‘p’; then ‘p if and only if q’ implies ‘. . . p . . . if and only if . . . q . . .’.” pg. 63-64 MOL
The schema mentioned in the quotation in 4 is only truth-functional.
The letters ‘p’ and ‘q’ are first introduced as part of the methods of logic in a way that is distinct from how they are dealt with in the chapter quoted from in 4.
“In logical studies it is convenient to adopt a single sign of negation, consisting of the prefix ‘-‘, applied to statements as wholes. Thus ‘-(Jones is away)’ means ‘Jones is not away’; the parentheses here serve to group, in a single whole, the statement to which ‘-‘ is applied. The sign ‘-‘ might be translated into words as ‘it is not the case that’; briefly it may be pronounced ‘not’. When a statement is represented as a single letter ‘p’, as is commonly done in logical discussion, the sign of negation will be placed above instead of in front; thus we shall write ‘???’ instead of ‘-p’ for the negation of ‘p’.” pg. 9 MOL
“Some philosophers, commendably diffident about positing propositions in this bold sense, have taken refuge in teh word ‘statement’. The opening question of this chapter illustrates this evasive use. my inveterate use of ‘statement’ in earlier books does not; I there used the word merely to refer to declarative sentences, and said so. Later I gave up the word in the face of the growing tendency at Oxford to use the word for acts that we perform in uttering declarative sentences. Now by appealing to statements in such a sense, instead of to propositions, certainly no clarity is gained. I shall say no more about statements, but will go on about propositions.” pg. 2 POL
*stā-, Proto-Indo-European root meaning “to stand, set down, make or be firm,” with derivatives meaning “place or thing that is standing.”
It might form all or part of: Afghanistan; Anastasia; apostasy; apostate; armistice; arrest; assist; astatic; astatine; Baluchistan; bedstead; circumstance; consist; constable; constant; constitute; contrast; cost; desist; destination; destine; destitute; diastase; distance; distant; ecstasy; epistasis; epistemology; establish; estaminet; estate; etagere; existence; extant; Hindustan; histidine; histo-; histogram; histology; histone; hypostasis; insist; instant; instauration; institute; interstice; isostasy; isostatic; Kazakhstan; metastasis; obstacle; obstetric; obstinate; oust; Pakistan; peristyle; persist; post (n.1) “timber set upright;” press (v.2) “force into service;” presto; prostate; prostitute; resist; rest (v.2) “to be left, remain;” restitution; restive; restore; shtetl; solstice; stable (adj.) “secure against falling;” stable (n.) “building for domestic animals;” stage; stalag; stalwart; stamen; -stan; stance; stanchion; stand; standard; stanza; stapes; starboard; stare decisis; stasis; -stat; stat; state (n.1) “circumstances, conditions;” stater; static; station; statistics; stator; statue; stature; status; statute; staunch; (adj.) “strong, substantial;” stay (v.1) “come to a halt, remain in place;” stay (n.2) “strong rope which supports a ship’s mast;” stead; steed; steer (n.) “male beef cattle;” steer (v.) “guide the course of a vehicle;” stem (n.) “trunk of a plant;” stern (n.) “hind part of a ship;” stet; stoa; stoic; stool; store; stound; stow; stud (n.1) “nailhead, knob;” stud (n.2) “horse kept for breeding;” stylite; subsist; substance; substitute; substitution; superstition; system; Taurus; understand.
It might also be the source of: Sanskrit tisthati “stands;” Avestan histaiti “to stand;” Persian -stan “country,” literally “where one stands;” Greek histēmi “put, place, cause to stand; weigh,” stasis “a standing still,” statos “placed,” stylos “pillar;” Latin sistere “stand still, stop, make stand, place, produce in court,” status “manner, position, condition, attitude,” stare “to stand,” statio “station, post;” Lithuanian stojuos “I place myself,” statau “I place;” Old Church Slavonic staja “place myself,” stanu “position;” Gothic standan, Old English standan “to stand,” stede “place;” Old Norse steði “anvil;” Old Irish sessam “the act of standing.””
The etymology of ‘sentence’ goes through ‘sense’ to *sent- which I already dealt with in #2025-0526-2141-the-first-two-words-of-paul-grahams-paper-good-writing
Neither ‘statement’ nor ‘sentence’ are effective, and as much is clear from Skinner’s “Verbal Behavior” (VB).
Like later Quine, I avoid ‘statement’ and if I am to speak of ‘utterences’ then I am already close enough to behavior to switch over entirely from the philosophic, logic, and gramamtical traditions to that of the radical behaviorists e.g. ‘form of response’, ‘topography of response’, or, better, ‘operanda’. Perhaps ‘operanda’ is the best: it points up the problem of describing responses, recording responses, and all else that comes up when comparing one occurrence with another by switching from the effectors of an organism to the effect on the laboratory environment.
This is not a note on behavior and I shall abandon this path of exploration.
I tend to use ‘sentence’, but ‘quotation’ is probably better. Let me look at its etymology.
quote(v.) late 14c., coten, “to mark or annotate (a book) with chapter numbers or marginal references” (a sense now obsolete), from Old French coter and directly from Medieval Latin quotare “distinguish by numbers, mark off into chapters and verses,” from Latin quotus “which in order? what number (in sequence)?,” from quot “how many,” from PIE *kwo-ti-, from pronominal root *kwo-.
*kwo- also *kwi-, Proto-Indo-European root, stem of relative and interrogative pronouns.
It might form all or part of: cheese (n.2) “a big thing;” cue (n.1) “stage direction;” either; hidalgo; how; kickshaw; neither; neuter; qua; quality; quandary; quantity; quasar; quasi; quasi-; query; quibble; quiddity; quidnunc; quip; quodlibet; quondam; quorum; quote; quotidian; quotient; ubi; ubiquity; what; when; whence; where; whether; which; whither; who; whoever; whom; whose; why.
It might also be the source of: Sanskrit kah “who, which;” Avestan ko, Hittite kuish “who;” Latin quis/quid “in what respect, to what extent; how, why,” qua “where, which way,” qui/quae/quod “who, which;” Lithuanian kas “who;” Old Church Slavonic kuto, Russian kto “who;” Old Irish ce, Welsh pwy “who;” Old English hwa, hwæt, hwær, etc.
From which it is clear that quotations spell things out e.g. morphemes, responses, etc.
Spelling and concatenating and listing are close to quotations.
spell(v.1) early 14c., spellen, “read letter by letter, write or say the letters of;” c. 1400, “form words by means of letters,” said in most etymology sources to be from Anglo-French espeller, Old French espelir “to mean, signify; explain, interpret,” also “spell out letters, pronounce, recite.”
This French word is from Frankish *spellon “to tell” or some other Germanic source, from Proto-Germanic *spellam (source also of Dutch spellen, Old High German spellon “to tell,” Old Norse spjalla, Gothic spillon “to talk, tell”).
The native cognate word is Old English spellian, spillian “to tell, talk, speak, discourse.” Only Barnhart seems to allow that the modern English word is partially from the Old English one, due to the difference in sense.
Klein’s sources suggest a possible origin for this Germanic group in a PIE root *spel- (2) “to say aloud, recite, speak with emphasis” and cognates in Greek apeilē “threat” among other words, but Beekes finds the suggestion “rather far-fetched.”
Also in early use speldren, from Old French espeldre, a variant of espelir. Related: Spelled; spelling.
In early Middle English still “to speak, preach, talk, tell,” hence such expressions as hear spell “hear (something) told or talked about,” spell the wind “talk in vain” (both 15c.). The meaning “form words with proper letters” is from 1580s.
Spell out “explain step-by-step” is recorded from 1940, American English. Shakespeare has spell (someone) backwards “reverse the character of, explain in a contrary sense, portray with determined negativity.”
“The quotation is just a name of the depicted sequence of letters and spaces, dissociated grammatically and semantically from the outlying text. The dissociation becomes still more graphic when we analyze the quotation explicity into names of the individual characters and spaces, linked by concatenation signs. …
We have a familiar word for this analysis of quotation: it is spelling. Spelled out, the words disappear and so do questions of substitutivity. Substitutivity within quotation is confusion of use with mention.” pg. 94-95 FSS
Ambiguous abbreviations are not a proper problem of any theory, but rather of the logic applied to it.
Russell’s contextual elimination of descriptions is ambiguous abbreviation: as much as such was noticed when he bound the context of a description in the same way a variable was bound by a quantifier.
Relative to a finite lexicon, contextual elimination of descriptions by minimal context is unambiguous in the theory, but still ambiguous in the logical methods e.g. schematic methods.
x is identical to (the F) if and only if some item is (y such that x is identical to y if and only if Fy).
Properly, a schematic premise is not a premise at all: each of its instances are.
Still, even here, ‘G(the F)’ abbreviates ‘some item is (x such that Gx and x is identical to (the F))’.
G---the F--- if and only if some item is (x such that G---x--- and x is identical to (the F).
---the F--- if and only if some item is (x such that ---x--- and each item is (y such that x is identical to y if and only if Fy).
..u{..v: ---..v---..x--- :..x}..y if and only if ---..u---..y---
before then introducing the premise of exstentionality for classes (and sets, sets are classes that belong to some class)
each item is (x, y, and z such that each item is (u such that u in x if and only if u in y) and x in z, only if y in z)
and before introducing the schematic premise of class abstraction
x in the class of {y such that Fy} if and only if Fx
and he never demonstrates the relation between class abstraction and descriptions in any theory with a premise of extensionality for classes:
the class of {x such that Fx} is identical to the (y such that each item is (x such that x in y if and only if Fx))
and, perhaps worse of all, in none of his texts, not even Mathematical Logic 4th edition (MOL), does he deal with the double life of BOTH predicate abstracts as class abstracts AND the copula of predication with the predicate ‘in’ of class membership (which I shall attempt to do here).
Much of the work required to bring predicate abstracts and class abstracts together, so that following Quine we can let predicate abstracts “moonlight” as names of their extensionality as classes, is found in “Set Theory and its Logic Revised Edition” (STL) by Quine.
“In formulating set theory in 1940 Godel appealed informally to fictious supplementary objects taht he called notions. They were like classes but were not values of variables. They were just a manner of speaking avoidable by cirumlocution. In my Brazlil lectures of 1942 I pressed this fiction. I presented what I called the virtual theory of classes and relations, in which I used the notation fo set theory as far as I could without assuming classes. See O Sentido da nova logica section 51. Martin was urging the idea concurrently. In Set Theory and Its Logic (1963, 1967) I made extensive use of the virtual theory of classes as an auxiliary to the real theory of classes. It proved to be a valuable axuiliary, but I persisted in presenting it as a mere prolegomenon to set theory and as a sly partial simulation thereof, failing, as we all did, to appreciate that it should stand squarely in elementary logic as a regimentation of a fundamental feature of language, the relative clause.” pg. 136 MOL
Predicate abstracts and functors render naive set theory as neither naive nor set theory. It becomes pure logic. The missing link between Frege and this surprising conclusion is von Neumann’s schematic methods.
There is no more to communication than effective action.
..x{..y: ---..y---..u--- :..u}..v if and only if ---..x---..v---.
In addition to committing to quantifiers when specifying the restrictions on substitution for predicate letters in valid schema to generate valid schema, as noted in the last entry, Quine also forgot to mention the problem of overlapping bondage in netsted abstracts with the same bound variables e.g. ‘{x: Fx and x{x: Gx only if xHy :y}x}’.
The example in 2 introduces a new abbreviation: ‘{..x: ---}’ for ‘{..x: ---:}’. Whether this shall be later advised is unknown.
Quine deals with the problem of overlapping bondage in “Mathematical Logic Revised Edition” in the way given by Hilbert and Ackermann: the inner most occurrence of bondage prevails over the others.
There remains the problem of an abstract whose binding prefix or postfix contains no variable free in its component. There also remains the problem of iterated abstracts e.g. ‘{x: {y: xFy}}’.
For now I adopt the two sided definition of predicate abstraction that most closely follows the patterns of development I uncovered in #a-stack-notation-for-predicate-functor-logic-2025-0414-1626.
..x{..y: ---..y---..u--- :..u}..v if and only if ---..x---..v---.
It is given with such strange notation as ‘---’ to indicate that the schematic premise itself is in fact an additional schematism which, though it can be given a more careful grammatical construction, is distracting at this moment.
“We may sum up the two restrictions symmetrically thus: Quantifiers of the substituted abstract must not capture variables of the schema in which the substitution takes place, and variables of the substituted abstract must not be captured by quantifiers of the schema in which the substitution takes place. These restrictions simply ward off confusions of variables which, if allowed, would cause substitution to deviate from tis intended purpose of interpreting term letters.” pg. 162 Quine, Methods of Logic 4th Edition (MOL)
Alas, this rule as given by Quine is needlessly tied up with quantifiers in large part so that fewer edits had to be made to weave in predicate abstraction as the logical importa of the grammarian’s relative clause without disrupting too much of the other text.
We may sum up the two restrictions symmetrically thus: abstracts of the substituted abstract must not capture variables of the schema in which the substitution takes place, and variables of the substituted abstract must not be captured by the abstracts of the schema in which the substitution takes place.
The main reason to promote the restrictions on substitution to pure predicate abstracts rather than to entertain quantifiers is to extinguish the mistake made by Frege when he introduced quantifiers and entangled them with the bondage and freedom of variables.
It also allows those who are interested in other predicate functors that might not conform to the old quantificational outlook to explore how deviant ones combine with the bound variables of predicate abstraction. For example, the whole problem of scope in the practices of computer programming are a consequence of the methods of predicate abstraction.
Note, the schematic premise of concretion introduces predicate abstracts, such as ‘{u: u pairs x with y :xy}’, as basic notation i.e. it ushers in a new gramamtical category by a new gramamtical construction and does so, by including the schematic premise of concretion in the contemplated theory, without enlarging the logic. Concretion assures us that anything said with abstracts can be said without them by eliminating each abstract for its concretion via the appropriate premise.
But, there is some care that must be taken and which I have not taken myself: predicate abstracts can be introduced as abbreviations for quantificational sentences, does the abbreviation correspond to the notation introduced by the schematic premise of concretion?
My repeated comments that Quine was not as careful with the problem of vacuous singular terms as he was with vacuous validity in #2025-0727-2102 was wrong: he makes repeated mention in, e.g. the chpater on identity, as he ramps up to his method presented in the chapter on the elimination of singular terms.
Among von Neumann’s many accomplishments (including the method of stored programs), perhaps his most concequential technique is his schematic method from his 1927 “Zur Hilbertschen Beweistheorie”.
“Es werden gewisse Schemata angegeben, d. h. gewisse Kombinationen von einfachen Zeichen, mit den Zeichen a,b,c,.. und xp, xq, .. sowie Subst derart, dass wenn wir fur a,b,c,… Formeln einsetzen und fur xp, xq,… Variable (unter Einhaltung der fur a,b,c,..; xp,xq,.. jeweils zu formulierenden Bedingungen) und dann die Subst ausfuhren, eine Normalformel entsteht.
Diese im folgenden anzugebenden Schemata sind die Axiomenschemata. Die Normalformeln, die so entstehen, dass isn einem Axiomenschema a,b,c,..; xp,xq,.. auf die oben beschriebene Art ersetzt und die Substausgefuhrt werden, sind die Axiome.” pg. 13-14 Mathematische Zeitschrift. 1927
“Certain schemata are given, i.e., certain combinations of simple symbols, with the symbols a, b, c,… and xp, xq, .. as well as substs such that if we substitute formulas for a, b, c,… and variables for xp, xq,… (subject to the conditions to be formulated for a, b, c,…; xp, xq,…) and then execute the substs, a normal formula results.
These schemata, to be given below, are the axiom schemata. The normal formulas that result from substituting a, b, c,…; xp, xq,… in an axiom schema in the manner described above and executing the substs, are the axioms.”
The distinction between schema and sentences (or, worse, sense) has raised a host of problems for the modern logician navigating a world awash with mathematical logic without any hint of its history and hurtles.
‘x is {y: Fy}’ for ‘some item is (z such that z=x and some item is (y such that y=z and Fy))’ (from pg. 17 of “Set Theory and its Logic Revised Edition” (STL) by Quine)
The abbreviation shown in 1 is from STL and is the full analysis of the various parts that variables play in a logic with predicate abstracts as primitive notation without actually introducing such basic notation.
Don’t forget to go through all of the canonical forms of truth functional compounds e.g. https://en.wikipedia.org/wiki/Boolean_function#Representation.
Here is a fresh attempt to gather together the methods of predicate abstracts and predicate functors from truth-functional and quantificational logic.
Perhaps unsurprisingly, the elimination of singular terms by singular descriptions is tied up with Wang’s single schematic premise for a theory of identity. I mention this to at least present some link between this note and the last one. There are many wonderful things occurring at once as a result of multiple mounds of work merging towards a single mountain of progress.
A quick note on terminology. Predicate functors break predicate logic into two grammatical categories: the recombic category and the alethic category.
Another note: all of predicate logic can be reduced to three predicate functors. I do not yet know if it is possible for fewer than three predicate functors to suffice for all of predicate logic. A reason to suspect this is the case is that Quine’s theory of abstraction and inclusion, like Church’s theory of abstraction and application, does the work of a theory of mathematical logic (in Quine’s case the theory is that of sets).
Here a clear distinction must be made between logic and mathematical logic. the grammatical particales ‘and’ and ‘or’ combine with component sentences in a grammatical construction that builds conjunctions and alternations as compound sentences. The unhappily named “boolean functions” designated by ‘and’ and ‘or’ are functions, they are singular terms, and they combine with other singular terms, such as those designated traditionally by ‘truth’ and ‘falsity’ as the so called ‘truth values’, to designate via a compound term like ‘truth and truth’ a truth value, in this example truth.
Inability to distinguish between ‘and’ as a grammatical particle and ‘and’ as a singular term designating a function lead to insufferable errors when arguments from predicate logic (where ‘and’ is a grammatical particle) are confalted with arguments from mathematical logic (where ‘and’ tends to designate a boolean function).
Surprisingly, much of truth-functional logic remains wedded to arguments which properly belong to mathematical logic e.g. whenever truth tables, trees, diagrams, or maps are invoked to establish the validity of a truth functional schemata. The logical alternative has been available for some time: full alternational normal form and full conjunctive normal form are both accessible through an argument by equivalents: truth functional validity is established outright by direct inspection.
Wang’s schematic premise of identity is “Fx if and only if some item is (y such that x is identical to y and Fy)”.
These most recent notes have mostly been born from discovery expedititions into Quine’s analysis of 1. singular terms, 2. identity and indiscernability, 3. descriptions, 4. predicate abstraction and concretion, 5. predication, and 6. predicate functors among other things.
“It is convenient sometimes to cite the axiom with its quantifier or quantifiers and soemtimes without, as the above examples shown.”
“For a more broadly applicable treatment of descriptions, then, I am turned back from Russell to Peano, who first brought descriptions into logic. It was Peano who introduced iota as well as epsilon into logic, both inverted and upright. For him singular description was premitive notation, and my definition of ‘the F = x’ was for him rather an axiom schema:
(the F)=x if and only if each item is (y such that Fy if and only if y=x)
Peano expressed the existence condition as ‘some item is (x such that (the F)=x)’ even as we, and invoked it as a condition whenever he used a singular description. David Hilbert followed him in this.
We can join company with Peano and Hilbert and Still adhere to my account of all singular terms as descriptions and my convention to use ‘a’, ‘b’, etc. as parameters for them all. Formally the resulting logic is the same, I believe, as Leblanc and Theodore Hailperin’s logic of free terms, provided that like those authors we extend the logical axioms of identity unconditionally to singular terms.” pg. 281-282 “Selected Logic Papers Enlarged Edition”
The extension hinted at in that last sentence is where the subitems of 10 most clearly come into play. It is the halmark of the schematic methods of logic in application to the analysis of singular terms to let free variables play the part of dummy singular terms in the same way that sentence letters play the part of dummy sentences and predicate letters play the part of dummy predicates. But, additional care must be taken when setting up terminology that deals with potentially empty singular terms i.e. the dummies must be managed in such a way that some interpretation may leave them without a designata!
This care was not quite taken by Quine overtly in MOL the same way that some care was taken overtly in MOL with the case of what I have called trivial validity which entertains the possibility of an empty universe by marking each existential as false and each universal as true and resolving the whole sentence by truth functional analysis to the truth value in the empty case.
Specifically, it is not until the third part of MOL that the problem of empty singular terms is even mentioned (in the first chapter/section of that part called “singular terms”). Then, unhappily still, it is not even hinted at as resolvable until three chapters later in “Elimination of Singular Terms” by a method of singular description to be further described here.
To link up MOL with Quine’s later paper on “Free Logics” I must give a bit of MOL and link what I give up with what I gave from that paper.
“Let us take up the problem, which has been looming for some time, of the truth value of such statements as ‘Cerberus barks’. Falsity, as a sweeping answer covering all statements containing ‘Cerberus’, would be over-hasty: first because the statement ‘There is no such thing as Cerberus’, at least, is true; and second, because whatever statements we adjudge false must admit of compounds, e.g., their negations, whcih will be true. Truth, as a sweeping answer, would encounter parallel difficulties.
Our deductive methods for singular terms throw no light on the question: for we already assume that the singular term names an object when we represent the singular term by a free varaiable, and we make the same assumption again when we adopt a descriptive premise for a description. Failing a named object, our methods show nothing, for what they purport to show rests then on a contrary-to-fact assumption. Common usage, moreover, likewise leaves us in the dark; for excepting such contexts as ‘There is no such thing as Cerberus’, a singular term is ordinarily used only when the speaker believes or cares to pretend that the object exists.” pg. 278 MOL
But, 19.1 and 19.2 come to be validities only in the crude way that what I have called trivial validities are explained away by Quine as having extralogical interest but yet also a logical resolution.
“Under ordinary usage, we saw (Chapter 3), truth values attach not to indicative conditionals as wholes but only to the consequents conditionally upon truth of the antecedents. Analogously, under ordinary usage truth values attach to contexts of singular terms for the most part only conditionally upon existence of the objects. But if we are to have a smooth logical theory we must fill such gaps, even though arbitrarily, in such a way that every statement comes to have a truth value. Thus it was that we conventionally extended teh concept of the conditionaly, in Chapter 3, so as to allow truth values generally to conditionals as wholes. An extension in the same spirit is needed now on the score of singular terms that do not name.” pg. 278 -279
Here I must note that the previous chapter on descriptions already introduced and eliminated a tricky problem that isn’t quite effectively distinguished from that mentioned in 18 and dealt with in spirit in 21.
It is the problem of descriptive (or described?) singular terms and nondescriptive (or undescribable?) singular terms as oppossed to the difference between empty and nonempty singular terms.
I am glad that I’ve written through these things because I had not made that distinction clear to myself before.
The definition in 25 can be simplified by noting that ‘each item is (y such that x=y if and only if Fy)’ implies ‘x=x if and only if Fx’ which is equivalent to ‘Fx’ so that when ‘F describes x’ for ‘each item is (y such that x=y if and only if Fy)’ then ‘F describes x’ implies ‘Fx’.
But note!!! The argument in 26 depends on a definition of validity (and consequently of implication) that presupposes each free variable designates some item!!! That is, it will be seen that Quine’s method of eliminating singular terms in MOL has to yield expected results.
Jumping ahead, predicates which belong to the lexicon when predicated of an empty singular term are taken as false. So, in an empty universe, the universal ‘each item is (x such that Fx)’ is false for any predicate substituting the predicate letter ‘F’ and ‘Fy’ when ‘F’ is substituted by a predicate of the lexicon is false, but when ‘F’ is substituted by a compound predicate …
There are two sets of notes to type up here: 1. the use of predicate abstracts in explaining free logics, and 2. a fuller and more compact exposition of the power of predicate functors to take over what was once relegated to (naive) set theory.
The first step from logic to set theory is the premise of extensionality of sets i.e. sets with the same members are the same.
But, before getting to logic, it’s time to deal with the problem of open sentences and free logics.
The phrase “free logic” comes through Hugues Leblanc and Rolf Schock according to Quine’s March 1994 paper “Free Logic, Description, and Virtual Classes” given on the occasion of Lebanc’s seventieth birthday in Montreal.
Free logic is then contrasted with logic where it is (errnoeously) said to include such premises as describe in 6.1 nad 6.2.
The problem of the empty universe is dealt with in standard predicate logic by the common method of “checking the empty universe” by marking each existential as false and each universal as true and resolving the whole by truth functional analysis.
Thus, logic retains the methods of prenexing and purifying by the rules of passage, and addresses the free logicians concerns about an empty universe.
I take this distinction a bit further than Quine and simply distinguish between trivial validity and nontrivial validity where “validity” comes to be short for “nontrivial validitiy”. A sentence is trivially valid if true from each interpretation in an empty universe e.g. by the method described in 8 of checking the empty universe. Then it is (nontrivially) valid if true from each interpretation in a nonempty universe i.e. by the familiar methods which embrace the rules of passage which otherwise are trivially nonvalid.
In “Methods of Logic 4th Edition” Quine gives a definition of validity outright relative to nonempty universes. I find my method far more inclusive in terminology which also emphasizes the outlook of, e.g., the free logician. So, if the free logician wishes to emphasize their outlook, they simply assert a sentence is trivially and nontrivially valid, where as the logician who is not particularly interested in trivial validity can stick with validities knowing that all free logicians are to read ‘validity’ as ‘nontrivially valid’.
Note, it is here where I find myself siding with those who criticize Quine for his tight restriction on what to count as methods of logic i.e. Quine left out the empty case and accomodated it by external invocation of truth-functional methods to “checking the empty universe”, but I keep it in logic as trivial validity.
The problem with singular terms which purport to designate is far more significant i.e. point 5.2 and its paraphrase in 6.2 are an opportunity for the methods of logic to shine in their accommodation of the free logicians concerns.
Singular terms which purport to designate but which don’t in a given theory are sometimes called ‘empty terms’ in parallel with ‘empty universes’. The classic example is ‘Pegasus’ or some other mythical creature which is presumably an empty singular term in any theory which purports to deal with this culturally significant singular term.
Quine’s analysis of singular terms is simple: they are all to be regimented as discriptions. Consequently, so called ‘free variables’ are to stand for some unspecified description. A description is formed from a descriptive predicate functor ‘the’ and is combined with a predicate abstract e.g. ‘the (x such that x is F)’ (where ‘is’ is taken as the copula of predication and NOT as a predicate e.g. not identity, not membership).
“Think of ‘a’ as a name, and think of ‘Fa’ as any sentence containing it. But clearly ‘Fa’ is equivalent to ‘some item is (x such that a=x and Fx)’. We see from this consideration that ‘a’ needs never occur except in the context ‘a=’. But we can as well render ‘a=’ always as a simple predicate ‘A’, thus abandoning the name ‘a’. ‘Fa’ gives way thus to ‘some item is (x such that Ax and Fx)’, where the predicate ‘A’ is true solely of the object a.
It may be objected that this paraphrase deprives us of an assurance of uniqueness that the name has afforded. It is understood that the name applies to only one object, whereas the predicate ‘A’ supposes no such condition. However, we lose nothing by this, since we can always stipulate by further sentences, when we wish, that ‘A’ is true of one and only one thing:
- some item is (x such that Ax)
- not some item is (x such that Ax, Ay, and not x is identical to y).
(The identity sign ‘is identical to’ here would either count as one of the simple predicates of the language or be paraphrased in terms of them).
The notation without names talks still of a and other objects, for they are the values of the quantified variables. An object can also be specified uniquely, still, by presenting some open sentence (in one variable) which that object uniquely satisfies. ‘Ax’ is such a sentence for teh object a. And the names can even be resotred at pleasure, as a convenient redundancy, by a convention of abbreviation. This convention would be simply the converse of the procedure by which we just now eliminated names. Each predication, let us say ‘Fa’, containing the name ‘a’, would be explained as an abbreviation of the quantification ‘some item is (x such that Ax and Fx)’. In effect this si somewhat the idea behind Russell’s theory of singular descriptions.” pg. 26 POL.
Quine hints at the next step but it is not yet clear from this explanation except that Russell’s method of eliminating singular descriptions is hinted at. Before moving onto the next step, I must note that Quine’s observation is profound: the method of abbreviating ‘some item is (x such that Ax and Fx)’ as ‘Fa’ is of grand consequence beyond it as an apparent intermediate step between singular terms (also called “names” in the quotation) and elimination of singular descriptions by Russell’s contextual method.
A bit more on notation that Quine did not get to: ‘..x is F to ..y’ is shortened to ‘..xF..y’ and predicate abstracts can be given “two sides” as in ‘{..x: ..xF..y :..y}’ which (I am now very excited) showed me that my method of using predicate abstracts to distinguish between variables as singular terms and variables as bound quantified colocaters is a key step to effective homogenization! (note, this insight does not require the left and right stack notation which I originally introduced here in Stack Notation for Predicate Functor Logic).
phew! let me take a slow step back because multiple things came together all at once there.
First, variables are used to abbreviate complex compound predicate functors in the following way.
They occur only and solely within a predicate abstract.
The steps in 17 are to be entirely discarded and replaced by the more carefully assembled predicate functors from here Stack Notation for Predicate Functor Logic.
By way of example, the predicate functor schema “(Drop F) only if (Push Drop F)” has a principle configuration of (1,1) since the principle configuration of a conditional is the max of left parts of the principle configurations of its components paired with teh max of the right parts of the principle configurations of its components, and the principle configuration of a drop is one plus the left part of the principle configuration of its component paired with the right part of the principle configuration of its component, and the principle configuration of a push is the left part of the principle configuration of its component paired with one plus the right part of its principle configuration. Finally, predicate letters have a principle configuration of zero paired with zero (unless they are given an explicit configuration by what I plan on introducing as a ‘rank’ operation like that of the J programming language but of a tree form).
A much easier way to find the principle configuration of a compound predicate functor schema is along the same pattern that I am about to demonstrate that goes from a compoud predicate functor schema to its abbreviation with variables.
For those familiar with truth-functional and quantification logic it is easier to explain first how predicate abstracts work and then how predicate functors work than it is to explain outright how predicate functors work.
This is because variables come to be bound by predicate abstracts and not by quantifiers in such a way that compound and complex predicates are easily constructed from sentences which only become open within a predicate abstract.
It also just occurred to me that predicate abstracts make it surprisingly obvious what the difference is between so called ‘free variables’ as singular terms and the variables bound by a predicate abstract.
It also occurs to me that Quine did not notice this distinction could be made with predicate abstracts.
An example: the sentence ‘x loves y’ is said, in classical truth functional and quantificational logic, to be open, and consequently, is often dealt with not as a sentence with singular terms ‘x’ and ‘y’, but rather as a predicate and not a sentence at all: ‘x loves y’ is said to be true of this paired off with that if and only if this loves that.
The accident here is of not recognizing the power of predicate abstracts!
The sentence ‘Tom loves Dick’ is as much an open sentence as ‘x loves y’, but because we are taught that such a sentence does not appear with respect to any old Tom or Dick, but with respect to some Tom and Dick that are otherwise uncovered by that all encompassing method of “context clues”, it is hard to see that ‘Tom’ and ‘Dick’ are variables just as ‘x’ and ‘y’ are.
This is made clear by the method of predicate abstraction where ‘Tom loves Dick’ and ‘x loves y’ are the same (they are complete sentences in that ‘x’ and ‘y’ are as much like ‘Tom’ and ‘Dick’ in that we’re not talking about any old x or y) and both are different from the predicates ‘{Tom,Dick: Tom Loves Dick}’ and ‘{x,y: x loves y}’ (which are in fact coextensive and hence quantificationally equivalent).
What’s more, when we look at the origin of the predicate abstract in the grammarians relative clause, we find that there are even in our native tongue a difference between ‘x loves y’ which is a perfectly fine closed sentence with items purportedly designated by the singular terms ‘x’ and ‘y’ and the incomplete sentence ‘x and y such that x loves y’.
In other words, predicate abstracts point up that what once went on as open sentences are actually incomplete sentences which have not been fully rendered as the relative clause which they purport to stand for.
This is far more profound than I had ever expected.
I am removing the following from [hints][#hints]:
** HOW TO CONTROL YOUR BEHAVIOR **
- Consequences select behavior.
- Control consequences to control behavior.
- Build new behavior from old behavior.
- Build new controls from old controls.
** HOW TO DEAL WITH ANHEDONIA AND UNCERTAINTY **
- Read Dalio’s “Life and Work Principles” when you feel anhedonia.
- Read Quine’s “Methods of Logic 4th Ed.” when you feel uncertain.
- Go to 1.
and I am removing 4 of
** HOW TO FIND A GOOD PSYCHOLOGIST **
- Never settle for a bad one.
- Good ones are exceptional people who happen to be psychologists.
- They are cultural liasons.
- You’re more likely to find a good psychologist through a friend than you are through a phycisian.
because I doubt its accuracy e.g. there may be families which point people to others who will only and solely reinforce their familial control and that this could be far more damaging than finding a psychologist through a phycisian.
I’ve made these edits because those hints do not help in the same way, e.g., that the hints on how prediate abstracts work or how indiscernibiliy works.
There are better hints to be made than these. The material contained within them will not be entirely dropped.
Another edit.
In [HOW TO MAKE IT LOGICAL][#how-to-make-it-logical] point 3 is stronger when given as ‘List true conclusions made from those few’. It prompts me to write point 4 as ‘Conclude as many from as few (premises) as you can.’ This introduces the important difference between premises and conclusions in the actual construction of a logical theory: everything that ‘goes into’ the construction is classed as a conclusion up to that point, it is only once the theory is assembled that the distinction between conclusions, premises, and truths.
Perhaps an even better way of doing this is to leave 3 as is “List ture sentences made from those few” and only add the parenthetical “(premises)” into point 4. I’ll try that out first: it follows the principle of minimal mutilation.
I just discovered that using backticks gives an inline quote under the rules of markdown. To get a literal back tick (for, e.g., use in writing out a javascript template literal) you must prepend a backslash to the backtick!
Instead of sectioning notes into days and days into entries timestamped down to the minute, I’ll try flattening everything down to the minute timestamped entries: I wasn’t very often making more than one entry a day except for right at the beginning and anything that can be done to flatten the organization seems like it is more likely to help than to hurt, but time will tell.
Last entry left off right as I was about to go over how what Quine calls ‘substitutivity of the biconditional’ or ‘substitutivity of covalence’ (and what is distinguishable from what he has called ‘subsitutivity of coextension’) gives a logical basis for the premises of a theory of identity.
I left off exactly at the moment that it occurred to me that in “Methods of Logic 4th edition” (MOL4) Quine made the distinction between ‘substitutivity of the biconditional’ and ‘substitutivity of the coextensional’ with the word ‘interchange’ and that he did so relative to truth-functional, monadic, and quantificational equivalence.
What is called ‘quantificational equivalence’ in MOL4 is elsewhere called ‘logical equivalence’, e.g. in ‘Philosophy of Logic 2nd Edition’ (POL2).
Most people are familiar with ‘logical equivalence’ as the counterpart to ‘logical implication’ (which is the most familiar) and, sadly, with ‘logical truth’ which, following the pattern from ‘equivalence’ and ‘implication’, is more helpfully called ‘logical validity’. In each case, the prefixed word ‘logical’ is dispensible if not just because there is no validity but that which is logical (a controvertial claim I am told) and it leaves room for distinguishing between different grades of validity e.g. truth-functional, monadic, quantificational, functorial, etc.
I’ll make this note here, though I may have already made it in the last entry but it can only help to repeat it, that the distinctions Quine makes in MOL4 are some of the most helpful that I’ve found in any of the books on math, logic, and science that I’ve read or been taught from.
This has largely to do with the fact that each sentence in MOL4 has evolved as a consequence of its context in the rest of the book.
Quine is highly adept at smoothly cross referencing forward and backwards within a text so as to tie together technical terminology into a critical mass of consequential distinctions.
It is only in “From Stimulus to Science” FSTS that there are perhaps more such distinctions: as difficult a text as it is, I can not recommend that anyone go on reading early Quine without cross referencing his past terminology with that in FSTS. Especially when navigating his nonlogical work (i.e. any work done outside the boundaries set up by MOL4).
The reason for this is that he drops certain distinctions that were once important in the evolution of his writing, but which were too sharp when literally applied. Subsequent revisions to terminological distinctions dulled and broadened the boundaries which softly separate concerns.
Perhaps the most important example is the change in terms from ‘Word and Object” (WAO) to FSTS: the later use of occasion, observation, and standing sentences is divorced more from their earlier commitment to what he then called ‘stimulus meaning’. In WAO he leaned too heavily on a crude grasp of verbal behavior as operant behavior which rendered it almost indistinguishable from the ancient analysis of verbal behavior under the stimulus-response psychology of Watson. Still, even in FSTS, Quine’s references to ‘dispositions’ and ‘global stimuli’ are an inadequate analysis of, e.g., Skinner’s “Verbal Behavior” in light of such work as in Skinner and Ferster’s “Schedules of Reinforcement”.
All that praise of Quine aside, the distinction he made in MOL4 was given as ‘Laws of Interchange’ rather than as ‘substitutivity of the biconditional’ or ‘substitutivity of the coextensional’.
The laws of interchange are given first in the chapter 9 “equivalence” pg. 63 of MOL4, where the equivalence of the chapter is specifically called truth-functional equivalence and contrasted with monadic and quantificational equivalence as already mentioned in 3 above.
“Substitution consists always in putting schemata for single letters, and for all recurrences of teh letters. When these restrictions are not met, the putting of one schema for another will be called not substitution but interchange. Thus interchange consists in putting one schema for another which need not be a single letter, and which need not be supplanted in all its recurrences.” pg. 63 MOL4
“Interchange of equivalents preserves validity, implication, equivalence, and inconsistency, and unlike substitution for letters, it even preserves consistency, nonvalidity, nonimplication, and nonequivalence.” pg. 64 MOL4
Here is another note I must make now: this just quoted sentence is an example of Quine using the terminology that I favor ‘nonvalidity’, ‘nonimplication’ and ‘nonequivalence’, but he mixes it with ‘inconsistency’ where, if he was to keep that method of prefixing, ‘invalidity’ can replace ‘nonvalidity’. I have settled on ‘nonconsistency’ and ‘non’ everything, because the definition of each such technical term is never given by me in my presentation of logic: their definition is implicit in taking ‘non’ as ‘not’ in any exposition where such technical terms occur.
What Quine calls “the third law of interchange” is what, following the other patterns he established with the phrases containing ‘substitution’, e.g. ‘substitution of the bicondtiional’, is either what I call “THE law of interchange” or “Interchange of Equivalents”.
Interchange of Equivalents covers two parts of Quine’s definition of ‘extensional contexts’ in FSTS: substitution of covalence (aka substitution of the biconditional) and substitution of coextensiveness (aka substitution of the coestensional). BUT, only when “interchange of equivalents” includes “Interchange of Truth-Functional Equivalents” and “Interchange of Quantificational Equivalents”.
The coextensional of a pair of predicates is the universal closure of their biconditional.
Quantificational equivalence is validity of the coextensional.
“Students of logic are familiar also with certain second-order generalizations regarding the validity of quantificational schemata. Conspicuous among these is the substitutivity of the biconditional, which may be represented thus:
each item is (..x, and y such that A if and only B) and CA, only if CB.
Here ‘A’, ‘B’, ‘CA’, and ‘CB’ represent any quantificational schemata such that the last two are alike except for containing as corresponding parts the respective schemata represented by ‘A’ and ‘B’; and ..’x’, and ‘y’ represent all variables , free in the schemata represented by ‘A’ and ‘B’, that are captured by quantifiers in the schemata represented by ‘CA’ and CB’ (See for example my Mathematical Logic section 18).” pg. 12 STL
The reference to section 18 of “Mathematical Logic Revised Edition” (ML) is another example of Quine’s deft cross referencing.
It is also a concrete example of how and when later and earlier writings of Quine can be interpreted with respect to the other. This is so important, and also this particular example is directly relevant to the general problem of establishing a logical basis for the premises of theories of identity (and establishing such links as there may be wtih theories of indiscernability) that I shall say mroe about section 18 of ML.
Section 18 of ML is titled “Substitutivity of the Biconditional” and it is from here that he borrows the same phrase in STL. The first paragraph gives what is there called “a principle” but what is in STL called “a second-order generalization”. In ML the principle is ultimately give as what is there called ‘a metatheorem’. Such invocations of the prefix ‘meta-‘ were dropped by Quine for the good reason that there it suggests a special status which is inappropriate e.g. it is still common for people to speak of ‘metatheories’, or perhaps even ‘metalogic’, as if they are some how above and beyond the reach of theory and logic. They are not in that they are just theories which purport to include some of the apparatuses of some theories as their items. Confusion on this issue is so great that an appendix had to be added by Quine to ML explaining the difference between a ‘metatheorem’ and a ‘theorem’: this appendix along with other problems involving the so called “higher order logics” entirely extinguished Quine’s use of such terminology in, e.g., MOL4.
“In section 13 a restriction was imposed according to which one formula can occur in another only in a context of quantification or truth-functional composition. This restriction gives rise, it will be found, to the following convenient substitutivity principle: if A and B are statements agreeing in truth value, then B can be substituted for any occurrences of A in any statement CA without affecting the truth value of CA. In other words, if the statements CA and CB are alike except that CB contains the statement B in places where CA contains the statement A, then CA and CB are alike in truth value if A and B are. In other words, (I) any statement of the form `${A} if and only if ${B}, only if ${CA} if and only if ${CB}` is true, where CB is like CA except for containing B in places where CA contains A. E.g., the conditional:
Smith met Jones if and only if Jones was in Omaha, only if each item is (x such that Smith met Jones or not Smith sold x to Jones) if and only if each item is (x such that Jones was in Omaha or not smith sold x to Jones)
is true.
I have been sitting for far too long i.e. since I started until now which is 2025 0722 2102.
I need a book holder.
There are a cluster of methods of logic that are almost unique to Quine but which are worthy of a braod inclusion in any presentation purporting to introduce a student to the study of logic.
The cluster of methods appears to crystalize around indiscernability.
It may also be said to crystalize around identity.
each item is (x and y such that x is identical to y if and only if each item is (F such that x is F if and only if y is F))
by the name ‘Leibniz’s Law’, and sometimes ‘Leibniz’s Biconditoinal Law’ when emphasizing that the connective between identity and the biconditional purporting to define ‘x is indiscernable from y’ (as the second order sentence ‘each item is (F such that x is F if and only if y is F) is itself the bicondtiional and not a conditional, with grammatical partical ‘only if’, or a converse conditional, with grammatical partical ‘if’.
Second order logic fails for the well known reasons, e.g. Lindström’s theorem, and must be abandoned wherever predicate logic prevails.
For more on the problems with putting predicates in positions suited for variables, see Quine’s chapter on “Deviant Logics” in “Philosophy of Logic 2nd Edition” (note, there are circumstances where a substitutional definition of quantification appears to carry out what n-order logics purport to deal with, but note that such a theory of logic is explained with predicate logic for exactly those reasons that primed Frege’s enterprise: extensionality).
each item is (x and y such that x is identical to y only if, x is F if and only if y is F)
is a premise of a theory subject to this law for (schematic) predicate letter ‘F’.
One more thing: since ‘F’ is not a schematic predicate letter to which only one variable is attached in the schema given in 7, the consequent of the universal conditional no longer carries the purported definition of indiscernability given by an allegidly second order sentence.
When the schematic letter ‘F’ is predicated of only one variable, then the predicates which are permitted to replace ‘F’, when forming an instance of a schema that contains it as a predicate letter, are only and solely one place predicates.
Indiscernability demands more than that each biconditional of a one place predicate with respect to one variable in the antecedent and another variable in the consequent be true. That is, it is not enough to somehow conclude ‘x is indiscernable from y’ from having shown that each instance of ‘Fx only if Fy’ is true (though, this is a line which has been taken in the past by those who are unable to distinguish between a predicate and its purported extension, or, worse, its perported property, attribute, essence, concept, quality, proposition, propositional function, or, in general, its intension as a conspicuously nonextensional substance or kind).
Note, to conclude ‘x is indiscernable from y’ from having shown each instance of ‘Fx only if Fy’ is true involves a theory in which predicates of truth, quotation, and substitution (or some other items which fascilitate the general method which Quine has called ‘semantic ascent’, but note that semantic ascent only involves quotation or its reflection, disquotation, and substitution may not be directly involved, though it appears unavoidably so, and to uncover this fully would require a digression from this digression so I shall not indulge it now).
Concretely, the conditional which paraphrases the conclusion of ‘x is indiscernable from y’ from the premise “each instance of ‘Fx only if Fy’ is true” is give as
each item is (x and y such that each item is (z such that z quotes a predicate only if the substitution of z for ‘F’ throughout ‘Fx only if Fy’ is true of (x,y)) only if x is indiscernable from y)
where ‘is true of’ is defined as in the consistent theory of predicates of ‘denotes’ from Quine’s chapter “Denotation and Truth” in “From Stimulus to Science” adapted to the careful method of extending ‘satisfaction’ to sequences in Quine’s chapter “Truth” from his “Philosophy of Logic”.
Thankfully, the method sketched in 13 is both like what must be done to paraphrase the traditional definition of ‘indiscernibility’ as ‘qualitative samness’ or ‘have all properties in common’ contemplated in the second order paraphrase of Leibniz’s law, and entirely obviated by a resolution of the problem pointed up in 11.
indiscernability is defined relative to the lexicon of a standard theory whose finite list of m, .., and n place predicates F, .., and G (respectively) as
and also seen not to depend on any particular feature of the natural numbers (though I admit to relying heavily on their convenience in pointing the way to systematically unraveling the abbreviations by ‘..’).
The full demonstration that this definition of indiscernability is adequate builds on the thurough demonstration by Quine in the first section of the revised edition of “Set Theory and its Logic”.
The demonstration there does not do full just to the problem posed by 11 here, which to reaffirm its relevance in this context, can be described as the problem of constructing an extensional paraphrase of Leibniz’s law.
Though I will not repeat Quine’s demonstration, I shall build upon it and emphasize the pieces relevant to the problem restated as in 17.
First, though Quine is not particularly interested in the larger problem of an extensional paraphrase of Leibniz’s law, the order of his presentation is surprisingly parallel to some arguments that have been presented to address this problem in the past.
Before I forget, identity and indiscernability are not to be confused with equality (which purports to be the extension of identity as a special kind of relation) and equivalence (as a predicate functor which when attached to, e.g., a two place predicate abbreviates the grammatical construction of the conjunction of attaching that same predicate to the predicate functors of reflexivity, symmetry, and transitivity, each of which should themselves not be confused for the narrower definitions of reflexive relation, symmetric relation, or transitive relation).
Sadly, I know of no good explanation of all the distinctions from 20 and the others used here: it is part because of that sad fact that I am even writing this now.
each item is (x, .., and y such that Fx..y if and only if Gx..y) and …F…, only if …G…
A footnote explains that the full demonstration of this can be found in the 13th section of Quine’s Mathematical Logic the revised edition (I can not overemphasize how important it is to get the latest edition of Quine’s books because they evolved in often drastic ways that change the consequences of his work from trivial to nontrivial).
Now, in Quine’s “Methods of Logic 4th edition”, there is a distinction made between substitutivity of the biconditional and substitutivity of the coextenstional (where the coextensional of a pair of predicates is the universal closure of their biconditoinal; it was originally introduced by Peano who attached what we now call the universally quantified variables to the mark of his biconditional (which was a three parallel equally lengthed lines stratled by its components, like an equal sign with one more line)).
Interestingly, in Quine’s definition of ‘exentional context’ in “from stimulus to science” he calls “substitutivity of covalence” what he can, and perhaps should, call “substitutivity of biconditional”! This is an interesting example of a common problem when it comes to dealing with logic: we are not the first to have arrived in such wilds and hence not the first to talk about all these beautiful and sometimes dangerous things).
Substitutivity of the biconditional, in Quines “Methods of Logic 4th edition”, is first established for truth-functional logic in chapter 9 on “equivalence”, and I am glad I went and hunted for it because I need to switch up my technical terms in order to avoid some deep pit falls.
The main distintion I’ve messed up is that between ‘substitution’ and ‘interchange’. The word ‘substitution’ is usually contrasted with the word ‘replacement’, and, often enough, defined with respect to it as “replace each occurrence of this in that with it”. Note, “replacement of occurrences” and the more carefully given “replacement of nonoverlapping occurrences” are precisely the problems that Quine addressed when he gave the first (surprising I know) definition of “substitution” in the 1936 paper (again I know that it is hard for people to accept that there was not a definition of “substitution” prior to 1936) “Definition of Substitution” to the “society of fellows, harvard university”.
Outside “Methods of Logic 4th edition” Quine picked the word ‘substitution’ rather than ‘interchange’ or ‘replacement’ because it is far more familiar to the audience of his books: ‘interchange’ is unfamiliar outside Quine’s “Methods of Logic” and that is why it is so valuable.
What I have come to call “substitutivity of the biconditional” and “substitutivity of the coextensional” are cast by Quine as “Laws of Interchange”.
Interchange comes up as distinct from substitution (and more suggestive than replacement) as a result of examining methods of generating validities from valid schema.
Sadly, I’ve been writing for a little under three hours straight (it’s 202507211944) and must stop here: hopefully I’ve laid enough foundation to pick up where I left off and get on to the cluster of methods from Quine (whcih include, identity as indiscernability, definition of predicate abstraction in a truth-functional and quantificational logic with a predicate of identity, and the elimination of singular terms as singular descriptions and the elimination of singular descriptions either by Russell’s contextual definition or by Peano’s premises, and the relation of those methods to the ones presented in methods of logic, where “descriptional premises” are overtly mentioned as part of theories which deal with items purportedly designated by singular descriptions, and the role of singular descriptions in the introduction of variables into a predicate functor theory and how they are related to predicate abstracts, and now it should be obvious why there is so much work done here and yet to be done here).
Don’t play the piano too much: it gets in the way of typing things up and the thinking that comes along with that.
I have not returned fully to my work on primitive recursive number theory, primitive recursive calculus, and primitive recursive analysis since I fully familiarized myself with Quine’s methods of logic. I’m returning to it fully now.
As Quine repeatedly showed throughout his later works: identity is coextensive with indiscernibility in a standard logic. A standard logic is one with a finite lexicon i.e. a finite list of basic predicates from which all grammatical categories are constructed via predication (with variables, which are also classed as part of the lexicon under technical complications, involving accentuation in the simplest case, but which are entirely dropped when predicate functors are taken as the only grammatical particles and predicates are the only items of teh lexicon).
There are many things that I finally wrote up that I have not copied here. Giving the definition of indiscernibility relative to a lexicon of predicates is one such thing.
In a truth functional and quantification logic, indiscernibility can be given as a predicate functor which takes a finite list of predicates and a pair of variables and constructs the relevant compound sentence.
Schematically, for predicate letters ‘F’, ..G, and ‘H’ of f, ..g, and h places (the number of places of a predicate letter tells us how many variables are to attach to any predicate which comes to substitute it: recall that each predicate has a shape, and, traditionally, it is given as a length from counting the number of variable letters attached to it e.g. ‘x parented y’ shows that ‘parented’ is a two place predicate). I use ‘..G,’ as a further schematism to avoid the trouble of introducing some scheme of subscripts, superscripts, or other indexical marks that tend to clutter methods presented for an unspecified but finite list of things (those who doubt there are definite methods of so indexing can start with predicate letters ‘F0’, ‘F1’, ‘F2’, …, ‘F’ concatenated with the decimal numeral of some number designated by ‘N’, but note that this tends to make it appear as if an argument or construction is the result of some number theory even though it is not; the only method by which all such doubts can be eliminated is to present a logical theory of logic but that is precisely what I am trying to avoid).
Write ‘x equals y with respect to F’ for ‘each item is (u,..v, and w such that Fxu..vw if and only if Fyu..vw, Fux..vw if and only if Fuy..vw, …, Fu..vxw if and only if Fu..vyw, and Fu..vwx if and only if Fu..vwy)’ where, all together, the list of distinct variable letters ‘u..vwx’ is of length f.
Write ‘x equals y with respect to F, ..G, H’ for ‘x equals y with respect to F, ..G, and each item is (u,..v, and w such that Hxu..vw if and only if Hyu..vw, Hux..vw if and only if Huy..vw, …, Hu..vxw if and only if Hu..vyw, and Hu..vwx if and only if Hu..vwy)’ where, all together, the list of distinct variable letters ‘u..vwx’ is of length h.
Other than by Quine’s method of quasiquotation, the mechanical production of the sentence abbreviated by ‘x equals y with respect to F, ..G, H’ can be got by the now standard string manipulation functions of modern computer programming languages. Note, it is strange to see Quine’s method of quasiquotation presented as ‘templates’ or ‘macros’ without directly mentioning their origins in Quine’s “Mathematical Logic”. The reason this is so strange, is that they were introduced for a specific reason, to make it all but impossible for a careless reader to falal into the use-mention trap, and yet that reason seems to go unheeded by many designers of programming languages and many programmers of such languages.
I must go out of my way to mention that Quine’s method of quansiquotation would save anyone from ‘what color is your function’, among many other strange methods that have cropped up recently in programming languages. Though, I must also go out of my way to mention that literally coloring functions, as in colorForth, uhdForth, or my own programming language is supported by Quine’s methods, so that might put people off because his methods are too liberal. To those so put off, I would have them restrict Quine’s method of quasiquotation rather than toss it out entirely as some ancient and useless relic of a bygone era.
A better way of introducing indiscernibility occurred to me while I was thinking to myself of better ways of getting it into a form amenable to inclusion in hints.
For the predicate letter ‘F’ write ‘=F’ for ‘{xy: each {u..v: Fxu..v if and only if Fyu..v, .., and Fu..vx if and only if Fu..vy}’ and ‘=(F,..,G)’, or just ‘=’, for ‘{xy: (=F)xy, …, and (=G)xy}’.
That method does depend on mastery of predicate abstracts, but there is no excuse for not mastering them when you belong to a culture that insists on the conspicuity of variables.
This note was supposed to deal with primitive recursive arithmetic, but it ended up making a new hint: how indicernibility works.
Notes on “A first-order axiomatization of the theory of finite trees” by Backofen, Rogers, and Vijay-Shanker 1995
This paper is related to my work on the theory of ordered pairs. The basic predicate in such a theory is ‘x pairs y with z’ and it is written ‘xPyz’ for short (technically, ‘P’ is short for ‘x, y, and z such that x pairs y with z’, or using Quine’s predicate abstract notation, ‘{xyz: x pairs y with z}’
Extensionality as given is a far cry from what most are used to when dealing with the basic principle of ordered pairs (as it is so often called) e.g. ‘(x,y)=(u,v) if and only if x=u and y=v’.
But, it corresponds directly to the more familiar principle of extensionality of sets e.g. ‘each item is (x such that x in y if adn only if x in z) and y in w, only if z in w’.
Even this is not so often noticed as the principle of extensionality of sets. In English “sets with the same members are equal” i.e. each item is (x such that x in y if and only if x in z) only if y equals z.
The parallel principle for orderd pairs is ‘each item is (u and v such that xPuv if and only yPuv) only if x=y’.
Old version of hint for “what to do” which was itself a newer version of the hint “what I must do before I die”.
WHAT TO DO
- Discover, predict, and control changes
in counts, rates, and accelerations- as selections from variations
on physical, chemical, biological, behavioral, and cultural scales- by making and maintaining strong practices
mediated by strong people marked by strong principles- from the sciences of
logic denotative, Boolean, functor
mathematics calculi, collections, categories
physics quantum, thermodynamic, gravitional
chemistry phyiscal, biophysical, biological
biology oranelles, organisms, environments
behavior biological, biosocial, social
culture history, science, technology
I have not worked on my stack based programming language for a while.
It is written in javascript: there’s no other language that seems as well known or as easy to know (and play with) than it. It will eventually be given as a logic program with Quine’s main method.
The garabage collector implements “A Linear Algorithm for Copying Binary Trees using Bounded Workspace” by K. P. Lee from March 1980.
In the past I took pictures of the code and posted it to Twitter. Now I shall do more to explain myself from the beginning.
let at = new Array(65536);
L = n => at[n*2];
R = n => at[n*2+1];
X = (n,L,R) => (at[n*2]=L, at[n*2+1]=R, n);
components of croppings denote when some item is the right of the left with the right of waow the compound denotes
WRONG!
An existential quantification consists of some sentence preceded by an existential quantifier whose variable is, say, the ith variable of the alphabet. This quantification, then is satisfied by a given sequence if and only if the constituent sentence is satisfied by some sequence that matches the given one except perhaps in its ith place. POL2 pg.39
This would change the definition of croppings as follows: croppings denote waow there is somewhere denoted by the component that matches it except perhaps at the right of the left of it.
The problem with this change is unfortunate because it does not make the compound a smaller shape than the component: it makes it the same shape. There is something important about making everything the same shape, and ultimately this may be the only option available to me because of precisely the problems Quine navigates with subtle talk of satisfaction by sequences rather than satisfaction by sequences of a given length on page 37 and 38.
For now I’m settling on the somewhat cryptic ‘Components of croppings denote the left with some item, with the right of waow the compound does.’
It is cryptic because it is complete yet, perhaps, incompressible.
Press shift and space to scroll up a page.
Once upon a time, a man grew up all alone. One day, a woman taught him to speak: she held up an apple, said ‘red’, he held it up, said ‘red’, she smiled, he held up another, said ‘red’, she frowned, held it up, said ‘green’, and so on.
Predicates do or don’t denote (are true or false of) where they occur. Constructions connect predicates, lexicons list atomic predicates, and grammars generate categories from recurrent constructions on lexicons. Logic reveals validities whose supplemented lexical substitutions (SLS) denote everywhere (Quine).
Constructions inheret where they denote from their components (Tarski) e.g. functionally, e.g.
joint denials denote waow each of their components don’t,
negations (self joint denials) denote waow their component does not,
alternations (negations of joint denials) denote waow some of their components do,
conjunctions (joint denials of negations) denote waow each of their components do,
alternative denials (alternations of negations) denote waow some of their components don’t; and
determinatively, e.g.
closures denote waow there is somewhere denoted by their component, and
components of croppings denote when some item is the right of the left with the right of waow the compound denotes; and
recombically, e.g. components of
drops denote the left of the left with the right,
pushes denote the left with the left of the right, with the right of the right,
hems denote the left with, the right of the left with the right,
of waow their compound does.
Predicates are
consistent waow their negation isn’t valid (some of their SLSs denote somewhere) ,
implied by others waow the conjunction of their self (the conclusion) with the negation of the other (the premise) isn’t consistent (each of their SLSs denotes where the same of the other does), and
equivalent to others waow they are mutually implicative (each of their SLSs donotes waow the same of the other does).
Examples:
Alternations of compounds with their negations are valid (they denote waow the compound does or its negation does, i.e. waow it does or does not, so, each SLS denotes everywhere).
Conjunctions of compounds with their negations are not consistent (they denote waow their compound does and its negation does, and, hence, waow it does and does not, so, each SLS of the conjunction does not denote everywhere i.e. the negation of the conjunction is valid).
Compounds are implied by and equivalent to their self.
Write a memon on Goodstein’s primitive recursive equation calculus and its notational extensions through to primitive recursive analysis e.g. the primitive recursive elementary functions and primitive recursive calculus (maybe even primitive recursive geometry through the Tarski/Hilbert form of the predicate ‘x between y and z’ or ‘x is colinear to y with z’ or ‘xyz colinear’ or, just, ‘colinear’.
The aim is primitive recursive quantum field theory i.e. primitive recursive functional analysis and operator theory.
Collect primary source I already have in my library, and collate with references from e.g. A Short Introduction to the Quantum Formalisms by Francois David.
Also, what’s up with the behaviorist Jack Michael? https://en.wikipedia.org/wiki/Jack_Michael He appears to have caused a lot of trouble for the science of behavior by building what he called “Applied Behavior Analysis”. The problem appears to be that the applied behavior analysts are not so much worried about the science of behavior as they are with its consequences in capitally controlled cultures i.e. they appear only to take advantage of those fragments of the science of behavior which they find most convenient without being subject to the scientific countercontrol which the science of behavior would otherwise enforce.
Once upon a time, a man grew up all alone. One day, a woman taught him to speak: she held up an apple, said ‘red’, he held it up, said ‘red’, she smiled, he held up another, said ‘red’, she frowned, held it up, said ‘green’, and so on.
Predicates do or don’t denote (are true or false of) where they occur. Constructions connect predicates, lexicons list atomic predicates, and grammars generate categories from recurrent constructions on lexicons. Logic reveals validities whose supplemented lexical substitutions (SLS) denote everywhere (Quine).
Constructions inheret where they denote from their components (Tarski). They sort into three bins: recombic, e.g. components of
drops denote the left of the left with the right,
pushes denote the left with the left of the right, with the right of the right,
hems denote the left with, the right of the left with the right,
of waow their compound does; functional e.g.
joint denials denote waow each of their components don’t,
negations (self joint denials) denote waow their component does not,
alternations (negations of joint denials) denote waow some of their components do,
conjunctions (joint denials of negations) denote waow each of their components do,
alternative denials (alternations of negations) denote waow some of their components don’t; and
determinative, e.g.
closures denote waow there is somewhere denoted by their component, and
components of croppings denote when some item is the right of the left with the right of waow the compound denotes.
Predicates are
consistent waow their negation isn’t valid (some of their SLSs denote somewhere) ,
implied by others waow the conjunction of their self (the conclusion) with the negation of the other (the premise) isn’t consistent (each of their SLSs denotes where the same of the other does), and
equivalent to others waow they are mutually implicative (each of their SLSs donotes waow the same of the other does).
Sometimes I get really mad at the world for letting so many bright stars burn out rather than helping them to shine on.
Components of lefts and rights, and the left and right components of Cartesian pairs, denote the corresponding parts of waow their compound does.
Wait, are these recombically complete? If there is no duplication then there is no way to construct the relfection of a predicate.
It just occurred to me that I’ve been down this road before when looking for a recombically complete set of predicate functors.
After you define the left, the right, and the Cartisian product functor, you see that you can’t construct reflections, so you add reflections constructions in e.g. reflections denote the ordered pair whose left and right part match where and only where (waow) their component does.
Then you’re left without a construction for the symmetric/swap of a prediate i.e. symmetrics denote the right with the left part of waow their component denotes.
Next, it is uncovered that there is no way to move items around the trees of items constructed from lefts, rights, cartesian pairs, reflections, and symmetrics, so the left and right shifts come up as a solution i.e. the component of the left shift denotes the left with the elft of the right, with the right of the right of waow the comopund denotes, and the comopnet of the right shift denotes the left with the left of the right, with the right of the right of waow the compound denotes.
Here, with the left and right shifts, we have the stack operations of push and pop, or in traditional forth notation ‘>R’ and ‘R>’ which are read aloud as ‘to ar’ and ‘ar from’. You can stick with these accumulated operations that are entirely ignorant of the duality of lists and stacks as the right and left parts of an ordered binary tree, but there are diminishing returns as it is uncovered that the left and right shifts which are called for inevitably, permit the construction of reflections and symmetrics that are dealt with in calculations with commensorate methods.
For example, components of reflections denote the left with, the right of the left with the right of waow their compound denotes, and components of symmetrics denote the left of the left with the left of the right, with the right of the left with the right of the right of waow their compound denotes.
The definitions of 4 and 5 are then of major reflections and symmetrics, and the definitions of 8 are their minor extensions.
You’re not bad for feeling things even though the things that you feel may feel bad.
Draft of definitions of recombically complete predicate functors for memo on logic:
The component of a
drop denotes the left of the left with the right,
push denotes the left with the left of the right, with the right of the right,
hem denotes the left with, the right of the left with the right,
of waow the compound does.
Finish collecting Quine quotations on singular terms in 2025 0629 1959.
Once upon a time, a man grew up all alone. One day, a woman taught him to speak: she held up an apple, said ‘red’, he held it up, said ‘red’, she smiled, he held up another, said ‘red’, she frowned, held it up, said ‘green’, and so on.
Predicates do or don’t denote (are true or false of) where they occur. Constructions connect predicates, lexicons list atomic predicates, and grammars generate categories from recurrent constructions on lexicons. Logic reveals validities whose supplemented lexical substitutions (SLS) denote everywhere (Quine).
Predicates are
- consistent waow their negation isn’t valid (some of their SLSs denote somewhere) ,
- implied by others waow the conjunction of their self (the conclusion) with the negation of the other (the premise) isn’t consistent (each of their SLSs denotes where the same of the other does), and
- equivalent to others waow they are mutually implicative (each of their SLSs donotes waow the same of the other does).
Constructions inheret where they denote from their components (Tarski) e.g.
- joint denials denote waow each of their components don’t,
- negations (self joint denials) denote waow their component does not,
- alternations (negations of joint denials) denote waow some of their components do,
- conjunctions (joint denials of negations) denote waow each of their components do,
- alternative denials (alternations of negations) denote waow some of their components don’t, and
- closures denote waow there is somewhere denoted by their component
Each substitution in valid predicates for (supplemented) lexicon denotes everywhere.
Logic Programming Notes by Frank Pfenning https://people.mpi-sws.org/~dg/teaching/lis2014/modules/lp-fp-07.pdf
It has been a while since I’ve worked on my calculus notes. I now have a clearer aim to put it into a memo.
Calculus, for me, is indistinguishable from primitive recursive analysis. In particular, the synthesis of algebra and geometry which hits the big tools of math in the rest of the sciences is precisely the synthesis of projective geometry and primitive recursive arithmetic that fits closely with Tarski and Givant’s metamathematical methods in geometry.
Sadly, there is no glory in diving into these methods of calculus without first finishing the memo on (predicate) logic. The exact primitive recursive methods, either in the geometric or algebraic constructions, are those which follow from the proper premises as in logic programming.
Logic programming is to be shown to follow immediately from Quine’s main method. Specifically, it follows immediately from the functional normal form of Quine’s main method where the purely notational tool of compound singular terms (which plays the logical part of Skolem’s functions).
Quine’s work on singular terms changed my outlook on logic: it turned it active where it had once been transcient.
The relevant definitions are these.
In terms of logical structure, what it means to say that the singular term “purports to name one adn only one object” is just this: The singular term belongs in positions of the kind in which it would also be coherent to use variables ‘x’, ‘y, etc. (or, in ordinary language, pronouns). pg.260
General terms, in contrast to singular ones, do not occur in positions appropriate to variables. pg. 261
The ‘x’ of an open sentence may refer to objects of any kind, but it is supposed to refer to them one at a time; and then application of ‘each {x:’ or ‘some {x:’ means that what the open sentence says of x is true of all or some objects taken thus one at a time. pg. 261
Just as the sentence letters in a schema stand as dummy sentences and the term letters ad dummy general terms, so the free variables may be seen as standing as dummy singular terms. pg. 262
It is usual in logic to write ‘the {x:’, with the inverted iota, to mean “the object x such that’. pg 274
Singular terms are called descriptions when written in this form. The singular terms of ordinary language which may be represented thus as descriptions begin typically with the singular ‘the’, but by no means necessarily so. pg. 274
In general a singular term purports to name one and only one object, and in particular a singular term of the form ‘the {x: Fx}’ purports to name the one and only object of which the general term represented by ‘F’ is true. Thus, if y is the object the {x: Fx}, then y must be such that y Fs and nothing-but-y Fs. this conjunction amounts to saying that, for each thing x, ‘F’ is true of x if x=y, and false of x otherwise. In short: each {x: Fx if and only if x=y}. pg. 274
If ‘F’ is true of nothing or of many things, then there is no such thing as the {x: Fx}. Actually the term appearing in the role of the ‘F’ of ‘the {x: Fx}’ in verbal examples from ordinary discourse very frequently needs supplementary clauses to narrow it down to the point of being true of only one object, but this situation can commonly be viewed merely as a case of the familiar practice of depending on context or situation to resolve ambiguities of ordinary language. Moreover, it would be unnatural to construe all use of the singular ‘the’ in this way; often a better account is simply pronominal. Commonly ‘the boy’, ‘the car’, serves merely as a pronoun whose grammatical antecedent is some name or description or perhaps some general term and quantifier. In logical notation such a pronoun might come through simply as a bound variable. pg. 275
arguments involving a singular term can be carried through by straight quantification theory with a free variable, say ‘y’, for the singular term, but that the application of the results depend on construing y as the object named by the singular term, and hence is contingent on existence of such an object. This construing of y, and the existence assumption on which it rests, figured nowhere in the schematism of the proof, but only in the informal step of application. pg. 275
the beauty of descriptions is that here the construing of y as the named object can itself be schematized quite explicitly as an additional premise of the form ‘each {x: Fx if and only if x=y}’. So our technique for arguments involving descriptions is as follows: we use free variables for the descriptions as for any singular term, but we also add a descriptional premise of the form ‘each {x: Fx if and only if x=y}’ for each description. an axiom of identity is also usually called for, because of the ‘=’ in the descriptional premise.
Now more than ever I am in a position to evolve a new experimental culture which, like the older tinier one, is governed by the principles of an experimental analysis of behavior.
The hints, memos, and notes made by members of such a new culture are only secondary to the concrete contingencies of the world from which they are selected.
Copleston’s history of philosophy might otherwise have been written as something like Foxfire.
In 2 of 2025 0629 1518 Quine’s initial definitions of ‘lexicon’ and ‘particle’ are given. In his book “Philosophy of Logic Second Edition” (POL2) they are defined as “words”. This is refined in “From Stimulus to Science”, indirectly, through his definition of ‘predicate’: “a predicate in my sense is always an integral word, phrase, or clause, grammatically a noun, adjective, or verb. Some are generated from others by grammatical constructions, notably the relative clause or, formally, predicate abstraction or predicate functors.” (pg. 61 From Stimulus to Science).
I want fast answers – that are sound, salient, and smooth – to apt questions. Rarely having them energizes me.
In “Philosophy of Logic Second Edition” Quine first defines lexical and partical items as “the words classed in the categories comprise the lexicon, whereas the words or signs that are not thus classified but are handled only as parts of specific constructions are the particles” (pg. 26-27).
The books I take notes from are hard to keep open. When I type I use both of my hands and have none free to hold open books. A bookstand may help me.
An experiment with a bookstand is like my experiment with earmuffs.
I’ve wasted so much time trying to figure out how spaces work with lists and sublists.
In working out how to alphabetize a schema the problem of freedom and bondage occurs. This is dealt with conspicuously by predicate abstracts and HOW PREDICATE ABSTRACTS WORK.
Quantifiers of the substituted abstract must not capture variables of the schema in which the substitution takes place, and variables of the substituted abstract must not be captured by quantifiers of the schema in which the substitution takes place.” (pg. 162 MOL4)
not (p and q), iff (not p) or not q
not (p or q), iff (not p) and not q
and with the law of double negation
not not p, iff p
yield the negational normal form where only predications are negated: push all negations inward and they pile up on predications until they are eliminated by hitting another already attached to a predication.
“an instance of a quantification exactly matches the old open schema that followed the quantifier, except that it may show a different variable in place of the recurrences of the variable of the quantifier. If it does show a different variable, it must show it in all the places (at least) where the old variable had been free in the old open schema that followed the quantifier. Moreover, it must show it free in those places. The reader will recognize in these requirements the effects of the restrictions on substitution for term letters.”[pg. 180 MOL4]
“A universal quantification implies each of its instances, and an existential quantification is implied by each of its instances.” [pg. 180 MOL4]
The definition in 3 of ‘an instance of a quantification’ misses the chance to define ‘instantial variable’ in the relevant context. Quine defines it here: “The instantial variable, which is substituted for the variable that was bound by the dropped existential quantifier, must be new. More accurately: it must be free in no line prior to this instantiation.” [pg. 190-191 MOL4]
There is a lot of relettering stuff that seems like it could be avoided by another normal form e.g. alphabetized normal form. Presumably this is a form where all free and quantified variables are replaced from left to right with variable letters in some alphabetical order.
More memo work
A basic technique in quantification theory is transformation of a formula in such a way as to bring all its quantifiers out to the beginning (prenexing) or, alternatively, to drive every quantifier in so that it governs only clauses in which its variable recurs (*purifying). The transformations depend on eight familiar equivalences called the rules of passage:
- some {x: p and Fx} if and only if (iff) p and some {x : Fx}
- each {x: p and Fx} iff p and each {x: Fx}
- some {x: p or Fx} iff p or some {x: Fx}
- each {x: p or Fx} iff p or each {x: Fx}
- some {x: p only if Fx} iff p only if some {x: Fx}
- each {x: p only if Fx} iff p only if each {x: Fx}
- some {x: Fx only if p} iff each {x: Fx} only if p
- each {x: Fx only if p} iff some {x: Fx} only if p
Four of these eight fail for the empty universe. Only 1, 4, 6, and 8 carry over.
A pure existential is valid if and only if we get a truth functionally valid schema by taking the alternation of the results of substituting the free variables for the existential ones in the matrix.
Where the matrix is the truth functional part of the prenex form (chop off the quantifiers and the matrix is what you’re left with). (I think the phrase ‘matrix’ is from Russell.)
The decision procedure for validity of pure existentials suggests that when prenexing always bring out universal quantifiars first in the hope that a pure existential shall be got.
each {x: some {y: each {z: Fxy, not Fxx, and Fyz, only if Fxz}}}
From a 1933 paper by Godel it is known that this prefix ‘each {x: some {y: each {z:’ is the simplest that a prenex infinity schema can have.
There is no complete proof procedure for showing schemata to be infinity schemata. For, if there were, we could add it to our complete proof procedure for finite consistency, and get a complete proof procedure for consistency. (pg. 215 of Methods of Logic 4th ed.)
where
The schemata that are consistent by not finitely so are the stubborn kind. They may be called infinity schemata. [pg. 215 of Methods of Logic 4th ed.]
The practical course of action is to identify grammatical categories which are just narrow enough to admit a decision procedure. That there is no complete proof procedure for consistency tells us that this practical course must always leave out some schemata that we may later be interested in. This is not a sadness because practical action also suggests we work on problems that are just narrow enough to be solved during an average human life span, or on the scale of the human species.
We see thus that our main method is a sound one. It is a rather natural one, moreover, despite the austerity of the argument above. The patter (1)-(12) could be verbalized, as a disproof of some conjunction of actual statements in lieu of (1)-(3), along the following line. According to (1), there is something that is F to everything. Very well, call it z. So we have (4). But (2) said that everything x is such that … . Well then in particular z will be that way. So we have (5): that there is something such that … . Call it t. Continuing thus, we get to the contradiction (7), (10)-(12). The premises (1)-(3) are thus disproved. One or two of them may be true, but not all three. [pg. 193 MOL4]
It is important to mention that this “verbalization, as a disproof” is natural in that it is the one which can be found in many of Socrate’s arguments as told by Plato. Furthermore, it is that method of argument which often has the strongest rhetorical punch: “so you say that such and such is the case, well let us assume it is, then there is such a thing as what you say, call it ‘x’, and so it is true of the other thing you’ve said, but x can not be both so and not so, we must have started wrong.”
The giant problem in firmly grasping the soundness of the main method is in the apparent absense of implication with respect to steps from existential premises to the lexically restricted instances. This is beautiful because it explains why the strictures of implication are not apparent throughout Plato. The role of implication is overblown by modern methods. Validity is too, but less so, and though proof theory has done the most to curb our enthusiasm, logical methods are, metaphorically, everything and nothing: they deal with everything because they deal with nothing in particular.
Once upon a time a man grew up in isolation. He only knew how to grunt. One day, a woman taught him to speak. She held up an apple, said ‘red’, he held it up, said ‘red’, and then she smiled. He held up another apple, said ‘red’, she frowned, held it up, and said ‘green’. On and on it went.
Predicates denote or don’t, i.e. are true or false of, the occasions on which they occur. Constructions pair component predicates with connectives. Lexicons list atoms, predicates without components. Grammars generate categories from repeated constructions on lexicons. Negations, caps, crops, dups, drops, pushes, pops, and swaps pair their component with ‘not’, ‘cap’, ‘crop’, ‘dup’, ‘drop’, ‘push’, ‘pop’, and ‘swap’ respectively. Conjunctions, alternations, and alternative denials pair their left and right components with ‘and’, ‘or’, and ‘nor’ respectively.
Constructed compounds inheret where they denote from their components. Negations denote where and only where (waow) their component doesn’t. Conjunctions/alternations denote waow each/some of their components do. Alternative denials denote waow each of their components don’t. Caps denote waow there is somewhere denoted by their component.
Occasions, like conjunctions, alternations, and alternative denials, split into left and right parts: the pile and the list. The list and pile do too: the right part of the pile is called the top, the left part of the list is called the head, the left part of the pile is called the rest of the pile and the right part of the list is called the rest of the list.
One occasion is the pop of an other waow
- the pile of the one is the rest of the pile of the other,
- the head of the one is the top of the other, and
- the rest of the list of the one is the list of the other.
One occasion is the push of an other waow
- the rest of the pile of the one is the pile of the other,
- the top of the one is the head of the other, and
- the list of the one is the rest of the list of the other.
One occasion is the drop of an other waow
- the pile of the one is the rest of the pile of the other and
- the list of the one is the list of the other.
One occasion is the dup of an other waow
- the top of the one is the top of the other,
- the rest of the pile of the one is the pile of the other, and
- the list of the one is the list of the other.
One occasion is the swap of an other waow
- the list of the one is the list of the other,
- the top of the one is the right part of the rest of the pile of the other, and
- the right part of the rest of the pile of other is the top of the other.
One occasion is the hem of an other waow
- the head of the one is the right part of the rest of the pile of the other,
- the rest of the list of the one is the list of the other, and
- the pile of the one is the pile of the other.
The pop, dup, and swap of an occasion can be defined from the push, drop, and hem of an occasion.
Two occasions with the same items can be described from the other with push, drop, and hem: recombic completeness.
Crops denote the drop of waow its component denotes. A component denotes
- the dup of waow its dup does,
- the drop of waow its drop does,
- the push of waow its push does,
- the pop of waow its pop does,
- the swap of waow its swap does, and
- the hem of waow its hem does.
Wore earmuffs while thinking in a chair today.
Added hint “how to deal with anhedonia and uncertainty”
The earmuffs appear to have helped.
While wearing them and thinking while sitting I was able to see that I had already written on the operations required to deal with the division of occasions into left and right parts.
I was able to then see that it connected exactly as Russell describes in “An Inquiry into Meaning and Truth” and “Human Knowledge”.
Then the sentences of the previous entry occurred to me, though only the first few since the rest were written at the computer with the earmuffs on.
I introduced the ‘tie up’ section at the beginning of notes. So far it lists under each loose end some of the entries that need to be tied together by, what I can only suppose, is a memo. It must be emphasized that this is all still very much a verbal discovery expedition.
Predicate memo work on the dichotomy of occasions
The left part of an occasion is called its pile, and the right part its list.
Occasions split into left and right parts: the pile and the list.
The list and pile do too: the right part of the pile is called the top, the left part of the list is called the head, the left part of the pile is called the rest of the pile and the right part of the list is called the rest of the list.
One occasion is the pop of an other where and only where (waow) the pile of the one is the rest of the pile of the other and the head of the one is the top of the other and the rest of the list of the one is the list of the other.
One occasion is the push of an other waow the rest of the pile of the one is the pile of the other, the top of the one is the head of the other, and the list of the one is the rest of the list of the other.
One occasion is the drop of an other waow the pile of the one is the rest of the pile of the other and the list of the one is the list of the other.
One occasion is the dup of an other waow the top of the one is the top of the other and the rest of the pile of the one is the pile of the other, and the list of the one is the list of the other.
One occasion is the swap of an other waow the list of the one is the list of the other, the top of the one is the right part of the rest of the pile of the other, and the right part of the rest of the pile of other is the top of the other.
The pop, dup, and swap of an occasion can be defined from the push, drop, and hem of an occasion.
One occasion is the hem of an other if the head of the one is the right part of the rest of the pile of the other, the rest of the list of the one is the list of the other, and the pile of the one is the pile of the other.
Two occasions with the same items can be described from the other with push, drop, and hem: recombic completeness.
Read the final two chapters of Russell’s “An Inquiry into Meaning and Truth” titled “Analysis” and “Language and Metaphysics”, and read “The Principle of Individuation” in “Human Knowledge: its scope and limits” also by Russell.
These both deal with the problem at hand: the division of occasions into parts and the influence, if any, this has on logical method.
Russell’s predicate is ‘x is part of y’ and mine is ‘x pairs y with z’.
For Russell, the question is if the predicate ‘x is part of y’ is or is not always definable in other terms of a given logical theory.
An example of something like this is Quine’s schematic methods in a schematic theory of indiscernibility/identity.
Somehow, in ways that are still beyond me, Russell concludes that predicates as opposed to just proper names/nouns are necessary in some sense which is presumably metaphysical.
Russell clings to qualities and other such stuff: it is inaccessible to me. It is only through carrying over his arguments into something like Quine’s “ordered sequence of active receptors” or “global stimulus” that I am likely to grasp what Russell “had in mind”.
Predicates denote or don’t, i.e. are true or false of, the occasions on which they occur. Constructions pair component predicates with connectives. Lexicons list atoms, predicates without components. Grammars generate categories from repeated constructions on lexicons. Negations, caps, crops, drops, pushes, pops, and swaps pair their component with ‘not’, ‘cap’, ‘crop’, ‘drop’, ‘push’, ‘pop’, and ‘swap’ respectively. Conjunctions, alternations, and alternative denials pair their left and right components with ‘and’, ‘or’, and ‘nor’ respectively.
Constructed compounds inheret where they denote from their components. Negations denote where and only where (waow) their component doesn’t. Conjunctions/alternations denote waow each/some of their components do. Alternative denials denote waow each of their components don’t. Caps denote waow there is somewhere denoted by their component.
Occasions, like conjunctions, alternations, and alternative denials, are said to have left and right parts. These parts can be divided further into left and right parts e.g. the left part of the left part of the occasion, the right part of the left part of the occasion, and so on. Whether occasions break into an endless dichotomy or whether they terminate into atomic parts is a problem I have not yet solved.
Just as the brave logician checks the empty universe as a degenerate special case, so too must the brave philosopher check the solipsistic universe as a degenerate special case.
Reading Bertrand Russells “Inquiry into Meaning and Truth” and “Human Knowledge” to deal with the problem of individuation, entification, reification, etc.
Does ‘this’ denote or designate?
Ordered pairs are identical if and only if their left parts are identical and their right parts are identical.
Predicates are true or false of the occasions on which they occur. Constructions pair component predicates with connectives. Lexicons list atoms, predicates without components. Grammars generate categories from repeated constructions on lexicons e.g. negations, crops, drops, pushes, pops, and swaps pair their component with ‘not’, ‘crop’, ‘drop’, ‘push’, ‘pop’, and ‘swap’ respectively; and conjunctions, alternations, and alternative denials pair their left and right components with ‘and’, ‘or’, and ‘nor’ respectively.
Predicates are true or false of the occasions on which they occur e.g. ‘red’ and ‘green’. Constructions make compound predicates from component predicates and connective particles e.g.
- the conjunction ‘red and green’ is made from ‘red’, ‘green’, and the particle ‘and’,
- the alternation ‘red or green’, and
- the negation ‘not red’.
Lexicons list particles and atoms – predicates without components. Grammars give lexicons and constructions.
The only way to control email is not to have it.
The previous drafts of the memo on predicate logic gave an inaccurate definition of predicate logic: they said that predicate logic sorts out which compounds are true or false of which occasions.
Predicate logic sorts out valid ones from the others.
Validity is defined from grammar and denotation (where ‘denotes’ is short for ‘true of’ so that each predicate denotes each seperate item it is true of; by contrast, a singular term purports to designate one and only one item).
How is validity defined from grammar and denotation?
Grammar not only sorts out what counts as a predicate and what doesn’t, it also explains how to construct and deconstruct a predicate into its most basic parts, and it is through these basic parts that denotation of compound predicates of a theory are defined (this being Tarski’s analysis of truth applied to a consistent theory of denotation).
What are the methods of grammar?
Sidorov S.A. (MSU), Shumakov M.N. (NIISI RAN) DSSP AND FORTH. COMPARE ANALYSIS
The methods of grammar relevant to logic are
Accept or reject the form of a response as that from a given language.
Recall, a language is a collection of the reinforcing practices of a verbal community: it is a social (verbal) environment composed of verbal contingencies that are built, maintained, and mediated by speakers and listeners.
Like the methods of any other effective science, grammar is primitive recursive: each of its steps conspicuously initiates and terminates with the same kind of control we have come to expect from swinging a hammer at a nail.
There is an unhealthy outlook in science and logic that follows primarly from misinterpretations of the absense of a decision procedure for predicate logic that I must correct.
Given a complete proof procedure for predicate logic it is known that each step prescribed by the procedure is mechanically given and one step is taken after another just as mechanically, but, on any given walk down the path(s) prescribed by the procedure, there is no guarentee of a terminus.
This is then interpreted as saying that, as with the halting problem of Turing’s machines, there is always a next step that must be taken after any given step and that once a walk has begun it is unending.
I’ve tried to sketch the situation using language similar to that of Zeno’s paradoxes in order to bring it up indirectly.
We do not, as logicians, get stuck, as in Zeno’s Dichotomy paradox, whenever go on the path set out by a proof procedure (note, I am, almost always, speaking with Quine’s main method in mind because it is a straight path and mechanically prescribed so there is no whiff of a problem like that of Frost’s road not taken).
At each step we may stop, and often do: note how frustrating it is to finally find a proof and to then try to explain how it occurred to you (this is another important feature of Quine’s main method: the mechanical method can be subverted immediately by a happy accident which can be run in a zig zag fashion with the grand and comprehensive procedure prescribed in the completeness proof of the main method).
That we stop at some step rather than get stuck on an endless walk makes all the difference between primitive recursive methods and those which purport to be general recursive.
The moral of the story, which I have not done a good enough job to teach, is that the metaphor of general recursion does not carry over to the other sciences in the same way that the metaphor of a convergent sequences does not carry over to a walk down Zeon’s Dichotomous lane.
The lexicon of a grammar lists the connective particles and atomic predicates which together with rules for connecting component predicates with connective particles into compound predicates, are said to generate the forms of response of a language grammatically.
Each construction rule is expected to have a unique deconstruction rule when dealing with a logical grammar: this eliminates ambiguities which florish in nonlogical langauges and speeds up the grammatical analysis of a compound into its particles and atoms.
Where many say “That’s how it is.”, scientists say “I don’t know.”
The Origins of Cognitive Thought by B. F. Skinner 1989
Once upon a time a man grew up in isolation. He only knew how to grunt. One day, a woman taught him to speak. She held up an apple, said ‘red’, he held it up, said ‘red’, and then she smiled. He held up another apple, said ‘red’, she frowned, held it up, and said ‘green’. On and on it went.
Phrases like ‘red’ and ‘green’ are predicates: they are true or false of the occasions on which they occur. Compound predicates are made from component predicates and grammatical particles.
Why are microcassette recorders so expensive? Some are hundreds of dollars.
More work on the first memo.
Once upon a time, a man grew up alone in the jungle. He only knew how to grunt. One day, a woman taught him to speak. She held up an apple, said ‘red’, he held it up, said ‘red’, and then she smiled. He held up another apple, said ‘red’, she frowned, held it up, and said ‘green’. On and on it went.
Phrases like ‘red’ and ‘green’ are called predicates because they are true or false of the occasion on which they are said. Compound predicates are built from component predicates and grammatical particles e.g. ‘red and round’ from ‘red’ and ‘round’ and the particle ‘and’. Atomic predicates have no components e.g. ‘red’, ‘parent of’, ‘takes’. Predicate logic sorts out which compounds are true or false of which occasions. The rest of the sciences sort out which atoms are true or false of which occasions.
Technically, one place predicates – e.g. ‘dog’, ‘round’, ‘runs’ – are true or false of items; two place predicates – e.g. ‘parent of’, ‘ontop of’, ‘less than’ – are true or false of ordered pairs of items; three place predicates are ture or false of ordered triples; and so on.
from Artificial Intelligence: Knowledge Representation and Reasoning
The problem to solve here is that the presentation by Deepak Khermani includes gems that are obscured by the clutter of murky logical methods. First I have to introduce the list notation that goes with the methods of logic when they developed within a LISP like programming language.
The aim is to pull these methods inside out: from logic programming to stack (or what I call pile) programming languages and then list (or what is properly called ‘sequence’) programming langauges. None of these sentences are satisfactory. frown
It occurs to me that 4 gives something like predicate functors until ‘such’ and ‘is’ are introduced.
(iff (is (..y) (such (..x) (is (..x) F))) (is (..y) F))
or
(iff (is (..y) (such (..x) (is (..x ..z) F))) (is (..y ..z) F))
where the length of ..y and ..x must be equal, or some convention like Quine’s in ‘Philosophy of logic’ for satisfaction by sequences, or my convention of remainder of division by length, or remainder of division by number of children and cycling back to root when indexing beyond an atom.
Another example with sequences – e.g. ‘(1 2 3 4)’ is short for ‘(1, (2, (3, (4, ()))))’ where ‘()’ is the empty binary tree – that demonstrates the remainder of the quotient by the length of the sequence. The following are equal (in this context):
Quine’s way is to repeat the last item as you do when you write unending decimal numbers that repeat the same digit over and over again e.g. zero point nine repeating (usually indicated by putting a line over the last digit that is to be repeated).
Once upon a time, a man grew up alone in the jungle and only learned to grunt. One day, a woman taught him to speak. She held up an apple, said ‘red’, he held it up, said ‘red’, and she smiled. He held up another apple, said ‘red’, she frowned, held it up, and said ‘green’. On and on it went.
HOW TO MAKE IT LOGICAL
- List the nouns, adjectives, and verbs you use to talk about it.
- Define as many as you can from as few as you can.
- List true sentences made from those few.
- Conclude as many as you can from as few as you can.
- Go to 1.
HOW TO CONTROL YOUR BEHAVIOR
- Consequences select behavior.
- Control consequences to control behavior.
- Build new behavior from old behavior.
- Build new controls from old controls.
I didn’t take a hint: instead of editing a fresh copy of the story I wrote in 1, I made edits to 1 directly. This even after having written 2. I’ve deprived myself and others of the opportunity to better grasp the origins of my verbal behavior by erasing past records. Don’t do that.
The memo on predicate logic may talk about the shapes of predicates: one, two, three, and many place predicates are a run of familiarly shaped predicates. Since we have names for numbers in the vernacular and we do not have names for trees in the vernacular, the many place predicates are more familiar than the tree shaped predicates. Instead of saying ‘many place predicates’ it is technically correct to say ‘sequence shaped predicates’, but such technical terms are not to be introduced lightly.
Technical terms weigh heavy on the reader: they slow things down rather than speed things up.
Drink more water than you think you need (largely because you can).
Primitive recursive analysis is significant if not just because it is that which follows almost immediately from the methods of logic programming.
Present the grammatical part first, what is otherwise mentioned unhappily as ‘well formed formula’, then present the denotative part, what is otherwise mentioned very unhappily as models and proofs.
Vernacular has number names but not tree names.
Just as we can speak of the shape of an array in the J programming language, so too can I speak of the shape of a predicate.
Predicates can be tree shaped, sequence shaped, and any other shape that can be set up from the introduction of ordered pairs as those items which are identical where and only where their left components are identical and the right components are identical.
Longer sequences can be defined by iteration: (x,y,z) as ((x,y),z), then (x,y,z,w) as ((x,y),z,w), and so on. But David Miller has shown me that this method leads to ambiguities of length if we admit sequences of different lengths in the same contexts, as I shall be doing.
and as Feferman mentions on page 4 of “Finitary Inductively Presented Logic” (with my paraphrasing)
Abbreviations. (x,y) := the pair whose left part is x and whose right part is y; (x) := x; (…x, y) := ((…x), y). Note that this representation makes each n-tuple (…x,…y) an m-tuple ((…x),…y) since (…x,…y) = ((…x),…y). Finite sequences will be represented below in a modified form in such a way that each finite sequence has definite length.
Once upon a time a man grew up grunting in the jungle until a woman taught him to speak. She held up an apple, said ‘red’, he held it up, slowly said ‘red’, and then she smiled. He held up another apple, said ‘red’, she frowned, held it up, and then said ‘green’.
The first memo must be on predicate (functor) logic.
It must combine Quine’s “Methods of Logic 4th edition”, “Philosophy of Logic 2nd edition”, and the 1954 entry in the Encyclopedia Americana on symbolic logic.
The encyclopedia entry is included in “selected logic papers Enlarged edition”.
It covers the entirety of an earlier edition of methods of logic in a few pages.
Many arguments are left out and since it was written in 1954 it does not contain Quine’s main method.
It also doesn’t contain the distinctions that Quine would later point up as distinctions between logic and mathematics.
I submit that the methods of logic can be completely, clearly, and effectively given in a memo of only a few pages.
The key is to entirely avoid the complexities of variables, schema, truth values, and other supplementary methods which obscure the role that grammar and denotation play in logic.
Once upon a time a man grew up in the jungle learning only how to grunt until one day a woman taught him to speak. She held up an apple, said ‘red’, the man help it up, grunted ‘red’, and she showered him with praises. He held up another apple, grunted ‘red’, she said ‘no’, held it up, and said ‘green’. So the story goes.
It occurred to me that a single paragraph of italic text could be put at the beginning of a memo to tell a story that makes it eaiser to find the memo and to set up its conclusions.
Sayings like ‘red’ and ‘green’ are called predicates because they are true or false of the occasion on which they are said. Technically, one place predicates – e.g. ‘dog’, ‘round’, ‘runs’ – are true or false of items; two place predicates – e.g. ‘parent of’, ‘ontop of’, ‘less than’ – are true or false of ordered pairs of items; three place predicates are ture or false of ordered triples; and so on.
It just occurred to me that when taking notes it is probably best that I number each phrase I compose rather than simply bullet point them; I just looked at a preview of this entry where bullet points where used instead of numbers and it was clear that numbers would help more than bullet points when finding and perhaps mentioning the specifics of this note in any note (including this one).
Something about the specificity of numbers makes them concrete markers when read as part of a page of text.
Bullet points just run together like a laundry list: this is an accurate report not a laundry list, and, ultimately, it aims to link conclusions with templates; do numbers help this? Yes. They enforce an order that is part of how the writing occurred.
It is a common mistake for people to say that their thoughts are multidimensional or tree like, as if there is a branch tree of choices that they are able to contemplate as a kind of superposition of comparisons; but that is inaccurate; we behave linearly and locally; it is only the mentalistic metaphors that point to some other world where the topology of concepts takes on supernatural shapes.
When writing numbered lists like these in markdown, put a space between each numbered item. If you don’t then what starts out as an easy to navigate number list turns into a big blob and your eye has to hunt for where the numbers are and there’s always the question if they are just the end of a sentence that coincidentally runs onto the beginning of the next line.
I should be able to look back on past notes and see the change in style from before and after I took the advice in the book “Writing that Works” by Roman and Raphaelson.
Another thing that writing with numbered lists allows me is to avoid complete sentences i.e. to avoid adding needless words; I can just go on to the next thing and leave the last thing behind; this is not the case when it comes to memos, but, there, numbered lists serve an even more powerful purpose: they can churn up effective action again and again, and if there is a particularly powerful item, its number can be learned and make finding it even that much easier.
Numbering items also makes it easier to break things into managable chunks.
Once upon a time a man grew up in the jungle. The only noises he made were grunts until one day a woman taught him to speak. She held up an apple, said ‘red’, he held it up, slowly said ‘red’, and she mirrored for him to take a bite out of it. He held up another apple, said ‘red’, she held it up, said ‘no’, then said ‘green’.
I’m apt to change ‘sayings’ to ‘phrases’ because ‘phrase’ has an etymology closer to that of verbal response.
In the past, ‘verbal response’ was what I would write when talking about logical behavior as a kind of verbal behavior, but it does not work when I’m talking to those who are not already involved in the science of behavior or who are unfamiliar with the details of its experimental analyses. Technical terms like ‘experimental analysis’ is unlikely to be read with the trappings of a concrete noun like ‘gadget’, although an experimental analysis can be described concretely as what goes on in a lab. Sadly, this then leaves the phrase further undefined: what happens in labs? Does the person reading this know what really goings on in labs? It is as part of an answer to this question that I think Skinner was primed to frequently refer to “Wundt’s brass instruments”.
Another practical reason to number these notes is that they are more likely than ever to be referenced in conversations with others. I am learning more about how best to connect with others without burdening them with time wasted or the promise there of. Most of my writing is of this sort, and whatever I can do to cut through it as early as possible, i.e. to make accurate and salient records of thinking and feeling, the faster I can get out effective memos, make effective conclusions, and build up a treasure trove of strong templates.
Phrases like ‘red’ and ‘green’ are predicates. They are true or false of the occasion on which they are said.
- When they are said, they are true or false of the occasion.
- They are true or false of where they are said.
- They are true or false where they are said.
The phrases ‘true of’ and ‘false of’ are for those who are unfamiliar with ‘denotes’ and for those who do not yet distinguish between ‘denotes’ and ‘designates’. Following Quine – who followed Mill – I say that predicates denote each thing they are true of and singular terms purport to designate one and only one thing.
The English relative clause ‘who loves Dick’ and the pidgin ‘x such that x loves Dick’ are uniformly paraphrased by the predicate abstract ‘{x: x loves Dick}’ which abstracts ‘Tom’ from ‘Tom loves Dick’ by binding the free occurrence of ‘x’ in the open sentence ‘x loves Dick’ with the prefix ‘x:’ so that the predication ‘{x:x loves Dick}Tom’ concretes to ‘Tom loves Dick’: whatever can be said of a thing can be said by predicating a predicate of it i.e. predicational completeness.
into
HOW PREDICATE ABSTRACTS WORK
- The English relative clause ‘who loves Dick’ and the pidgin ‘x such that x loves Dick’ are uniformly paraphrased by the predicate abstract ‘{x: x loves Dick}’.
- It abstracts ‘Tom’ from ‘Tom loves Dick’ by binding the free occurrence of ‘x’ in the open sentence ‘x loves Dick’ with the prefix ‘x:’.
- The predication ‘{x:x loves Dick}Tom’ concretes to ‘Tom loves Dick’.
- predicational completeness whatever can be said of a thing can be said by predicating a predicate of it
Phrases like ‘red’ and ‘green’ are predicates. They are true or false of where they are said. Specifically, one place predicates – e.g. ‘dog’, ‘round’, ‘runs’ – are true or false of items; two place predicates – e.g. ‘parent of’, ‘ontop of’, ‘less than’ – are true or false of ordered pairs of items; three place predicates are ture or false of ordered triples; and so on. Technically, there are types of predicates for every shape of nesting of ordered pairs e.g. nested ordered pairs are, in general, called binary trees, and n-tuples are taken as sequences which are themselves special binary trees which terminate down their left or right branch with the empty tree (a unique binary tree which has no left or right components).
I was delighted when I read Quine’s encyclopedia article on ‘symbolic logic’ to see that he immediately emphasized alternational normal forms before introducing conditionals or biconditionals. In his presentation he has access to the truth values and interpretations of dummy sentence letters as truth values (and then the dummy predicate letters are having assigned a truth value for each tuple of items in the prescribed “universe of discourse”). These methods fail to emphasize the key to modern logical methods: Tarski’s analysis of truth.
Tarski’s analysis of truth points up exactly those ways in which logical methods are derivative of the methods of grammar and the methods of denotation. Most people are familiar with the methods of grammar: they are taught that complete sentences have a subject and a predicate and that compound sentences are composed from component sentences that are compounded with grammatical particles like ‘and’, ‘not’, or ‘or’. Most people are unfamiliar with the methods of denotation: they are not taught how verbal behavior is shaped by the reinforcing practices of a verbal community as in the story about the man who grunted in the jungle until he met a woman who spoke and listened in English.
Just lost like 10 minutes to twitter: picking it up to send off a tweet is followed by scrolling. The origins of picking up the phone to send the tweet in the first place are in that elusive schedule of reinforcement called “posting”. Just as babies appear to be born babblers, adults appear to be born tweeters [groucho face].
I recently became interested in how subtitles and captions are composed. I watch almost everything with subtitles on, and happened to notice that those of BritBox were exemplary. Here are some hints on making captions work: Captioning Tip Sheet and Captioning Key
Went away from the keyboard to make some book recommendations for a friend; got lost reading on evolutionary biology; fear I’m letting my free time slip away unproductively (making a book recommendation is productive, getting lost reading something that is not scheduled is unproductive).
Should I start a new timestamped note after a certain number of items have been listed? after I’ve gotten up from the keyboard?
I wrote some notes on ‘how to sit and think’ and ‘how to walk and think’, but I can’t find them now. I’m mad.
Some rejected designs for the hint on what I must do before I die:
WHAT I MUST DO BEFORE I DIE
- Discover, predict, and control changes–in counts, rates, and accelerations
- as selections from variations–on physical, chemical, biological, behavioral, and cultural scales
- by making and maintaining strong practices–mediated by strong people marked by strong principles
- from the sciences:
- logic–denotative, Boolean, functor
- mathematics–calculi, collections, categories
- physics–quantum field theory, statistical thermodynamics, gravity
- chemistry–phyiscal, biophysical, and biological
- biology–oranelles, organisms, environments
- behavior–biological, biosocial, social
- culture–history, science, technology
WHAT I MUST DO BEFORE I DIE
- Discover, predict, and control changes
- in counts, rates, and accelerations
- as selections from variations
- on physical, chemical, biological, behavioral, and cultural scales
- by making and maintaining strong practices
- mediated by strong people marked by strong principles
- from the sciences
- logic–denotative, Boolean, functor - mathematics–calculi, collections, categories - physics–quantum field theory, statistical thermodynamics, gravity - chemistry–phyiscal, biophysical, and biological - biology–oranelles, organisms, environments - behavior–biological, biosocial, social - culture–history, science, technology
These dated entries are now officially notes in the following sense.
HOW TO DO IT
Conclude from contemplation on templates:
Memos and hints evolve from notes.
I’m now writing under the control of
HOW TO WRITE WELL 99% OF THE TIME
HOW TO WRITE WELL THE OTHER 1% OF THE TIME
Journal your thoughts and feelings.
HOW TO WRITE WELL 99% OF THE TIME
HOW TO WRITE WELL THE OTHER 1% OF THE TIME
Journal your thoughts and feelings.
HOW TO SOLVE MATH PROBLEMS WITH MY NEPHEWS OVER THE PHONE
FROM TWITTER
Evolving sentences.
Quotations record verbal responses without recording their occasion or their consequence.
Became:
Quotations show you what was said without showing you who said it, when they said it, where they said it, or what hapened after they said it e.g. ‘This is red.’ does not show that a child said it of a blue ball to an adult who then said ‘No, that is blue.’
Now for some live work.
Notes on “the brain is not a computer”.
The brain is not a computer: it is not a consequence of centuries of science and logic. It evolved, like any other organ, by natural selection by consequences from variations.
It evolved without anything like the deliberate designs which brought about modern digital computers and their analogue predecessors. We can say that the brain evolved as part of a behaving organism and that the practices which brought about computers evolved as social environments brought more and more of their deferred consequences under scientific and logical control.
A common mistake, especially among brain scientists, is to say that the mind is what the brain does. That different parts of the mind are the computations which run on different parts of the brain as biochemical computers.
What I finally wrote (and I regret sending it):
The brain is not a computer in the same way that the physical world is not an equation. Computers are a consequence of centuries of science and logic, but the brain is a consequence of some 4 billion years of natural selection. Zeno’s paradoxes teach how such metaphors miscarry.
HOW TO CONTROL YOUR BEHAVIOR
Your behavior is under the control of its consequences. The more control you have over the consequences of your behavior, the more (self) control you have over your behavior.
HOW TO MAKE A THEORY LOGICAL
HOW TO ACT
Conclude from contemplation on templates:
If memos, notes, and rules were lights of a traffic signal they would be colored red, yellow, and green.
Logic is the axis on which the lever arm of language turns. Without logic, all the work in the world is without fixed purpose.
This continues work on my paper on logic from 2025 0521 1232.
In order to better show how old notation gives way to better methods by way of the substitutional definition of validity, and to take a big step towards finishing my paper on logic, I shall write out my thinking as I prove the theorems which, at first, appear to decidedly take truth functional (denotative functional) logic into truth functional and existential closure logic of one place predicates (Boolean closure logic).
This corresponds to the arguments from chapter 18 of Quine’s “Methods of Logic 4th edition”. First a review of terms:
Compounding is (denotative) functional when, exclusively, each like compound denotes or each like compound does not denote where and only where (waow), exclusively, each like component denotes or each like component does not denote. Chains are compounds compounded functionally e.g. joint denials denote waow each of their components do not, negations are self joint denials (they denote waow their component does not), alternations are negations of joint denials (they denote waow some of their components do), and conjunctions are joint denials of negations (they denote waow each of their components do).
Subcompounds of compounds are their self or those of their components. Subsitutions of like compounds for like nonchain subcompounds are (denotative) functional. Compounds are (functionally) valid waow each of their functional substitutions denotes everywhere, consistent waow their negation is nonvalid (some of their substitutions denote somewhere), implied by others waow the conjunction of their self (the conclusion) with the negation of the other (the premise) is nonconsistent (each of their functional substitutions denotes where the same of the other does), and equivalent to others waow they are mutually implicative (each of their functional substitutions denotes waow the same of the other does).
The last addition made to the evolving paper was the definition of (existential) closures and what I call Boolean chains and what Quine called “Boolean statement schemata”. Quine could call them statements, or, as he tended to in his later work, sentences because they were built directly from the truth functional combinations of sentences as instances of truth functional schema of sentence letters e.g. ‘It is raining or it is cold’ is an instance of the truth functional schema ‘p or q’ where ‘p’ and ‘q’ are sentence letters and ‘p’ is substituded by the sentence ‘it is raining’ and ‘q’ is substituted by the sentence ‘it is cold’.
Here, there are only predicates, and the shift from ‘is true’ to ‘is true of’ is coupled with a disavowal of schematic methods entirely for purely grammatical ones e.g. ‘rain or cold’ is the alternation of the predicates ‘rain’ and ‘cold’. Thus, the alternation of ‘rain’ and ‘cold’ is denotes where and only where ‘rain’ denotes and ‘cold’ denotes i.e. ‘rain or cold’ is true where and only where ‘rain’ or ‘cold’ is. To be absolutely clear in traditional terms: there is somewhere, call it x, such that ‘rain or cold’ denotes x if and only if ‘rain’ denotes x or ‘cold’ denotes x. I am hesitant to put things in these terms because they do not properly carry the methods at hand.
They posit some item (here called ‘x’) which poses as an individuated or reified object, but what I have in mind is much more like an undifferentiated occasion or, now speaking with overt metaphors, a global stimulation or a kind of passing show or, unhappily, a unitary essence (though I would never agree to anything like this last blatently metaphysical description).
The plan is to attack the problem of individuation and reification through the abstract items known as ordered pairs and subject to the single rule that ordered pairs are identical if and only if their corresponding components are. This simple plan is actually much more difficult to execute than it would seem because the problem of individuation and reification are difficult. They are difficult in that individuation and reification are remote from talk of them: they are not easy to examine by speaking and listening alone. The wedge of simplicity that I plan on pushing through this problem is that talk of the left and right part of an ordered pair is sufficiently similar to what is called for by individuation that it can sturdy any drift into metaphysical speculation.
This long prelude to what is ultimately a report on my discovery expedition into a proof about Boolean chains is warrented by the problem posed by introducing the principle to prove. Quine was able to give the law as follows
If a Boolean term schema is consistent, then in any universe containing a given object there is an interpretation fo term letters that makes the schema come out true of that object. [pg. 115]
A Boolean term schema is just a truth functional composition of one place predicates which Quine then paraphrases into truth functional sentence schema by distributing the name of an object, e.g. ‘x’, down the truth functional connectives until it is predicated of each one place predicate. The steps are given by the first paragraph of chapter 18:
To say of an object x that it is an F, we write ‘Fx’. Here then is a new sort of sentence schema: ‘Fx’, ‘Gx’, etc. These may be compounded by truth functions; e.g. ‘Fx and not Gx, or Gx only if Hx’. We may conveniently abbreviate such compounds by extracting the ‘x’ everywhere and putting it at the end, thus: ‘[F and not G, or G only if H]x’. We arrive in this way at schematic representations of certain complex terms: schemat such as ‘not G’, ‘F and G’, ‘G only if H’, ‘F and not G, or G only if H’, etc. They will be called Boolean term schemata. [pg. 114]
Since I begin with predicates rather than sentences, I have no recourse to the logic of sentences and their truth functional structure. I also do not have access to methods of truth value analysis that Quine uses throughout his text. The consequence of these restrictions is a prestine simplicity: what comes out as truth functional analysis in Quine’s text comes out as equivalence to succesive alternational developments of subcomponents. I’ve not gone through the details of this development in the paper yet because it would break the law of premature optimization.
Most importantly, I do not have names for things: variables play no part at all nor do singular terms nor do singular descriptions. All of these carry over into singular predicates which denote one and only one item: all methods of designation reduce to this use. Predication is achieved by, e.g., an existential cropping of a conjunction of a singular predicate with a one place predicate. Existential croppings are straight out of Quine’s presentations of predicate functor logic. Croppings do not crop up anywhere in the Boolean parts of my paper on logic.
Returning now to the law at hand in Quine’s words. The consistency of the Boolean term schema of the law guarentees that there is some object of some universe under some interpretation that it is true of. What the law says is something more than what consistency directly implies: pick an object in any universe whatsoever and there is an intrepretation that makes that consistent Boolean term schema true of that object in that universe. Quine’s argument is short and simple in ways that are not accessible directly by my methods:
For, being consistent, the schema resolves to the true under some substitution of the true and the false for the term letters. We have merely to interpret each term letter as true or false of the given object according as its substitute was the true or the false. [pg. 115]
One note on my quotation of Quine’s words: I’ve replaced quotations of a special bold letter tee and its upside down reflection with the phrases ‘the true’ and ‘the false’ even though Quine otherwise mentioned these notations as short for ‘truth’ and ‘falsity’ as truth values. This is primarly to draw a closer connection between Frege’s outlook and Quine’s (it is mostly to remind my future self that there is more to say about this later).
Anywhere is denoted by some functional substitution of a consistency.
That is the problem to solve. It is not solved here because I stopped here writing before I solved it.
Henkin’s theorem is presented in Ebbinghaus, Flum, and Thomas as
Let capital phi be a consistent set of formulas which is negation complete and contains witnesses. Then, for all phi,
the term interpretation of capital phi satisfies phi if and only if capital phi proves phi.
This reveals one of the weaknesses of Ebbinghaus, Flum, and Thomas’ book (and a lot of books on mathematical logic): a needlessly complex sequence of symbols and definitions are lofted atop each other in haste to establish a narrow statement of an otherwise general result.
There are a number of things that I have to get out of the way in order to write on this:
Almost all books on mathematical logic (and some on logic) make a grand dichotomy between syntax and semantics as if these are two complementary and comprehensive classifications of the methods of logic i.e. as if a method of logic is syntactic or semantic (or perhaps both, although usually only by some equivalence). This yields the myopic view that there are two classifications of logical methods: model theoretic and proof theoretic. In EFT model theoretic methods are a kind of mathematical semantics and proof theoretic methods are a kind of matheamtical syntax. But this is not really just a pecularity of EFT’s book on mathematical logic, it can be found throughout books on mathematical logic e.g. the classic by Joseph R. Shoenfield.
This is a mistake of history and says more about the insensitivity of mathematical logicians to logicians than it does anything about logic or mathematical logic.
2.
Well, there’s a fragment of writing that I’ll never finish.
There is a grand misconception that since logic is not about anything in particular then logic is not about the world and hence logic, as with science, is without ethical or moral values. Logic is part of the world. It “deals with” or “is about” the world through each of its primary origins: grammar and denotation (as ‘is true of’ and not as ‘designates’).
Ethical and moral contingencies (as consequences from responses on occasions e.g. a door opens from a push on a lever, an electron emits from a photon on an atom, etc.) are as much a consequence of logic as they are grammar and denotation. For most people, it is easier to note the part played by denotation in ethics than it is the part played by grammar. Grammar is traditionally dealt with as something secondary to denotation as if denotation covered meaning rather than meaning following as a trivial kind of reference accomodated entirely by denotation of a special sort (just as designation is a special case of denotation e.g. denotation by a one place singular predicate).
The origins of grammatical and denotative verbal behavior dispell such traditional fictions and fictious explanations leaving a concrete link between logic and the rest of the world: the autoclitics of Skinner’s “Verbal Behavior”.
…
The ethics of logic works much like the grammar of logic in that the logic of grammar strengthes the grammar of logic and so on like so many other human practices. Just as the logic of grammar is a consequence of the gramamr of logic, so to is the logic of ethics a consequence of the ethics of logic.
…
Logic is the fulcrum of languages.
the ethics of extensionality
the ethics of validity
the predicates of ethical theories
This continues my work on my little lisp from way back in 2025 0514 2226.
It has been a while since I last worked on my little lisp, and, as is so often the case when I return to what I left incomplete, I must begin by collecting the entirety of what was done. As a reminder: I do my programming in a browser by first going to the url “about:blank” and then opening the developer tools on it. In Google’s Chromium, script snippets are my programming environment. This is not because it is “the best programming environment”, but because this happens to be the easiest way I can program on whatever computer happens to be available to me.
// pairs
let theEmptyPair={}
, isEmpty = it => it == theEmptyPair
, pairOf = (it, that) => [it, that]
, leftOf = it => isEmpty(it) ? it : it[0]
, rightOf = it => isEmpty(it) ? it : it[1];
// sequences as pairs
let theEmptySequence = theEmptyPair
, isEmptySequence = isEmpty
, singletonSequenceOf = it => pairOf(it, theEmptySequence)
, headOf = leftOf
, restOf = rightOf
, concatOf = (it, that) => isEmptySequence(it) ? that
: pairOf(headOf(it), concatOf(restOf(it), that))
, prependSingletonOf = (it, that) =>
concatOf(singletonSequenceOf(it), that)
, appendSingletonOf = (it, that) =>
concatOf(that,singletonSequenceOf(it));
// stacks as pairs
let theEmptyStack = theEmptyPair
, isEmptyStack = isEmpty
, singletonStackOf = it => pairOf(theEmptyStack, it)
, pushOf = pairOf
, dropOf = leftOf
, topOf = rightOf
, secondOf = stack => topOf(dropOf(stack))
, drop2Of = stack => dropOf(dropOf(stack))
, prependOf = stack => pushOf(drop2Of(stack)
, prependSingletonOf(topOf(stack), secondOf(stack)))
, appendOf = stack => pushOf(drop2Of(stack)
, appendSingletonOf(topOf(stack), secondOf(stack)));
// letters, strings, concatenations and runes
let isConcatenationOf = (x,y,z) => x == y.concat(z)
, isEmptyConcatenation = x => isConcatenationOf(x,x,x)
, theEmptyConcatenation = ''
, isIdenticalConcatenation = (x,y) =>
isConcatenationOf(x,y,theEmptyConcatenation)
, theConcatenationOf = (x,y) => x.concat(y)
, stringOf = (...letters) => theEmptyConcatenation.concat(...letters)
, firstLetterOf = letters => isEmptyConcatenation(letters) ? theEmptyConcatenation : letters[0]
, restLettersOf = letters => isEmptyConcatenation(letters) ? theEmptyConcatenation : letters.slice(1)
, theAlphabet = '() 0123456789abcdefghijklmnopqrstuvwxyz'
, runeHelpOf = (it, abc) =>
isEmptyConcatenation(abc) || isIdenticalConcatenation(it, firstLetterOf(abc)) ? theEmptyPair
: pairOf(theEmptyPair, runeHelpOf(it, restLettersOf(abc)))
, runeOf = it => runeHelpOf(it, theAlphabet)
, letterHelpOf = (it, abc) => isEmptyConcatenation(abc) ? theAlphabet
: isEmpty(it) ? firstLetterOf(abc)
: letterHelpOf(rightOf(it), restLettersOf(abc))
, letterOf = it => letterHelpOf(it,theAlphabet)
, runesOf = letters => isEmptyConcatenation(letters) ? theEmptySequence
: prependSingletonOf(runeOf(firstLetterOf(letters))
, runesOf(restLettersOf(letters)))
, lettersOf = runes => isEmptySequence(runes) ? theEmptyConcatenation
: stringOf(letterOf(headOf(runes)), lettersOf(restOf(runes)))
// read and print sequences
let theOpenRune = theEmptyPair
, isOpenRune = isEmpty
, readOpenRuneOf = (stack, runes) =>
readSequenceOf(pushOf(stack,theEmptySequence), restOf(runes))
, theCloseRune = pairOf(theEmptyPair, theOpenRune)
, isCloseRune = it => !isEmpty(it) && isEmpty(leftOf(it)) && isOpenRune(rightOf(it))
, readCloseRuneOf = (stack, runes) =>
readSequenceOf(appendOf(stack), restOf(runes))
, readDefaultRune = (stack, runes) =>
readSequenceOf(appendOf(pushOf(stack,headOf(runes))), restOf(runes))
, readSequenceOf = (stack, runes) =>
isEmptySequence(runes) ? headOf(topOf(stack))
: isOpenRune(headOf(runes)) ? readOpenRuneOf(stack,runes)
: isCloseRune(headOf(runes)) ? readCloseRuneOf(stack, runes)
: readDefaultRune(stack, runes)
, readOf = runes => readSequenceOf(theEmptyStack, runes);
let parenOf = runes => prependSingletonOf(theOpenRune
, appendSingletonOf(theCloseRune, runes))
, printSequenceOf = sequence =>
isEmptySequence(sequence) ? theEmptySequence
: concatOf(printOf(headOf(sequence))
, printSequenceOf(restOf(sequence)))
, printOf = sequence =>
isEmptySequence(sequence) ? parenOf(theEmptySequence)
: parenOf(printSequenceOf(sequence))
, read = letters => readOf(runesOf(letters))
, print = sequence => lettersOf(printOf(sequence));
The definition of ‘isConcatenationOf’ aspires to more than it can reach in Javascript: a three place predicate in the lexicon of a theory of concatenation. The definition of ‘theConcatenationOf’ most clearly reveals the weakness (to anyone who wasn’t already confused by the invocation of ‘.concat’ in the definition of ‘isConcatenationOf’): it is not short for a singular description. The three place functional predicate ‘is the concatenation of’ is constructed with the help of singular descriptions as
x, y, and z are such that x is identical to the (w such that w concatenates y with z)
where ‘is identical to’ is coextensive with ‘is indistinguishable from’ as in a schematic thoery of identity. This is the basic context within which Russell’s elimination of singular descriptions occurs:
x, y, and z are such that some item is (w such that x is identical to w and each item is (u such that u is identical to w if and only if u concatenates y with z)).
I shall leave that behind and go on to what occurs to me in response to copying the latest iteration here.
I’ve been looking for a better way to weave together the work so that it is cumulative without imposing additional constraints which are extraneous (as was done in an older iteration which built stacks from sequences rather than from pairs).
The problem remains printing and reading, as much as these are likely to stick around in later iterations. A new solution occurred to me which is no more foreign to me than any other solutions that have occurred to me (that is they were already present in some work or another that I was exposed to and can not even remotely be attributed to me except perhaps for passing through me).
By including a default function which catches what the defined reader doesn’t, i.e. as in the default case of a case statement in a standard imperitive programming language, readers and printers can be combined like LEGOs to accomidate evolving occassions that must be dealt with e.g. as when the items of a sequence are dealt with as something other than sequences.
The names of functions defined are already far longer than they would ever be if I was programming privately, perhaps this is for the better or perhaps this is for the worst, I can not yet tell. I’m inclined to introduce the following obtuse names “readStringAsRune”, “readStringAsRunes”, etc. so that then all is readers e.g. “readPairAsString” is the new name of the old function “printAsString”. It may even be more convenient then to just name the functions “letterAsPair”, “lettersAsPair”, “lettersAsSequence”, etc. where, e.g., “lettersAsSequence” happens to be identical to the composition of the functions designated by “lettersAsSequence” and “sequenceAsPair”. But this doesn’t make sense.
The distinction between reading and writing is an io problem. So, “inputAsSequence” or “readAsSequence” makes sense because everthing after “inputAs” is, by explicit design, a pair.
Aha, ignore all that. I uncovered a potential solution that simplifes and unites. Instead of creating a seperate set of functions for letters, strings, and the like, I can just go straight to a sequence of runes.
Darn, I’ll have to come back and fill this out later. It has been very hard to find time to write out what I have to say as it occurs to me, even with respect to these little steps along this programming project.
The following incomplete and uncommented code is left here as a record of what was done but not as what was explained.
, where = (is, seq) => isEmptySequence(seq)
|| is==headOf(seq) ? theEmptySequence
: prependSingletonOf(theEmptySequence, where(is, restOf(seq)))
, here = (at, seq) => isEmptySequence(at) ? headOf(seq)
: itemAt(restOf(seq), restOf(at))
, each = (f, seq) => isEmptySequence(seq) ? theEmptySequence
: prependSingletonOf(f(headOf(seq)), each(f, restOf(seq)))
, over = (f, f0, seq) => isEmptySequence(seq) ? f0
: f(headOf(seq), over(f, f0, restOf(seq)))
, til = (stop, base, step, part, rest, it) => stop(it) ? base
: step(part(it), til(stop, base, step, part, rest, rest(it)))
, sequenceOf = items => til(x=>x.length==0, theEmptySequence, prependSingletonOf, x=>x[0], x=>x.slice(1), items)
, here = (at, seq) => til(isEmptySequence, headOf, x=>x )
, theAlphabet = sequenceOf('() 0123456789abcdefghijklmnopqrstuvwxyz')
, readRune = rune => where(rune, theAlphabet)
, printRune = pair => here(pair, theAlphabet)
, readRunes = runes => each(readRune, sequenceOf(runes))
, printRunes = pair =>
It is unclear if taking up these methods of functions on functions is sufficiently smooth to warrent admission into this discovery expedition. It occurs to me that since I have only recently begun sharing public notes that those reading these may not know how familiar I am with the scope and limits of different methods of programming and their mathematical theories. Ultimately this is of no concern to me because if I can avoid burdening others with all that I have learned and subsequently passed over, then they will be left with that much more energy to learn all that I never shall in my life.
R.P. shared a delightful paper with me upon reading my last entry on Paul Graham’s “Good Writing”: The Philosophy of Composition by Edgar Allan Poe. It is to be compared to B. F. Skinner’s talk “On Having a Poem” a recording of which can be heard here: https://www.bfskinner.org/bf-skinners-lecture-on-having-a-poem/
Standards of records required to provide explanations for the origins of poetic behavior, or any verbal behavior whatsoever, are to be found in Skinner’s “Cumulative Record Definitive Edition” on page 115 titled “A Case History of Scientific Method” which appeared originally in around 1955. The details mentioned by Skinner in that paper are not sufficient, but his larger work “Schedules of Reinforcement” with Ferster provides sufficient conditions for explaning the behavior of pigeons under controlled laboratory conditions. Similar methods are at least required to explain human behavior (however inconvenient they might be).
I do not have time to write on all these things now, and have left this note here in the off chance that I may return to it later.
But, here is what I have to say of what I have already read of Poe’s paper. The meat of the matter is found in the eigth paragraph. The earlier paragraphs simply set the reader next to Poe as if all that was read had been just penned before you. It is a delightful trick, and much of the paper itself follows the same methods as those who purport to be only comedians and yet have surprisingly strong political effects.
The eigth paragraph begins after all the essential features of a behavior analysis would otherwise have ended:
Let us dismiss, as irrelevant to the poem, per se, the circumstance—or say the necessity—which, in the first place, gave rise to the intention of composing a poem that should suit at once the popular and the critical taste.
Such a dismissal is perhaps the most profound deception in the entirety of Poe’s paper. The ‘per se’ pricks the vail of deception and provides Poe with permission, from the reader who continues to read, to be as sly as is his want. This is a perfect example of what I was talking about when I said in my paper on Paul Graham’s “Good Writing” that
the initial ambiguity and the consequent incompatible methods could have been charitably explained away as nonrhetorical literary language e.g. as a linguistic ignis fatuus.
Where Graham failed, Poe has instructively succeeded.
I very much appreciate R.P. bringing this beautiful example of good writing to my attention.
Here I sincerely analyze Paul Graham’s latest paper on “Good Writing” (I transcribe the snarky analysis I made on twitter at the end for completeness).
The first sentence
There are two senses in which writing can be good: it can sound good, and the ideas can be right.
begins with the ambiguous phrase “There are”. Vernacular expands it to either “there are at least” or “there are exactly” (there are exactly three ways “there are” can be taken and the third is as “there are at most” as in “there are at most three Stooges” but this is not carried by the vernacular). Logicians, having noted this ambiguity, regiment “there are” as “there are at least” e.g. “There are at least two senses in which writing can be good”. This regimentation is not merely a convenience. It is the difference between existence and uniqueness e.g. “there is exactly one even prime number” and “there is at least one even prime number”. Lawyers also noted this ambiguity and regiment “there are” as “there are exactly” under the rule “Expressio unius est exclusio alterius”. Between logicians and lawyers the language of the home leaves the first sentence of Paul Graham’s essay woefully underdetermined.
The strong reader rallies with contextual clues. But, the first clue is very far from the proximate context. In the eleventh paragraph the lawyers finally win: “So yes, the two senses of good writing are connected”. But then, before the sentence can end, everyone looses. The full sentence is “So yes, the two senses of good writing are connected in at least two ways.” and it spits on the languages of lawyers, logicians, and locals. For, “the two senses” implies the lawyers rule but “in at least two ways” breaks it since two and only two connections are given.
Had Graham’s “Good Writing” been on something other than good writing, then the initial ambiguity and the consequent incompatible methods could have been charitably explained away as nonrhetorical literary language e.g. as a linguistic ignis fatuus. But, charity extends to the paper as an example of good writing in exactly those ways explained: “it can sound good, and the ideas can be right.” In other words, the ambiguous and incompatible methods are covert rather than accidental. The deferred cosequences of covert methods are rarely helpful i.e. the ninth item of Russell’s decalogue, “Be scrupulously truthful, even if the truth is inconvenient, for it is more inconvenient when you try to conceal it.”
The etymology of ‘sense’ likely goes to the PIE root *sent- for “to go” and is the “source also of Old High German sinnan “to go, travel, strive after, have in mind, perceive,” German Sinn “sense, mind,” Old English sið “way, journey,” Old Irish set, Welsh hynt ‘way’” (https://www.etymonline.com/word/sense). Replacing ‘sense’ with ‘way’ as in “There are two ways in which writing can be good” uncovers the metaphorical role of ‘sense’ in any sentence in which it occurs. But, a reason to avoid “way” in Graham’s sentence is because “two ways” is near “diverging paths” and the point is made later in the essay that these ways are not independent. Alternatively, different ways can just as well diverge and yet later intersect and in general, the nearer a pair of people are to a destination the fewer ways there are to finish their trip i.e. the more likely they are to take the same way.
This negative outlook can be countered with the German’s “sinn” which is most famous as a single word from Frege’s “Über Sinn und Bedeutung”. It is clear from the remainder of this essay that Graham is committed to a mentalistic outlook, one that follows the German ‘sinn’ through ‘sense’ to ‘mind’. Thus, the occurrence of ‘sense’ primes talk of mental things. Ideas, as in the vernacular (or that special part of the vernacular called folk psychology), are objects of the mind (charitably, the mind is what the brain does), expressed by words. Then, writing is the a process by which ideas are expressed through words. This is the same story as in hoodoo rootwork where a spell is written on a slate and washed into a glass of water which somehow expresses the cure and is absorbed by the drinker.
More charitably, the mind is what the body does, and the body as much has ideas as it has behaviors. In fact, the intermediate talk of ideas which are later expressed by behaviors then dissolves and writing as a repertoire is said to be good or bad in as far as it is reinforced by prevailing cultural practices. Then, to identify the key features of good writing is to identify the reinforcing practices which build and maintain one repertoire of writing rather than some other. This is not the same as concluding with a kind of cultural relativism nor with a kind of cultural darwinism (“survival of the fittest” is so often interpreted as “‘fit’ is equivalent to ‘survive’” that any such short story on the process of selection by consequences from variation on cultural scales is bankrupt).
Good verbal behavior presumably contributes to the survival of the culture within which it occurs. It presumably takes into account the deferred consequences of its appearence in a culture sensitive to contributions of the deferred consequences of its practices to its survival. It no more helps to supplant these descriptions by “sounds good and gets ideas right” than it does to supplant the principles of aerospace engineering with Kelly Johnson’s “If it looks good, it will fly well” (which I was unable to find a record of from a search with DuckDuckGo or Google).
The result of my snarky analysis is copied here from Twitter (one of the worse places to write on the internet):
The whole paper is paraphrased as follows:
Good writing is sound, salient, and, essentially, slick. Why essentially? Lo, Paul Graham knows it is, and thou shalt not show Him otherwise. Soundness and salience are essential too. Why? Hark! Thou shalt know it as He does for thou art made as Him. He shakes the heavens and the earth with his grace and all becomes beautiful in its perfection: “In the beginning was the Word, and the Word was with God, and the Word was God.”
Books on my reading shelf:
Books on my desk:
the 4th edition was published in 2019 (maybe get that)
get Richard J. Evan’s “Third Reich Trilogy”:
- “The Coming of the Third Reich” 2003
- “The Third Reich in Power” 2005
- “The Third Reich at War” 2008
the updated edition was published in 2001
I ask myself “What would I tell my younger self?” as a way of checking on how I am doing and how I might do better sooner rather than later. For example, would I tell my younger self to work on the practices of logic as much as I have been over these past months? How would I even explain logic to my younger self? Also, what does my younger self really have to do with my present self or my future self? These are just some questions which flow from me like an endless stream of unstoppable curiosity. Do I have answers for most of my questions? No. Do I have answers for a few of them? Only very few.
So, what would I tell my younger self about logic? Would telling my younger self to do something different with respect to logic really change his behavior in any way? Are counterfactual conditionals like these even helpful? They are relative to a theory of the world and it is relative to such a theory that logic would have the greatest impact, even for my younger self.
Do we say that young people have theories of the world, ones that evolve over time as they are tested and as trials and errors pile up for or against this or that variation on some working theory of the world? That’s a long question. Long questions rarely have short answers that are also self contained i.e. answers that encapsulate the question without posing it.
Rather than start with how logic would help my younger self I’ll start with how logic has helped me, the older self to my younger self. Logic helps me deal with the world and all the problems it presents to me. Putting a problem into words is the first step and paraphrasing those words into a logical theory helps me factor some large theory of the world into tiny parts that I can see at a glance. It focuses me on the predicates of a theory and the potential premises whose implied conclusions are so often fated to be overthrown by experiment. None of this is what a young me would care about or notice.
When I help my nephews with their math homework I almost always do it over the phone so that we have to talk out a solution without simply sending each other pictures of things we’ve written down. The first thing I tend to teach them is to read out the problem aloud. To most people this may seem like a tiny thing: don’t people know how to read aloud? No, a lot of people don’t. They were barely taught to read selections from books aloud to each other, and are rarely asked to read math problems aloud. How do you read most of the notation anyway?
That there are effective ways of reading mathematical notation aloud makes all the difference between noticing a parenthesis and skipping it (and messing up the rest of the problem). What does this have to do with logic? Logic takes our exposure to our native tongue and shapes it up into a verbal machine that introduces mechanical advantages to a language. But, what does that mean to a youngling?
There are different ways you can talk about the world. Some of them are helpful and some of them are hurtful. If you want the way that you talk about the world to help you and not hurt you then speak logically. Logic is made for dealing with what is true. As much as dealing with what is true helps you to deal with the world, logic helps.
How do you speak logically? First you have to notice that the way you talk about the world affects how you deal with it. Lots of teachers struggle to teach this. We know they struggle because so many kids say “Will this be on the test?” or “What does this have to do with the real world?”. If teachers were honest they would tell kids “It has nothing to do with the real world. This is a game that we made. We make you play it because we don’t know what else to do. We don’t know how else to help you deal with the world.” Instead, they threaten students with bad grades. Bad grades are not what education avoids: they are just side effects of teachers not knowing what to do. It is the bad consequences of dealing with the world poorly that actually threaten the young and old alike.
Education is not overtly presented as “the way to avoid dealing poorly with the world and the pains that follow from doing so”. It is presented as some sort of obligation or insurance or self contained arena of competition. As if the educational environment is of independent interest relative to the rest of the world. The practices of education have survived because they just barely make the educated person more effective than the uneducated person in almost everything that they do. “Just barely” is all it takes: those who talk up 10x or 100x multipliers are credit/debit shills. The snowball of fundamental productivity plows through the ups and downs of short sighted accelerationists (who now, more than ever, invoke the “long term” as nothing more than a verbal hedge or buzzword).
So what have I actually said of logic itself? There is little to say, especially now that we have Quine’s main method of proof for quantificational logic: a sentence is inconsistent if when put into prenex form a truth functional inconsistency is accumulated from instantiation so that existentials always receive new variables. Logic links truth with grammar through validity either as that which can be derived from Quine’s main method (as inconsistency of negation) or that which is true under all interpretations (where interpretation can be taken semantically or syntactically if at least the premises of placevalue notation are present).
From these technical conclusions there are vernacular conventions e.g. very few sentences are of any consequence what so ever since validity is no kind of honorific but rather an expedient of verbal linkage. Where logic links up the essential sentences of a theory, it binds their fate in any practical experiment. Deferred consequences of logical links come out in surprising ways that are fatal to a fledgling theory. Those insensitive to logic are more easily spotted as its practices become more and more familiar. Such logical deviants break up logical links either out of careless insensitivity or out of coercive intervention.
Yesterday I finally got out about five sentences that open up the world of existential closures to my methods of logic. I also carried through the subtle shift from “denotes” in the definition of validity to “denotes everywhere”.
Much self reflection occurred yesterday and today, but it is not written here. Most of it was silent.
Today is the day that I make a big step with the continued work on my logic paper from 2025 0514 1345.
Sometime last week I compressed eight cards of outline on my logic paper into three cards. Those new notes do not include the serious edit of adding “denotes everywhere” wherever I had written “denotes” and carrying through the rest of the paper with the corresponding predicate “denotes somewhere”. Here I shall write out the new notes with new edits and follow through the consequences of editing the definition of denotative functional validity as “each functional substitution denotes everywhere” rather than the old version “each functional substitution denotes”. I shall add comments on what is written or what is to be written (as a record of my thinking at the time of making the revisions) using the standard markdown notation for quotations (on most sites this is displayed as text in a box of a different color).
Logic is the science of validity and validity is a consequence of grammar and denotation.
Compounding is (denotative) functional when, exclusively, each like compound denotes or each like compound does not denote where and only where (waow), exclusively, each like component denotes or each like component does not denote. Chains are compounds compounded functionally e.g. joint denials denote waow each of their components do not, negations are self joint denials (they denote waow their component does not), alternations are negations of joint denials (they denote waow some of their components do), and conjunctions are joint denials of negations (they denote waow each of their components do).
The phrase “where and only where” is critical here: it links up the practices of logic with talk of the rest of the world. This turns on a geometric interpretation of spacetime which I do not mention except through the occurrence of ‘where’ and ‘when’. A deferred consequence of this spacetime trick is that when I change “this” from a pronoun to a predicate it is easier to navigate how individuation is bound up with logic and the rest of the world. In general, I take ‘it’ as designating and ‘this’ as denoting chunks of spacetime. That does not imply that I take spacetime as prior to or primitive relative to what I write on logic. It is an outrider to my logical methods: it is on the periphery and only gently touches the outer rim of logical methods.
Subcompounds of compounds are their self or those of thier components. Subsitutions of like compounds for like nonchain subcompounds are (denotative) functional. Compounds are (functionally) valid waow each of their functional substitutions denotes everywhere, consistent waow their negation is nonvalid (some of their substitutions denotes somewhere), implied by others waow the conjunction of their self (the conclusion) with the negation of the other (the premise) is nonconsistent (each of their functional substitutions denotes where the same of the other does), and equivalent to others waow they are mutually implicative (each of their functional substitutions denotes waow the same of the other does).
This is where the big change occurs. The addition of the single word “everywhere” to the definition of validity, which now reads “Compounds are (functionally) valid waow each of their functional substitutions denote everywhere”, carries on into the definition of consistency as “Compounds are (functionally) consistent waow their negation is nonvalid” the fundamental consequence of which is that “Compounds are consistent waow some of their functional substitutions denote somewhere.” This is key to making arguments about the validity or consistency of an existential (or univeral) closure of a chain as defined. I’ve stuck to the spatiotemporal “where”s and occasional “when”s so that there is a constant reminder that in explaining the practices of logic, there is nothing like the logic explained which is yet available. This is hard to get out in a way that doesn’t churn up more confusion than the verbal constructions can bear.
I had planned to avoid this early mention of “somewhere” and “everywhere” by getting away with the connectives “where” and “only where” as well as the conspicuous “each” and “some” which carry over firmly from this language to logic. The phrases “somewhere” and “everywhere” are sometimes adverbs and sometimes nouns in their native language of English. Since my aim is on logic and not English and its grammar, I’ve not done as Quine did in “word and object” and gone out of my way to link up the elementary school grammar of English with the technical grammar of Logic. But, here I am confronting the problem head on in this comment.
The “somewhere”s and “everywhere”s as they occur in these definitions are picked to play a grammatical part in English which is between adverb and noun. It is from Logic and not from English that I got this usage: the phrase “everywhere” is an implicit universal quantification and the phrase “somewhere” is an implicit existential compound. The method of implicit quantification, in this form, is Dreben’s: the construction of implicit existential compounds from their implicit universal components depends, with Dreben’s method, upon the scope of purported quantifiers. Quine’s method of functional normal forms makes the simplifying assumption that all problems of scope are established by assuming prefix normal form. In logic programming, the historical origin of both Quine and Dreben’s method is emphasized by referring directly to either implicitly universally quantified variables or to Skolem constants, or to Skolem functions of implicitly universally quantified variables and Skolem constants, and so on.
In the definition of validity then “valid where and only where each of their functional substitutions denote everywhere” there are only universals, and carry over to predicate abstract notation as
each {xyz : x is valid if and only if y is a functional substitution of x and y denotes z}
under Quine’s interpretation, and as
each {x: x is valid if and only if each {y : y is a functional substitution of x and each {z : y denotes z}}}
under Dreben’s (more or less). But, in rendering this definition as a sentence in predicate abstract logic it is important to note that the letter to the right of ‘denotes’ does not effectively carry the complex whereabouts which ultimately come through the later development. In an implicit quantifier notation of some logical programming languages, this definition carries over to the sentence
?x is valid if and only if ?y is a functional substitution of ?x and ?y denotes ?z
but none of these examples shows the notation for implicit existential compounds.
The carry over of the fundamental consequence of the definition of consistency “consistent waow some functional substitutions denote somewhere” does. Quine’s and Dreben’s interpretations are both then
each {x : x is consistent waow y_(x) is a functional substitution of x and y_(x) denotes z_(y_(x))} which emphasizes with the underscore that Quine gives the notation with subscripts rather than with additional notation like underscores and parentheses.
Since neither variables nor quantifiers are part of predicate logic (they are part of quantificational logic), none of this actually concerns us here.
Alternations of compounds with their negations are valid (they denote waow the compound does or its negation does, i.e. waow it does or does not, so, each functional substitution denotes everywhere). Conjunctions of compounds with their negations are nonconsistent (they denote waow their compound does and its negation does, and, hence, waow it does and does not, so, each functional substitution of the conjunction does not denote everywhere i.e. the negation of the conjunction is valid). Compounds are implied by and equivalent to their self.
I’m uncertain if I should include these supporting arguments in something more than the parenthetical. That compounds are implied by and equivalent to their self is shown by the same step by step reduction to previous and accumulating results that have been built up thus far. A compound implies itself when the conjunction of it with its negation is nonconsistent and that is precisely what was just shown and the same conjunction occurs twice in the case of self equivalence.
Functional substitutions in
Interchanges of equivalents in a compound are equivalent to that compound (each functional substitution of a compound matches the same of its interchange, except perhaps for the same of the equivalents which otherwise denote in tandem, so each denotes waow the other does i.e. they are equivalent).
Interchanges of equivalents in
I still don’t feel quite right about the argument for nonvalidities. I just fixed it by distinguishing between the step of equivalence and the step of the negation denoting somewhere from some functional substitution (so that the interchange of the negation, being equivalent to the negation, denotes somewhere from the same functional substitution). Why avoid notation though? Notation undermines the sensitivities conditioned by carefully designed sentences. There are very likely better sentences than the ones I’ve made thus far, but these are the ones I have and they work well enough.
Compounds are equivalent to
The equivalents of associativity yield the many component alternations and conjunctions which are equivalent to iterated nestings of alternations or conjunctions down their left or right components.
Validity is
Each clause is equivalent to a disjoint or empty clause. Each chain is equivalent to
(Existential) closures denote waow there is somewhere denoted by their component. Closed chains are closures of (open) chains.
Boolean chains are closed or are chains of Boolean chains. Without whereabouts closures denote nowhere since there is nowhere denoted by their component i.e. they are nonconsistent. Only nontrivial results are contemplated here. Check the trivial case by interchanging each closure with a nonconsistency.
The alternation of the closure of a compound with the closure of the negation of that compound is (nontrivially) valid and trivially nonconsistent.
Consequences of closure conditionals are closed chains and their antecedents are conjunctions of them. Consequences or antecedents of degenerate closure conditionals can be empty.
This is a little experiment where I try to get out the distinction between the two specializations of ‘denotes’ which are ‘denotes somewhere’ and ‘denotes everywhere’. Usually when I write on something like this, the context is already available to me in that I write after having looked exactly where the problem I am solving actually occurred. In this case, I have not looked at my notes on my paper on logic. Thus, I am without context other than that which I bring to this keyboard unsupplemented by recent exposure to the problem.
The way the paper is designed now, it introduces denotative functional methods of compounding components in the same way that truth functional methods of compounding components is introduced along classical lines except it avoides the most familiar appeal to truth values. The definition of truth functional compounding is usually given in the following way: a method of compounding components is truth functional if the truth value of the compound is a function of the truth values of the components. So, in a sentence like “It is red and it is round.” the compound is the whole sentence and it is compounded, from a logical grammar of which ‘and’ is an atom, of the components ‘it is red’ and ‘it is round’ and the truth value of ‘it is red and it is round’ is a function of the truth values of ‘it is red’ and ‘it is round’. Assigning truth values to the components is one way of interpreting them, and calculating the truth value of the compound with reference to a truth table is one way of specifying the truth function which carries the truth values of the components to the truth value of the compound.
Whether ‘it is red’ is assigned the value of truth or falsity is often confronted as a problem of interpretation, though when the sentence is given directly and not through the indirect methods of schematisms, e.g. ‘x is F’ or ‘it is F’ or just ‘Fx’, then the problem of interpretation is seen in its more familiar form e.g. selecting a specific item which ‘x’ or ‘it’ designates and specifying a particular predicate which is substituted for ‘F’, in this case substituting ‘red’ for ‘F’ in ‘it is F’ gives ‘it is red’ and taking ‘it’ as designating a red thing results in the assignment of truth to ‘it is red’. If it so happens that under this same interpretation ‘it’ designates a round thing then the compound ‘it is red and it is round’, which is an instance of the schemata ‘it is F and it is G’ or ‘Fx and Gx’ where ‘F’ is substituted by ‘red’ and ‘G’ is substituted by ‘round’, then since both the components ‘it is red’ and ‘it is round’ are assigned truth and since the connective ‘and’ marks the compound as receiving the value of truth where and only where each of its components are assigned the value truth, then the compound ‘it is red and it is round’ is interpreted as true.
Under these methods, the truth functional methods that go out of their way to introduce abstract objects called truth values and functions there of, each assignement of truth values to the truth functional components of a compound fixes the truth value of the compound. Now, having gone out of my way to describe this method, I am left saying that this is not the path taken here. Nor, and this is the crux of the matter, is it the strongest path that can be taken through the methods of logic. In fact, taking this path has come to weaken logical and mathematical practices in ways which are largely unseen but which are of great consequence: they’ve all but broken the most powerful links between language and logic. This is unexpected for many reasons, but the most relevant reason is this: those who are most familiar with the links between language and logic are precisely the ones who are most likely to break up the links between them and least likely to point up the most powerful links.
How this has come to pass is a pressing problem and one whose solution is well within my sights, but it reaches beyond the scope of language and logic and is best put aside as I have just as well put aside how it has come that different philosophers go in and out of style.
So what is the alternative that I submit is so much more powerful and important? It is well known but is not carefully dealt with in the same systematic way that any mathematical logician would quickly demand of an introduction to model theory or proof theory (or, beyond an introduction, to the concerns of mathematical logic as a whole). It is the method explicated by Quine in his “Philosophy of Logic” which can be briefly and artfully stated as “Logic chases truth up the tree of grammar.” and “Logic is the sum of two components: grammar and truth.”
The predicates of grammar are coupled with the predicates of truth (that I must always explain myself when I say “predicates of truth” as if it was not shown almost a century, now, ago that there is no consistenty theory of a predicate of truth which satisfies the minimal requirement of disquotation, or that I have to go through the way in which disquotation is an instance of the maxim of minimal mutilation, should be alarming enough to those who know better). When presenting logic, i.e. when teaching it to a fresh learner, there is nothing but the verbal communities with which that learner has been in contact from which to teach them. Logical behavior is taught to those without any logical behavior: this is far more controversial than it looks.
That logical behavior can be taught to an organism without any logical behavior often comes as a surprise to most people, especially logicians and mathematicians. This is largely because of the historical principles which guide both philosophers and logicians away from anything like the experimental analysis of behavior. Logic, and its methods, are to transcend our material world through some metaphysical mist, or so it is said. That we, as mere mortals, are in touch with some otherworldly essence of logic is merely an aftereffect or perhaps even an uninteresting collateral effect of some more fundamental metaphysical principles.
This is something which must be explained, but which would take me too far from this path. To be clear, those who hold to metaphysical practices must be accounted for in an effective theory of the world, though, like most effective theories, it may not account for them in the way they might hope.
I shall also not go into the methods by which a pigeon can be conditioned to behave logically without, in the case of a pigeon, clearly having any logical behavior like that so often attributed to even the most isolated human (that a pigeon can be made to “do logic” is often astounding to people who are mostly unfamiliar with what little there ultimately is to logic in itself since they are unable to seperate logic from its application, which is another one of the key problems caused by traditional methods).
The grammatical part of logic can be introduced with some basic predicates, though they are not introduced as the application of logic so they can not be relied upon to bring appropriate behavior of the learner under control without being a tiny variation on some nonlogical behavior. The nonlogical behavior often relied upon when introducing logic is that of the vernacular. An example of a minimally viable repertoire is that what is written here can be read: much more work must be done to make this as precise as it is, e.g., in a text like Skinner’s “verbal behavior” (for all that many have said to extinguish its control over scientific practice).
Sentences are taken, tentatively, as nothing more than quotations i.e. as concatenations of letters as in the familiar strings of computer science. This is displeasing for many people who hold the form of a given response as sacred and who are unaware of the contributions of the consequences and the occasions upon which a response of a given form is emitted. I can not go out of my way to explain that here now. As far as the following description goes, I make no commitment to quotations and you can pick your favorite stand in items.
A single grammatical predicate does the work of the truth functional part of a language: x jointly denies y with z. Joint denial is not the only such predicate e.g. ‘x alternatively denies y with z’ is another such. I prefer joint denial because ‘nor’ is a sentence connective in my native tongue i.e. those who speak in a language with ‘nor’ are already that much closer to behaving logically with respect to it.
With this grammatical predicate a predicate of truth can be combined and a tiny example of Quine’s outlook can be demonstrated in a familiar way. In general, it is supposed that the predicate ‘is true’ is an abbreviation for the highest of the truth predicates among those in the inductive hierarchy of Tarski’s analysis which belongs to the lexicon of our present language.
The following is what I call the synthetic categorical of joint denial (this is not quite correct, as the categoricals are to be distinguished from the universal closures of conditionals, but I have assumed that logic is being taught to a learner without a history of exposure to a community of logicians and hence if they can read this sentence at all then it will bring them under the control characteristic of categoricals):
each item is (x, y, and z such that x jointly denies y with z, y is not true, and z is not true, only if x is true)
The analytic categorical is
each item is (x, y, and z such that x jointly denies y with z and x is true, only if y is not true and z is not true)
The bicategorical is
each item is (x, y, and z such that x jointly denies y with z, y is not true, and z is not true, if and only if x is true)
The synthetic categorical chases truth up the tree of grammar as Quine said, and the analytic categorical chases truth down the tree, and the bicategorical chases it both ways.
Looking back on the traditional way of introducing joint denial, it is seen that no method of truth values or truth functions or models or proofs have been used to explain joint denial. Using the method of predicate abstracts, which I’ve mentioned many times before and which can be quickly found with a simple text search among previous entries, it can easily be explained how it comes that Frege began talk of “truth functions”.
The predicate (abstract)
(x, y, and z such that x jointly denies y with z, y is not true, and z is not true, only if x is true)
is almost functional in its first place. Note, schematically
‘Functional F’ is short for “each item is (w,x,y, and z such that Fxyz and Fwyz, only if x is indiscernible to w)” where indiscernibility is given schematically (I’m not sure if I’ve explained that elsewhere or not, but I’m pretty sure I have).
The predicate ‘x jointly denies y with z’ is where functionality goes faulty since there may be many different ways that a compound may be formed from its components into a joint denial of its components e.g. “neither Tom loves Dick nor Dick loves Tom” or perhaps “Tom loves Dick nor Dick loves Tom” or perhaps “not Tom loves Dick or Dick loves Tom”.
This is where the all important methods of paraphrasing are revealed as of foundational importance. Logical practices restrict the predicate ‘(x,y, and z such that x jointly denies y with z)’ so that it is functional and then the predicate
(x, y, and z such that x jointly denies y with z, y is not true, and z is not true, only if x is true)
is functional. Note, without the early constraint that the learner of logic have some language (as the reinforcing practices of a verbal community) in which they speak, there is little that can be done from simply presenting the synthetic and analytic conditionals.
Also, note, that paraphrasing is linked with indiscernibility by the logicians insistence of paraphrasing so that “Functional (xyz : x jointly denies y with z)” is true.
Sadly I must end here without going from ‘is true’ to ‘is true of’.
Another great recommendation from R.P. arrived yesterday: Scott Atran’s 1990 “Cognitive foundations of natural history: towards an anthropology of science”.
The introduction unveils the inspiration for the rest of the text: the theories of Chomsky and Piaget are nonconsistent. A nonconsistency is constructed of science from Chomsky’s premise that ‘each fundamental type of human-knowledge arises from a specialized cognitive aptitude” (pg. x Artran ‘cognitive foundations’) and Piaget’s premise that “the innate and universal foundations of human thinking reduce to an undifferentiated intelligence, which is responsible in the same way for all cognitive operations” (pg. ix) from which talk of an intelligent scientist goes sour.
It is already clear from teh introduction that those items of Chomsky’s theory which Atran keeps upon rejecting the contrasting principle from Paiget’s theory are those which root themselves in teh behaviorist tradition i.e. the contingencies of the world, and Earth in particular, are such that global uniformities are selected without any centralized coordination so that the practices of science are as much a consequence of the evolved tendancy of organisms to bring the world (of which other members of their species are a part) under their control. This permits a diversity among practices of presciences, and admits of evolution therein, while explaining those features across cultures that prompt those like Piaget to speak of a unifying intellect.
It is the contingencies of the world which select uniformities (for reasons that may be initially capricious) and which maintain them: not some ephemeral intellect. The problem for Chomsky, and hence for Atran in support of him, shall be looking beyond the individual to the contingencies of natural, operant, and social selection and variation that explain how bits of behaving biology build and maintaining scientific communities.
The first chapter points to the mistake which is easily corrected by the radical behaviorist: it is not some essence of common sense that haunts the brains of the masses, but rather the vernacular as the language of the household where language is nothing more, nor less, than teh reinforcing practices of a verbal community. When a person is said to be “using common sense” we can look at their speach and trace its origins through the listeners that selected it from its past variations. Correspondance between common sense as speaking in the vernacular is easily established by examining the world upon which we speak. It is one that is filled with other members of our species and which has been sufficiently stable so as to not only permit the survival of the species but also the survival of those verbal practices which compose cohesive cultures.
The practices between cultures do not always carry over perfectly because the world is not so uniform, e.g., as to select exactly the same form of response across great distances. Geography is not sufficiently uniform so as to strengthen vernacular which is diverse in some places, e.g. as when it was established in 2010 in Chapter 16 entitled “Franz Boas and Inuktitut Terminology for Ice and Snow: From the Emergence of the Field to the ‘Great Eskimo Vocabulary Hoax’ by Igor Krupnik and Ludger Müller-Wille from the 2010 book “SIKU: Knowing Our Ice Documenting Inuit Sea Ice Knowledge and Use” that there are a larger number of root words for their world of ice and snow than in other languages of the world.
Vocal musculature is a critical uniformity among organisms of a given species that permits verbal communities to acquire new members whether they are old or young.
As I mentioned in my last entry, there are only tiny chunks of writing that I can do right now and though that frustrates me I am at least glad to have gotten this bit out.
Last night I watched more of “The Great Philospohers” from 1987 which I mentioned in Bryan Magee’s “Men of Ideas” and “The Great Philosophers”. Frederick Copleston spoke on Schopenhauer and Magee mentioned Copleston’s series “A History of Philosophy” which is now eleven volumes but which was published originally as nine between 1946 and 1975. Here is the link to the wikipedia page: https://en.wikipedia.org/wiki/A_History_of_Philosophy_(Copleston).
It will take some time for me to get through it all, as it takes me so very long to read even one book in the way that I do (like a loom weaving threads from dozens of books at a time). In anticipation I shall do as I do and write out the volumes titles and dates of publication:
They are sold as a series of eleven books though they were originally published as a series of nine. The last two were added into the series in 2003 for reasosn that I do not yet know.
Another scrapped entry from yesterday that I was unable to complete and which went unpublished. It continues work on my logic paper, but sadly rests as a mere record of my past about which I know so very little even this single day later. Recently, I have struggled to find the large chunks of time which tend to release the greatest potential from any writing that I might do. As much as there is a temporary moment of frustration it is so common that it almost goes without saying except in these peculiar circumstances where detailed records of responses are demanded.
2025 0517 1625 A lot of work on my paper on logic is occuring at once and I have not been able to complete it all. I am doing my best to make a record of all the responses which appear with respect to it. This continues the last entry on the paper 2025 0516 1352.
Quantificational logic, as opposed to predicate (functor) logic, defines validity with respect to the truth of its sentences (relative to provisions for grammatical constructions with free variables e.g. treating them as dummy singular terms or as abbreviations of their universal closures). Predicate logic defines validity with respect to the denotation of its predicates: there are no variables and hence no provisions for free variables (and, happily, no complex rules of substitution for bound variables).
Quantificational logic seems to predicate the quotation of a sentence of the predicate ‘is true’ as in “‘it is raining or it is not raining’ is true”. By Tarski’s analysis, this sentence is equivalent to “‘it is raining’ is true or ‘it is not raining’ is true” which is itself equivalent to, by the same analysis, “‘it is raining’ is true or ‘it is raining’ is not true” whihc, as an instance of a truth functionally valid schema, is true (we also know from Tarski’s analysis that the predicate ‘is true’, which occurs repeatedly throughout the quoted equivalents, is not coextensive with the predicate occuring right before this parenthetical). Note, the use of ‘predicate’ and ‘sentence’ picks out nothing but what is spelled out by a quotation e.g. ‘is true’ is a predicate and ‘it is raining’ is a sentence.
The sentence “it is raining or it is not raining” is itself an instance of a truth functionally valid schema and hence it is true without recourse to the intermediate steps of the argument by equivalents in the previous paragraph. On the other hand, the sentence “it is raining” is not an instance of a valid schema (either when ‘it’ is taken as a dummy singular term or when the whole is taken as an abbreviation for “each item is it such that it is raining”) i.e. whether, exclusively, ‘it is raining’ is true or ‘it is not raining’ is true, or, directly, whether, exclusively, it is raining or it is not raining (exclusivity has nothing in particular to do with this example, whether ‘or’ is read exclusively or not in the previous paragraph). These two examples demonstrate the difference between a triviality and the potential urgency of an umbrella.
Predicate logic, at first, seems to predicate the quotation of a predicate and some chunk of the world of the two place predicate ‘is true of’, aka ‘denotes’, as in “‘rain or not rain’ denotes this” which is equivalent to “‘rain’ dentoes this or ‘not rain’ denotes this”. Equivalently, “‘rain’ denotes this or ‘rain’ does not denote this” Now, instead of being an instance of a truth-functionally valid schema, the compound predicate “‘rain’ denotes this or ‘rain’ does not denote this”
This is an incomplete and, it would appear, incoherent note that continues work on my paper on logic from 2025 0514 1345.
Logic is the leverarm of language, and those who speak in it can lift the weight of the world with their words. There are multiple ways of specifying valid verbal behavior, but only two are mentioned to most modern students: the model theoretic and the proof theoretic. Under appropriate assumptions, these methods are shown to be equivalent. In most cases, both methods are presented schematically so that upon establishing one or the other validity there is a secondary step to projecting one or the other definition of validity to the verbal behavior that occurs in logical practice.
It is with great sadness that the model theoretic methods are called out as semantic and the proof theoretic methods are called out as syntactic. Neither the word ‘semantic’ nor the word ‘syntactic’ are worthy of a place in logic except from its application e.g. in formalisms, perhaps in those special formalisms called “axiomatic systems”, of model theories (set theories within which the semantic methods are derived by interpretation) or of syntactic theories (concatenation theories within which the syntactic methods are derived by calculation).
Logic is linguistic like grammar. It tells us something about the reinforcing practices of verbal communities. Principles aimed at verbal communities generally are said to be transcendent, and those aimed at a single community are said to be immanent.
On a given occasion the response “Red” and the response “Round” are reinforced by a listener, or punished,
To say of a denotative functional compound that it denotes somewhere or that it denotes everywhere is to do something like the implicit quantifier notation used in logic programming
The definition of validity of a functional compound must be ammended to explicitly mention “denotes everywhere” where “denotes” was always short for in previous editions. The reason for this is that “consistency” as “some functional substitutions denote” must be altered to “some functional substitutions denote somewhere” so that problems of existential closure can be dealt with uniformly through what Quine calls “Boolean Term Schema”.
Compound Whereabouts and Quine’s Functional Normal Form x_(y_(z)uv_(ws)) becomes ((z),u,(w,s)) this is important to Hilbert and Bernay’s completeness, especially as presented in Kleene’s “Introduction to Metamathematics”. write on quine’s explanation of hilbert and bernay’s completeness in “selected papers on logic”
This fragment of a note was written yesterday as I began to read the 1994 book “The Bell Curve” by Richard J. Herrnstein and Charles Murray. I stopped writing on it because it no longer presented itself as scientific literature. It became obvious that it is a part of nationlistic literature more so than anything scientific. Thus, a deep reading has been delayed for when I address the collection of nationalistic writers e.g. Thomas Sowell, Milton Friedman, and Francis Fukuyama.
Though their writings contain copious references to what clearly purports to be scientific literature, there is nothing in the design of their documents which brings up science as the primary source of control over what they have to say. Much of what is written is of this sort, and even much of what is written here by me is of this sort.
We, as humans, speak under the control of a wide variety of social environments, and, through them, have contact with a wide variety of contingencies of the world. Not all of them are scientific. Not all of them purport to be under scietific control. Some do, and among those, there are some which do so to share, by association, in the strengths of scientific practice. Whether they are the result of scientific practices, as in a theoretical analysis of the social consequences of a science of behavior, is important to me because of the control that scientific practices have over my own behavior as a reader and writer.
A copy of Richard Herrnstein’s 1973 “I.Q. in the Meritocracy” is arriving soon and it purports to be, like “The Bell Curve”, under the control of scientific practices. Until it arrives I have stopped writing on “The Bell Curve” and have simply been reading it as time has allowed.
Here is the fragment:
2025 0515 1911 The Bell Curve For much of the past year I have avoided digging into Herrnstein’s work because there is so much of it which is incompatible with the science of behavior as a science of behavior rather than a science of cognition (which is nothing more than a reincarnation of a science of mind but with that ever present seasoning of “computational” or “informational” sprinkled about). But, the genetic control of behavior continues to be poorly dealt with, as in Charles Murray’s 2020 book “Human Diversity: The biology of gender, race, and class”, and the less I read the less I’m likely to know.
The general problem is simple: there is no room in scientific practices for events happening in other dimensions other than those under investigation. People are no more haunted by more or less intelligent minds than houses are haunted by more or less malevolent ghosts. To say so is to already
This continues my work on my little lisp from 2025 0512 1400.
Much as you do when you are solving a problem in physics, I shall write down the basic programs that have evolved thus far. This note adds a layer to the onion of grammars growing out the primitive parenthetical language.
// pairs
let theEmptyPair={}
, isEmpty = it => it == theEmptyPair
, pairOf = (it, that) => [it, that]
, leftOf = it => isEmpty(it) ? it : it[0]
, rightOf = it => isEmpty(it) ? it : it[1];
// sequences as pairs
let theEmptySequence = theEmptyPair
, isEmptySequence = isEmpty
, singletonSequenceOf = it => pairOf(it, theEmptySequence)
, headOf = leftOf
, restOf = rightOf
, concatOf = (it, that) => isEmptySequence(it) ? that
: pairOf(headOf(it), concatOf(restOf(it), that))
, prependSingletonOf = (it, that) =>
concatOf(singletonSequenceOf(it), that)
, appendSingletonOf = (it, that) =>
concatOf(that,singletonSequenceOf(it));
// stacks as pairs
let theEmptyStack = theEmptyPair
, isEmptyStack = isEmpty
, singletonStackOf = it => pairOf(theEmptyStack, it)
, pushOf = pairOf
, dropOf = leftOf
, topOf = rightOf
, secondOf = stack => topOf(dropOf(stack))
, drop2Of = stack => dropOf(dropOf(stack))
, prependOf = stack => pushOf(drop2Of(stack)
, prependSingletonOf(topOf(stack), secondOf(stack)))
, appendOf = stack => pushOf(drop2Of(stack)
, appendSingletonOf(topOf(stack), secondOf(stack)));
Eventually these basic functions will settle down and an instructive sequence of tiny steps will lead the novice through these methods of programming. What follows are the new basic functions on concatenations as javascript strings. A small change in initial defintions is given to better follow the theory of concatenations like those of Tarski.
// letters, strings, concatenations and runes
let isConcatenationOf = (x,y,z) => x == y.concat(z)
, isEmptyConcatenation = x => isConcatenationOf(x,x,x)
, theEmptyConcatenation = ''
, isIdenticalConcatenation = (x,y) =>
isConcatenationOf(x,y,theEmptyConcatenation)
, theConcatenationOf = (x,y) => x.concat(y)
, stringOf = (...letters) => theEmptyConcatenation.concat(...letters)
, firstLetterOf = letters => isEmptyConcatenation(letters) ? theEmptyConcatenation : letters[0]
, restLettersOf = letters => isEmptyConcatenation(letters) ? theEmptyConcatenation : letters.slice(1)
Now, the alphabet is introduced as a special concatenation.
let theAlphabet = '() 0123456789abcdefghijklmnopqrstuvwxyz';
The following new definitions for making runes and letters skip over the intermediate steps of working with tallies. Even though such methods otherwise would be instructive, the definitions given are basic methods of “finding where a thing is at” and are more than good enough.
let runeHelpOf = (it, abc) =>
isEmptyConcatenation(abc) || isIdenticalConcatenation(it, firstLetterOf(abc)) ? theEmptyPair
: pairOf(theEmptyPair, runeHelpOf(it, restLettersOf(abc)))
, runeOf = it => runeHelpOf(it, theAlphabet)
, letterHelpOf = (it, abc) => isEmptyConcatenation(abc) ? theAlphabet
: isEmpty(it) ? firstLetterOf(abc)
: letterHelpOf(rightOf(it), restLettersOf(abc))
, letterOf = it => letterHelpOf(it,theAlphabet)
, runesOf = letters => isEmptyConcatenation(letters) ? theEmptySequence
: prependSingletonOf(runeOf(firstLetterOf(letters))
, runesOf(restLettersOf(letters)))
, lettersOf = runes => isEmptySequence(runes) ? theEmptyConcatenation
: stringOf(letterOf(headOf(runes)), lettersOf(restOf(runes)))
Then an example, which deviates from the prior example by including a ‘not’:
lettersOf(runesOf("(()())() this should not be ignored"))
(()())() this should not be ignored
Which is a good sign (sadly, not a proof).
The change in alphabet has led to a slight change in the placement and form of the prior definitions of “theOpenRune” and “theCloseRune”: it is now with the definitions of reading and printing sequences.
// read and print sequences
let theOpenRune = theEmptyPair
, isOpenRune = isEmpty
, readOpenRuneOf = (stack, runes) =>
readSequenceOf(pushOf(stack,theEmptySequence), restOf(runes))
, theCloseRune = pairOf(theEmptyPair, theOpenRune)
, isCloseRune = it => !isEmpty(it) && isEmpty(leftOf(it)) && isOpenRune(rightOf(it))
, readCloseRuneOf = (stack, runes) =>
readSequenceOf(appendOf(stack), restOf(runes))
The default method of reading runes which were not open or close runes was to skip over them. Now any nonparenthetical runes are added to the end of the sequence under construction. This makes for minor alterations of the method of reading sequences.
let readDefaultRune = (stack, runes) =>
readSequenceOf(appendOf(pushOf(stack,headOf(runes))), restOf(runes))
, readSequenceOf = (stack, runes) =>
isEmptySequence(runes) ? headOf(topOf(stack))
: isOpenRune(headOf(runes)) ? readOpenRuneOf(stack,runes)
: isCloseRune(headOf(runes)) ? readCloseRuneOf(stack, runes)
: readDefaultRune(stack, runes)
, readOf = runes => readSequenceOf(theEmptyStack, runes);
The rest is as it was.
let parenOf = runes => prependSingletonOf(theOpenRune
, appendSingletonOf(theCloseRune, runes))
, printSequenceOf = sequence =>
isEmptySequence(sequence) ? theEmptySequence
: concatOf(printOf(headOf(sequence))
, printSequenceOf(restOf(sequence)))
, printOf = sequence =>
isEmptySequence(sequence) ? parenOf(theEmptySequence)
: parenOf(printSequenceOf(sequence))
, read = letters => readOf(runesOf(letters))
, print = sequence => lettersOf(printOf(sequence));
The examples let it be shown that nothing has changed from when they were last run.
lettersOf(parenOf(runesOf('(()()())')))
((()()()))
print(read('()'))
()
print(read('(()(())())'))
(()(())())
print(read('(())(()(())())'))
(())
print(read('(()(())())((()))'))
(()(())())
print(read('))))(('))
()
None of these include a space rune so that has yet to be treated as any special case. But, the next example with other than close or open runes paves the way for the next layer of the onion mentioned at the beginning of this note.
print(read('(this is a test)'))
((()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()())(()()()()()()()()()()()()()()()()()()()())(()()()()()()()()()()()()()()()()()()()()())(()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()())(()())(()()()()()()()()()()()()()()()()()()()()())(()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()())(()())(()()()()()()()()()()()()())(()())(()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()())(()()()()()()()()()()()()()()()()())(()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()())(()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()))
The method of making runes is now revealed as a way of reinforcing the parenthetical reader and printer: nonparenthetical runes are read as sequences of of empty sequences. This opens up the following more sophistocated methods without entailing further alterations to the parenthetical reader and printer.
letterOf(headOf(read('(this is a test)')))
t
letterOf(headOf(restOf(read('(this is a test)'))))
h
lettersOf(read('(this is a test)'))
this is a test
This took far longer to finish than I had aimed at, but I’m happy to be this far. Though it is easy to go back and see that I’ve been near a similar target in past iterations, they relied more on peculiarities of javascript than on the basic logic of a design of a lisp like language which does not depend on its implementation.
Quine’s main method corresponds, in a significant way that I have not fully grasped, to Aristotle’s principle of noncontradiction: this also happens to be the crux of the methods of logical programming. The demonstration of the relevant inconsistency constructs, by conjunctive description that comes out to a generalization of ‘the primitive recursive value of a the functional predicates of a given theory’, the item whose existence is presumably implied by noncontradiction.
This entry ended up contributing to the work on my paper on logic from 2025 0510 1238.
The past two days have been filled with private conversations and wayward contemplations that were sadly ungoverned by the work done here. It happens, and yet, the more often it happens, the less often I stumble into a tiny success in any of my little endevours.
For example, I’ve failed to note my responses to watching more of the series I mentioned in Bryan Magee’s “Men of Ideas” and “The Great Philosophers”. I’ve also failed to note the arrival of Kir, Raven, and Schofield’s “The Presocratic Philosophers”, or the multiple rereadings of key parts of Quine’s fourth edition of “Methods of Logic” (for all the times I’ve read the book from cover to cover there is still more than I can carry with me without referencing it).
I see each of these unnoted events as failures to hit where I have aimed. Thankfully, a few bits of minor productivity do carry me onward if not wayward.
As mentioned in my last note on the evolution of my paper on logic, I can not wait to finish the sections covering denotative functional duality prior to starting on the new grammatical category of existential closures. Yet, such duality is more a labor saving device (it multiplies the consequences of any result on functional compounding by two) than a milestone. The milestone, under the grammatical definition of valdity as substitution on supplementation of lexicon (by some finite list of previously unintroduced predicates), is the equivalence of each funtional compound to one in the grammatical category of alternational normal forms and conjunctive normal forms. It yields a decision procedure for validity: a n component chain is valid waow its full alternational normal form is compounded of n to the power of two clauses. This retreives, without oblique references to items purported by the phrase “denotative values” or “truth values”, the breadth and depth of methods of truth tables and, much more importantly, truth trees, though, to be clear, the law of validity just stated is the carry over of truth tables, and the grammatical import of truth tree methods has yet to be described conspicuously as such (by anyone).
The grammatical import of truth trees are a consequence of alternational and conjunctive development which are already presented in the parts of the paper on logic that I’ve produced thus far. Such methods are generalizations of Gentzen’s sequents along the lines of Quine’s method of existential conditionals which is a decision procedure for Boolean statement schemata. I introduce sequents as special conditionals long before reaching out from Quine’s main method to those of natural deduction, which I expect to show are nothing more than the application of Quine’s main method to sequents so intorduced (for me, a sequent is to be a denotative functional compound whose antecedents components are conjoined as the antecedent of a conditional whose consequent is the alternation of the consequent components of the sequent). The parallels between implications and conditionals are the smooth passage from Quine’s main method applied to sequents and the methods of natural deduction.
This continues work on my little lisp from 2025 0509 1722.
It occurred to me in the shower that I can use the defined operations for manipulating concatenations of letters as javascript strings to give some interesting and important definitions of basic predicates and functions on concatenations. It also occurred to me that this fits perfectly into both Quine’s “Concatenation as a Basis for Arithmetic” as well as my outlook on bit strings and binary trees.
The first function which hit me was one that checks whether a letter is in a concatenation of letters. This led me to a definition that either constructs directly the appropriate rune, constructed step by step, and then to Quine’s definition of tallies in a theory of concatenation with at least two distinct atoms.
Since there can be, and are, many distinct atoms in the concatenations of letters as javascript strings, a basic predicate of identity of letters is unavoidable, or at least unadvisable at this moment. The following sketch of code appeard.
let isIdenticalLetter = (it, that) => it == that
, isLetterOf = (it, that) => !isEmptyLetter(that)
&& (isIdenticalLetter(it,firstLetterOf(that))
|| isLetterOf(it,restLettersOf(that)))
, theAlphabet = '() 0123456789abcdefghijklmnopqrstuvwxyz'
, isLetter = it => isLetterOf(it, theAlphabet)
, isIdenticalLetters = (it, that) =>
(isEmptyLetter(it) && isEmptyLetter(that))
|| (isIdenticalLetter(firstLetterOf(it), firstLetterOf(that))
&& isIdenticalString(restLettersOf(it), restLettersOf(that)))
The alphabet is no longer a javascript array each item of which is a javascript string: it is now just a javascript string.
There are two next steps that appeared: 1) start with the basic predicate ‘x concatenates y with z’ in javascript and build the rest from there so that it is closer to traditional theories of concatenation (the newer theories focus on a family of functions that prepend or appened a single atom/letter to its argument, and while something like that is easier with modern programming langauges to set up, it does not leverage the power of pure predicate logic like the older techniques which have their origins through Tarski’s theories of concatenation), and 2) define the extended versions ‘letterOf’ and ‘lettersOf’ and ‘tallyOf’ etc.
The predicate ‘x concatenates y with z’ is defined as
let isConcatenationOf = (x,y,z) => x == y.concat(z);
then, the empty string, or empty letter as I’ve been calling it, is the item denoted by the following predicate.
let isEmptyString = x => isConcatenationOf(x,x,x);
Somewhere while I was coding these things up it finally occurred to me that ‘Of’ is actually just a mark of a relative predicate, and not a good indicator of what is classically taken as a function value: that is done by ‘the’ as in singular descriptions, but these are not to be confused with functions and I am reluctant to conflate matters further. But I shall, just to see how it goes.
let theConcatenationOf = (x,y) => x.concat(y)
, theEmptyConcatenation = '';
Social commitments ended this entry prior to any conclusions.
The work that must be done on my paper on logic can not be done here. It has done best when worked on with B. F. Skinner’s Thinking Aid. The thinking aid is nothing more nor less than a way of arranging index cards onto a sturdy flap of a folder with masking tape so that each part of a paper can be flipped to so as to cover up all others and leave labels for each section visible. Items can be easily moved around or replaced.
Right now I am doing a grand compression of those elements that have settled down and not undergone many edits e.g. definitions and examples of denotative functional compounds. Another change is in the method of arranging the cards. In my first experiment with Skinner’s thinking aid, I only put the cards on the outside of a folder whose contents were loose notes which primed and prompted the cards I made. Now I am arranging the cards on the inside of the folder so that I can get two full pages when laid flat.
Upon getting eight cards down to three I couldn’t stand writing anymore.
This continues work on my paper on logic from 2025 0507 1421.
Upon reading my latest work on my logic paper I made the following notes:
Functional completeness is also a special problem of equivalent grammatical categories: it deals with the problem of grammatical supplementation. For example, the current presentation takes advantage of the purported equivalence between conjunction as a denotative functional method of compounding components and a compound of joint denials of the same components to introduce conjunction as a compound of joint denials. Each denotative functional method of compounding components is introduced as some compound of joint denials. This is not the same as showing that joint denials are functionally complete. Functional completeness contemplates a grammar with a method of denotative compounding which is newly called ‘conjunction’ and the compounds of which must be shown to be functionally equivalent to some compounding of joint denials.
The problem of functional completeness is the one of indefinite supplementation and subsequent equivalence of some one grammatical category (such as that of pure negations and conjunctions) with an other (such as that of joint denial). A general method can be introduced which puports to demonstrate how, under any supplementation of the grammar of a language by a functional method of compounding, a corresponding compound in a canonical category (such as that of pure joint denials) can be constructed. Wittgenstein mistook the indefiniteness for limitlessness when he failed to establish quantificational sentences as unending truth functional compounds in the Tractatus.
There is an other way of approaching the problem of functional completeness: translation or paraphrasing. This solves the problem in that it disolves it: the foreigner is now the responsible party in that the fruits of the grammar of the native is open to them only in so far as they pick out their method of functional compounding among those of the native. The native is concerned by such problems of paraphrasing in the same way as the foreigner: if the native is to enjoy the advantages of some foreign grammar then the foreign must be superimposed upon the native. This is the line that I’ve taken when, e.g., negations are given as self joint denials.
That there are only finitely many n component methods of functional composition and that the defining characteristic of a functional composition is that its compounds exclusively denote or do not denote and that whether the compound does or does not denote is implied by whether each of its components exclusively do or do not gives the basis for any satisfactory method of paraphrasing from a foreign grammar or proving equivalence of a supplemented grammatical category to some initial and cononical category.
A few of the sentences at the beginning of some sections, e.g. laws of validity, consistency, implication, and equivalence, are now redundant since they repeat what was paraphrased immediately following the introduction of the relevant term.
The problem of individuation is likely to be case as the transition from existential closures ‘some‘as “there is somewhere” to existential croppings ‘some’ as “there is something”. The ultimate resolution of ‘some’ as “some items are” or perhaps ‘some*’ as a specific iteration of ‘some’s depends significantly upon the specific recombic functors that are introduced. It occurred to me yesterday that my Stack Notation for Predicate (Functor) Logic solves a great problem with smoothing the transition from those unfamiliar with predicate functor logic and yet familiar with quantificational logic e.g. ‘xFy’ a instance of which is ‘x is a member of y’ or, in short, ‘x in y’ is familiar for having a lowercase (variable) letter on the left and on the right, and it is this basic division between left and right which not only paves the way (following Russell’s “Human Knowledge: scope and limits”) for some method of individuation while also smoothly supporting the analysis of “somewhere” as a binary tree of “somethings” which extends the ordered tuple methods which are already familiar in the carrying over of n place predicates to n place relations and which shall come to be the carry over of tree shaped predicates to tree shaped relations: this is already a standard method in many programming langauges, e.g. Paul Graham’s bel, and is called ‘destructuring assignment’.
If I wait to get to existential closures until I am done with denotative functional duality, then I may never get on with the more pressing matter of presenting Quine’s main method in its autonomous predicate functor form.
Logic is the science of validity and validity is a consequence of grammar and denotation.
Compounding is (denotative) functional when, exclusively, each like compound denotes or each like compound does not denote, where and only where (waow), exclusively, each like component denotes or each like component does not denote. Chains are compounds compounded functionally.
Joint denials denote waow each of their components do not. Negations are self joint denials: they denote waow their component does not. Alternations are negations of joint denials: they denote waow some of their components do. Conjunctions are joint denials of negations: they denote waow each of their components do.
Subcompounds of compounds are their self or those of their components. Substitutions of like compounds for like nonchain subcompounds are (denotative) functional. Functional substitutions of functional substitutions of compounds are functional substitutions of their self.
Compounds are (functionally)
Alternations of compounds with their negations are valid (they denote waow the compound does or its negation does, i.e. waow it does or does not, so, each functional substitution denotes). Conjunctions of compounds with their negations are nonconsistent (they denote waow their compound does and its negation does i.e. waow it does and does not, so, each functional substitution does not denote). Compounds are implied by and equivalent to their self.
Functional substitutions in
Interchanges of equivalents in a compound are equivalent to that compound (each functional substitution of a compound matches the same of its interchange, except perhaps for the same of the equivalents which otherwise denote in tandem, so each denotes waow the other does i.e. they are equivalent).
Interchanges of equivalents in
Compounds are equivalent to
The equivalents of associativity yield the many component alternations and conjunctions which are equivalent to iterated nestings of alternations or conjunctions down their left or right components.
Validity is
Exclusive alternations are negations of the biconditionals of their components.
Each clause is equivalent to a disjoint or empty clause. Each chain is equivalent to
I may need to work more on this with Skinner’s Thinking Aid.
This continues work on the little lisp from the last entry 2025 0509 1429.
I was tempted to introduce functions that took a letter and turned it into the corresponding rune as a slower introduction to the method of encoding letters by their numeric index in a given alphabetization and then encoding that number as a tally (where a tally is the proper name for a sequence of empty pairs which is equivalent to the iterated pairing of the empty pair on the left of the empty pair). Now that I write of that temptation I am prepared to indulge it rather than leap to the ‘index of’ methods. These grammatical methods, though they may seem trivial to someone already familiar with the breadth and depth of algorithms from a classical education in computer science, are of great consequence in the traditional methods of arithmetizing syntax.
let run=code=>console.log(code,'\n',eval(code)) // for examples
// pairs
let theEmptyPair={}
, isEmpty = it => it == theEmptyPair
, pairOf = (it, that) => [it, that]
, leftOf = it => isEmpty(it) ? it : it[0]
, rightOf = it => isEmpty(it) ? it : it[1];
// sequences as pairs
let theEmptySequence = theEmptyPair
, isEmptySequence = isEmpty
, singletonSequenceOf = it => pairOf(it, theEmptySequence)
, headOf = leftOf
, restOf = rightOf
, concatOf = (it, that) => isEmptySequence(it) ? that
: pairOf(headOf(it), concatOf(restOf(it), that))
, prependSingletonOf = (it, that) =>
concatOf(singletonSequenceOf(it), that)
, appendSingletonOf = (it, that) =>
concatOf(that,singletonSequenceOf(it));
// stacks as pairs
let theEmptyStack = theEmptyPair
, isEmptyStack = isEmpty
, singletonStackOf = it => pairOf(theEmptyStack, it)
, pushOf = pairOf
, dropOf = leftOf
, topOf = rightOf
, secondOf = stack => topOf(dropOf(stack))
, drop2Of = stack => dropOf(dropOf(stack))
, prependOf = stack => pushOf(drop2Of(stack)
, prependSingletonOf(topOf(stack), secondOf(stack)))
, appendOf = stack => pushOf(drop2Of(stack)
, appendSingletonOf(topOf(stack), secondOf(stack)));
// letters and Runes
let theEmptyRune = theEmptyPair
, isEmptyRune = isEmpty
, theOpenRune = pairOf(theEmptyPair, theEmptyRune)
, isOpenRune = it => !isEmpty(it) && isEmpty(leftOf(it)) && isEmptyRune(rightOf(it))
, theCloseRune = pairOf(theEmptyPair, theOpenRune)
, isCloseRune = it => !isEmpty(it)&&isEmpty(leftOf(it))&&isOpenRune(rightOf(it))
, theEmptyLetter=''
, isEmptyLetter = it => it == theEmptyLetter
, theOpenParen = '('
, isOpenParen = it => it == theOpenParen
, theCloseParen = ')'
, isCloseParen = it => it == theCloseParen
, stringOf = (...letters) =>
letters.length ? letters.shift() + stringOf(...letters) : theEmptyLetter
, firstLetterOf = letters => isEmptyLetter(letters) ? theEmptyLetter : letters[0]
, restLettersOf = letters => isEmptyLetter(letters) ? theEmptyLetter : letters.slice(1)
, runeOf = letter => isOpenParen(letter) ? theOpenRune
: isCloseParen(letter) ? theCloseRune
: theEmptyRune
, letterOf = rune => isOpenRune(rune) ? theOpenParen
: isCloseRune(rune) ? theCloseParen
: theEmptyLetter
, runesOf = letters => isEmptyLetter(letters) ? theEmptySequence
: prependSingletonOf(runeOf(firstLetterOf(letters)), runesOf(restLettersOf(letters)))
, lettersOf = runes => isEmptySequence(runes) ? theEmptyLetter
: stringOf(letterOf(headOf(runes)), lettersOf(restOf(runes)))
run('lettersOf(runesOf("(()())() this should be ignored"))');
// read and print sequences
let readOpenRuneOf = (stack, runes) =>
readSequenceOf(pushOf(stack,theEmptySequence), restOf(runes))
, readCloseRuneOf = (stack, runes) =>
readSequenceOf(appendOf(stack), restOf(runes))
, readSequenceOf = (stack, runes) =>
isEmptySequence(runes) ? headOf(topOf(stack))
: isOpenRune(headOf(runes)) ? readOpenRuneOf(stack,runes)
: isCloseRune(headOf(runes)) ? readCloseRuneOf(stack, runes)
: readSequenceOf(stack, restOf(runes))
, readOf = runes => readSequenceOf(theEmptyStack, runes);
let parenOf = runes => prependSingletonOf(theOpenRune
, appendSingletonOf(theCloseRune, runes))
, printSequenceOf = sequence =>
isEmptySequence(sequence) ? theEmptySequence
: concatOf(printOf(headOf(sequence))
, printSequenceOf(restOf(sequence)))
, printOf = sequence =>
isEmptySequence(sequence) ? parenOf(theEmptySequence)
: parenOf(printSequenceOf(sequence))
, read = letters => readOf(runesOf(letters))
, print = sequence => lettersOf(printOf(sequence));
run("lettersOf(parenOf(runesOf('(()()())')))")
run("print(read('()'))")
run("print(read('(()(())())'))")
run("print(read('(())(()(())())'))")
run("print(read('(()(())())((()))'))")
run("print(read('))))(('))")
This appears to have simplified definitions while also pointing to general methods e.g. the definitions of empty, open, and closed runes are ripe for inductive generalization as was done in past versions that gave the alphabet as a javascript array on which functions of search were invoked.
Here is the completed code from yesterday’s work 2025 0508 2207 on my little lisp. I also put the appropriate printer code that was missing from the last note into that note.
let run=code=>console.log(code,'\n',eval(code)) // for examples
// pairs
let theEmptyPair={}
, isEmpty = it => it == theEmptyPair
, pairOf = (it, that) => [it, that]
, leftOf = it => isEmpty(it) ? it : it[0]
, rightOf = it => isEmpty(it) ? it : it[1];
// sequences as pairs
let theEmptySequence = theEmptyPair
, isEmptySequence = isEmpty
, singletonSequenceOf = it => pairOf(it, theEmptySequence)
, headOf = leftOf
, restOf = rightOf
, concatOf = (it, that) => isEmptySequence(it) ? that
: pairOf(headOf(it), concatOf(restOf(it), that))
, prependSingletonOf = (it, that) =>
concatOf(singletonSequenceOf(it), that)
, appendSingletonOf = (it, that) =>
concatOf(it,singletonSequenceOf(that));
// stacks as pairs
let theEmptyStack = theEmptyPair
, isEmptyStack = isEmpty
, singletonStackOf = it => pairOf(theEmptyStack, it)
, pushOf = pairOf
, dropOf = leftOf
, topOf = rightOf
, secondOf = stack => topOf(dropOf(stack))
, encatOf = stack => pushOf(dropOf(dropOf(stack))
, appendSingletonOf(secondOf(stack), topOf(stack)));
// letters
let theEmptyLetter=''
, isEmptyLetter = it => it == theEmptyLetter
, stringOf = (...letters) =>
letters.length ? letters.shift() + stringOf(...letters) : theEmptyLetter
, firstLetterOf = letters => isEmptyLetter(letters) ? theEmptyLetter : letters[0]
, restLettersOf = letters => isEmptyLetter(letters) ? theEmptyLetter : letters.slice(1)
, theOpenParen = '('
, isOpenParen = it => it == theOpenParen
, theCloseParen = ')'
, isCloseParen = it => it == theCloseParen
, runesOf = letters => isEmptyLetter(letters) ? theEmptySequence
: isOpenParen(firstLetterOf(letters)) ? prependSingletonOf(theOpenRune, runesOf(restLettersOf(letters)))
: isCloseParen(firstLetterOf(letters)) ? prependSingletonOf(theCloseRune, runesOf(restLettersOf(letters)))
: runesOf(restLettersOf(letters))
, lettersOf = runes => isEmptySequence(runes) ? theEmptyLetter
: isOpenRune(headOf(runes)) ? stringOf(theOpenParen, lettersOf(restOf(runes)))
: isCloseRune(headOf(runes)) ? stringOf(theCloseParen, lettersOf(restOf(runes)))
: lettersOf(restOf(runes));
run('lettersOf(runesOf("(()())() this should be ignored"))');
// read and print sequences
let theOpenRune = theEmptyPair
, isOpenRune = isEmpty
, readOpenRuneOf = (stack, runes) =>
readSequenceOf(pushOf(stack,theEmptySequence), restOf(runes))
, theCloseRune = pairOf(theEmptyPair,theEmptyPair)
, isCloseRune = it => !isEmpty(it)&&isEmpty(leftOf(it))&&isEmpty(rightOf(it))
, readCloseRuneOf = (stack, runes) =>
readSequenceOf(encatOf(stack), restOf(runes))
, readSequenceOf = (stack, runes) => isEmptySequence(runes) ? headOf(topOf(stack))
: isOpenRune(headOf(runes)) ? readOpenRuneOf(stack,runes)
: isCloseRune(headOf(runes)) ? readCloseRuneOf(stack, runes)
: readSequenceOf(stack, restOf(runes))
, readOf = runes => readSequenceOf(theEmptyStack, runes);
run("isEmpty(read('()'))")
let parenOf = runes =>
prependSingletonOf(theOpenRune, appendSingletonOf(runes, theCloseRune))
, printSequenceOf = sequence =>
isEmptySequence(sequence) ? theEmptySequence
: concatOf(printOf(headOf(sequence))
, printSequenceOf(restOf(sequence)))
, printOf = sequence =>
isEmptySequence(sequence) ? parenOf(theEmptySequence)
: parenOf(printSequenceOf(sequence))
, read = letters => readOf(runesOf(letters))
, print = sequence => lettersOf(printOf(sequence));
run("lettersOf(parenOf(runesOf('(()()())')))")
run("print(read('()'))")
run("print(read('(()(())())'))")
run("print(read('(())(()(())())'))")
run("print(read('(()(())())((()))'))")
run("print(read('))))(('))")
This entry shall be different than prior ones. Some time in the near future I shall look back on what I have done and select what has worked from what has not. The only way I know how to carry out that task effectively is to print out what I have written and go over it with a pen in my hand and some note paper near by, but this will be the first time that I attempt to do this in public.
There are a lot of firsts among what is written here. This is another first not entirely divorced from the more detailed review of what I have done that is yet forthcoming. Originally I had aimed at scheduling a public review at the end of last month. As the days go by I shall miss the mark more and more, but I do not expect to be that far off. Perhaps I am simply being optimistic.
This entry is a response to what I have to say about what I have done that I recall without looking back over a print out of what is recorded here. There are three threads that are woven between my writings:
For all the philosophy that is here, I have never been a big fan of it. The moment I was exposed to the science of behavior, as outlined in any one of B. F. Skinner’s books, I left philosophy as anything more than a narrow science of a narrow kind of verbal behavior (what I have since come to call “the science of smooth dialogue”).
To my surprise, the science of behavior has yet to take up much space here. As much as I’ve tried to keep all my episodes of reading on the record here, I have failed to do so on many occasions. In particular, I have read selections from Skinner’s “Cumulative Record” without noting what my responses to them were here. Even when I carry around a notepad and a pen in my pocket, I still read without making any record of what occurs to me while reading. There is an excuse looming; something about thinking without writing being faster and easier than writing without thinking.
I’ve spent far more time working on my little lisp than I ever planned. This is partially becuase I am frequently reminded of the missing link between Feferman’s Finitary Inductively Presented Logics and something like LISP. It is also because I must make sure that there is nothing I have missed about each step in the construction and conventions of LISP that might later leave me blind when I finally come to compare and contrast it with FORTH or some other concatenative language.
Another reason I’ve spent so much time on it is because I see each obstacle at some later stage of any longer work as an opportunity to introduce a foundation whose deferred consequences entirely avoid such obstacles. The more obstructions I confront the more likely I am to examine the link between whatever premises I’ve overtly identified and how presently obstructing conclusions follow from them. This is rarely a matter of formal deduction, but almost always a result of methods with a strong familial similarity to logic. Almost always, major problems are the result of a poorly selected lexicon. The basic predicates with which a problem is stated and in which its solutions are proposed almost always decide my fate.
In the past LISP has been saddled with lambda calculi. No such theories are likely to turn up here: lambda calculi are nonconsistent without some crippling and complex system of typing. Furthermore, the methods of lambda calculi are merely the theoretical echos of better logical methods e.g. predicate abstraction and concretion are the logical operations that permit talk of lambda abstraction and application. This has been known since Quine and Church traded reviews of each others works in the early half of the 20th century. The moment Quine revealed that application could be reinterpreted as the collective relate (an operation on relations uncovered by Russell among his short summaries of work he and Whitehead had done), it should have been clear that no result obtained from the lambda calculus should be taken as special to it. Little did either Quine or Church know that Quine’s later methods of predicate abstraction and concretion revealed the logically universal way of dealing with the multitude of theories mathematicians may propose (be they lambda calculi, relational calculi, set theories etc.)
Thus, I’ve spent much time fussing over tiny distinctions between various attempts at getting out something like LISP: the methods are to be closer to those of Goodstein’s equation calculi than to any lambda calculus. But, it is more than that. Ordered pairs are the most modest abstract objects that may be admitted to any theory and which release most of the grammatical methods that have found greatest use in the analysis of carefully designed languages. This is the premise not only revealed at the end of Quine’s “Word and Object” but also a turning point in Feferman’s F.I.P.L.s.
Up until very recently this was the path I planned on following in order to reach out from logic to the practices of programming. It is only upon uncovering “Logic Programming” that the entirety of programming practice can be reduced to the predicate “is valid”, or, as the case often is, the kindred predicate “is inconsistent”. Had this been shown to me earlier I may have avoided much of my work on primitive recursive functions of ordered pairs. But, I’m still stuck in a world of Forths and Lisps rather than a world of logic.
What little I have learned of Prolog has no promise to me. It works on a limited slice of logical methods and Datalog doesn’t appear to do much better. The rush to monetize logic programming and its fundamental relation to database systems generally ruined the logic. The modern world, one where large masses of capital are committed to the matrix methods of machine learning, is technically one where logic programming is primed to reign supreme for the following simple reason: any path which leads to inconsistency works but some paths are faster than others for no reason in particular. Pro- and data- log aimed to only permit constructions which admit short chains of relevant implications. The historic obsession with implication as in formal deduction has blinded modern computer scientists and logicians to the role of logical methods prior to any science e.g. mathematical or computational. It is only a matter of time before the window for logical programming closes once again for the same reason it did in the 80s.
Through my little lisp I still plan on exploring the methods of logic programming. This shall be done in parallel with my other programming language which is much like uhdForth.
The problem I’m posed with upon writing these paragraphs is this: can I juggle more while also building a new experimental culture?
This continues my work on my little lisp from 2025 0507 2029.
I’m hopeful that whatever logical errors in the code I produced yesterday will be solved today: I am more awake now than I was when I was working on this last night. It seems that I forgot about the output at the end of the reader: instead of just taking the top of the stack as the appropriately constructed sequence, I must actually take the head of the sequence atop the stack. This is due to how a close parenthesis (or close rune) works in the degenerate case where there are less than two items on the stack. In this degenerate case the last parenthesis (of a sequence of runes with balanced parenthesis, as I have not checked the unbalanced cases yet) actually encats the singleton sequence of the top of the list to what it takes as an empty sequence, but which is really the result of a degenerate case of taking the top of the empty stack: it returns the empty sequence.
As I went to copy the code over here I again fell prey to coding rather than copying. It occurred to me that there are errors that have slipped in because I have once again not properly factored out the boundaries between languages that are under construction.
Thanks to my prior work I have happily seperated the defining features of sequences from those of lists. Though it is not a traditional distinction, it is consistent with Feferman’s Finitary Inductively Presented Logics use of sequences as special ordered pairs and lists as a sequence grammatically tagged as a sequence. The analysis of a LISP list as a sequence tagged grammatically as a sequence is key to general grammatical methods at work in modern programming practices and which are sadly conflated with problems of types as an abstract concern.
My aim now is to simplify the reader and printer to work entirely with sequences in order to sus out any logical errors at the level of a purely parenthetical language rather than one that also induces the complexity of runes and strings.
The new starting point is this code:
// pairs
let theEmptyPair={}
, isEmpty = it => it == theEmptyPair
, pairOf = (it, that) => [it, that]
, leftOf = it => isEmpty(it) ? it : it[0]
, rightOf = it => isEmpty(it) ? it : it[1]
// sequences as pairs
, theEmptySequence = theEmptyPair
, isEmptySequence = isEmpty
, singletonSequenceOf = it => pairOf(it, theEmptySequence)
, headOf = leftOf
, restOf = rightOf
, concatOf = (it, that) => isEmptySequence(it) ? that
: pairOf(headOf(it), concatOf(restOf(it), that))
// stacks as pairs
, theEmptyStack = theEmptyPair
, isEmptyStack = isEmpty
, singletonStackOf = it => pairOf(theEmptyStack, it)
, pushOf = pairOf
, dropOf = leftOf
, topOf = rightOf
, secondOf = stack => topOf(dropOf(stack))
, encatOf = stack => pushOf(dropOf(dropOf(stack))
, concatOf(secondOf(stack), singletonSequenceOf(topOf(stack))));
The only part that still needs a bit of explanation is what the function designated by ‘encatOf’ actually does and why it keeps occurring in the code (or why some function like it continues to reoccur). My explanation for now is that it is the least intrusive way to take advantage of the power of stack based methods without bringing every defined operation on pairs and sequences to stacks. To do so would be to create a stack based language like FORTH, also called a concatenative language. That is not the aim here.
What the function designated by ‘encatOf’ does is to help with how to interpret a closed parenthesis. This gets us to the new function being defined. For now it is designated by ‘readSequenceOf’. It aims to take a sequence of runes, each of which is an open or close rune, and construct the corresponding iterated sequences of empty sequences. The open rune begins a new sequence by pushing the empty sequence onto the stack and the close rune adds the item at the top of the stack to the end of the sequence that is second from the top of the stack.
Here is my first attempt based on all that I’ve learned so far.
// read sequences from a sequence of open and closed runes
let theOpenRune = theEmptyPair
, isOpenRune = isEmpty
, readOpenRuneOf = (stack, runes) =>
readSequenceOf(pushOf(stack,theEmptySequence), restOf(runes))
, theCloseRune = pairOf(theEmptyPair,theEmptyPair)
, isCloseRune = it => !isEmpty(it)&&isEmpty(leftOf(it))&&isEmpty(rightOf(it))
, readCloseRuneOf = (stack, runes) =>
readSequenceOf(encatOf(stack), restOf(runes))
, readSequenceOf = (stack, runes) => isEmptySequence(runes) ? headOf(topOf(stack))
: isOpenRune(headOf(runes)) ? readOpenRuneOf(stack,runes)
: isCloseRune(headOf(runes)) ? readCloseRuneOf(stack, runes)
: readSequenceOf(stack, restOf(runes))
, readOf = runes => readSequenceOf(theEmptyStack, runes);
Now only a small number of specialty operations on javascript strings appear to be needed to interface with these functions as defined.
// letters
let theEmptyLetter=''
, isEmptyLetter = it => it == theEmptyLetter
, stringOf = (...letters) =>
letters.length ? letters.shift() + stringOf(...letters) : theEmptyLetter
, firstLetterOf = letters => isEmptyLetter(letters) ? theEmptyLetter : letters[0]
, restLettersOf = letters => isEmptyLetter(letters) ? theEmptyLetter : letters.slice(1)
, theOpenParen = '('
, isOpenParen = it => it == theOpenParen
, theCloseParen = ')'
, isCloseParen = it => it == theCloseParen;
They are the same old same old, except for the last two that single out two particular letters. It finally occurred to me to add the following two definitions to the end of the definitions of functions for treating ordered pairs as sequences:
let prependSingletonOf = (it, that) =>
concatOf(singletonSequenceOf(it), that)
, appendSingletonOf = (it, that) =>
concatOf(it,singletonSequenceOf(that));
Then, the definition of the function designated by ‘encatOf’ becomes just as follows.
let encatOf = stack => pushOf(dropOf(dropOf(stack))
, appendSingletonOf(secondOf(stack), topOf(stack)));
Construction of a sequence of runes from a string of letters of which only the parenthesis are of concern proceeds as follows.
let runesOf = letters => isEmptyLetter(letters) ? theEmptySequence
: isOpenParen(firstLetterOf(letters)) ? prependSingletonOf(theOpenRune, runesOf(restLettersOf(letters)))
: isCloseParen(firstLetterOf(letters)) ? prependSingletonOf(theCloseRune, runesOf(restLettersOf(letters)))
: runesOf(restLettersOf(letters))
Then the function that gets us back from a sequence of runes to letters is as follows.
let lettersOf = runes => isEmptySequence(runes) ? theEmptyLetter
: isOpenRune(headOf(runes)) ? stringOf(theOpenParen, lettersOf(restOf(runes)))
: isCloseRune(headOf(runes)) ? stringOf(theCloseParen, lettersOf(restOf(runes)))
: lettersOf(restOf(runes))
A brief example demonstrates that the functions from letters to runes and vice versa work.
lettersOf(runesOf("(()())() this should be ignored"))
(()())()
This assures us that when we send a sequence of runes to the reader it corresponds to the parentheses in the given string.
Next the sequence printer. It should print out a sequence of runes that, when put back into the reader and printed should output an identical sequence of runes. The notation for sequences adopted in the reader has an open and close rune around its contents.
let parenOf = runes => prependSingletonOf(theOpenRune, appendSingletonOf(runes, theCloseRune))
An example:
lettersOf(parenOf(runesOf('(()()())')))
((()()()))
The sequence printer assumes that the pair it is passed is to be printed as a pure sequence. If asked to print the empty sequence it prints an open rune followed by a closed rune. Otherwise it encloses the print of each item in the sequence, one after the other, in open and closed runes.
let parenOf = runes =>
prependSingletonOf(theOpenRune, appendSingletonOf(runes, theCloseRune))
, printSequenceOf = sequence =>
isEmptySequence(sequence) ? theEmptySequence
: concatOf(printOf(headOf(sequence))
, printSequenceOf(restOf(sequence)))
, printOf = sequence =>
isEmptySequence(sequence) ? parenOf(theEmptySequence)
: parenOf(printSequenceOf(sequence))
, read = letters => readOf(runesOf(letters))
, print = sequence => lettersOf(printOf(sequence));
The following examples test out the reader and the printer.
isEmpty(read('()'))
true
print(read('()'))
()
print(read('(()(())())'))
(()(())())
print(read('(())(()(())())'))
(())
print(read('(()(())())((()))'))
(()(())())
The last two examples indicate conspicuously that the case in the function denoted by ‘readSequenceOf’ where the rune sequence is empty and returns the head of the top of the stack on which the corresponding sequence was being constructed.
This parenthesis notation is the infix version of the postfix notation from Bit Strings and Binary Trees but instead of interpreting everything as binary trees they are interpreted as pure sequences i.e. iterated sequences of the empty sequence.
This is the basic method upon which my earlier attempts at constructing a reader and printer for past versions of my list are based.
This continues the work on my little lisp from 2025 0505 1725.
Last note left on a question: were all these distinctions between stacks as LISP lists as sequences and sequences as pairs helpful? What I uncovered upon further reflection is that the answer to this question is yes and no. These distinctions are still of critical importance, but not in the straight line from pairs up through stacks as LISP lists. What occurred to me is that there are two languages that I’ve built upon native javascript and that I’ve failed to untangle them from each other.
The lowest language is the programmable part of the arithmetic of ordered pairs, and the language above that is my little LISP. Thus there are two readers: one that constructs LISP lists and the one for my little LISP (which is more like a calculator or evaluator).
Sequences and stacks are then implemented as before the last entry: atop the arithmetic of ordered pairs.
// pairs
let theEmptyPair={}
, isEmpty = it => it == theEmptyPair
, pairOf = (it, that) => [it, that]
, leftOf = it => isEmpty(it) ? it : it[0]
, rightOf = it => isEmpty(it) ? it : it[1]
// sequences as pairs
, theEmptySequence = theEmptyPair
, isEmptySequence = isEmpty
, singletonSequenceOf = it => pairOf(it, theEmptySequence)
, headOf = leftOf
, restOf = rightOf
, concatOf = (it, that) => isEmptySequence(it) ? that
: pairOf(headOf(it), concatOf(restOf(it), that))
// stacks as pairs
, theEmptyStack = theEmptyPair
, isEmptyStack = isEmpty
, singletonStackOf = it => pairOf(theEmptyStack, it)
, pushOf = pairOf
, dropOf = leftOf
, topOf = rightOf
, secondOf = stack => topOf(dropOf(stack))
, encatOf = stack => pushOf(dropOf(dropOf(stack))
, concatOf(secondOf(stack), singletonSequenceOf(topOf(stack))));
All is back to how it was a few notes ago but with a few better names for the different operations. These name changes also introduce some later distinctions that will come as the result of grammatical (type) analysis. I’ve gone back to not focusing on a specific alphabetization of runes. The reader is now assembled from simpler functions for reading the special open, close, and space runes.
There are a few other tricks of the trade that have been included e.g. javascript defaults to giving negative one when it does not find the index of an item you are looking for in an array, but it greatly simplifies a theory, in almost all cases, when it returns the length of the array. Another very helpful trick is to return the entire alphabet when you give it an index that is outside the range of the alphabet. Both of these are examples of degenerate cases that can be made to play nicely with the rest of a theory.
// letters
, theEmptyLetter=''
, isEmptyLetter = it => it == theEmptyLetter
, stringOf = (...letters) =>
letters.length ? letters.shift() + stringOf(...letters) : theEmptyLetter
, firstLetterOf = letters => isEmptyLetter(letters) ? theEmptyLetter : letters[0]
, restLettersOf = letters => isEmptyLetter(letters) ? theEmptyLetter : letters.slice(1)
// letter as number
, abc =[...'() 0123456789abcdefghijklmnopqrstuvwxyz']
, alphabeticalIndexOf = letter =>
abc.includes(letter) ? abc.indexOf(letter) : abc.length
, alphabeticalLetterOf = index =>
0 <= index && index < abc.length ? abc[index] : abc
// number as tally (as pair)
, tallyOf = n => n>0 ? pairOf(theEmptyPair, tallyOf(n-1)) : theEmptyPair
, countOf = x => isEmpty(x) ? 0 : 1 + countOf(rightOf(x))
// rune as tally
, runeOf = letter => tallyOf(alphabeticalIndexOf(letter))
, letterOf = rune => alphabeticalLetterOf(countOf(rune))
// letters as sequences of runes
, runesOf = letters =>
letters.length ? concatOf(
singletonSequenceOf(runeOf(firstLetterOf(letters)))
, runesOf(restLettersOf(letters)))
: theEmptySequence
, lettersOf = runes =>
isEmptySequence(runes) ? theEmptyLetter
: stringOf(letterOf(headOf(runes)), lettersOf(restOf(runes)))
// recursive definition of identity of pairs
, id = (it, that) =>
(isEmpty(it) && isEmpty(that))
|| (!(isEmpty(it) || isEmpty(that))
&& id(leftOf(it), leftOf(that))
&& id(rightOf(it), rightOf(that)))
// reader
, theOpenRune = runeOf('(')
, isOpenRune = it => id(it, theOpenRune)
, readOpenRuneOf = (stack, runes) =>
readerOf(pushOf(stack, theEmptySequence), restOf(runes))
, theCloseRune = runeOf(')')
, isCloseRune = it => id(it, theCloseRune)
, readCloseRuneOf = (stack, runes) =>
readerOf(encatOf(stack), restOf(runes))
, theSpaceRune = runeOf(' ')
, isSpaceRune = it => id(it, theSpaceRune)
, readSpaceRuneOf = (stack, runes) => !isEmptySequence(topOf(stack))
? readerOf(pushOf(encatOf(stack), theEmptySequence), restOf(runes))
: isEmptySequence(secondOf(stack)) ? readerOf(stack, restOf(runes))
: readerOf(pushOf(stack, theEmptySequence), restOf(runes))
, isRune = it => countOf(it) < abc.length
, readRuneOf = (stack, runes) =>
readerOf(encatOf(pushOf(stack, headOf(runes))), restOf(runes))
, readerOf = (stack, runes) =>
isEmptySequence(runes) ? topOf(stack)
: isOpenRune(headOf(runes)) ? readOpenRuneOf(stack, runes)
: isCloseRune(headOf(runes)) ? readCloseRuneOf(stack, runes)
: isSpaceRune(headOf(runes)) ? readSpaceRuneOf(stack, runes)
: isRune(headOf(runes)) ? readRuneOf(stack, runes)
: readerOf(stack, restOf(runes))
, readOf = runes => readerOf(theEmptyStack, runes)
This reader is increadibly powerful. Here is a tiny printer that anticipates the method of lists as sequences which begin with the symbol “list”.
// printer
, printSequenceOf = sequence => isEmptySequence(sequence) ? theEmptySequence
: concatOf(printOf(headOf(sequence))
, concatOf(singletonSequenceOf(theSpaceRune)
, printSequenceOf(restOf(sequence))))
, parenOf = runes => concatOf(singletonSequenceOf(theOpenRune)
, concatOf(runes, singletonSequenceOf(theCloseRune)))
, printParenSequenceOf = sequence => parenOf(concatOf(singletonSequenceOf(theSpaceRune), printSequenceOf(sequence)))
, isList = sequence => id(headOf(sequence), runesOf('list'))
, printListOf = list => printParenSequenceOf(list)
, printOf = item => isList(item)? printListOf(item) : item
// external read and print
, read = letters => readOf(runesOf(letters))
, print = item => lettersOf(printOf(item))
Neither the reader nor the printer seem to be working as expected e.g. reading a space rune goes wrong when there are lots of parentheses bunched up.
This continues my work on my paper on logic from 2025 0412 1422.
The following presentation of logic is unique in a number of ways. Primarily it is quantificational logic. It does away with quantifiers as in “(each item is x such that)(x is human only if x is mortal)” by skipping over the intermediate methods of Quine’s predicate abstracts as in “(each item is)(x such that x is human only if x is mortal)” which clearly factors out the predicate functor ‘each item is’ from the predicate ‘x such that x is human only if x is mortal’ by doing away entirely with the classical role of variables as in ‘each human only if mortal’.
Predicate functor logic is a foreign language to those who speak natively with all the cross referential power of pronouns, whether they be free and taken as dummy singular terms or bound and taken as dummies of qunatification. Though the grammars of (the unhappily named) natural languages are burdened with only indirect analyses of cross reference, the carefully designed language of predicate functor logic eliminates all referential oddities by explaining their cross referential operations as the result of the (logical) equivalence of a predicate with one in a special grammatical category (those constructed from the so called recombic predicate functors).
Although at most three predicate functors are sufficient to assemble all the compound predicates from the lexicon of a logical langauge, the predicate functor foreigner would find themselves lost in construction. Thus, full predicate functor logic is introduced on a schedule which is sufficiently similar to classical treatments to admit little or no alarm among those already familiar with the path to quantificational logic from truth functional logic and through the Boolean closure logics of many place predicates (this is an urgent reminder that Boolean logic is the logic of existential closures of truth functional compounds of basic predicates, and that Boolean logic is not to be confused, as it so often is, with the algebra of truth functions as functions on the set whose sole members are the true and the false).
In addition to the unique methods of predicate functor logic, there is an additional and equally profound difference in this presentation of logic: it follows from the grammatical definition of validity as, broadly, substitution for lexicon. Just as Quine failed to provide an autonomous proof procedure for predicate functor logic (one that did not depend on translation to quantificational logic and back from the same), so too did Quine fail to bring his grammatical definition of validity as truth of substitution under supplementation of lexicon to predicate logic. The substitutional basis for the grammatical definition of validity is established by carefully combining the works of Hilbert and Bernay’s in their proof of completeness of quantificational logic in the second book of their Grundlagen as a strengthening of Tarski’s model theoretic definition of validity originally from his analysis of truth in his 1933 paper “The concept of truth in the languages of the deductive sciences” but conspicuously from his 1956 paper “Arithmetical extensions of relational systems” with Robert Vaught, and the completeness theorems of Skolem, Herbrand, and Godel. For more details on the delicate combination of these fundamental components of logical practice see Quine’s “Philosophy of Logic second edition” section entitled “Adequacy of Substitution”.
So the distinguishing features of this presentation are predicate functors and substitutional validity. There are secondary features which are perhaps just as important, but which are not so easy to explain. Traditional logical methods deal with schema and functions of truth values by building up schematic sentence letters with connectives parallel to functions from truth values to truth values as in truth tables, or, much better, as truth trees. Following the guidence of grammatical methods, no such methods are found here. They are secondary in that each branch of a truth tree is an otherworldly mirror on the grammatical structure of their alternational development and reduction by equivalents of identity. This stricture has the benefit of eliminating confusion caused by schematic letters e.g. they are sometimes taken as more than dummies and made to refer or mean in one or another mysterious way.
I am not full convinced of the utility of these self contained grammatical methods, but shall push them as far as I can without losing track of how well the schedule upon which quantificational practices are built is carried over.
What follows is merely an outline. It charts out the path that has not yet been taken based on the best available reports that have come back from those who have survived their trip into the jungle of logic without having been too harshly damanged by its wilds. Russell cleared much of the way, but was left with unhappy impressions that, thankfully, do not seem to have harmed Quine into a frightful silence: Quine wrote on logic till his final days.
These notes are terse and a not inconsiderable amount of effort on the part of the reader is required to effectively navigate them: many of the later conclusions lack supporting arguments. The key to taking each step, e.g. from your native tongue to the truth functional part of predicate functor logic or from the truth functional part to the Boolean closure part, is Quine’s method of paraphrasing inwards. No matter what the inner components of a sentence may be we can analyze their outermost grammatical structure while leaving their inner parts as yet unanalyzed. But, now we must take our sentences as those of a predicate functor community i.e. as one where primitives and compounds are predicates as integral words or phrases such as “father of”, “is man”, or just “man”.
I submit that the sentences of a predicate functor community are like those we attribute to early humans, e.g. “each human only if mortal” rather than the aristotelian universal “All humans are mortal” or the quantificational “Everything is such that it is human only if it is mortal”, and that this is no mere accident of expedient analysis.
A final note on my methods of logic: I have tried to defer the introduction of as many distinguishing features as far as possible. For example, conditionals and biconditionals are introduced far after functional equivalence. This method of spreading out concerns is invaluable in any endevour to simplify and elucidate without adding comments of simplification adn elucidation.
Logic is the science of validity and validity is a consequence of grammar and denotation.
Compounding is (denotative) functional when, exclusively, each like compound denotes or each like compound does not denote, where and only where (waow), exclusively, each like component denotes or each like component does not denote. Chains are compounds compounded functionally.
Joint denials denote waow each of their components do not. Negations are self joint denials: they denote waow their component does not. Alternations are negations of joint denials: they denote waow some of their components do. Conjunctions are joint denials of negations: they denote waow each of their components do.
Subcompounds of compounds are their self or those of their components. Substitutions of like compounds for like nonchain subcompounds are (denotative) functional. Functional substitutions of functional substitutions of compounds are functional substitutions of their self.
Compounds are (functionally)
Alternations of compounds with their negations are valid (they denote waow the compound does or its negation does, i.e. waow it does or does not, so, each functional substitution denotes). Conjunctions of compounds with their negations are nonconsistent (they denote waow their compound does and its negation does i.e. waow it does and does not, so, each functional substitution does not denote). Compounds are implied by and equivalent to their self.
Functional substitutions in
Interchanges of equivalents in a compound are equivalent to that compound (each functional substitution of a compound matches the same of its interchange, except perhaps for the same of the equivalents which otherwise denote in tandem, so each denotes waow the other does i.e. they are equivalent).
Interchanges of equivalents in
Compounds are equivalent to
The equivalents of associativity yield the many component alternations and conjunctions which are equivalent to iterated nestings of alternations or conjunctions down their left or right components.
Chains are nonvalid waow they are not valid and, hence, not where each of their functional substitutions denotes i.e. where some of its functional substitutions do not denote.
Validity is
Exclusive alternations are negations of the biconditionals of their components.
Each clause is equivalent to a disjoint or empty clause. Each chain is equivalent to
Each chain is equivalent to a compound compounded only by
That’s all I have the strength to do for now.
This continues my work on my little lisp from 2025 0504 0140.
In the last entry it occurred to me that the only distinction between types of items that is implicit in LISP is that between lists and nonlists (where an example of a nonlist is a symbol). Consequently, I introduced the following definitions:
// pairs as lists
, isList = item => !isEmpty(item) && isEmpty(carOf(item))
, singletonListOf = item => consOf(theEmptyPair, item)
, theEmptyList = singletonListOf(the)
, prependHelpOf = (x,y) => isEmpty(x) ? (isList(y) ? cdrOf(y) : y)
: consOf(carOf(x), prependHelpOf(cdrOf(x),y))
, prependOf = (x,y) => isList(x) ? prependHelpOf(x,y)
: prependHelpOf(singletonListOf(x),y)
The problem with this is that there is a big difference between a binary tree (as an iterated pairing of the empty pair) as a sequence and as a LISP list. This difference broke the beauty of the symmetry of definitions that take binary trees as stacks on the one hand and sequences on the other. While I was reading Feferman’s “Finitary Inductively Presented Logics” it occurred to me that by distinguishing between sequences and lists I can further factor the underlying arithmetic of binary trees out from the implementation details of my little lisp.
The empty sequence is identical to the empty pair and a pair is the empty sequence when it is the empty pair. The left part of a pair is called the head of the sequence. It is also called the first item of the sequence. The right part of a pair is called the rest of the sequence. Sequences are constructed from other sequences by concatenation.
The code which corresponds to this specification is as follows.
// basic operations on pairs
let theEmptyPair={}
, isEmpty = it => it == theEmptyPair
, pairOf = (it, that) => [it, that]
, leftOf = it => isEmpty(it) ? it : it[0]
, rightOf = it => isEmpty(it) ? it : it[1]
// sequences as pairs
, theEmptySequence = theEmptyPair
, isEmptySequence = isEmpty
, singletonSequenceOf = it => pairOf(it, theEmptySequence)
, headOf = leftOf
, restOf = rightOf
, concatOf = (it, that) => isEmptySequence(it) ? that
: pairOf(headOf(it), concatOf(restOf(it), that))
A sequence whose head is the empty pair is now a LISP list. Thus a pair is a LISP list if it is not the empty pair and if its left part is. Promote a sequence to a list by concatenating the singleton sequence whose sole item is the empty pair with the sequence. So, the empty list is the list of the empty sequence, and the singleton list of an item is the list of the singleton sequence of that item. The first item of a list is called its ‘car’ and the list of the rest of the items in the sequence of the list is called the ‘cdr’. Lists shall be constructed by consing rather than concatenating so that nonlist items are automatically promoted to the appropriate singleton list.
// LISP lists as sequences
, isList = it => !isEmpty(it) && isEmpty(leftOf(it))
, listOf = sequence => pairOf(theEmptyPair, sequence)
, theEmptyList = listOf(theEmptySequence)
, isEmptyList = it => isList(it) && isEmptySequence(rightOf(it))
, singletonListOf = it => listOf(singletonSequenceOf(it))
, carOf = list => headOf(rightOf(list))
, cdrOf = list => listOf(restOf(rightOf(list)))
, consOf = (it, that) =>
!isList(it) ? consOf(singletonListOf(it),that)
: isList(that) ? listOf(concatOf(rightOf(it),rightOf(that))
: consOf(it, singletonListOf(that)))
In the past, I built stacks directly from pairs, but now I shall build them from LISP lists since they are needed as part of my little lisp and not as part of the arithmetic of pairs.
The empty stack is the empty list, an item is pushed onto a stack by consing it to the stack, the top of the stack is got with car, and the top is dropped from the stack with cdr. The second from the top of the stack is the top of the drop of that stack. Finally, the top two items of the stack are “enconsed” by replacing them with the cons of the second from top with the singleton list of the top: this is used mostly for implementing the reader.
// stacks as LISP lists
, theEmptyStack = theEmptyList
, isEmptyStack = isEmptyList
, pushOf = consOf
, dropOf = cdrOf
, topOf = carOf
, secondOf = stack => topOf(dropOf(stack))
, enconsOf = stack => pushOf(dropOf(dropOf(stack))
, consOf(secondOf(stack), topOf(stack)))
Is this a good way of doing things though? The old way clung to the arithmetic of binary trees as iterated ordered pairs of the empty pair. This way goes back to the earlier method of distinguishing between runic lists and nonrunic lists, or runic symbols and nonrunic symbols. The difference is that here, symbols are pairs rather than javascript strings. This doesn’t eliminate the problem of dealing with javascript strings, we still need to carry them over into some sequences of symbols.
Has this accomplished anything or is it like when you make a substitution into an algebra problem and end up with ‘zero equals zero’.
My interests in philosophy are largely historic: the literature of philosophers tells us something about what was going on from its compatibility and incompatibility with cultural practices uncovered by archeology. Bryan Magee (1930-2019) presented two series of television shows which gave the public a instructive look into the work of philosophers past and present:
The first is composed of the following speakers and topics:
The second is composed of the following speakers and philosophers:
Both series can be easily found on YouTube and are well worth a long watch, especially when the speaker is speaking on their own work e.g. Quine.
This begins my read of Peter Adamson’s “Classical Philosophy: a history of philosophy without gaps” from 2014.
It was recommended to me by R.P. when he noticed that I was reading some other histories of philosophy in addition to Durant’s “Story of Philosophy” which I last wrote about in 2025 0430 2324.
As the preface explains, this is one in a series of books that was written from a podcast on the history of philosophy. Unlike the other histories of philosophy that I have on hand, this one purports to exhaust the nooks and crannies of philosophers throughout the world, hence the “without gaps”. More than anything, I look forward to the list of references promised: a good reading list can save a lifetime of woe.
Classical philosophy is very narrow in its scope, and the book is broken into the traditional three parts:
The ‘Early Greeks’ are sometimes called the “presocratics” as if Socrates was the saviour of philosophy. Each chapter of the part on early greek philosophy deals with a specific philosopher and, in a few places, with groups of them. The last two parts deal exclusively with Socrates, Plato, and Aristotle.
Andamson says he links up the philosophers to their social environments and that he doesn’t see a clean seperation between the concerns of the philosopher and those of the culture of which they were a part. His claim that philosophy does not occur without affluence or some kind of social saftey net is common enough to be somewhat beyond question, but I do question it. I submit that the philosophy which has survived to our day, e.g. as a corpus of literature if I am pressed to give a concrete account of philosophy as a part of our social environment, is a collateral effect of any verbal community and that what has survived is simply that. Said humorously: the absense of evidence is not evidence of absense.
Philosophy is a consequences of the reinforcing effects of a verbal community which practices a kind of self control i.e. the deferred consequences of the reinforcing practices of the verbal community are taken into account by it. Any culture which comes to teach its people to care for their own survival and the survival of their culture produces philosophers in that there are those sensitive to the contribution that verbal behavior can make to that survival.
When Socrates passes through Plato it is ethics that reigns supreme as “ the study of ideal conduct; the highest knowledge, said Socrates, is the knowledge of good and evil, the knowledge of the wisdom of life” [pg. 3 Durant “The Story of Philosophy”]. Here, ideal conduct is that which does more than avoid the threats and damages which doom the practices of a culture: it is that which contributes most to its survival for whatever reason no matter how consistent or inconsistent it may be with prevailing verbal practices.
The history of philosophy is, to me, the history of a conspicuous kind of cultural self control. It is no surprise that the sciences descend from philosophy and that philosophy can be, in my measure, best taken as a kind of science. In the past I have suggested that it is the science of smooth discourse in as much as it eeks out where and when verbal behavior works and where and when it does not.
The first chapter is “Everything is Full of Gods: Thales”. The Dorians invaded mainland Greece around 1100 BCE and those Athenians who escaped eastward over the Aegean Sea landed under the leadership of Ion. Born c. 626/623, Thales grew up in the southern Ionian colony of Miletus. He is seen by many as the first Greek philosopher.
As when I read Durant’s description of Athens in the “Context of Plato” in 2025 0420 2247, Miletus was a nexus of cultural practices from its eastward and westward neighbors. Thales may have visited Egypt, picked up some Babylonian astronomy, and used it to predict the solar eclipse of 585 BCE. Reports on astronomical phenomena give us ways of calculating the ordering of various events in addition to the timing of the report itself (assuming such reports are not manufactured post hoc).
As with my conviction that the stories of Socrates are to be taken as I take most of history, that is as folklore, perhaps even that stronger folklore called just ‘lore’, so are the stories of Thales taken by me. This outlook is also echoed in the next few pages of Adamson’s sketch.
It is dawning on me that for many of these philosophers there are simply a list of basic stories that we can list and cling to without having to commit ourselves to their historic accuracy. But, my interest in world history (and the role of archeology in setting up and taking down the histories we have inherited) prevents me from resting easy with such foggy commitments. The urgency of writing on world history continues to press itself upon me.
The main stories of Thales are these
Adamson contrasts Thales with
as we do scientific behavior with literary behavior.
The chapter ends with a tiny argument assembled by Adamson as an example of a possible argument Thales may have made:
“everything is full of gods, and I’ll show you this using the example of the magnet. It seems to be lifeless, but it must have a soul, because it can initiate motion. So, by extension, you should at least be open to the idea that everything has a soul, which is divine.”[pg. 7-8]
He then admits the evidence for this argument is weaker than the argument is.
This continues my work on my little lisp from 2025 0502 2048.
An alternate design decision occurred to me: the only implicit distinction betwen teh part played by ordered pairs in a lisp is that between list and nonlist. Thus, rather than checking if a pair is a symbol or not, it is more general to check if it is a list or not. Then the rest of the distinctions can be made however the programmer desires.
This also has the benefit of, in the simple case where an item is either a list or a symbol, of identifying a symbol by its first rune which can never be empty because of the alphabetization adopted. This also allows space to serve its traditional function without adding greatly to the complexity of the definition of the reader. It also allows for a more uniform presentation of pairs as lists followed by pairs as stacks.
// basic operations on pairs
let theEmptyPair={}
, isEmpty = x => x == theEmptyPair
, consOf = (x,y) => [x,y]
, carOf = x => isEmpty(x) ? x : x[0]
, cdrOf = x => isEmpty(x) ? x : x[1]
// left and right singletons
, enlistOf = x => consOf(x,theEmptyPair)
, enstackOf = x => consOf(theEmptyPair,x)
// pairs as lists
, isList = item => !isEmpty(item) && isEmpty(carOf(item))
, singletonListOf = item => consOf(theEmptyPair, item)
, theEmptyList = singletonListOf(the)
, prependHelpOf = (x,y) => isEmpty(x) ? (isList(y) ? cdrOf(y) : y)
: consOf(carOf(x), prependHelpOf(cdrOf(x),y))
, prependOf = (x,y) => isList(x) ? prependHelpOf(x,y)
: prependHelpOf(singletonListOf(x),y)
// pairs as stacks
, theEmptyStack = theEmptyPair
, isEmptyStack = isEmpty
, pushOf = consOf
, popOf = carOf
, topOf = cdrOf
, secondOf = stack => topOf(popOf(stack))
, drop2 = stack => popOf(popOf(stack))
, enpendOf = stack =>
pushOf(drop2(stack)
,prependOf(secondOf(stack)
,enlistOf(topOf(stack))))
// reader
, isOpenParen = isEmpty
, isCloseParen = x => !isEmpty(x)&&isEmpty(carOf(x))&&isEmpty(cdrOf(x))
, isSpace = x => !isEmpty(x)&&isEmpty(carOf(x))&&isCloseParen(cdrOf(x))
, readHelperOf = (stack,runes) => isEmpty(runes) ? carOf(topOf(stack))
: isSpace(carOf(runes)) ?
( isEmpty(topOf(stack)) ? readHelperOf(stack,cdrOf(runes))
: readHelperOf(pushOf(enpendOf(stack),theEmptyPair), cdrOf(runes)))
: isCloseParen(carOf(runes)) ? readHelperOf(enpendOf(stack), cdrOf(runes))
: isOpenParen(carOf(runes)) ? readHelperOf(pushOf(stack,theEmptyList), cdrOf(runes))
: readHelperOf(enpendOf(pushOf(stack,carOf(runes))), cdrOf(runes))
, readOf = runes => readHelperOf(theEmptyStack, runes)
I’ll test this code and figure out the printer next time.
Failure of cross reference does not require a referent.
Today is a Bob Ross and Bosch day. While I do not seem to enjoy reading Michael Connelly, I do enjoy watching his stories on screen. They are that perfect drip of mid. Law & Order is another.
It is better to have gone out of the way to note the source of a quote than to go out of your way to find where you sourced it.
This continues the work on my paper on logic. It also contains more on my philosophy of logic as well as a careful analysis of Quine’s proof that no logical theory of the world is beyond supplementation (which itself contains an alternate presentation of Quine’s analysis of denotation as a generalization of Tarski’s analysis of truth).
My methods of logic are unique in that they present Quine’s predicate functor methods independently of quantificational logic. Quantificational logic is then constructed from predicate functor logic by way of Quine’s predicate abstracts. The difference between predicate functor logic and quantificational logic is the difference between ‘is true’ and ‘is true of’.
A word on the words ‘denote’ and ‘predicate’ from Quine’s “From Stimulus to Science”:
“For what follows we must come squarely to terms with ‘denote’. Since it is often used interchangeably with ‘designate’, and a singular term normally designates one and only one object, readers are apt to think of denotation as relating a predicate likewise to a single object, namely the class of all thoe things it is true of, or a property shared by them. In my use of ‘denote’, as in John Stuart Mill’s [Mill, ChapterII $5], a predicate denotes rather each separate thing of which it is true. The class or property is not involved.
For years, to obviate confusion, I avoided ‘denote’ altogether in favor of ‘true of’; but that evasion woudl be impracticable in these pages where denotation is becoming the center of action. Unlike Mill, I still withhold the word from singular terms; they are well served by ‘designation’.
A word of caution is in order regarding ‘predicate’ too. Some logicians take a predicate as a way of building a sentence aroudn a singular term or, more concretely, as what Peirce called a rheme [Volume 2, paragraph 95], a sentence with blanks in it, these being distinctively marked in the case of a many-place predicate. This version covers, implicitly, the potential output of predicate abstraction or predicate functors. But a predicate in my sense is always an integral word, phrase, or claus, grammatically a noun, adjective, or verb. Some are generated from others by grammatical constructions, notably the relative clause or, formally, predicate abstraction and predicate functors.”[pg. 60-61 Quine “From Stimulus to Science”]
One of the virtues of this method is that less is spoken of than ever before. This has the benefit of eliminating explanations that are ultimately offshoots of predicate logic itself. In the past, and perhaps even in these passages, there are parts of the world which intruded upon logic and muddied its methods. The most famous of which is mathematics. There is much to be recommended from mathematical methods in matheamtical logic, but there is less to say about its ultimate contributions to logic itself. For example, formalism continues to confuse those unable to seperate logic from mathematical logic as if the formalisms were the subject of logic itself. One wonders by what methods such formalists make conclusions about their formalisms.
Logic no more has formalisms as its subject matter than it has anything in particular as its items of interest. Some, like Frege, have mistaken the schematics which fascilitate the articulation of logical principles and practices as part of logic itself rather than as a technical tool for fascilitating an exacting analysis of the same.
Specifically, Frege mistakes the mathematical metaphor of functions for functional predicates (and he never actually deals with the functional predicates relevant to his analysis of logic e.g. the predicate ‘{xy: some item is {uv: x is the truth value of u, y is the truth value of v, and u negates v}}’ is confused with its purported extension).
Other times, like Frege, logic was so far removed from schematic methods that it bled into thoughts, judgements, senses, meanings, and other possessions of the minds which haunt human bodies and perhaps even that amazing organ we call the brain.
I submit that all these mistakes are not mistakes: they are each a piece in a larger puzzle which has yet to be pictured as a complete whole. That practice which has done the most to put together all these pieces as if they fit into a larger whole is science. It is only recently that the science of philosophy was so far divorced from the other sciences. The philosophical revolt against psychology as a science of mind is justified in that it was easy to critique and its conclusions were in stark contrast with those of our stronger sciences. Cognative scientists continued to pluck what they could from philosophy and any failures they were confronted with were attributed to weaknesses not in their own science but in some philosophical or logical method: the philosophers were punished for the past transgressions against psychology as the science of mind.
The science of behavior and the philosophy of radical behaviorism mends many such mistakes by revealing not only a sharper picture of what science has assembled but also an inclusive explanation of how such mistakes were so easy to make:
“Mentalism kept attention away from the external antecedent events which might have explained behavior, by seeming to supply an alternative explanation. Methodological behaviorism did just the reverse: by dealing exclusively with external antecedent events it turned attention away from self-observation and self-knowledge. Radical behaviorism restores some kind of balance. It does not insist upon truth by agreement and can therefore consider events taking place in the private world within the skin. It does not call these events unobservable, and it does not dismiss them as subjective. It simply questions the nature of the object observed and the reliability of those observations.
The position can be stated as follows: what is felt or introspectively observed is not some nonphysical world of consciousness, mind, or mental life but the observer’s own body. This does not mean, as I shall show later, that introspection is a kind of physiological research, nor does it mean (and this is the heart of the argument) that what are felt or introspectively observed are the causes of behavior. An organism behaves as it does because of its current structure, but most of this is out of reach of introspection. At the moment we must content ourselves, as the methodological behaviorist insists, with a person’s genetic and environmental histories. What are introspectively observed are certain collateral products of those histories.”[pg. 18-19 Skinner “About Behaviorism”]
Under this rubric truth quickly distinguishes itself from validity:
“Truth. The truth of a statement of fact is limited by the sources of the behavior of the speaker, the control exerted by the current setting, the effects of similar settings in the past, the effects upon the listener leading to precision or to exaggeration or falsification, adn so on. There is no way in which a verbal description of a setting can be absolutely true. A scientific law is derived from possibly many episodes of this sort, but it is similarly limitied by the repertoires of the scientists involved, The verbal community of the scientist maintains special sanctions in an effort to guarentee validity adn objectivity, but again, there can be no absolute. No deduction from a rule or law can therefore be absolutely true. Absolute truth can be found, if at all, only in rules derived from rules and here it is mere tautology.”[pg.150 Skinner “About Behaviorism”]
The word “tautology” descends through Wittgenstein’s failed attempt to reduced quantificational logic to a kind of never ending truth-functional logic (sometimes unhappily called “sentence logic”). The popularity of the word is unfortunate because it reinforces defective distinctions between clear cut definitions of validity which are very rarely mentioned and yet when mastered frequently simplify the methods of logic. They are described by Quine as follows
“It was remarked at the end of Chapter 7 that validity may be ascribed not only to truth-functional schemata but also, by extension, to the sentences whose forms those schemata depict; but that it is well then to add the qualifier ‘truth-functional’. COrrespondingly a sentence obtainable by substitution in a valid quantificational schema is quantificationally valid. Such a sentence is true, or true for all values of its free variables. But it may or may not be truth-functionally valid; its truth may depend solely on its truth functional structure, or it may depend partially on how the quantifiers are arranged.
We may also note an intermediate grade, monadic validity. A sentence is quantificationally valid if it can be got by substitution in a valid quantificational schema; it is monadically valid, more particularly, if it can be got by substitution in a valid quantificational schema which is monadic; and it is truth functionally valid if it can be got by substitution in a valid truth-functional schema.” [pg. 176 Quine “Methods of Logic 4th Edition”].
A warning for those who are tempted, like Frege or Russell, to occasionally take schemata as the items of logic: all this talk of substition into valid schema to obtain valid sentences can be skipped by resorting directly to sentences and their grammatical structure. The grammatical structure with which validity is concerned is specifically designed to avoid the miscarriage of conclusions.
“Sentences have the same grammatical structure when they are interconvertible by lexical substitutions. Our definition of logical truth, then, can also be put thus: *a logical truth is a truth that cannot be turned false by substituting for lexicon. When for its lexical elements we substitute any other strings belonging to the same grammatical categories.”[pg. 58 Quine “Philosophy of Logic Second Edition”]
There are two things of note here:
Whatever promise there is to this grammatical definition of quantificaitonal validity is lost when the grammatical part of variables is reflected upon as Quine does from having read Gilbert Harman’s 1971 review in Metaphilosophy. Quine’s corrective is to contemplate only substitution for predicates and not for variables which are otherwise part of the lexicon of logical languages (Harman suggests taking variables as belonging to atomic grammatical elements which are more like logical connectives than items of the langauges lexicon).
Such restricted substitution no longer leads from one sentence to an other of the same grammatical structure: they may differ in the variables which they contain and no substitution for predicates alone can change one variable into an other. There are further troubles that, when confronted, lead to a stronger definition of quantificational validity by grammatical method:
“Our definition of logical truth in terms of substitution of lexicaon confronts also another and graver difficulty: it appeals only to substitution of predicates for predicates, as against substitution of sentences for simple sentences. We saw early in this chapter how the sentential approach coudl issue in a stronger, narrower concept of logical truth, screening out some cases that woudl have slipped through if only predicate substitution had been called for. IT can even be shown that the version in terms of predicate substitution will inevitably be too weak, as long as our stock of predicates is finite [This was pointed out to me by Mark L. Wilson]. The natural remedy, then is to exploit the indefiniteness of our category of predicates: to admit substitution not only of the predicates in some imagined list, but also of any predicates that might be added. Thus adjusted, then, the abstract version rusn as follows: a logical truth is a sentence that cannot be turned false by substitution for lexicon, even under supplementation of lexical resources.”[pg. 59 Quine “Philosophy of Logic 2nd Edition”]
Here then is where predicate functor logic reigns supreme and reveals the intimate relationship between validity (or what Quine and so many others speak of as “logical truth”), grammar, and truth. Since the only items of the lexicon of a predicate functor logic are predicates then there is no question of how substitution is to covey one sentence to an other with the same grammatical structure. It is also clear that supplementation of lexicon is required from contemplation of a lexicon of only one place predicates as Quine did in “Word and Object”:
“But may we not still aspire to the discovery of some fundamental lexicon on the basis of which all traits and states of everything could in principle be formulated? No; we can prove that opennes is unavoidable, as long anyway as the predicates of a theory are included as items of the theory.
For, let S1, S2, … be the sentences in the notation of a theory that have ‘x’ as their sole free variable. Each of them is, for each object of the universe as value of ‘x’, true or false of that object; hence each of them, being also an object of the universe, is true or false of itself. We easily show that no general term definiable in the theory is true of exactly those of S1, S2, … that are false of themselves. (For, if ‘F’ were such a term, then ‘Fx’ would be true of itself if and only if false of itself. [This argument is in principle Cantor’s. The form I have given it is reminiscent also fo Grelling’s paradox, and the use made of it is reminiscent of Tarski.]) Such a term can be added, irreducibly supplementing the theory.”[pg. 231-232 Quine “Word and Object”]
Note, I have changed a few words in the above quote to more closely bring what Quine said to bear on terminology used here. It falls well within Quine’s principles of paraphrasing.
This argument puzzled me at first, and as a result of the following analysis I was able to produce an even more surprising argument that directly reveals the immediate limitations on any consistent theory of denotation. It also brings together the principles of prediate abstraction, predicate concretion, disquotation and (to my surprise) the deduction theorem while also just giving the exact predicate that is missing.
Under the assumption that among the items of the theory are the predicates themselves it is possible, following Quine’s generalization of Tarski’s analysis of truth to denotation (which I shall state), to define a predicate which, in anticipation of the definition to follow, shall be written out as ‘degree 0 denotes’. It is defined just as Quine specifies on pg.63 of “From Stimulus to Science” but with predicate functors as predicate abstracts in a quantificational logic as described in An Incomplete Sketch of My Philosophy of Logic):
A detailed explanation of the construction of predicates which permit the manipulation of quotations as above can be found in the last chapter of Quine’s “Mathematical Logic” on his theory of protosyntax. Feferman’s Finitary Inductively Presented Logics are perhaps a simpler alternative but the presentation does not address this construction in particular.
Now, a predicate coextensive with ‘degree 0 denotes’ as defined above (rather than taking it as an abbreviation as has been done with the phrase ‘for’ these are changed to sentences of coextension) can not belong to such a theory i.e. such a predicate must be added as a primtive to “irreducibly supplement the theory” (the case where it is already among the primitives is covered after this). For, assume it does, then so does the predicate ‘{x: not (x degree 0 denotes x)}’.
Suppose, ‘{x: not (x degree 0 denotes x)}’ degree 0 denotes ‘{x: not (x degree 0 denotes x)}’. By disquotation (which is the defining principle given above), {x: not (x degree 0 denotes x)}’{x: not (x degree 0 denotes x)}’. By concretion, it is not the case that ‘{x : not (x degree 0 denotes x)}’ degree 0 denotes ‘{x: not (x degree 0 denotes x)}’. Hence, by the deduction theorem, ‘{x: not (x degree 0 denotes x)}’ degree 0 denotes ‘{x: not (x degree 0 denotes x)}’ only if it is not the case that ‘{x : not (x degree 0 denotes x)}’ degree 0 denotes ‘{x: not (x degree 0 denotes x)}’.
Now, suppose it is not the case that ‘{x : not (x degree 0 denotes x)}’ degree 0 denotes ‘{x: not (x degree 0 denotes x)}’. By disquotation, it is not the case that {x: not (x degree 0 denotes x}’{x: not (x degree 0 denotes x)}’. By concretion, it is not the case that it is not the case that (yes, the same phrase is repeated twice there) ‘{x : not (x degree 0 denotes x)}’ degree 0 denotes ‘{x: not (x degree 0 denotes x)}’ i.e. ‘{x : not (x degree 0 denotes x)}’ degree 0 denotes ‘{x: not (x degree 0 denotes x)}’. Hence, by the deduction theorem, it is not the case that ‘{x : not (x degree 0 denotes x)}’ degree 0 denotes ‘{x: not (x degree 0 denotes x)}’ only if ‘{x : not (x degree 0 denotes x)}’ degree 0 denotes ‘{x: not (x degree 0 denotes x)}’ i.e. ‘{x : not (x degree 0 denotes x)}’ degree 0 denotes ‘{x: not (x degree 0 denotes x)}’ if it is not the case that ‘{x : not (x degree 0 denotes x)}’ degree 0 denotes ‘{x: not (x degree 0 denotes x)}’.
Therefore, ‘{x: not (x degree 0 denotes x)}’ degree 0 denotes ‘{x: not (x degree 0 denotes x)}’ if and only if it is not the case that ‘{x : not (x degree 0 denotes x)}’ degree 0 denotes ‘{x: not (x degree 0 denotes x)}’, a contradiction.
Though that mays seem like a mouthfull of nit picking steps, it gives each principle that, step by step, permits the given contradiction.
The above contradiction followed by assuming that ‘degree 0 denotes’ was not a part of the lexicon of the theory but was coextensive with some compound constructed from the lexicon. What’s interesting is that there is nothing wrong with supposing ‘degree 0 denotes’ is a basic predicate, now added to the theory, which is defined as above (that is each instance of the corresponding coextensive schema) with respect to each OTHER basic predicate except for itself! Read that very carefully because it is very slipery.
So, in that case, the case where you have ‘degree 0 denotes’ as stipulated, we are left again where Quine’s argument starts, but now, we contemplate a new predicate ‘degree 1 denotes’ which is defined from the basic predicates, one of which is ‘degree 0 denotes’, and which, by the same argument, can not be coextensive with any compound of the basic predicates, one of which is ‘degree 0 denotes’. Thus, ‘degree 1 denotes’ can be added as a basic predicate defined from the other basic predicates by the inductive definition above.
Now, for any theory, there is a finite lexicon, and thus there is some smallest ‘n’ for which ‘degree n denotes’ does not belong to the lexicon i.e. there is no end to how a theory might be supplemented.
Quine’s argument involving one place predicate glosses over the problem of denotation, as Quine frequently did, with the phrase “is true of”, but it is precisely on the analysis of this predicate that Quine’s argument (and Frege’s work, for ‘is true’ is a degenerate case of ‘is true of’ i.e. ‘denotes’) hinges.
When I started this not I had an aim of typing up the rest of my sketch of my paper on logic i.e. the concrete consequences of arguments like those above. That did not happen because now I’m going to switch to work on my little lisp before I forget what occurred to me with respect to it earlier today.
This continues my work on my little lisp from 2025 0501 1754.
I made great progress in the last entry and between the end of the last note on this project and now I was able to solve the problem of going back and forth between lisp and forth. I wish there was an easy way to write while sleepy or tired or whatever happened last night while I failed to put together the pieces as in the simple example that occurred to me while I was watching a show called “The Sticky”. It is common for me to work until exhaustion. Sadly, the exhaustion which comes from reading, writing, and thinking does not look like it does when it is from lifting. You are not a sweaty mess nor are your muscles fatigued in that way that makes it hard to walk up or down steps or get out of a chair (which, even writing about them brings that feeling too me).
My lingering on this problem of exhaustion is important: it is probably the biggest problem to solve when you must solve so many problems before you die. There is no clear indication that I shall do all that I must before I die and the bottleneck to most of what I must do is largely a matter of dealing with the pains of weak thinking. Perhaps the biggest part of this big problem is encapsulated in one of my most important principles: “Stop at the sign.”
It can be hard to notice when thinking has gone weak. This seems to be the result of prescientific practices around thinking, reading, and writing generally. Rather than seeing a schedule of reinforcement that maintains the behaviors that have led to the most profound of scientific and logically discoveries, prescientific peoples see the individual as a willing agent or some other such mentalistic haunties of the brain by the mind. But, if I waste more time now solving this problem then I will not get out the solutions to problems that I have already solved. It is difficult to say both that “here we have the biggest problem and the biggest bottleneck to solving problems” and then “but now I will leave that to work on lesser problems.” It is precisely in being able to say both these things that schedules of reinforcement resolve what appears to be a contradiction.
Here is an exact replica of what occurred to me while watching “The Sticky” earlier today and which leads to the radical simplification of any reader for any lisp like language while also explaining exactly how it is that lisp is a forth:
((a)(bc))
0 | (a)(bc))
0,0 | a)(bc))
0,(a,0) | )(bc))
((a,0),0) | (bc))
((a,0),0),0 | bc))
((a,0),0),(b,0) | c))
((a,0),0),(b,(c,0)) | ))
((a,0),((b,(c,0)),0)) | )
(((a,0),((b,(c,0)),0)),0) |
Believe it or not, this shows that my sleepy claim that a simple alphabetization radically simplifies the reader of any language that purports to be some sort of parenthetical or tree like form, but which can be immediately interpreted as a string of stack operations. This exactly generalizes the methods from Bit Strings and Binary Trees.
Recall that the reader was previously composed of steps involving lexing and parsing: this new method goes straight from a list of runes to the list they purport to designate. All of that complexity is eliminated by a change alphabatization (I’ve gone ahead and reproduced all the code for the current state of this project).
Ooops! Once I opened the tab to bring the code over here I started coding and totally forgot about writing out what I was doing here.
Here’s where I’m starting from:
let run=code=>console.log(code,'\n',eval(code)) // for examples
// basic operations on pairs
let run=code=>console.log(code,'\n',eval(code)) // for examples
// basic operations on pairs
let theEmptyPair={}
, isEmpty = x => x == theEmptyPair
, consOf = (x,y) => [x,y]
, carOf = x => isEmpty(x) ? x : x[0]
, cdrOf = x => isEmpty(x) ? x : x[1]
// left and right singletons
, singletonListOf = x => consOf(x,theEmptyPair)
, singletonStackOf = x => consOf(theEmptyPair,x)
// to and from tallies
, tallyOf = n =>
n>0 ? singletonStackOf(tallyOf(n-1)) : theEmptyPair
, countOf = x => isEmpty(x) ? 0 : 1 + countOf(cdrOf(x))
// the alphabet
, abc =[...'()0123456789abcdefghijklmnopqrstuvwxyz ']
, alphabeticalIndexOf = letter => abc.indexOf(letter)
, alphabeticalLetterOf = index => abc[index]
// to and from rune
, runeOf = letter => tallyOf(alphabeticalIndexOf(letter))
, letterOf = x => alphabeticalLetterOf(countOf(x))
// basic string operations
, theEmptyLetter=''
, isEmptyLetter = x => x == theEmptyLetter
, concatenationOf = (...strings) =>
strings.length ? strings.shift() + concatenationOf(...strings) : theEmptyLetter
, firstLetterOf = letters => isEmptyLetter(letters) ? theEmptyLetter : letters[0]
, restLettersOf = letters => isEmptyLetter(letters) ? theEmptyLetter : letters.slice(1)
// to and from runes
, runesOf = letters =>
letters.length ? consOf(runeOf(firstLetterOf(letters))
, runesOf(restLettersOf(letters)))
: theEmptyPair
, lettersOf = x =>
isEmpty(x) ? theEmptyLetter
: concatenationOf(letterOf(carOf(x)),lettersOf(cdrOf(x)))
// prepending pairs as lists
, prependedListOf = (x,y) =>
isEmpty(x) ? y : consOf(carOf(x), prependedListOf(cdrOf(x),y))
// pairs as stacks
, theEmptyStack = theEmptyPair
, isEmptyStack = isEmpty
, pushOf = consOf
, popOf = carOf
, topOf = cdrOf
, secondOf = stack => topOf(popOf(stack))
, drop2 = stack => popOf(popOf(stack))
, enpendOf = stack =>
pushOf(drop2(stack)
,prependedListOf(secondOf(stack)
,singletonListOf(topOf(stack))))
// recursive definition of identity
, id = (x,y) =>
(isEmpty(x) && isEmpty(y))
|| (!(isEmpty(x) || isEmpty(y))
&& id(carOf(x),carOf(y))
&& id(cdrOf(x),cdrOf(y)))
// reader
, isOpenParen = x => id(runeOf('('),x)
, isCloseParen = x => id(runeOf(')'),x)
, isSpace = x => id(runeOf(' '),x)
, readHelperOf = (stack,runes) =>
isEmpty(runes) ? carOf(topOf(stack))
: isCloseParen(carOf(runes)) ? readHelperOf(enpendOf(stack), cdrOf(runes))
: isOpenParen(carOf(runes)) ? readHelperOf(pushOf(stack,theEmptyPair), cdrOf(runes))
: isSpace(carOf(runes)) ? readHelperOf(stack,cdrOf(runes))
: readHelperOf(enpendOf(pushOf(stack,carOf(runes))), cdrOf(runes))
, readOf = letters => readHelperOf(theEmptyStack, runesOf(letters))
// printer
, parentheticalOf = x => isEmpty(x) ? runesOf('()') : prependedListOf(runesOf('('),prependedListOf(x,runesOf(')')))
, printListOf = list =>
isEmpty(list) ? theEmptyPair
: prependedListOf( printHelperOf(carOf(list)), printListOf(cdrOf(list)))
, printHelperOf = item =>
isEmpty(item) ? runesOf('()')
: isEmpty(carOf(item))? parentheticalOf(prependedListOf(runesOf('()'),cdrOf(item)))
: parentheticalOf(printListOf(item))
, printOf = item => lettersOf(printHelperOf(item));
Everything is simpler than ever, but it is no longer clear that the specific alphabetization is as important as I thought, but there is still some opportunity.
I did so much that it is hard to go back and write through it. Since there is still more that I must do I’m going to just start writing from here and hopefully get back to where I started.
It looks like the reader and the printer are still more complex than they need to be and that they can be simplified if I can figure out the appropriate stack operations. As it stands the reader interprets each rune as a stack operation:
Enpending is something that happens often but which doesn’t often get its own name: it’s when you enlist an item, i.e. put it into a singleton list, and then append it to some other list of items. There are distinctions for operations like this in the J programming language, but that is for another time.
Well this is one heck of a fragmentary note.
Surprisingly, I was able to entirely defer the introduction of the identity function and finally push off all the work that goes into external reading and printing so that everything comes together quickly. Here it does end up being helpful to have a specific alphabetization: it is key to the simplicity of the internal reader and printer. This is all quite exciting and there is so much more to come.
// basic operations on pairs
let theEmptyPair={}
, isEmpty = x => x == theEmptyPair
, consOf = (x,y) => [x,y]
, carOf = x => isEmpty(x) ? x : x[0]
, cdrOf = x => isEmpty(x) ? x : x[1]
// left and right singletons
, singletonListOf = x => consOf(x,theEmptyPair)
, singletonStackOf = x => consOf(theEmptyPair,x)
// prepending pairs as lists
, prependOf = (x,y) =>
isEmpty(x) ? y : consOf(carOf(x), prependOf(cdrOf(x),y))
// pairs as stacks
, theEmptyStack = theEmptyPair
, isEmptyStack = isEmpty
, pushOf = consOf
, popOf = carOf
, topOf = cdrOf
, secondOf = stack => topOf(popOf(stack))
, drop2 = stack => popOf(popOf(stack))
, enpendOf = stack =>
pushOf(drop2(stack)
,prependOf(secondOf(stack)
,singletonListOf(topOf(stack))))
// reader
, isOpenParen = isEmpty
, isCloseParen = x => !isEmpty(x)&&isEmpty(carOf(x))&&isEmpty(cdrOf(x))
, isSpace = x => !isEmpty(x)&&isEmpty(carOf(x))&&isCloseParen(cdrOf(x))
, readHelperOf = (stack,runes) => isEmpty(runes) ? carOf(topOf(stack))
: isSpace(carOf(runes)) ? readHelperOf(stack,cdrOf(runes))
: isCloseParen(carOf(runes)) ? readHelperOf(enpendOf(stack), cdrOf(runes))
: isOpenParen(carOf(runes)) ? readHelperOf(pushOf(stack,theEmptyPair), cdrOf(runes))
: readHelperOf(enpendOf(pushOf(stack,carOf(runes))), cdrOf(runes))
, readOf = runes => readHelperOf(theEmptyStack, runes)
// printer
, theOpenParen = theEmptyPair
, theCloseParen = consOf(theEmptyPair,theEmptyPair)
, isSymbol = item => !isEmpty(item)&&isEmpty(carOf(item))
, theSymbolPair = theEmptyPair
, parenOf = runes => consOf(theOpenParen, prependOf(runes, singletonListOf(theCloseParen)))
, printListOf = list => isEmpty(list) ? theEmptyPair
: prependOf(printOf(carOf(list)),printListOf(cdrOf(list)))
, printOf = item => isSymbol(item)? parenOf(prependOf(printOf(theSymbolPair),cdrOf(item)))
: parenOf(printListOf(item))
// the alphabet
, abc =[...'() 0123456789abcdefghijklmnopqrstuvwxyz']
, alphabeticalIndexOf = letter => abc.indexOf(letter)
, alphabeticalLetterOf = index => abc[index]
// to and from tallies
, tallyOf = n =>
n>0 ? singletonStackOf(tallyOf(n-1)) : theEmptyPair
, countOf = x => isEmpty(x) ? 0 : 1 + countOf(cdrOf(x))
// to and from rune
, runeOf = letter => tallyOf(alphabeticalIndexOf(letter))
, letterOf = x => alphabeticalLetterOf(countOf(x))
// basic javascript string operations
, theEmptyLetter=''
, isEmptyLetter = x => x == theEmptyLetter
, concatenationOf = (...strings) =>
strings.length ? strings.shift() + concatenationOf(...strings) : theEmptyLetter
, firstLetterOf = letters => isEmptyLetter(letters) ? theEmptyLetter : letters[0]
, restLettersOf = letters => isEmptyLetter(letters) ? theEmptyLetter : letters.slice(1)
// to and from runes
, runesOf = letters =>
letters.length ? consOf(runeOf(firstLetterOf(letters))
, runesOf(restLettersOf(letters)))
: theEmptyPair
, lettersOf = x =>
isEmpty(x) ? theEmptyLetter
: concatenationOf(letterOf(carOf(x)), lettersOf(cdrOf(x)))
// external read and print
, read = letters => readOf(runesOf(letters))
, print = item => lettersOf(printOf(item))
// recursive definition of identity
, id = (x,y) =>
(isEmpty(x) && isEmpty(y))
|| (!(isEmpty(x) || isEmpty(y))
&& id(carOf(x),carOf(y))
&& id(cdrOf(x),cdrOf(y)));
For reasons that escape me I’ve had a few different youtube playlists queued up in the background the past few days. The first, and perhaps the most important is Groucho Marx’s “You Bet Your Life”. There is nothing like Groucho. He is certinaly one of my idols in that I mention him as often to those near me as I do Skinner and Quine to everyone else.
I don’t know how accurate the title is. Next, of all things, Marathon lore:
Last, and most certainly least, a playthrough of an old game I recall from my childhood:
Such things are part of modern folklore, though they very often lack the warmth that I associate with well worn folklore. Perhaps such stories from games are nothing more than the folklore of my life.
Sometimes I need silence in order to solve problems. Other times I need there to be something going on in the background. I need to at least see that someone else at some other time was making progress on something they set out to do even if I might not be at this exact moment.
This continues the work on my little lisp from 2025 0426 1851.
The basic definitions are now
let run=code=>console.log(code,'\n',eval(code)) // for examples
// basic operations on pairs
let theEmptyPair={}
, isEmpty = x => x == theEmptyPair
, consOf = (x,y) => [x,y]
, carOf = x => isEmpty(x) ? x : x[0]
, cdrOf = x => isEmpty(x) ? x : x[1]
// recursive definition of identity
, id = (x,y) =>
(isEmpty(x) && isEmpty(y))
|| (!(isEmpty(x) || isEmpty(y))
&& id(carOf(x),carOf(y))
&& id(cdrOf(x),cdrOf(y)))
// left and right singletons
, singletonListOf = x => consOf(x,theEmptyPair)
, singletonStackOf = x => consOf(theEmptyPair,x)
// to and from tallies
, tallyOf = n =>
n>0 ? singletonStackOf(tallyOf(n-1)) : theEmptyPair
, countOf = x => isEmpty(x) ? 0 : 1 + countOf(cdrOf(x))
// the alphabet
, abc =[...'0123456789abcdefghijklmnopqrstuvwxyz ()']
, alphabeticalIndexOf = letter => abc.indexOf(letter)
, alphabeticalLetterOf = index => abc[index]
// basic string operations
, theEmptyLetter=''
, isEmptyLetter = x => x == theEmptyLetter
, concatenationOf = (...strings) =>
strings.length ? strings.shift() + concatenationOf(...strings) : theEmptyLetter
, firstLetterOf = letters => isEmptyLetter(letters) ? theEmptyLetter : letters[0]
, restLettersOf = letters => isEmptyLetter(letters) ? theEmptyLetter : letters.slice(1)
// to and from rune
, runeOf = letter => tallyOf(alphabeticalIndexOf(letter))
, letterOf = x => alphabeticalLetterOf(countOf(x))
// to and from runes
, runesOf = letters =>
letters.length ? consOf(runeOf(firstLetterOf(letters))
, runesOf(restLettersOf(letters)))
: theEmptyPair
, lettersOf = x =>
isEmpty(x) ? theEmptyLetter
: concatenationOf(letterOf(carOf(x)),lettersOf(cdrOf(x)))
// prepending and appending pairs as lists
, prependedListOf = (x,y) =>
isEmpty(x) ? y : consOf(carOf(x), prependedListOf(cdrOf(x),y))
, appendedListOf = (x,y) => prependedListOf(y,x);
Symbols, which I once took as native javascript strings, are now to be pairs whose left part is the empty pair:
let isSymbol = x => isEmpty(carOf(x));
Then, as a matter merely of convention, the right part of a symbol is taken as runes which spell out the name of the symbol. Recall, way back in the first entry on this line of work, I mentioned that a candidate way of defining atoms was by making them identical to their left part. After that I asked what of the right part? To which I answered: this might help us to more easily, as humans, differentiate between symbols. But, note, here, symbols are not atoms in this sense, they are just a special case of ordered pair.
This method of distinguishing symbols from nonsymbols is itself a convention imposed upon the particular designs of my pure little lisp: an other may come along and devise a method which behaves nothing like the symbols of LISPs past. Such an eample is not as far off as it might seem to those wedded to LISP e.g. I’ve already mentioned in my list of links on concatenative languages that there is a calculus of trees which purports to do more with less from a foundation of ordered pairs as that contemplated here by me. That, like so many other things that I mention among these notes, is for another time.
The introduction of symbols as special pairs interferes in one place with the broader convention of taking pairs as lists. What of the difference between a pair playing the part of a list whose first item (its left part) is itself the empty list? It is by answering this question that we find the first use of symbols: the symbol symbol.
let theSymbolSymbol = consOf(theEmptyPair, theEmptyPair);
But, “the symbol symbol” is a mouthfull and there is a name I am more fond of which is short and which came to me originally through my study of the J programming language:
let ace = theSymbolSymbol;
A more incremental presentation of the above and one that rids the design of its commitment to a particular pair as the symbol pair is as follows.
First, add a letter for the symbol pair to the alphabet. Here I pick ‘$’ as that letter because it resembles the letter ‘S’ in ‘Symbol’:
abc =[...'0123456789abcdefghijklmnopqrstuvwxyz ($)'];
Then the symbol rune is the symbol pair
let theSymbolRune = runeOf('$');
or just
let $ = theSymbolRune;
so that the symbol symbol, ‘ace’ for short, is
ace = consOf($,$);
But, this too breaks the convention that the right part of a symbol is to be taken as spelling out the name of a particular occurrence of the symbol. This can be remedied directly with
ace = consOf($,singletonListOf($));
This is unsatisfactory for many reasons, the least of which is that it makes it hard to see the convention being enforced (i.e. that ‘ace’ designates the symbol symbol and that the name of the symbol symbol is designated by “runesOf(‘$’)”. So it is that there is one last alternation that can be made now and which not only satisfies this convention but eases reading and printing: it is taking the symbol pair as the runes of the item designated by “’$’” (and yes the double quotes enclosing the single quoted occurrence of ‘$’ is needed).
let theSymbolPair = runesOf('$');
isSymbol = pair => id(pair, theSymbolPair);
theSymbolSymbol = consOf(theSymbolPair, theSymbolPair);
$ = theSymbolPair;
ace = theSymbolSymbol;
There is some temptation to designate the symbol symbol by ‘$$’ but I’ll avoid that for now.
Surprisingly, the way in which this convention aids in reading and printing imposes its own convention on the grammar of programs in this little lisp of mine. It also happens to be the grammar that helps to bring LISP and FORTH closer together: they are already extremely close because LISP is a FORTH that is ashamed of its stacks.
The convention is this: programs are space seperated spellings of symbols. Said another way: the names of the symbols of a program are spelled out by whatever is seperated by spaces. There is more, because we have said that the symbol symbol is marked as a symbol by having the symbol pair as its left part, we have introduced the convention of marking the part of speech played by a list with its left part. It shall be shown that the distinctino between symbols and lists is enough to mark out any grammatical distinctions that may later be used to the benefit of the programmer.
But, this is not quite right either because it is the job of the reader to build lists and symbols. Thus, the program grammar is actually the result of further contemplation on what the reader must do and how it must do it.
Since this is a LISP and not a FORTH, parenthesis play a special part in the reader from the beginning. The reader can be taken as a tiny interpreter/compiler (there is no distinction really between these things other than how we speak of them when we speak of them asperationally). It reads a list of runes and builds a symbol or a list of symbols or a list each item of which is either a symbol or a list of symbols, and so on, from it.
The commands that go to the reader are the ones that build the pairs upon which the lisp works, whether it directly evaluates them or alters them by way of some grammatical rule of transformation (commonly known as a macro). As said before, the commands are seperated by spaces, so there is actually another even tinier interpreter/compiler that sits between a list of runes and the reader (that is when you discard the native javascript transformation from a javascript string into a list of runes!)
This tinier interpreter/compiler is called a tokenizer or a lexer. But, this is only appropriate when it is spoken of in relation to some other language. Otherwise, it is just another programming language that takes a pair as a list of runes and builds a list of lists of runes from it! The name “lexer” is the most suggestive of what is to be done: it is to break a complex compound into its lexical atoms. The lexicon of a language is what you get if you grab that language by its grammatical trees and shake it until all its words fall out.
The alphabet of a language is what you get when you grab its lexemes, i.e. items of its lexicon, and shake them until its letters fall out. It is from thes letters, or runes, that the lexer builds lexemes. So, to recap, the lexer builds lexemes, and the rest of the reader builds symbols, lists of symbols, and so on from the list of lexemes made by the lexer.
So, on to the lexer! Our convention of spelling lexemes by seperating their spellings with spaces divides the lexer into two operations: skipping spaces and making words from their spellings. First, a method of detecting spaces:
let theSpaceRune = runeOf(' ')
, isSpace = x => id(x, theSpaceRune);
Next a way of taking a list of runes and trimming off any spaces from the front.
let trimSpacesOf = x => isEmpty(x) ? x : isSpace(carOf(x)) ? trimSpacesOf(cdrOf(x)) : x;
This is the easy step. The hard step is making words from their spellings and putting them all together as a list of runes. There are two paths that are before us: one where stacks are formally introduced or one where continuations are formally introduced. The simpler path is stacks, but they are both the same thing (continuations are stacks in disguise, and this part of the “lisp is a forth ashamed of its stacks” thing I said earlier). Also, stacks are the best, better even than lists (which are just the mirror of stacks, or stacks in reverse, which is itself foreshadowing).
The empty pair is taken as the empty stack, pushing an item onto a stack is making an ordered pair whose left part is the old stack and whose right part is the new top of stack, popping the empty stack gives the empty stack and popping a nonempty stack gives its left part, and, finally, peeking, or looking at the top of a stack returns its right part or the empty pair if you peek the empty pair.
let theEmptyStack = theEmptyPair
, isEmptyStack = isEmpty
, pushOf = consOf
, popOf = carOf
, topOf = cdrOf;
Notice, these stack operations are all just different names for the same old operations on pairs that were introduced in the beginning. Stacks are simply a different perspective on pairs, lists are another, and there are others still. Ordered pairs are probably the most important abstract objets that humans have introduced into the world.
The plan for the lexer is to make a function that takes three arguments:
Thus the lexer will just be a way of setting up the helper function with an empty stack, an empty list, and the list of runes to be lexed. The first step is to trim off any spaces that happen to be at the front of the list. So, the helper can assume that the list of runes to be lexed is already trimmed: this must now be enforced by the design of the helper itself. One option is to make a helper for the helper, one that does nothing more than take the same three arguments as the helper and simply trim the third and then pass them onto the helper. This seems excessive and can be accomplished by just making sure to trim any initial spaces from the third argument once a space is confronted. At the same time, confronting a space is a command that the current lexeme under construction is no longer under construction and can be pushed onto the stack of lexemes lexed thus far. All that remains is to take the first rune of the list to be lexed, check if it is a space or not, if it is not then append it to the lexeme under construction, and if it is then push the constructed lexeme onto the stack of lexemes lexed and trim the list of runes to be lexed before passing everything onto the helper again.
Oh! It just occurred to me that I do not even need to worry about all that stack stuff! I can just append lexemes to the list of lexemes because I already gave the definitions for append and prepending. So, forget all that stack stuff for now.
let lexerHelperOf = (lexemes, lexeme, runes) =>
isEmpty(runes) ? (
isEmtpy(lexeme) ? lexemes
: prependedListOf(lexemes,singletonListOf(lexeme)))
: isSpace(carOf(runes)) ?
lexerHelperOf(prependedListOf(lexemes,singletonListOf(lexeme))
,theEmptyPair,trimSpacesOf(runes))
: lexerHelperOf(lexemes
,prependedListOf(lexeme,singletonListOf(carOf(runes))),cdrOf(runes))
, lexOf = runes =>
lexerHelperOf(theEmptyPair,theEmptyPair,trimSpacesOf(runes));
The rest of the reader is usually called the parser for very good reason:
“parse(v.): 1550s, in grammar, “to state the part of speech of a word or the words in a sentence,” a verbal use of Middle English pars (n.) “part of speech” (c. 1300), from Old French pars, plural of part “a part,” from Latin pars “a part, piece” (from PIE root *pere- (2) “to grant, allot”) in the school question, Quae pars orationis? “What part of speech?” Transferred (non-grammatical) use is by 1788. Pars also was a common plural of part (n.) in early Middle English. Related: Parsed; parsing.” https://www.etymonline.com/word/parse
From the list of lexemes the parser deduces the part of speach to which each lexeme belongs and, from doing so, constructs the grammatical tree from which they fell. There are only two parts of speech for the items of my little LISP: symbol or list. So the parser either builds lists or symbols based on the lexemes it is fed, one by one. An open parenthesis starts the construction of a list, a closing parenthesis ends the construction of a list under construction, and everything else is a symbol. The conventions established so far make it so that the any pair whose left part is not the symbol pair is a list. This can be made overt rather than covert by introducing a general convention that the way of dealing with the right part of a pair is to be determined from its left part.
This is a general method e.g. FORTH’s dictionaries are designed precisely in this way so that native code on the left tells the machine how to deal with the stuff on the right. Since it appears as if symbols can be used to defer the introduction of this general method (so that the programmer of my little lisp can set such grammatical concerns up as they see fit), I shall keep with the implicit way of taking pairs as lists.
So, there are two things the parser must do: build symbols and build lists. Parsing a symbol is easy: just make a pair whose left part is the symbol pair and whose right part is the lexeme which names the symbol being parsed.
let parseSymbolOf = lexeme => consOf($,lexeme);
Note that this automatically produces the symbol symbol when the lexeme is the symbol pair. This is an example of one of the conveniences which follows from the conventions evolved here.
Parsing a list is only slightly harder. Now, all that stuff about stacks can be unforgotten. When the first lexeme of the list of lexemes to be parsed is an open parenthesis then whatever we were parsing up to that point goes on the stack of things we have to go back to parsing.
It’s slightly easier to grasp this when you start parsing with an open parenthesis. This tells you to start with an empty list, i.e. the empty pair, and build up the rest of the list based on the remaining lexemes. Next suppose there is neither a closed parenthesis or an open parenthesis as the next lexeme. Then it must be a symbol. So you parse that symbol and then append it to the list under construction. This is all fine and good until you run into another open parenthesis. You need to start making a new list, but, when you’re done making it, you have to append what you’ve made onto the end of the last list you were working on and continue parsing with whatever lexemes are left over.
If you’re lucky, you have a helper that has an argument for the list you’re going to get back to constructing, an argument for the list currently under construction, and the rest of the lexemes you have to go through to finish everything up. But, since you can run into an unknown number of open parenthesis before reaching your first closed one, you can’t just keep hopping you have helpers for helpers. For me, stacks are the obvious answer. They occur to me so readily because I deal with the world all the time with stacks. There are stacks and stacks of things that I’m going to come back to after I finish working with the thing that is at the top of the stack.
Now, when you run into an open parenthesis you push whatever list is under construction onto the stack of lists under construction, and start with the empty list. If the next lexeme is neither an open or closed parenthesis you append its parsed symbol to the list under construction. If it’s a closed parenthesis then you are done constructing the current list and append it to the list atop the stack of lists under construction.
let isOpenParen = x => id(runesOf('('),x)
, isCloseParen = x => id(runesOf(')'),x)
, parseHelperOf = (stack, list, lexemes) =>
isEmpty(lexemes) ? list
: isOpenParen(carOf(lexemes)) ?
parseHelperOf(pushOf(stack,list),theEmptyPair,cdrOf(lexemes))
: isCloseParen(carOf(lexemes)) ?
parseHelperOf(popOf(stack)
, prependedListOf(topOf(stack),singletonListOf(list))
, cdrOf(lexemes))
: parseHelperOf(stack
, prependedListOf(list
, singletonListOf(parseSymbolOf(carOf(lexemes))))
, cdrOf(lexemes))
, parseOf = lexemes => parseHelperOf(theEmptyPair, theEmptyPair, lexemes);
There are a few things that seem as if they are not going to work out well based on the above design:
To answer that last question is to get to the heart of the matter I submit. So, rather than parsing symbols and such I’ll just take every lexeme that is not a parenthesis directly over to the list under construction:
parseHelperOf = (stack, list, lexemes) =>
isEmpty(lexemes) ? list
: isOpenParen(carOf(lexemes)) ?
parseHelperOf(pushOf(stack,list),theEmptyPair,cdrOf(lexemes))
: isCloseParen(carOf(lexemes)) ?
parseHelperOf(popOf(stack)
, prependedListOf(topOf(stack),singletonListOf(list))
, cdrOf(lexemes))
: parseHelperOf(stack
, prependedListOf(list
, singletonListOf(carOf(lexemes)))
, cdrOf(lexemes));
By simply dropping the step that parses symbols something important has been revealed: this is almost the same form of program as the primtiive encoder and decoder cooked up in Bit Strings and Binary Trees.
The following supplementary functions make it clearer:
let secondOf = stack => topOf(popOf(stack))
, drop2 = stack => popOf(popOf(stack))
, appendOf = stack => pushOf(drop2(stack)
, prependedListOf(secondOf(stack), topOf(stack)));
parseHelperOf = (stack, lexemes) =>
isEmpty(lexemes) ? topOf(stack)
: isOpenParen(carOf(lexemes)) ? parseHelperOf(pushOf(stack,theEmptyPair), cdrOf(lexemes))
: isCloseParen(carOf(lexemes)) ? parseHelperOf(appendOf(stack), cdrOf(lexemes))
: parseHelperOf(appendOf(pushOf(stack,carOf(lexemes))), cdrOf(lexemes));
parseOf = lexemes => parseHelperOf(theEmptyPair, lexemes);
Here, the open parenthesis plays the part of the zero of the encoder, the close parenthesis plays the part of the one of the encoder, and the default case appends the first of the lexemes onto the list at the top of the stack.
This suggests a certain ordering to the alphabet in order to make the similarities stronger along with a complete overhall of method. It seems like it is worth it and that it also brings the similarities between lisp and forth even closer together while simplifying going between one and the other (and without eliminating lisp as a language). It also suggests a pretty obvious alphabetization.
I was unable to see through the rest of what I had seen to my self. Meditation and sleep are the most likely solutions to this problem.
This continues my read of Durant’s “Story of Philosophy” from 2025 0420 2247.
Last time I cracked open the spine of this old book I had just read that “Critias was a pupil of Socrates, and an uncle of Plato”. Critias is worthy of his own story from what little I read of his history outside of what less Durant mentioned. He is also surprisingly relevant to the world we are presently faced with: the Sophists were selling aristocracy by bashing democracy and Critias ended up the very brief head of Athen’s Oligarchical Party. There is not enough written by Durant or that I have already read as part of my work on the history of the world to deduce to what extent “democracy” was anything that existed strongly enough to even be denounced e.g. what written records we have are not even enough to make too many commitments to the stories that happen to have survived to our day.
Archeology is a countercontrol to the division of the past into history and prehistory where prehistory is simply a weak stand in for “unrecorded” or “having no surviving records written in a language as the reinforcing practices of a verbal community”. It is quite clear that there is more to be got from simulating past environmental conditions and watching how humans navigate such contrived worlds in many cases than to take the word of a textual record that just so happens to have found its way to our modern world. The validity of such methods depends on the difference between the primary reinforcers of the past and those now. As far as I can tell they are still the same old same old: sex, food, water, violence. There may be a few others, but these are the primary reinforcers from which most of the organisms are likely to evolve or condition the much larger collection of secondary reinforcres that come to be called social because they involve an other organism (perhaps of the same species) in some way key to the conditioning contingency.
Alas, archeology is not what I am reading, though I would love to know more about the archeological work that has been done on Athens during the times of Socrates and Plato. That is for a different time.
The second section of the first part is titled “Socrates”. His bust bursts onto the scene as if a wart on his philosophic reputation: bald, round,
“it was rather the head of a porter than that of the most famous of philosophers.”[pg. 7]
There are various sculptures of Socrates purported from his time and their accuracy is probably more asperational than scientific e.g.
and, last but not least, a collection of them as part of an article on the big question “Who was Socrates?”
One bust or the other is a concrete connection with the ever ephemeral Socrates.
He spoke and listened in ways that differed remarkably from those of his time, and, presumably, of our time as well. Just as the young and curious of our day are apt to ask “why? why? why?”, whether as inquiries or objections, Socrates queried his fellow Athenians in search of something beyond question.
Perhaps this is all that was different about Socrates’ verbal behavior and perhaps this was all that was required to bring a cross section of young and old back to his arena of discourse. What is known of Socrates is second hand: no writing of his survives if ever it was made. We shall never have the records we might need to recreate what made his verbal behavior so remarkable that descriptions of it have made their way to us as something beyond myth and history.
The stories of Socrates are, for me, folklore. Most of what is called history is folklore in my measure. There’s little beyond the dutiful work of archeologists that allows me to speak with any certainty on the intersection of history and folklore. When taken as folklore it becomes easier to see that there are certain stories that are told with a character called Socrates and, just as knowing the stories of Shakespeare can alert you to fitting surprises, knowing the stories of Socrates can strengthen otherwise remote cultural connections.
I for one cherish Shakespeare far more than the stories of Socrates, but can not doubt the effect that Plato’s tales of Socrates have had on European philosophy, just as there is even less doubt of the effect Shakespeare has had on European literature.
Two delightful links (from a friend) on a ternary stack machine:
My first reading of Begriffsschrift was a while ago when I first got the invaluable text “From Frege To Godel” Third Printing edited by van Heijenoor. At that time I was mainly interested in von Neumann’s outlook on set theory and logic. But, it was already clear that Frege had made his Begriffsschrift and all else was commentary upon it.
R.P. and I have disagreed on this matter. Last Friday I received my copy of “The Frege Reader” edited by Michael Beaney and began a close and careful analysis of Frege’s writings without them being of secondary interest as they were when I was studying Quine’s “Set Theory and its Logic” and was led to von Neumann’s 1925 “An axiomatization of set theory”.
Not only do I have a new respect for Frege. I am more certain than ever that he shared my own outlook on his commentaries. Today, I finally started reading Frege’s 1891 “Function and Concept” and was delighted to see that he had this to say
“There can be no question of setting forth my Begriffsschrift in its entirety, but only of elucidating some fundamental ideas.”[pg. 131 Frege “The Frege Reader” edited by Michael Beaney]
So that the result of his comments upon Begriffsschrift can be taken merely as elucidation and nothing more. The etymology of the root is here from https://www.etymonline.com/word/*leuk-
“*leuk-. Proto-Indo-European root meaning “light, brightness.”
It might form all or part of: allumette; elucidate; illumination; illustration; lea; leukemia; leuko-; light (n.) “brightness, radiant energy;” lightning; limn; link (n.2) “torch of pitch, tow, etc.;” lucent; lucid; Lucifer; luciferase; luciferous; lucifugous; lucubrate; lucubration; luculent; lumen; Luminal; luminary; luminate; luminescence; luminous; luna; lunacy; lunar; Lunarian; lunate; lunation; lunatic; lune; lunette; luni-; luster; lustrum; lux; pellucid; sublunary; translucent.
It might also be the source of: Sanskrit rocate “shines;” Armenian lois “light,” lusin “moon;” Greek leukos “bright, shining, white;” Latin lucere “to shine,” lux “light,” lucidus “clear;” Old Church Slavonic luci “light;” Lithuanian laukas “pale;” Welsh llug “gleam, glimmer;” Old Irish loche “lightning,” luchair “brightness;” Hittite lukezi “is bright;” Old English leht, leoht “light, daylight; spiritual illumination,” German Licht, Gothic liuhaþ “light.”
We can see this is consistent with Frege’s prior sentence:
“Today I should like to throw light upon the subject from another side, and tell you about some supplementations and new conceptions, whose necessity has occurred to me since then.”[pg. 130-131 Frege “The Frege Reader” edited by Michael Beaney]
This sentence appears, to some, to say that there are necessary changes that must be made to Begriffsschrift in light of his work on the forthcoming “Grundgesetze der Arithmetik, Volume I” published in 1893. But, a careful reading of Begriffsschrift or a careful reading of the introduction or a careful reading of the rest of the paper reveals this can not be the case and that Frege is simply doing as he says e.g. supplementing and newly conceptualizing.
“It is my intention, in the near future, as I have indicated elsewhere, to explain how I express the fundamental definitions of arithmetic in my Begriffsschrift, and how I construct proofs from these solely by means of my symbols. For this purpose it will be useful here to be able to refer to this lecture so as not to be drawn then into discussions which many might condemn as not directly relevant, but which others might welcome.”[pg.130 Frege “The Frege Reader” edited by Michael Beaney]
His supplements and conceptualizations do not alter the Begriffsschrift. It is besides the point if they are largely mistaken when they are taken literally.
Thus, where his elucidations are incompatible with his Begriffsschrift we can, by his own admission, take his elucidations in jest, error, or, what is perhaps most valuable, as metaphor.
But there is more! Frege has already in the introduction of Begriffsschrift explained his metaphorical methods and their contribution to his designs! He repeats the same in “Function and Concept”. We are, from the very introduction of Begriffsschrift, told of its arithematical origins and of the exact metaphorical extensions which smooth the steps from the arithmetic of numbers to that of thought or concept.
For those who have not yet undertaken an endevor like Frege’s it may seem strange to treat his latter writings with such apparent disregard. But, the disregard is only apparent, as it is otherwise familiar to those of us who have come to understand (or, as the case may be, to misunderstand) our younger self.
So it is that Frege can be taken as metaphor where he is incompatible with the strictures of Begriffsschrift, and all his elucidations remain as elucidating as ever, but now as metaphor. Said another way, if later commentaries were to be given as foundational alternatives to Begriffsschrift then the entirety of the new foundation would be required as part of subsequent publications: this is known from the fact that Frege did not present Begriffsschrift as an alteration of his earlier attempts at such notation which are merely mentioned at the end of section three of Begriffsschrift.
I am once again drawn to entry 2025 0416 2358 in which I finally made a public record of being unable to keep up with my self. There is more I have written than I have been able to write out here. Thankfully, I am less likely to lose what it is I wrote than I ever was thanks to a social experiment I did some four or so years ago.
Together with two friends I designed a tiny culture around a collection of strict contingencies set up with the help of a somewhat elaborate sequence of index cards and index card holders. The strict method was evolved until it was eventually dropped, but it had done its work on my friends and my self in ways which were mostly within the reach of my designs e.g. the often complex contingencies linking time, reading, and writing into a social environment where collective action was a major conclusion to any such work were revealed more clearly than any of us had ever seen before.
Whatever was dropped of the evolved practices of reading and writing continued to control all our subsequent interactions. We had all come to keep a notebook in our pockets and to bring them out whenever we sat down to work with an other (especially when the other was one of the members of our tight knit group.
Somewhere along the way we landed on “The Pocket Notebook” by the Portage company as they appeared to, at the time, have the right price point. Previously I had been working mostly with the beautiful “Field Notes” by the company of the same name: John Gruber had mentioned them more than once and they were sturdy and simple in those ways that I cherish.
There are still many of those pocket notebooks around and I occasionally buy replacements when I run out of them. Spiral bound is important when you’re really “on the go” but it is not effective when you’re building cumulative records. The choice between one or the other is largely a matter of “where am I?” Pages can be torn from a spiral notebook and displayed as was once a major part of the tiny culture run via index cards. They can also be kept in place. When it comes to long form bound notes that are written as in a cumulative record in an experimental environment, I go with hardback books bound for artists.
Recently I’ve experimented with binding my own books as a further way of simplifying. There is still promise in such methods if not just because I haven’t tried them out in detail.
The first concrete hint of the transition from sense to schematism is in the first section (again I keep my etymological principles by presenting both translations of this selection from the books which I have in hand):
“I adopt this fundamental idea of distinguishing two kinds of symbols, which unfortunately is not strictly carried through in the theory of magnitude, [consider 1, log, sin, Lim], in order to make it generally applicable in the wider domain of pure thought. I therefore divide all the symbols I use into those by means of which one can represent different things and those that have a quite determinate sense. The first are the letters, and these should serve primarily to express generality. For all their indeterminancy, it must be insisted that a letter retain in the same context, the meaning once given to it.”[pg. 52 Frege ‘The Frege Reader’ edited by Michael Beaney]
“I adopt this basic idea of distinguishing two kinds of signs, which unfortunately is not strictly observed in the theory of magnitudes [consider 1, log, sin, lim], in order to apply it in the more comprehensive domain of pure thought in general. I therefore divide all signs that I use into those by which we may understand different objects and those that have a completely determinate meaning. The former are letters and they will serve chiefly to express generality. But, no matter how indeterminate the meaning of a letter, we must insist that throughout a given context the letter retain the meaning once given to it.”[pg. 10-11 Frege ‘From Frege to Godel’ Third Printing edited by van Heijenoor]
The former come immediately over to the variable letters as a grammatical category perhaps constructed from a single lowecase letter to which further members are added by iterated accentuation e.g. “x’” and “x’’’’”.
Although it is somewhat out of place here, I shall mention overtly that there are many details to the problems of grammatical analysis which are better explaiend in Quine’s “Philosophy of Logic 2nd Edition” than in any explanation I may give. In general, the methods of the grammarian can be applied to the methods of the logician and have been so with great effect. The result is that the methods of logic can then be applied to the methods of the grammarian and so on with ever greater effect (although there is presumably some kind of diminishing return). Thus it is that the grammar of logic can be spoken of logically. This is what is out of place here, but it is not so far out of place as to block me from further comment.
Talk of “the grammatical category of variables” can be dropped with all of its commitments of a theory of grammatical categories by giving a logical theory one predicate of which is ‘is a variable’. Here a warning is given about reading “predicate” which I repeat from Quine:
“A word of caution is in order regarding ‘predicate’ too. Some logicians take a predicate as a way of building a sentence aroudn a singular term or, more concretely, as what Peirce called a rheme [Volume 2, paragraph 95], a sentence with blanks in it, these being distinctively marked in the case of a many-place predicate. This version covers, implicitly, the potential output of predicate abstraction or predicate functors. But a predicate in my sense is always an integral word, phrase, or claus, grammatically a noun, adjective, or verb. Some are generated from others by grammatical constructions, notably the relative clause or, formally, predicate abstraction and predicate functors.”[pg. 61 Quine “From Stimulus to Science”]
This is given in the chapter “Denotation and Truth”. Note, for later reference, this restriction includes the results of predicate abstraction and predicate functors e.g. ‘{x : x loves Dick}’ and ‘Refl love’ (which goes into ‘love oneself’) are predicates. A reminder that an explanation of predicate abstraction from quantificational logic and predicate functors from predicate abstraction is to be found in my 2025 0422 2322 An Incomplete Sketch of My Philosophy of Logic.
Under the control of this convention the predicate ‘loves Dick’ is coextensive with the predicate abstract ‘{x : x loves Dick}’ which is to be distinguished from ‘x loves Dick’ which may be taken in one of three ways:
‘x loves Dick’ is like the sentence ‘He loves Dick’ were ‘He’ is a pronoun which purports to designate one and only one item. It is then for the rest of the sentences in a given theory to establish the existence (or non-existence) of the item purportedly designated by ‘x’ or ‘he’.
This is what has been called a Free Logic, but under the methods imposed by Quine in his paper “Free Logic and Virtual Classes” of 1994 and more carefully integrated into the methods of logic in his 1982 “Methods of Logic 4th edition”. Leblanc and Schock had different methods for dealing with what they called free logics which are clumsy compared to Quine’s e.g. rather than simply alter the definition of ‘traditional logic’ to demand first contemplating a nonempty universe of discourse and only as a special case going through and marking existentials as false and universals as true in the degenerate case of an empty universe, they would demand traditional logic distinguish between those where an instance of a schematic premise such as “each item is {x : Fx only if Fx}” be introduced at every turn lest the theory fall to the degenerate case.
The thrust of Quine’s arguement for his method is two fold: 1) the rules of passage do not hold in the empty case but are part and parcel of traditional logic, 2) admiting singular terms which merely purport to designate one and only one item clears the way for a careful analysis of the descriptional premises which are so often invisible to those who invoke singular terms as if they could not even be so invoked without designating some unspecified item.
This is my favorite method and it is also a delightful way of noticing how smoothly singular terms disolve into singular descriptions: a more detailed explanation of this dissolution and its combination with predicate abstracts as purported designations of their extensions shall not be found in this note.
‘x loves Dick’ is short for its universal closure ‘each item is {x such that x loves Dick}’ with predicate functors and predicate abstracts or, in crusty quantificational languages, ‘(each item is x such that)(x loves Dick)’. It is this way of taking ‘x loves Dick’ that is most familiar from the practices of arithmetic where an equation with variables such as ‘x+x = 2x’ is said to be “an identity” in that “each item is {x : x +x = 2x}” is true.
It is also this method which is commonly used in logic programming where question marks are prepended to the variable letters that are under universal quantification e.g. ‘?x loves Dick’ or ‘?x loves ?y’. This is because of the way that some logic programming langauges take such sentences as querries upon which they generate substitutions that satisfy the sentence based on the contents of the database being querried.
In general, this is the beginning of the method of Skolemization or its notational variation as Quine’s functional normal form (which does away with Skolem functions while retaining their logical import). Quine’s functional normal form introduces compound variable-like letters which are either a classic variable letter or are a classical varible letter subscripted with a list of compound variable-like letters. Then the variable letters which have no subscript are implicitely universal quantified and those with subscripts are implicitely existentially quantified and the order of these quantifications can be constructed from the structure of the subscripting. At a probably much later time I shall give the algorithm for converting to and from functional normal form since it greatly simplifies logic programming with Quine’s main method.
‘x loves Dick’ is not a sentence nor a predicate nor an unmarked universal closure but rather a kind of sentence under construction. For example, the place marked by ‘x’ in ‘x loves Dick’ is waiting to be replaced by a proper singular term e.g. replace ‘x’ by ‘Tom’ in ‘x loves Dick’ to get ‘Tom loves Dick’ which is a complete sentence. This is my least favorite method.
I’ve taken so much time to describe these different ways of taking ‘x loves Dick’ because it is the multitude of incompatible methods that were presented to Frege and which he had to navigate in order to arrive at his Begriffsschrift as something coherent, systematic, and, above all, extensional. His inspiration was arithmetic, and in that way he went with the second method whenever he said that a sentence like ‘x loves Dick’ is used to express a generality like that expressed from arithmetic in ‘x(y+z) = xy + xz’, and it is in Quine’s functional normal form that there are the most similarities between Frege’s methods and those of arithmetic e.g. as when he says
“The first are the letters, and these should serve primarily to express generality. For all their indeterminancy, it must be insisted that a letter retain in teh same context, the meaning once given to it.”[pg. 52 Frege ‘The Frege Reader’ edited by Michael Beaney]
“The former are letters and they will serve chiefly to express generality. But, no matter how indeterminate the meaning of a letter, we must insist that throughout a given context the letter retain the meaning once given to it.”[pg. 10-11 Frege ‘From Frege to Godel’ Third Printing edited by van Heijenoor]
But, alas, none of these three methods actually coincide with Frege’s usage or his own explanations of his own usage which is consistent throughout Begriffsschrift and wherever Begriffsschrift is directly invoked.
The methods of Frege are schematic and it is the way in which they are schematic which leads to sentences in Frege’s texts that, if read carelessly or in isolation, can lead to a great deal of confusion at many different times e.g. at some times the schematism is part of a theory of schematisms and at other times there is no such theory under investigation and the schematic methods are purely logical (as in talk of ‘valid schema’ through which validities are got by substitution but with which no items or objects are concerned).
The subject of “At Plataea the Persians were defeated by the Greeks” is “the Persians”. The predicate is “{x: At Plataea x were defeated by the Greeks}” via Quine’s predicate abstract notation.
Here is my single sentence summary of Quine’s predicate abstract notation (for more on its definition in a quantificational logic see 2025 0422 2322 An Incomplete Sketch of My Philosophy of Logic):
The English relative clause ‘who loves Dick’ and the pidgin ‘x such that x loves Dick’ are uniformly paraphrased by the predicate abstract ‘{x: x loves Dick}’ which abstracts ‘Tom’ from ‘Tom loves Dick’ by binding the free occurrence of ‘x’ in the open sentence ‘x loves Dick’ with the prefix ‘x:’ so that the predication ‘{x:x loves Dick}Tom’ concretes to ‘Tom loves Dick’: whatever can be said of a thing can be said by predicating a predicate of it i.e. predicational completeness.
The object of “At Plataea the Persians were defeated by the Greeks” is “the Greeks”. The most general predicate is “{x,y,z: At x y were defeated by z}” from which the sentence can be factored conspiucously into its predicate, subject, and objects as “{x,y,z: At x y were defeated by z} Plataea, the Persians, the Greeks”. This analysis is of major consequence to explaining Frege’s main paragraph in Begriffsschrift which I first mentioned here.
Rules of inference are the manufactured products of logical methods: they mark the end of logical heavy lifting and the beginning of logical economy for all. Those obsessing over the by products of this or that collection of rules of inference are only indirectly engaged in logical methods. Inference rules are the mechanical proxies of logical work.
The translation of Frege (1848–1925) through Quine (1908-2000) begins with John Horne Tooke’s (1736-1812) suggestion that
“you substitute the composition, &c, of terms wherever he has supposed a composition of ideas”[pg.6 Quine “From Stimulus to Science”].
(Note, Tooke’s “The Diversions of Purely” played a part of mention in Skinner’s (1904 –1990) theoretical analysis of verbal behavior.)
This technique evolved through Bentham’s (1748–1832) contextual fictions (I have not been able to yet identify if Bentham read Tooke), then Boole’s (1815–1864) concrete differential notation, and directly on through Russell’s (1872–1970) triumphant contextual definition of singular description where ‘the {x such that Fx}’ is elimianted (with major restrictions like those in Quine’s “Free Logic and Virtual Classes” of 1994 and their projection back and forth through Quine’s “Methods of Logic 4th edition” of 1982) as follows:
‘G the {x such that Fx}’ for ‘some item is {y such that Gy and each item is {x such that x is identical to y if and only if Fx}}’.
Frege, Peirce (1839-1914), Peano (1858-1932), Dedekind (1831-1916), and Cantor (1845-1918) are in the firmament from which Russell’s theory of descriptions appeared (I am simply using it here as a concrete checkpoint).
As an aside, Russell and Frege dreamed of deducing mathematical truth from logical truth where their logic contained set membership and is what we would today call a mathematical logic: it was Gödel (1906-1978) in 1931 who finally severed the methods of mathematics from the methods of logic (Quine submits that Russell’s paradox of 1902 is an earlier component of the full severence). It is the analysis of the ancestral of a relation and its confusion with the transitive closure of a predicate that prompted Frege and Russell to easily mix mathematics and logic: they mixed copuli of predication with predicates of membership. The confusion between mathematical logic and (predicate) logic can be forgiven when carrying Frege through Quine because it is a confusion which almost all mathematicians and logicians continue to make either out of laziness or ignorance: the lazy maintain the ignorant.
Quine continues the Tookeian tradition through his method of semantic ascent, and it is with such methods that the bulk of Frege can be happily translated. It is not effective to go straight to semantic ascent first though because there are a number of distinctions which Frege makes which are interdependent with each other in ways that are factored out by Quine in different ways that don’t depend on semantic ascent.
First I shall collect the relevant selections from Quine and then I shall collect and collate the relevant sections from Frege. But, it is impossible not to allude to Frege for there is so much which he already noticed and which is only slightly obscure relative to Quine’s distinctions. In fact, I shall already break my plan here at the beginning because there is a beautiful distinction which Frege makes and which is otherwise obscure to me but in light of the following key distinction by Quine:
“A context is extensional if its truth value cannot be changed by supplanting a component sentence by another of the same truth value, nor by supplanting a component predicate by another with all the same denotata, nor by supplementing a singular term by another with the same designatum. Succinctly, the three requirements are substitutivity of covalence, of coextensiveness, and of identity, salva veritate. A context is intensional if it is not extensional.”[pg. 90 Quine “From Stimulus to Science”].
The division between extensional and intentsional context is given by Frege in the third section of the first part of Begriffsschrift. Since I know of no way of seperating Frege’s overlapping concerns I must present the section in its whole each time I explain one dimension of distinction that Quine makes which is relevant to it. I also have access to two different translations of Begriffsschrift and must put both of them right after each other because of my etymological principles.
“A distinction between subject and predicate finds no place in my representation of a judgement. To justify this, I note that the contents of two judgements can differ in two ways: either the conclusions that can be drawn from one when combined with certain others also always follow from the second when combined with the same judgements, or else this is not the case. The two propositions ‘At Plataea the Persians were defeated by the Greeks’ and ‘At Plataea the Persians were defaeted by the Greeks’ differ in the first way. Even if a slight difference in sense can be discerned, the agreement predominates. Now I call that part of the content that is the same in both, the conceptual content. Since only this has significance for the Begriffsschrift, no distinction is needed between propositions that have the same conceptual content. If it is said, ‘The subject is the concept with which the judgement is concerned’, then this applies also to the object. It can therefore only be said: “The subject is the concept with which the judgement is primarily concerned’. The linguistic significance of the position of the subject in the word-order lies in its marking the place where what one particularly wants to draw the attention of the listener to is put. (See also $9.) This can have the purpose, for example, of indicating a relation between this judgement and others, thereby fascilitating the listener’s grasp of all the interconnections. Now all those features of langauge that result only from the interaction of speaker and listener— where the speaker, for example, takes the listener’s expectations into account and seeks to put them on the right track even before a sentence is finished— have no counterpart in my formula language, since here the only thing that is relevant in a judgement is that which influences its possible consequences. Everything that is necessary for a valid inference is fully expressed; but what is not necessary is mostly not even indicated; nothing is left go guessing. In this I closely follow the example of the formula language of matheamtics, in which subject and predicate can also be distinguished only by violating it. Imagine a language in which the proposition ‘Archimedes was killed at the capture of Syracuse’ is expressed in the following way: ‘The violent death of Archimedes at the capture of Syracuse is a fact’. Even here, if one wants, subject and predicate can be distinguished, but the subject contains the whole content, and the predicate serves only to present it as a judgement. Such a language would have only a single predicate for all judgements, namely ‘is a fact’. It can be seen that there is no question here of subject and predicate in teh usual sense. Our Begriffsschrift is such a language and the symbol |— is its common predicate for all judgements.
In my first draft of a formula language I was misled by the example of ordinary language into constructing judgements out of subject and predicate. But I soon convinced myself that this was an obstacle to my particular goal and only led to useless proxlixity.”[pg. 54 Frege ‘The Frege Reader’ edited by Michael Beaney]
I have read this section multiple times, and have written it out twice now: it is probably one of the most important paragraphs in human history. It combines practically every major distinction which separates the practices of logic from all other human practices. Here is the same selection from a different translation (note the difference in spelling from the last’s ‘judgement’ to ‘judgment’):
“A distinction between subject and predicate does not occur in my way of representing a judgment. In order to justify this I remark that the contents of two judgments may differ in two ways: either the consequences derivable from the first, when it is combined with certain other judgments, always follow also from the second, when it is combined with these same judgements, [[and conversely,]] or this is not the case. The two propositions “The Greeks defeated the Persians at Plataea” and “The Persians were defeated by the Greeks at Plataea” differ in the first way. Even if one can detect a slight difference in meaning, the agreement outweighs it. Now I call that part of the content that is the same in both the conceptual content. Since it alone is of significance for our ideography, we need not introduce any distinction between propositions having the same conceptual content. If one says of the subject that it “is the concept with which the judgment is concerned”, this is equally true of the object. We can therefore only say that the subject “is the concept with which the judgment is chiefly concerned”. In orderinary language, the place of the subject in the sequence of words has the significance of a distinguished place, where we put that to which we wish especially to direct the attention of the listener (see als $9). This may, for example, have the purpose of pointing out a certain relation of the given judgment to others and thereby making it easier for the listener to grasp the entire context. Now, all those peculiarities of ordinary language that result only from the interaction of speaker and listener— as when, for example, the speaker takes the expectations of the listener into account and seeks to put them on the right track even before the complete sentence is enunciated— have nothing that answers to them in my formula language, since in a judgment I consider only that which influences its possible consequences. Everything necessary for a correct inference is expressed in full, but what is not necessary is generally not indicated: nothing is left to guesswork. In this I faithfully follow the example of the formula language of mathematics, a language to which one would do violence if he were to distinguish between subject and predicate in it. We can imagine a language in which the proposition “Archimedes perished at the capture of Syracuse” would be expressed thus: “The violent death of archimedes at the capture of Syracuse is a fact”. To be sure, one can distinguish between subject and predicate here, too, if one wishes to do so, but the subject contains the whole content, and the predicate serves only to turn the content into a judgement. Such a language would have only a single predicate for all judgments, namely “is a fact”. We see that there cannot be any question herer of subject and predicate in the ordinary sense. Our ideography is a language of this sort, and in it the sign |— is the common predicate for all judgments.
In the first draft of my formula langauge I allowed myself to be misled by the example of ordinary language into constructing judgments out of a subject and predicate. But I soon became convinced that this was an obstacle to my specific goal and led only to useless prolixity.[pg.12-13 Frege ‘From Frege to Godel’ Third Printing edited by van Heijenoor]
Another reason I have presented this paragraph twice is becuase it can only really be taken as a whole the way that Frege presents it, and to factor it into its component parts, which is what can be done once each of Quine’s methods is in hand, would be to dismantle the innerworkings of vintage Cortébert pocket watch!
To make the concrete connection then between Quine’s extensional and nonextension contexts and Frege’s distinction between conceptual content and nonconceptual content of a judgement:
“the contents of two judgements can differ in two ways: either the conclusions that can be drawn from one when combined with certain others also always follow from the second when combined with the same judgements, or else this is not the case.” [pg. 54 Frege ‘The Frege Reader’ edited by Michael Beaney]
“the contents of two judgments may differ in two ways: either the consequences derivable from the first, when it is combined with certain other judgments, always follow also from the second, when it is combined with these same judgements, [[and conversely,]] or this is not the case.”[pg.12 Frege ‘From Frege to Godel’ Third Printing edited by van Heijenoor]
I did not grasp the distinction that Frege was making without interpreting what he wrote through Quine. First, the distinction is between the complementary ways that the contents of a judgement can differ. That is, he distinguishes between different (yet complementary) differences: this is why it is a bit confusing. We take a pair of judgements, say there is a judgement to your left and a different judgement to your right, and look at each conclusion that is implied by the content of the left judgement under a potentially empty collection of supplementary judgements and look at each conclusion that is implied by the content of the right judgement under the same collection of supplementary judgements. Then, if there is for each conclusion from the left judgement the same conclusion from the right judgement under the same collection of supplementary judgements, the different contents of the left and right judgements are different only in their non-conceptual content.
The example Frege gives for the different contents of two judgements that differ only in their non-conceptual content is as follows:
“The two propositions ‘At Plataea the Persians were defeated by the Greeks’ and ‘At Plataea the Persians were defaeted by the Greeks’ differ in the first way.”[pg. 54 Frege ‘The Frege Reader’ edited by Michael Beaney]
“The two propositions “The Greeks defeated the Persians at Plataea” and “The Persians were defeated by the Greeks at Plataea” differ in the first way.”[pg.12 Frege ‘From Frege to Godel’ Third Printing edited by van Heijenoor]
This example, as given, does not emphasize that the contents are to be of a judgement: this is because Frege has said he shall only contemplate contents which are judgeable. Following through the example in detail: 1) the two quotations are spelled differently i.e. they are not identical as concatenations of letters, 2) thus they differ in at least one way (as contents) and the question is how they differ with respect to the distinction Frege has made between conceptual and nonconceptual content, 3) Frege claims that there is no conclusion that can be got from the one that can not be got from the other with respect to the same supplementary judgements, 4) therefore they have the same conceptual content and differ only in their non-conceptual content, one part of which is, presumably, their spelling as quotations.
Now note that Frege, in this example, goes far beyond the difference in content that I have mentioned as a difference in the spelling of their quotations: he wishes to go beyond propositions as sentences that may differ in spelling or, that complex difference in spelling called a difference in phrasing. This is seen in the sentence following the sentence with the examples:
“Even if a slight difference in sense can be discerned, the agreement predominates.”[pg. 54 Frege ‘The Frege Reader’ edited by Michael Beaney]
“Even if one can detect a slight difference in meaning, the agreement outweighs it.”[pg.12 Frege ‘From Frege to Godel’ Third Printing edited by van Heijenoor]
Here was also see our first divergence between translations with “sense” in about the same place as “meaning” is in the other. I do not own the German editions (yet), but a bit of etymology obviates any problems of translation from there (I can not get into this in greater detail without become derailed entirely in Skinner’s analysis of verbal behavior).
When I worked through Frege’s example I neither mentioned the sense nor the meaning of a sentence: I stuck with the difference in phrasing and took that to be slight in that, following Frege, there is not a conclusion from the one phrasing that can not be got from the other phrasing under identical supplementation of auxiliary judgements.
This radical translation is the hallmark of my presentation here. Though I may not land on any singular method of radical translation from Frege’s words and sentences, I shall demonstrate how, through Quine and Skinner, nothing like what Frege or most philosophers expect, as in senses and meanings, is actually at work throughout Frege’s writings: so much so that nothing need be edited within his texts besides the already well known slips and such. I see this as being in the same vein as Tooke, Bentham, Boole, Russell, and Quine, but wish to make it clear that I feel no affinity to a sort of nominalism (it shall be hard for some to accept this last point but the only way I could make this clear is by going over Quine’s early stint as a nominalist with Goodman in 1947 which comes a cropper in ways parallel to Carnap’s Aufbau).
Thus I mention now Quine’s outlook on propositions in 1995:
“So it is in standing sentences that the notion of meaning goes shaky. But this is the very locus of the philosophic notion of proposition: the meaning of a sentence of fixed truth value. Many philosophers have seen propositions as abstract objects that statements served to express. They have seen them as the bearers of truth values; sentences were true and false only in the sense of expressing true or false propositions.
There is indeed a usage of ‘proposition’ that is useful and unobjectionable. It can be construed as denoting the sentences themselves, rather than their meanings, but it is used instead of ‘sentence’ when we are concerned with the sentences as an object of belief (as we shall be in Chapter VIII) rather than with its morphology and syntax. I deny myself this convenient usage, for fear of beclouding issues; but it carries no commitment to sentence meanings.”[pg.77 Quine “From Stimulus to Science”]
I make no such commitment as there outlined at this time, but give it as an example of a principle governing my contemplation.
It is in the next sentence, where Frege defines the conceptual content of a proposition, that the quote from Quine, where he defines extensional contexts, activates:
“Now I call that part of the content that is the same in both, the conceptual content. Since only this has significance for the Begriffsschrift, no distinction is needed between propositions that having the same conceptual content.”[pg. 54 Frege ‘The Frege Reader’ edited by Michael Beaney]
“Now I call that part of the content that is the same in both the conceptual content. Since it alone is of significance for our ideography, we need not introduce any distinction between propositions have the same conceptual content.”[pg.12 Frege ‘From Frege to Godel’ Third Printing edited by van Heijenoor]
This is another way of saying that the contexts of Frege’s propositional conceptual contents are extensional: conceptual content is defined by its invarience over implication which is sufficient to guarentee extensionality in such contexts. While Carnap is known for his principle of extensionality (later demoted to the mere conjecture that intensional contexts can be paraphrased by extensional contexts), it is Quine’s “Word and Object” that I have in mind when I say that invarience over implication guarentees extensionality.
It is because the contexts of Frege’s propositional conceptual contents are extensional that I can present the following elaboration of transition from sense/meaning to schematics.
The next fragmentary step of my view of Frege is here where the examples are further analyzed An Analysis of “At Plataea the Persians were defeated by the Greeks”
Or you can jump over that and just go here: 2025 0429 1424 A Preliminary Outline of The Method of Bringing Frege Through Quine.
First editorial comment added to an older note: the shortener at the end of Bit Strings and Binary Trees does not work in a degenerate case that is now described there!
I skimmed Frege’s paper on thought. Almost everything he says is covered by B. F. Skinner’s theoretical analysis in “Verbal Behavior” 1957.
A theory of behavior may admit abstract items designated by a description of a contingency of reinforcement e.g. “a door opens from a push on a lever” can be made to designate an abstract ordered pair whose first component is a concrete chunk of doorish spacetime and whose second component is an ordered pair whose first component is a slice of spacetime which intersects the doorish chunk and the lever chunk through the chunk of behaving organism (perhaps narrowed by setting up a mechanical device called a detector or cumulative recorder) and whose second component is the leverish chunk. This ordered pair is abstract in that its existence does not depend on a mental or physical theory: the components are singular descriptions and not singular terms. That is, the ordered pair exists whether the components do or not.
You can do all of this and bring Skinner’s analysis of verbal behavior into the abstract realm of orderd pairs to get something beyond ‘x from y on z’ with descriptive premises for ‘x’, ‘y’, and ‘z’ (as in Quine’s “Free Logic and Virtual Classes” of 1994). I set this up to shift from propositions to ordered pairs where the theory of ordered pairs is clear: they are identical when their components are. What both Quine and Skinner do is to make all these excursions to simulate some abstract object templated on a mental or physical analysis moot. You can skip them entirely and nothing sacred disappears.
It may continue to feel as if the sacred has been paraphrased away, but if you stick with Frege then that would be letting psycholgoy intrude upon philosophy.
The word ‘sadness’ contains ‘adn’ which is also one of my common mispellings of ‘and’.
I am forgoing my traditional methods of reading Frege for the following reasons.
I can state my outlook on Frege concretely: it is the interpretation of Frege by Quine in “On Frege’s Way Out” of 1955 in light of all that Quine later worked out in, e.g., “From Stimulus to Science” of 1996. What I can do is paraphrase all that can be brought through Frege without breaking our best logical practices. I have already noted such latter day methods on Twitter when I wrote of schematic theories of predicates of extensionality. I will do my best to write that up and read from Frege.
The key to this transition was provided by B. F. Skinner’s theoretical analysis in “Verbal Behavior” of 1957 which ends with chapters on logic, science, and thinking. Whereas Quine indulges the attitudinists with their properties, propositions, concepts, information, attributes, and other nonextensional contexts long enough to dismiss them, he does not explain them away (as many have subsequently griped). Skinner provides the details of such an explanation and there is some path from Quine’s method of semantic ascent and Skinner’s analysis of autoclitic verbal behavior which covers past efforts by Frege, Russell, Whitehead, and the moderns like Per Martin-Löf primed by their ill fated allegiance to their fellow attitudenists.
Quine’s paper “On Frege’s Way Out” points up Frege’s position as one where attributes/concepts/propositions/information/essences/properties (and whatever new nonextensional contexts have been cooked up while I was writing this sentence) are secondary to extensional contexts: it is Russell and whitehead who based their theory of extensional contexts of classes on the nonextensional contexts of attributes.
The transition is one from sense to schematics. It is the one Quine promoted and which, I submit, is also promoted by Frege, though I accept that it is perhaps only through Quine’s later work that the paraphrase can be judged as satisfactory and read as perfectly parallel to Frege in Begriffsschrift. Frege is to be judged solely on Begriffsschrift: the rest of his work is commentary and is like barnacles that must be cleaned from the hull of humanity’s greatest achievement.
There is one last reason for the urgency of this all: as verbal machines become more articulate they shall make it hard for most people to cling to nonextensional contexts. Furthermore, those who dispense with them sooner rather than later, as prescribed by Skinner, are more likely to survive: they can skip intermediate ruminations on attributes. This saves time and extends the reach of extensioanl contexts which are among our greatest verbal technologies.
For those who may not be familiar with the distinction between extensional and nonextensional contexts here is Quine’s definition:
“A context is extensional if its truth value cannot be changed by supplanting a component sentence by another of the same truth value, nor by supplanting a component predicate by another with all the same denotata, nor by supplementing a singular term by another with the same designatum. Succinctly, the three requirements are substitutivity of covalence, of coextensiveness, and of identity, salva veritate. A context is intensional if it is not extensional.”[pg. 90 Quine “From Stimulus to Science”].
This continues my work on my little lisp from 202504211546. If you haven’t been following along then this is a good place to start, but, like the rest of the notes I’ve made thus far, it assumes that you know basic javascript and how to define functions by recursion.
Simplification does not defer design decisions: it pushes them out of the problem space entirely. Upon approaching the design of my little lisp’s reader, such simplifications occurred to me and, hence, solved major problems by disolving them rather than resolving them into simpler problems with simpler solutions. This is a halmark of factoring as a problem elimination process rather than a problem solving process. A problem can very often be factored out of existence faster than a solution can be found to it.
A pure lisp is one without any atoms. The empty pair takes over the part played by the item previously designated by ‘nil’. The empty pair shall here be designated by ‘theEmptyPair’. Though I have a theory of ordered pairs in the works, I am committed here to javascript as the lingua franca of the internet.
let theEmptyPair={}
, isEmpty = x => x == theEmptyPair
, consOf = (x,y) => [x,y]
, carOf = x => isEmpty(x) ? x : x[0]
, cdrOf = x => isEmpty(x) ? x : x[1];
Identity is no longer a primitive:
let id = (x,y) =>
(isEmpty(x) && isEmpty(y))
|| (!(isEmpty(x) || isEmpty(y))
&& id(carOf(x),carOf(y))
&& id(cdrOf(x),cdrOf(y)));
There is no longer the troublesome distinction between proper and dotted lists, between symbols or runes, and we are free to build such distinctions, if we so desire, from this simpler starting point.
Note, I was reluctant to start with a pure LISP because I am already familiar with the theory of ordered pairs upon which it may be based and had hoped to learn something new from the purported simplicity of more familiar implementations of LISP. The design of the reader made it clear that there is nothing to be gained from mixing the construction of native and foreign items (the items of javascript and lisp respectively).
This switch to a simpler design also brings my little pure lisp into closer contact with my little concatenative language designed from simplifying Charles H. Moore’s uhdForth (it is sad that so far there is no one who has written on uhdForth but me: the reason for this is that almost everything about uhdForth has to be deduced from short yet logically complete presentations in garbled youtube videos).
Under this simplification every item is a rune, a runic list, a proper list, and a number of dotted lists (no greater than the length of the proper list). Thus, there is no longer a printer, but rather a collection of printers to be used depending on the context of the program as written. This simplification also switches the emphasis from the printer to the reader: why?
There are no strings that cross the boarder from javascript to the pure lisp: strings are left behind! Rather than wrestling with problems of peculiar implementations, we are left to establish whatever expedient conventions work for now. They may be changed later, but such conventions do not change the lisp: they change how we set it up. It shall later become clear that these conveniences are not about lisp at all, but rather the theory of ordered pairs. An example of one such theory is Finitary inductively presented logics by Solomon Feferman.
There is one kind of item which crosses the boarder between javascript and my lisp which I have not mentioned all this time: the boolean values designated by ‘true’ and ‘false’ in javascript. I’ve mentioned them in descriptions of past functions, but not as naturalized citizens of my lisp. The traditional line is to take the empty pair as either the mark of truth or the mark of falsehood. This can not be done with javascript without digging into late manglings of the language (though it is hard to tell when something that is already mangled is mangled more). The reason is one which I have also not mentioned overtly, but which I shall mention now: I’ve been using the javascript ternary conditional notations ‘p ? t : f’ and hidden versions of the same in ‘p && q’ and ‘p || q’.
In javascript, such expressions control the execution of their subparts e.g. if ‘p’ designates the same as ‘true’ then ‘t’ is executed in ‘p ? t : f’ and ‘f’ is executed otherwise. Mixing ‘executed’ and ‘designated’ and all such things here has gone against my better principles, but these are not the things I aimed at addressing here so I’m moving on without further comment.
The recursive definition of the function designated by ‘id’ above is not simple, but the following functions are and they establish a key convention:
let singletonListOf = x => consOf(x,theEmptyPair)
, singletonStackOf = x => consOf(theEmptyPair,x);
Almost all of the conventions that I shall adopt here are from my concatenative language that I have yet to present here.
The singleton functions provide a simple way of tallying, and hence transforming a native javascript number into a foreign tally:
let tallyOf = n =>
n>0 ? singletonListOf(tallyOf(n-1)) : theEmptyPair;
Then, to read an external character is to find its index in some alphabet and take the tally of that:
let abc =[...'0123456789abcdefghijklmnopqrstuvwxyz ()']
, alphabeticalIndexOf = letter => abc.indexOf(letter)
, alphabeticalLetterOf = index => abc[index];
There is an even more devilish way of going back and forth acrosst he boarder between javascript and my little lisp. Rather than use tallies, a bit based method can be introduced e.g. only a bit at a time crosses the boarder one way or the other. I leave that for a few code fragements later.
The old names for functions can now be used anew (after introducing the familiar string functions under slightly different names):
let theEmptyLetter=''
, isEmptyLetter = x => x == theEmptyLetter
, concatenationOf = (...strings) =>
strings.length ? strings.shift() + concatenationOf(...strings) : theEmptyLetter
, firstLetterOf = letters => isEmptyLetter(letters) ? theEmptyLetter : letters[0]
, restLettersOf = letters => isEmptyLetter(letters) ? theEmptyLetter : string.slice(1);
Oops! Forgot the inverse function to the one that makes tallies:
let countOf = x => isEmpty(x) ? 0 : 1 + countOf(cdrOf(x));
Now old names can be used anew!
let runeOf = letter => tallyOf(alphabeticalIndexOf(letter))
, runicListOf = letters =>
letters.length ? consOf(runeOf(firstLetterOf(letters))
, runicListOf(restLettersOf(letters)))
: theEmptyPair
, letterOf = x => letterAtIndexOf(countOf(x))
, stringOf = x =>
isEmpty(x) ? theEmptyString
: concatenationOf(letterOf(carOf(x)),stringOf(cdrOf(x)));
Some much needed examples:
countOf(tallyOf(3))
3
letterOf(runeOf('a'))
a
lettersOf(runesOf('this is a test'))
this is a test
Now for some strange examples that more closely reveal what I hinted at long ago about the joys of letting degenerate cases flourish:
letterOf(tallyOf(3))
3
lettersOf(runeOf('a'))
0000000000
countOf(runesOf('this is a test'))
14
For those who may have been put off by the definitions of the functions designated by ‘tallyOf’ and ‘countOf’ since they seemed to favor singleton stacks over singleton lists, I hope the last example calms your nerves: these conventions are selected by their native logic in javascript (and in any other language which purports to cover arithmetic and concatenation for that matter). Said poetically: economize on externalities.
A further consequence of these simplifications is that I can bring together the work done from 2025 0413 1513 Bit Strings and Binary Trees and explain how to cross the boarder between langauges bit by bit without leaving our native language:
let theEmptyStack = theEmptyPair
, isEmptyStack = isEmpty
, pushOf= (stack,item) => consOf(stack,item)
, popOf = carOf
, peekOf = cdrOf
, topOf = stack => peekOf(stack)
, secondOf= stack => peekOf(popOf(stack))
, pop2Of = stack => popOf(popOf(stack))
, pairUpOf = stack => pushOf(pop2Of(stack),consOf(secondOf(stack),topOf(stack)))
, pairUpEachOf = stack =>
isEmpty(popOf(stack)) ? topOf(stack)
: pairUpEachOf(pairUpOf(stack))
, encodeHelper = (digits, stack) =>
isEmptyLetter(digits) ? pairUpEachOf(stack)
: isZeroDigit(firstLetterOf(digits)) ?
encodeHelper(restLettersOf(digits), pushOf(stack,theEmptyPair))
: isOneDigit(firstLetterOf(digits)) ?
encodeHelper(restLettersOf(digits), pairUpOf(stack))
: encodeHelper(restLettersOf(digits),stack)
, encode = digits => encodeHelper(digits,theEmptyStack);
with some examples (notice the last one)
decode(encode('001'))
001
decode(encode('0000'))
0000111
decode(encode('10100'))
001010011
decode(encode('00101011'))
000101011
decode(runeOf('a'))
000000000001111111111
lettersOf(encode(decode(runesOf('this is a test'))))
this is a test
That’s enough for today.
This uses notation from A Stack Notation for Predicate Functor Logic.
Here I give a few quick examples of how to translate a sentence of quantificational logic to one of predicate functor logic. As I finish up my little lisp I’ll write out mechanical algorithms for accomplishing each of these steps and look for a few shortcuts.
As an aside, it is possible to reduce all of predicate logic to just three predicate functors:
I’ll save that reduction for another time maybe.
The basic predicate functors in this presentation of homogenization will be then
In the linked note I show how to get
from those basic six starting functors. With the appropriate combination of swaps, pops, and pushes, it is possible to move the variables of any predicate into any position:
Note, for a predicate functor designated by ‘f’ the notation ‘f^0’ designates the same as ‘f’ and ‘f^(n+1)’ designates the same as ‘f^n f’ so that each predicate functor issues in its iterates in the expected way.
The pidgin ‘ret’ is short for ‘retrojection’ which I take from Quine’s explanation of the permutational part of homoginization.
Now predicates in quantificational logic are said to be of degree n when they have n variables attached to them (most often on the right hand side) e.g. ‘Fxyxxzu’ is of degree six and ‘Gyxyxy’ is of degree five.
The first step in translating a closed sentence of quantificational logic into a predicate functor is to attach the degree of each predicate in the sentence to its name as a numeral (this does not actually involve arithmetic in the notation any more than counting the number of variables attached to a predicate of quantificational logic does)e.g. ‘Fxyxxzu’ becomes ‘F6xyxxzu’ and ‘Gyxyxy’ becomes ‘G5yxyxy’. This must be done for each predicate in the sentence being transformed. Sentence letters (e.g., often ‘p’, ‘q’, ‘r’) get carried over to degree zero predicates.
Next, to simplify matters, before or after attaching these numerals, the sentence is to be put into one of its equivalents which is compounded of only existential quantifiers, conjunctions, and negations.
Now, negations get taken over simply enough by bringing it into predicate functor form e.g. ‘not Fxyz’ becomes ‘(not F3)xyz’ and ‘some item is {x such that (not Fxyz) and (not Gux)}’ becomes ‘some item is {x such that (not F3)xyz and (not G2)ux}’.
Where ever there are duplicate occurrences of a variable they can be eliminated by retrojections, pops, pushes, and dups. A mechanical way of doing this to alphabatize the variables attached to a predicate and to put it in stack form (which favors the left hand side of a predicate) e.g. ‘Fxyxzyz’ becomes ‘F6xyxzyz’ then ‘xxy(swap pop push^3 F6)zyz’ to ‘xxyyzz(pop swap pop^2 swap push^3 F6)’ then to
‘(push dup push dup push dup pop swap pop^2 swap push3 F6)xyz’
which avoids retrojections and is adhoc really.
Now assuming this is done then it is only homogenization that is left and it only occurs when confronted by conjunctions e.g. given
‘Faxy and Gbz’
missing letters can be brought in by ‘drop’ e.g.
‘bz(drop^2 F3)axy and ax(drop^2 G2)bz’
and alphabetized
‘(..f drop^2 F3)abxyz and (..g drop^2 G2)abxyz’
by some appropriate sequence of functors designated by ‘..f’ and ‘..g’ respectively: the functors ‘(..f drop^2)’ and ‘(..g drop^2)’ are said to homogenize ‘F’ and ‘G’ respectively. This can obviously be generalized to connectives of more components than the two of primitive conjunction. Finally then, the ‘and’ is brought in so that
’((..f drop^2 F3) and (..g drop^2 G2))abxyz’.
Last, but not least, the variable of an existential quantifier is brought to the top of the left stack by appropriate pops, then ‘some’ is brought in e.g.
‘some item is {z such that Fxyz}’
becomes
‘some item is {z such that xyz(pop^3 F3)}’
and then
‘xy(some pop^3 F3)’.
Together these complete the translation to and from predicate functor logic.
This is a record of my reading of “The Frege Reader” edited by Michael Beany. It was recommended to me by R.P. (@ResonantPyre). Since this is my first public reading of a primary text (I do not read secondary texts— like I have Durant’s “Story of Philosophy”— as I might Plato’s dialogues) I shall say a bit on what I do when I read.
First, this is a record of my verbal responses to my exposure to the text. To read is to do more than listen with my eyes. It is built upon gazing at the page and the marks upon it. The narrower repertoire of looking, upon which the complex repertoire of seeing is built and upon which reading is built, bring the marks into focus and I can begin to see words, perhaps letters, sentences, and, the key feature of seeing, all that I see to myself which is not to be found on the page or in its marks.
To read is to go beyond seeing as listening with my eyes. It is speaking in response to sights of, sounds of, feelings of, etc. By uncovering what I have to say I am more likely to discover what I am thinking and feeling. There is still much I have to learn about how best to grasp my thinking, my feeling, and what of past and present environments may be controlling it now.
Without further ado!
I opened to the page with all the publishing informatin on it. It was first published in 1997. Immediately, the first name of the dedication caught my eye “Peter Geach”. It occurred to me that he is mentioned by Quine, perhaps in “From stimulus to Science” where Quine discusses pronouns of laziness. After checking the page (I can see to myself where in the book Quine would have written about that), it is found that yes, Geach is where Quine got the phrase “pronouns of laziness” (they are, for Quine, a focal point of reification).
Now I turn to the contents of “the Frege reader”, and, as I very often do, I write them down:
It contains selections from his larger works which I shall get at a later time (though I do wish there was a single complete volume with all his texts and relevant major correspondances). They are presented in chronological order and are mostly extracts: I do not yet know what the difference is between an extract and a selection.
Next, the preface, though I turned to appendix 2 on Frege’s logical notation to take a look at its layout. I take Frege’s notation as one of the greatest accomplishments in human history and was likely drawn to it from how often I ahve been reinforced by it in the past.
The preface says that Frege’s works can not be allowed to “speak for themselves” to a student for there are subtleties that may easily go unnoticed but which are key to firmly grasping Frege’s collective works.
I am impatient and shall skip any further introductory matters and go straight to Frege’s first work “Begriffsschrift: a formal language of pure thought modeled on that of arithmetic”. Again, I am met with more commentary that I simply skip to get to Frege’s “preface”:
“The recognition of a scientific truth generally passes through several stages of certainty.”[pg. 48]
Already, the developmentalistic metaphors which plague psychology and the other nonbiological sciences appears: “passes through several stages” just as a child is said to pass through grades from kindergarten onwards and upwards. For those interested in precise and accurate reports on changes in human behavior— which ‘recognition’ purports to be— it is not the passage of time between stages that matters (or, the metaphorical extension of the same to “certainty”), but rather what happens as time passes.
In the end, any developmentalist, when pressed, can give some description of purportedly characteristic features of any performance to be found within a deliniated stage. Whether such characteristic descriptions are maintained or whether a given sequence of corresponding stages seperated by changes in characteristic performances are maintained is a problem of statistical aggregation. For one aimed at something more foundational than flexable predictions relative to the overlap of an individual’s behavior among differing performances in differing stages of development there is little of consequence to be kept besides the vague gesturings to unexamined environments.
It is the ‘recognition’ that “generally passes through stages of certainty”, not the “scientific truth” itself. Might “recognition” just as well be replaced by “cognitition”? Are there nonscientific truths? Is this law of recognition itself subject to a staged development? If so, what stage is it in? What stages are there? Is it a spectrum from uncertainty to certainty as the word “certainty” suggests? Is the generality here scientific or logical?
Following Skinner’s principle “Etymology is the archeology of thought.” from his paper “The Origin of Cognitive Thought” I shall show the etymology of “recognition” back to its proto-indo-european roots and say a bit more about why this is relevant: (each of the following selections are from https://www.etymonline.com/ !)
recognition(n.)— mid-15c., recognicion, “knowledge (of an event or incident); understanding,” from Old French recognition (15c.) and directly from Latin recognitionem (nominative recognitio) “a reviewing, investigation, examination,” noun of action from past-participle stem of recognoscere “to acknowledge, know again; examine” (see recognize).
Sense of “acknowledgment of a service or kindness done” is from 1560s. Sense of “formal avowal of knowledge and approval” (as between governments or sovereigns) is from 1590s; especially acknowledgement of the independence of a country by a state formerly exercising sovereignty (1824). The meaning “a knowing again, consciousness that a given object is identical with an object previously recognized” is by 1798 (Wordsworth). The literary (especially stage) recognition scene “scene in which a principal character suddenly learns or realizes the true identity of another character” is by 1837 (in a translation from German).
recognize(v.)— early 15c., recognisen, “resume possession of land,” a back-formation from recognizance, or else from Old French reconoiss-, present-participle stem of reconoistre “to know again, identify, recognize,” from Latin recognoscere “acknowledge, recall to mind, know again; examine; certify,” from re- “again” (see re-) + cognoscere “to get to know, recognize” (see cognizance).
With ending assimilated to verbs in -ise, -ize. The meaning “know (the object) again, recall or recover the knowledge of, perceive an identity with something formerly known or felt” is recorded from 1530s.
cognizance(n.)— mid-14c., conisance, “device or mark by which something or someone is known,” from Anglo-French conysance “recognition,” later, “knowledge,” from Old French conoissance “acquaintance, recognition; knowledge, wisdom” (Modern French connaissance), from past participle of conoistre “to know,” from Latin cognoscere “to get to know, recognize,” from assimilated form of com “together” (see co-) + gnoscere “to know” (from PIE root *gno- “to know”).
Meaning “knowledge by observation or notice, understanding, information” is from c. 1400. In law, “the exercise of jurisdiction, the right to try a case” (mid-15c.). Meaning “acknowledgment, admission” is from 1560s. The -g- was restored in English spelling 15c. and has gradually affected the pronunciation, which was always “con-.” The old pronunciation lingered longest in legal use.
*gno-, Proto-Indo-European root meaning “to know.” It might form all or part of: acknowledge; acquaint; agnostic; anagnorisis; astrognosy; can (v.1) “have power to, be able;” cognition; cognizance; con (n.2) “study;” connoisseur; could; couth; cunning; diagnosis; ennoble; gnome; (n.2) “short, pithy statement of general truth;” gnomic; gnomon; gnosis; gnostic; Gnostic; ignoble; ignorant; ignore; incognito; ken (n.1) “cognizance, intellectual view;” kenning; kith; know; knowledge; narrate; narration; nobility; noble; notice; notify; notion; notorious; physiognomy; prognosis; quaint; recognize; reconnaissance; reconnoiter; uncouth; Zend.
It might also be the source of: Sanskrit jna- “know;” Avestan zainti- “knowledge,” Old Persian xšnasatiy “he shall know;” Old Church Slavonic znati “recognizes,” Russian znat “to know;” Latin gnoscere “get to know,” nobilis “known, famous, noble;” Greek gignōskein “to know,” gnōtos “known,” gnōsis “knowledge, inquiry;” Old Irish gnath “known;” German kennen “to know,” Gothic kannjan “to make known.”
Most people are stuck on linguistics and philosophy when trying to figure out what something means. The science of behavior skips over them and goes straight to the contingencies (a contingency is a consequence from response on occasion e.g. a door opens from a push on a lever) of which we now only remotely speak. It is through “The Diversions of Purely” by John Horne Tooke that Skinner and so many others have come to better appreciate the role of etymology in the analysis of verbal behavior.
The control that each part of a text has over the responses of a reader are almost entirely the result of etymology as the history of the characteristic consequences of verbal responses of given form on characteristic occasions. We now know that the wheel, horses, and most language descend from the bronze-age peoples of the eurasian steppes (see “The Horse, the Wheel and Language” by David W. Anthony 2007). Most cultures of the world remain under the control of circumstances which are not very different from those of the bronze age people, and each word of a modern language which purports to mention or refer to a mentalistic object such as an idea or thought can be traced through to its PIE root which reveals the concrete contingencies from which we speak though we are oh so remote from such ancient speakers.
Thus, going from “recognize” to “*gno-“ and back again through “can” and “could” as “have power to”, “be able to”, or through “acquaint” as “to be near to” or “to bring near” or “intimate with”, we reach out to the concrete contingencies which control the verbal behavior that is a consequence of them. B. F. Skinner has the following to say as an introduction to “knowing”:
“We say that a newborn baby knows how to cry, suckle, and sneeze. We say that a child knows how to walk and how to ride a tricycle. The evidence is simply that the baby and child exhibit the behavior specified. Moving from the verb to the noun, we say that they possess knowledge, and the evidence is that they possess behavior. It is in this sense that we say that people thirst for, pursue, and acquire knowledge.
But this brings us at once to the question of what it means to possess behavior. … to say that a response is emitted does not imply that it has been inside the organism. Behavior exists only when it is being executed. Its execution requires a physiological system, including effectors and receptors, nerves, and a brain. The system was changed when the behavior was acquired, and it is the changed system which is “possessed.” The behavior it mediates may or may not be visible at any given moment. There are parallels in other parts of biology. An organism “possesses” a system of immune reactions in the sense that it responds to invading organisms in a special way, but its responses are not in existence until it is being invaded. It is often useful to speak of a repertoire of behavior which, like the repertoire of a musician or a company of players, is what a person or company is capable of doing, given the right circumstances. Knowledge is possesed as a repertoire in this sense.”[pg. 152 Skinner “About Behaviorism”]
And then, jumping ahead to Skinner’s section “Knowledge as Power and as Contemplation”
Much of what is called contemplative knowledge is associated with verbal behavior and with the fact that it is the listener rather than the speaker who takes action. We may speak of the power of words in affecting a listener, but the behavior of a speaker in identifying or describing something suggests a kind of knowledge divorced from practical action. Verbal behavior plays a principal role in contemplative knowledge, however, because it is well adapted for automatic reinforcement: the speaker may be his own listener. There are nonverbal behaviors having the same effect. Perceptual responses which clarify stimuli and resolve puzzlement may be automatically reinforcing. “Getting the meaning” of a difficult passage is similar. … Contemplation of this kind would be impossible, however, without a previous exposure to contingencies in which action is taken and differentially reinforced.”[pg. 155-156 Skinner “About Behaviorism”]
So it is that I return to the first sentence from Frege
“The recognition of a scientific truth generally passes through several stages of certainty.”[pg. 48]
The word “recognition” is replaced with “*gno-“ and the rest comes through a careful analysis of what people do when they are said “to know”. It is not passing through “several stages of certainty” that strengthens the possession of a scientific truth as a kind of knowledge or cognitition or even a conception (of which I expect there to be much talk later), but rather the concrete contingencies to which the knowing organism is exposed and which result in the contemplative knowing which purportedly reaffirms itself as in automatic reinforcement of a trained “knower of scientific truths”.
I shall not take so long with each sentence as I submit that most of them cluster around a few threads of etymology that bring us back to the science of behavior and then follow from there. There is much more to go yet I shall leave this here for now.
Tomorrow I may even get to the second sentence :)
My interest in the history of the world is largely the result of my interest in the history of science and technology. Technologies are concrete artifacts of cultural practices. They mediate conspicuous behaviors and are an easy access point to the study of social/cultural selection and variation.
The sewing machine is my goto example. The mechanization of the behaviors which once dominated the practices of sewing make it so that the consequences of sewing appear far more often without the direct participation of individaul sewers. Repertoires of sewing are almost extinguished as the remote control of sewing machines eliminates the contingencies that once taught the people of the world to sew.
Washing machines are perhaps better known for the labors they are said to save. While it may come easy to us to say that labors are saevd by washing machines, especially when the push of a button has the same consequences which were once the result of a long chain of responses, it is more accurate to say that a change in the environment changed prevailing behaviors and to seperate such changes from teh value judgements that are so quickly provided by a majority of our ethical practices.
The primary reason to stick with an accurate report of behavior rather than amending it with a value judgement is that the ethical practices which bring such judgements to the judge are not yet themselves the result of a conclusion from contemplation on accurate templates of the world. More than that, they are very rarely so: the science of behavior is new and has had little to no effect on prevailing ethical practices in any of the cultures of the world. It is to history that I look to see the changes in cultural practices that shaped modern ethical practices specifically and human practices generally.
Our world is now one with much more than just sewing and washing machines. We now have verbal machines that do for verbal behavior what sewing and washing machines do for sewing and washing. Fewer people are required than ever before to respond verbally in ways which once prevailed.
Computers have long since provided remote listeners (called electrical/computer engineers) with control over the behavior of computer programmers. Circuits, once elaborate networks of relays, mediate the consequences of a listerner’s response to the speech of programmers: the programmer sets up bits as a consequence of the verbal responses we call “computer programming” and the computer mediates the consequences of having exposed teh circuit designer to those bits without the circuit designer having any direct contact with the programmer as a contemporaneous listener.
This may all seem needlessly detailed or obtuse to anyone unfamiliar with the science of behavior and, specifically, the experimental analysis of behavior in laboratory environments. Many are apt to say that what I wrote above is already a functional part of the vernacular and goes without saying. But, it can not be so, especially in light of human history, where we so often find similarly authoritative proclamations about this or that kind of control over the world and our ethical responses to it. The science of behavior has had little to no effect on the practices of programming, much less on our ethical practices in response to them.
So it is that my interest in history is the systematic reevaluation of reports on history in light of the science of behavior through the experimental analysis of behavior.
R.P. (@ResonantPyre) has given me a few great book recommendations over the past few months:
I have not yet bought “The English and their History”. Wilson’s “Wandering Significance” was the fist that I bought up on R.P.’s recommendation. It was recommended as a result of some conversations we had on Quine and physics. Wilson is not the kind of writer that I’m used to: he takes a winding path around every single point rather than sticking it in your face. One of the reasons I so love B. F. Skinner’s writings is that he goes out of his way to stick your nose in everything he has to say.
“Wandering Significance” is presented as a very long essay on the origins and operations of meaningful language (though it is often much much more than that). There was very little of it that I read and am likely to repeat: specific examples in physics didn’t line up with what I had repeatedly witnessed physicists do (inside the lab and outside it). May sections felt like the long delays you see in a movie where a person is rushing to finish a homecooked meal before the jig is up. Though I may come back to it later, what little it has to say about Quine and Skinner combined with its strange outlook on the practices and behaviors of phycisists left me uninterested in reading more. That has not stopped me in the past though, so the future of this book is open.
I just got the first volume of “A history of philosophy with gaps” on “classical philosophy” by Adamson. I’ve flipped through some pages, read an introduction, and was excited to see that there were many references to primary texts that were prominently placed and mentioned. This made me very happy: I am looking for a more comlpete and comprehensive yet singularly organized outline of major philosophers and their texts and this may just be it.
As much as I enjoy clicking through a few Wikipedia articles to jog my memory when I’ve fogotten some detail I once knew, I can not recommend it for anything much more than that. This seems to especially be the case when it comes to philosophy. There’s just too much that people have said, and too little cohesion between entries. As I get through Durant’s “Story of Civilization” I’ll have much more to say about monolithic efforts that have a kind of grand unification without presumming cosmic authority (this problem was already brought up from my reading of Bourbaki which is a series built by committee but governed by logical and mathematical practices which do much of the heavy lifting when it comes to cosmic unifications and simplifications).
My readings of Frege are limited to “The Foundations of Arithmetic” and what can be found in “From Frege to Godel” (a text that any student of logic must have: it is truly indispensible). “The Frege Reader” says it covers his major writings: I still must buy his other books to read them in full.
Books are some of my favorite things.
While I had hoped to write every day without breaking my streak, it has been two days since I last wrote something here. My only excuse, if there is even a need for one (which I think there is not), is that I had some wonderful conversations with people about things I will soon be writing on (I aim to summarize two of the conversations I’ve had), but also I have been recovering from a workout injury.
For most of my life I did little more than math, physics, piano, and clarinet. As I have grown older I have made more friends than I ever imagined. Some of them are so kind and so generous with their time that they teach me new things. A few months ago I was lucky enough that a family friend had found a new way to spend his retirement by being my weightlifting coach. He spent most of his life weightlifting and competing in Highland games and he had a gym in his basement.
All of these things came together and because of what he has learned through his life I have been slowly strengthening my body in ways similar to how I have strengthend my mind. Sadly, I swung a kettlebell the wrong way and was given an a cute pain in my back which I am still recovering from. Thankfully, it has gotten better by the day and I have found it easier and easier to get into and out of chairs.
I let myself indulge in a sorta lavish relaxation all in the name of “letting my back heal”. Though I did not stop writing and reading, I did stop working here at the computer keyboard. It will be interesting to see what other gaps occur in my writings as the blizzard of life envelops me anew each day.
The signifcance of Quine’s predicate functors can not be understated by me. He showed, by way of predicate functors, that variables are not a neccessary part of predicate logic, and hence, baring the embrace of some weaker practices purporting to undermine predicate logic, no logical theory involves variables neccessarily.
Thus, given a theory and its predicate logic, the predicate functors of Quine give a way, once and for all, to rid ourselves of the lgoical import of pronomial cross reference. This eliminates, then, all problems which turn up from such cross referencing e.g. all problems of scope and bondage.
It also turns our focus back to the probelms to solve: what is the logic of the theory under investigation? In predicate (functor) logic this means settling what predicates belong to the lexicon of the theory. The requirement that the lexicon be mentioned by listing the predicates is preformally stated as “the lexicon is finite”, though the emphasis here is on the preformality of that statement otherwise the invocation of the predicate ‘is finite’ quickly demands its own logical theory.
Now, with the lexicon given, the paucity of premises is among the next steps. I say “among the next steps” because it is often through a sequence of steps that the predicates and premises of a given theory evolve from the conclusions of their implication.
It is here, at implication, where the supremacy of predicate logic is without question except for those who are wedded to deviant methods that always liven the outer realms of any inquiry.
But, let me linger here lest I be seen as some immoral logical supremacist! Thanks to Quine’s main method, implication can be explained entierly by two short sentences “for those abreast of the jargon, it is as follows. To prove that a given set of premises implies a contemplated conclusion, prove that the premises are inconsistent with the negation of that colclusion. Do so by putting the premises and the negated conclusion into prenex form and then accumulating a truth functional inconsistency by persistent instantiation of the universal and existential quantifiers, taking care to use a new variable for each existential instantiation.”[pg. 51 Quine “From Stimulus to Science”].
Whatever else anyone says they actually mean by implication can only be grasped natively by foreign import i.e. paraphrase into a logical theory as above.
This is staunch. It is also repulsive to many. I accept all such responses to such declarations as relevant to a full theory of logical and scientific behavior. What can not be accepted are methods which, e.g., break extensionality of a theory by recourse to some deviant logic: “A context is extensional if its truth value cannot be changed by supplanting a component sentence by another of the same truth value, nor by supplanting a component predicate by another with all the same denotata, nor by supplementing a singular term by another with the same designatum. Succinctly, the three requirements are substitutivity of covalence, of coextensiveness, and of identity, salva veritate. A context is intensional if it is not extensional.”[pg. 90 Quine “From Stimulus to Science”].
Without extensionality, intersubjective agreement by qualified witnesses is verbally bankrupt. It is already hard enough to grasp a theory with its necessary (but not sufficient) predicate logic. But, to do so without predicate logic is to plunge the world into the caprice of uncontemplated cooperation. There can be no verbal coordination without extensionality.
Note, none of this says that a theory is its predicate logic. Whatever more a theory may be is hard to say, in the same way that it is hard to say when a given definition of verbal behavior is appropriate.
With all of that wawa by the way, let me give you a quick path from familiar and traditional quantificational logic through Quine’s predicate abstracts (and their principle of concretion), to predicate functors so that you might more firmly grasp where I am coming from.
First, take ‘{x:Fx}y’ as short for ‘some item is {u such that u=y and some item is {x such that u=x and Fx}}’. Elsewhere I may come back and explain how it comes that the two place predicate of identity, written as ‘=’ short for ‘is identical to’, is indiscernability in that they are coextensive and indiscernability is in hand from any lexicon like those here contemplated e.g. in a lexicon with one one place prediate ‘F’ and one two place predicate ‘G’ (these are not to be taken as schematic predicate letters as are used in the remainder) ‘x is indiscernable from y’ is short for ‘Fx if and only if Fy, and each item is {u such that Gux if and only if Guy, and Gxu if and only if Gyu}’.
The English relative clause ‘who loves Dick’ and the pidgin ‘x such that x loves Dick’ are uniformly paraphrased by the predicate abstract ‘{x: x loves Dick}’ which abstracts ‘Tom’ from ‘Tom loves Dick’ by binding the free occurrence of ‘x’ in the open sentence ‘x loves Dick’ with the prefix ‘x:’ so that the predication ‘{x:x loves Dick}Tom’ concretes to ‘Tom loves Dick’: whatever can be said of a thing can be said by predicating a predicate of it i.e. predicational completeness.
Predicate abstracts as the logical import of the relative clause are the import of pronomial reference and respective problems of freedom and bondage.
The principle of concretion comes along for free as follows (assuming either that the lexicon of the language is finite or there is a predicate of identity provided subject to the following familiar constraints in either case). Godel’s premises of identity are 1) identity is reflexive, i.e. ‘x=x’, and 2) identity is substatutive, i.e. each instance of the schema of substitutivity ‘x=y and Fx, only if Fy’. They are equivalent to Wang’s premises, i.e. each instance of Wang’s schema ‘Fx iff some item is {y such that x=y and Fy}’. By two instances of Wang’s schema, ‘some item is {u such that u=y and Fu} iff some item is {u such that u=y and some item is {x such that u=x and Fx}}’ and ‘some item is {u such that u=y and Fu} iff Fy’, ‘{x:Fx}y iff Fy’ is true.
In a truth functional and quantificational logic of predicates without a predicate of identity ‘{x:Fx}y’ is introduced by the schema of concretion ‘{x:Fx}y iff Fy’. Concretion ties relative clauses as predicate abstracts into a truth functional and quantificational logic of predicates.
Predicate abstracts of many places are defined with the help of ‘..x’ for ‘x.0 x.1 .. x.length(x)-1’ where ‘x’ is a list of variables and ‘x..y=u..v’ for ‘x=u and ..y=..v’ (the case where ‘y’ and ‘v’ are of different lengths is handled by ‘..x=..y’ for ‘x.0..x.min(len(x),len(y)) = y.0..y.min(len(x),len(y))’ so that ‘{..x:F..x}..y’ for ‘some item is {..u such that ..u = ..y and some item is {..x such that ..u = ..x and F..x}}’ with the schema of generalized concretion following as ‘{..x:F..x}..y iff F..y’.
Predicate abstracts like ‘{xy:Fxy}’ are predicates just like ‘father (of)’ or ‘older than’. They are not items like sets or relations. It so happens that extensionality along with a few meager assumptions on the existence of sets– e.g. that they yield a serviceable theory of ordered pairs– often permits predicate abstracts to moonlight as designators of abstract items like sets. Nothing like that is found here, but shall eventually be found elsewhere when I eventually get to that.
(Note that the present methods do not incorporate the stack based notation that I’ve recently uncovered. I have yet to carry that efficiency through this sequence of arguments.)
The basic recombic predicate functors can now be introduced as abbreviations for the appropriate predicate abstract:
The basic logical predicate functors are then
Now for the predicate functors that are usually introduced as part of set theories:
Some of the predicate functors from the theory of relations
therefore ‘equiv =’ is true. The next functors bridge something like the gap between “the relational part of a set” and their origins in predicate logic:
Unlike relations, the complement ‘-F’ of a two place predicate ‘F’ is a two place predicate, and ‘-F’ is equivalent to ‘(-F)`’, but where predicates come to do doulbe duty as names of their extension, the complement of a relation need not be the two place part of its complement. This is a tiny example of the way in which theories like the calculus of relations obscure their (stronger) logic. With the generalization of the k place part comes the generalization of composition. Is it possible to introduce the k place part from kplace composition cross products and confinements?
component confinement follows from iterated left and right confinements. Together a left and right confinement is a sliced confinement
The following definitions introduce functional predicates and all that is often said to follow from theories of functions. Here, generalized composition is avoided, and, hence, the predicates operated on by the predicate functors are now two-place predicates.
The various theorems about functions then turn out to be applications of schema of functional predicates e.g. ‘Func void’, ‘Func =’, ‘(Func F) and (Func G), only if (Func (F of G))’, and ‘(Func F) only if (Func (F and G))’ Talk of correlations comes up often and is an application fo the logic of correlative predicates:
Then similarities are relative to a corraltive predicate
Somewhere around this point in my exploration of predicate functor logic it became clear that what goes by the name ‘naive set theory’ is neither naive nor set theory: it is just logic, pure predicate logic. The notational abbreviations afforded by predicate functors reveal this This part of my exploration is where I saw the damage done by Tarski’s success e.g. with set theoretic simjulations of logic and with comingling predicate logic and the calculus of relations. The entire affair is summed up by the difference between ‘denotes’ and ‘designates’.
I’ve still avoided introducing the singleton of a predicate. It placys the part of singular terms in a pur predicate (functor) logic. It’s just russell’s theory of descriptions adapted to these variableless methods.
Singular predicates play the part once played by singular terms.
But, it is not a singular predicate which exactly plays the part once played by singular terms, it is another predicate which, for now, I call ‘the singleton’
these carry over a lot of what went as finite set theory
Now the components of a predicate come through the image of the singleton (this is then an operation that generalizes ‘application and the ‘collective relate’)
The arguments of a two place predicate are those items of its second component which are true of that predicate with respect to only one first component. The definition is beautiful
Just as there is a serviceable theory of identity embedded in any logic with only finitely many predicates– the universal closure of the conjunction of each biconditional of each parallel permutation of primitive predicates– so to is there a serviceable theory of denotation.
Much earlier, i defined the left and right confinement from a predicate of equality. That was unnecessary. Those definitions become logically equivalent to the following when a predicate of equality is present.
Descending from the beautiful definition of argument of a predicate is the notorious method of function abstraction (aka lambda abstraction)
note ‘lambda lambda F’ is equivalent to ‘lambda F’. At first this seemed like an error: we can function (lambda) abstract again and again can’t we? Then it dawned on me: this theorem covers degenerate lambda abstractions in the classical case. An example of a degenerate lambda abstraction in the classical case is ‘lambda x lambda x f’. A carefully designed grammar eliminates any such degenerate ‘expression’. The equivalence of ‘lambda^2 F’ to ‘lambda F’ suggests a collapse to ‘lambda x f’ i.e. a rule of evaluation.
Happily, ‘func lambda F’ is true i.e. the function abstraction of a predicate is a functional predicate.
From function (lambda) abstraction, the problem is posed as to what definition of ‘function application’ fits. I’m torn between two alternatives. The one assumes function application is always going to occur on a functional predicate:
The traditional definition of function application (with “F’x” as ‘the y such that Fyx’) dissolves into my prior definition of ‘component’ which generalizes function application and the collective relate.
All this being said, there is a lot of work to be done to bring these things together. But that is somewhat secondary to the larger problem to solve.
My methods of logic are to be pure predicate logic: predicates and predicate functors. The challenge is to make a story of logic which is autonomous from quantificational logic. Another part of the challenge is to free the world from the confines of Tarski’s theory fo models which too strongly weds logic to mathematical logic and mathematical logic to theories of sets.
The prevailing solution has been to abandon predicate logic for deviant logics and deviant foundations which purport to be unbeholden to the constraints of, e.g., smooth discourse, e.g., fascilitated by extensionality. In such methods I simply see theories from theoraticians who have yet to present their predicates and premises.
Category theorists and type theorists are the most familiar of the gang as far as I know. Some computational theorists enter upon this territory, but, e.g., automata have yet to present themselves as foundational instruments in the same way as categories and types have. The closest is calculi, but there is not yet a general flag underwhich they fly (I’d propose Fefermann’s Finitary Inductively Presented Logics if pressed).
I had to stop my work on the latest little lisp entry in order to make note of a profoundly interesting interview that was just shared with me. Australian identical twins Bridgette Powers and Paula Powers were recorded after they witnessed a car crash that turned into something more. What is interesting is that they speak in a way that presumably makes it easy to hear when each of their verbal responses is strong or weak.
When they are speaking “in sync” they are emitting, what I can only guess is, the same operant, the same unit of response. When they speak “out of sync” you get to listen to the verbal behavior that is presumably equally strong but which, if equally strong within the same skin, could not be emitted at the same time.
I do not have time now to do more research on them, but I submit that they have been interviewed prior to this most recent event.
Today was a hammond b3 organ jazzy kinda day.
This continues my work on my little lisp from 202504201615.
and this is a reminder that my aim is to release a method of logic programming from Quine’s main method by implementing each step to it he makes in “Methods of Logic 4th Edition”.
In the last entry I did not give any examples of the functions defined running because I have not yet checked if they even work in such example cases. First, a complete summary of the code that I’m working with (eventually each of the suprisingly disparate entries will come together under a unified collection of verbal practices which, ideally, approach something like Leibniz calculus ratiocinator and characteristica universalis, but I’ll have more to say on that later).
let run=code=>console.log(code,'\n',eval(code)) // for examples
// some basic lisp functions
, consOf = (car,cdr) => [car,cdr]
, carOf = cons => cons[0]
, cdrOf = cons => cons[1]
, nil = 'nil'
, isIdentical = (x,y) => x==y
, isNil = x => isIdentical(x,nil)
, isPair = x => Array.isArray(x)
, isAtom = x => !isPair(x)
// proper and dotted lists
, isProperList = x => isNil(x) || (isPair(cdrOf(x)) && isProperList(cdrOf(x)))
, isDottedList = x => !isProperList(x)
// some basic javascript string functions
, emptyString=''
, isIdenticalString = (x,y) => x==y
, isEmptyString = string => isIdenticalString(string, emptyString)
, concatenationOf = (...strings) =>
strings.length ? strings.shift() + concatenationOf(...strings) : emptyString
, firstCharOf = string => isEmptyString(string) ? emptyString : string[0]
, restCharsOf = string => isEmptyString(string) ? emptyString : string.slice(1)
, isString = x => 'string' == typeof x
// basic rune functions
, runeMark = '^'
, isRuneMark = x => isIdenticalString(x,runeMark)
, isRune = x => isString(x) && isRuneMark(firstCharOf(x))
, isRunic = list => isNil(list) || (isRune(carOf(list)) && isRunic(cdrOf(list)))
// from strings to runic lists and back again
, runeOf = char => concatenationOf(runeMark,char)
, runicListOf = string =>
isEmptyString(string) ? nil
: consOf(runeOf(firstCharOf(string)), runicListOf(restCharsOf(string)))
, charOf = rune => restCharsOf(rune)
, stringOf = runicList =>
isNil(runicList) ? emptyString
: concatenationOf(charOf(carOf(runicList)), stringOf(cdrOf(runicList)))
// prepending proper lists clears the way for buliding
// runic lists from runic lists
, prependedProperListOf = (properList1, properList2) =>
isNil(properList1) ? properList2
: consOf(carOf(properList1)
,prependedProperListOf(cdrOf(properList1), properList2))
// how to make atoms and dotted lists proper lists
, singletonListOf = x => consOf(x,nil)
, properListOf = x =>
isNil(x) ? nil
: isAtom(x) ? singletonListOf(x)
: consOf(carOf(x), properListOf(cdrOf(x)))
// generlization of prepepending proper lists to atoms and lists
, prependedListOf = (x,y) =>
prependedProperListOf(properListOf(x),properListOf(y))
, appendedListOf = (x,y) => prependedListOf(y,x)
// printing atoms i.e. symbols
, isSymbol = x => isString(x)
, atomicPrintOf = atom =>
isNil(atom) ? atomicPrintOf('()')
: isSymbol(atom) ? runicListOf(atom)
: atomicPrintOf('!?');
Now, examples for the latest functions defined in the latest entry
isRunic(prependedListOf(runicListOf('this is a test'),runeOf('!')))
true
stringOf(prependedListOf(runicListOf('this is a test'),runeOf('!')))
this is a test!
stringOf(atomicPrintOf(nil))
()
stringOf(atomicPrintOf('x'))
x
stringOf(atomicPrintOf(3))
!?
That solves the problem of printing atoms where atoms in my little lisp are each symbols (notice that the last example gave ‘!?’ as a default when the atom printer doesn’t know how to print the item which purports to be an atom: in a different design I have no symbols, but that is closer to the little uhdForth-like language I’ll show off after I get this little lisp done).
Next is the problem of printing proper lists and dotted lists. First I’ll work on proper lists because that seems simpler. When a proper list gets printed it starts with ‘(‘ and ends with ‘)’. Everything between those parenthesis has to be printed by the main printer that dispatches the appropriate kind of printer on the argument given i.e.
let printOf = x =>
isAtom(x) ? atomicPrintOf(x)
: isProperList(x) ? properListPrintOf(x)
: isDottedList(x) ? dottedListPrintOf(x)
: atomicPrintOf('!?')
, properListPrintOf = properList =>
prependedListOf(runeOf('(')
, prependedListOf(properListPrintEachOf(properList), runeOf(')')));
Dotted lists can wait. The problem is reduced to printing each of the items in the proper list.
let properListPrintEachOf = properList =>
isNil(properList) ? nil
: prependedListOf(runeOf(' ')
, prependedListOf(printOf(carOf(properList))
, properListPrintEachOf(cdrOf(properList))));
As much as I do not like these long and camel case names, it is all I have for now. With these definitions we should be able to test out printing everything but dotted lists.
Oh! But, it just occurred to me that there is a special case of proper list that must be handled: runic lists! There is also another special case that just occurred to me: the runes of ‘(‘ and ‘)’ must be distinguishable from the parenthesis printed around the printings from the list!
let openParenthesis = runeOf('(')
, closeParenthesis = runeOf(')')
, isParenthesis = x => isIdentical(x,openParenthesis) || isIdentical(closeParenthesis);
atomicPrintOf = atom =>
isNil(atom) ? atomicPrintOf('()')
: isParenthesis(atom) ? prependedListOf(runeOf('^'),runicListOf(atom))
: isSymbol(atom) ? runicListOf(atom)
: atomicPrintOf('!?');
let runicListPrintOf = runicList =>
prependedListOf(runeOf("'")
, prependedListOf(runicList,runeOf("'")));
printOf = x =>
isAtom(x) ? atomicPrintOf(x)
: isRunic(x) ? runicListPrintOf(x)
: isProperList(x) ? properListPrintOf(x)
: isDottedList(x) ? dottedListPrintOf(x)
: atomicPrintOf('!?');
Now for some much needed examples:
stringOf(printOf(nil))
()
stringOf(printOf('thisIsABigAtom'))
thisIsABigAtom
stringOf(printOf(runicListOf('this is a runic list')))
'this is a runic list'
So far so good. Next to test printing proper lists. A much needed function for making lists is added to the mix.
let listOf = (...x) => x.length ? consOf(x.shift(), listOf(...x)) : nil
The function designated by ‘listOf’ takes any number of arguments and returns a proper list each item of which is the appropriate argument. Thankfully, I ran into an error when I tried to run
stringOf(printOf(listOf(nil,'thisIsABigAtom',runicListOf('this is a runic list'))))
It said that ‘dottedList’ was not defined. My definition of proper list was missing a clause! Here is the corrected version:
isProperList = x => isNil(x) || isNil(cdrOf(x)) || (isPair(cdrOf(x)) && isProperList(cdrOf(x)));
Another crash and another correction! The code for detecting parenthesis was missing an argument:
isParenthesis = x => isIdentical(x,openParenthesis) || isIdentical(x,closeParenthesis);
Now for a big example that almost works exactly as expected:
stringOf(printOf(listOf(nil,'thisIsABigAtom'
,listOf(closeParenthesis, 'x', openParenthesis)
, runicListOf('this is a ) runic ( list'))))
( () thisIsABigAtom ( ^^) x ^^() 'this is a ) runic ( list')
Everything in there was expected except for the double ‘^^’ in front of the parentheses. Because I handled runic lists by wrapping them in single quotes I avoided the problem of having to prepend an extra escape character to atoms that begin with them: the function designated by ‘runicListOf’ escapes the escape character as expected! So a quick correction:
atomicPrintOf = atom =>
isNil(atom) ? atomicPrintOf('()')
: isSymbol(atom) ? runicListOf(atom)
: atomicPrintOf('!?');
and the big example becomes
stringOf(printOf(listOf(nil,'thisIsABigAtom'
,listOf(closeParenthesis, 'x', openParenthesis)
, runicListOf('this is a ) runic ( list'))))
( () thisIsABigAtom ( ^) x ^() 'this is a ) runic ( list')
which is very satisfying, but there is a space missing before a closing parenthesis:
let spaceRune = runeOf(' ');
properListPrintOf = properList =>
prependedListOf(openParenthesis
, prependedListOf(properListPrintEachOf(properList)
,prependedListOf(spaceRune,closeParenthesis)));
The big example is then
stringOf(printOf(listOf(nil,'thisIsABigAtom'
,listOf(closeParenthesis, 'x', openParenthesis)
, runicListOf('this is a ) runic ( list'))))
( () thisIsABigAtom ( ^) x ^( ) 'this is a ) runic ( list' )
All that is left is to deal with dotted lists. They seem to be just like proper lists but instead of checking whether the relevant item is the item designated by ‘nil’ check if the relevant item is an atom:
let dottedListPrintOf = dottedList =>
prependedListOf(openParenthesis
, prependedListOf(dottedListPrintEachOf(dottedList)
,prependedListOf(spaceRune,closeParenthesis)))
, dotRune = runeOf('.')
, dottedListPrintEachOf = dottedList =>
isAtom(dottedList) ?
listOf(spaceRune, dotRune, spaceRune, atomicPrintOf(dottedList))
: prependedListOf(spaceRune
, prependedListOf(printOf(carOf(dottedList))
, dottedListPrintEachOf(cdrOf(dottedList))));
Now for the simplest example to check for (hopefully) easy to fix bugs:
stringOf(printOf(consOf(closeParenthesis, openParenthesis)))
( ^) . ^(,nil )
Oh, wow, what an interesting mistake. Everything works right up to the space after the dot, but then something goes wrong. That thing is also something that often goes wrong when working with lists: prepending is not the same as listing out. Prepending gets rid of trailing empty lists when putting each of its arguments together. This should correct things.
dottedListPrintEachOf = dottedList =>
isAtom(dottedList) ? prependedListOf(
listOf(spaceRune, dotRune,spaceRune), atomicPrintOf(dottedList))
: prependedListOf(spaceRune
, prependedListOf(printOf(carOf(dottedList))
, dottedListPrintEachOf(cdrOf(dottedList))));
The example:
stringOf(printOf(consOf(closeParenthesis,openParenthesis)))
( ^) . ^( )
Great! Now to put everything together into one really big example:
stringOf(printOf(consOf(listOf(nil,'thisIsABigAtom'
,listOf(closeParenthesis, 'x', openParenthesis)
,runicListOf('this is a ) runic ( list')))),'x')))
( ( () thisIsABigAtom ( ^) x ( ^) . ^( ) ^( ) 'this is a ) runic ( list' ) . x )
Wonderful. That’s a big problem solved. The next problem to solve is the reader: it must take strings like the ones displayed from the printer and build the appropriate lists and atoms.
Since there was a lot of editing along the way, here are all the function definitions that were finally accepted. I changed some names in order to avoid collisions e.g. ‘listOf’ by ‘properListOf’ and ‘properListOf’ by ‘properOf’ and now ‘listOf’ designates a function that can be used to produce both dotted and proper lists depending on whether the last argument is the empty proper list or some other atom. Everything below was checked on the latest big example.
let run=code=>console.log(code,'\n',eval(code)) // for examples
// some basic lisp functions
, consOf = (car,cdr) => [car,cdr]
, carOf = cons => cons[0]
, cdrOf = cons => cons[1]
, nil = 'nil'
, isIdentical = (x,y) => x==y
, isNil = x => isIdentical(x,nil)
, isPair = x => Array.isArray(x)
, isAtom = x => !isPair(x)
// proper and dotted lists
, isProperList = x => isNil(x) || isNil(cdrOf(x)) || (isPair(cdrOf(x)) && isProperList(cdrOf(x)))
, isDottedList = x => !isProperList(x)
// some basic javascript string functions
, emptyString=''
, isIdenticalString = (x,y) => x==y
, isEmptyString = string => isIdenticalString(string,emptyString)
, concatenationOf = (...strings) =>
strings.length ? strings.shift() + concatenationOf(...strings) : emptyString
, firstCharOf = string => isEmptyString(string) ? emptyString : string[0]
, restCharsOf = string => isEmptyString(string) ? emptyString : string.slice(1)
, isString = x => 'string' == typeof x
// basic rune functions
, runeMark = '^'
, isRuneMark = x => isIdenticalString(x,runeMark)
, isRune = x => isString(x) && isRuneMark(firstCharOf(x))
, isRunic = list => isNil(list) || (isRune(carOf(list)) && isRunic(cdrOf(list)))
// from strings to runic lists and back again
, runeOf = char => concatenationOf(runeMark,char)
, runicListOf = string =>
isEmptyString(string) ? nil
: consOf(runeOf(firstCharOf(string)), runicListOf(restCharsOf(string)))
, charOf = rune => restCharsOf(rune)
, stringOf = runicList =>
isNil(runicList) ? emptyString
: concatenationOf(charOf(carOf(runicList)), stringOf(cdrOf(runicList)))
// how to make lists of lists and atoms
, listOf = (...x) => x.length > 1 ? consOf(x.shift(),listOf(...x)) : x.shift()
// how to make proper lists of lists and atoms
, properListOf = (...x) => x.length ? consOf(x.shift(),properListOf(...x)) : nil
// prepending proper lists clears the way for buliding
// runic lists from runic lists
, prependedProperListOf = (properList1, properList2) =>
isNil(properList1) ? properList2
: consOf(carOf(properList1)
,prependedProperListOf(cdrOf(properList1), properList2))
// how to make atoms and dotted lists proper lists
, singletonListOf = x => consOf(x,nil)
, properOf = x =>
isNil(x) ? nil
: isAtom(x) ? singletonListOf(x)
: consOf(carOf(x), properOf(cdrOf(x)))
// generlization of prepepending proper lists to atoms and lists
, prependedListOf = (x,y) =>
prependedProperListOf(properOf(x),properOf(y))
, appendedListOf = (x,y) => prependedListOf(y,x)
// printer dispatch
, printOf = x =>
isAtom(x) ? atomicPrintOf(x)
: isRunic(x) ? runicListPrintOf(x)
: isProperList(x) ? properListPrintOf(x)
: isDottedList(x) ? dottedListPrintOf(x)
: atomicPrintOf('!?')
// how to print atoms i.e. symbols
, isSymbol = x => isString(x)
, atomicPrintOf = atom =>
isNil(atom) ? atomicPrintOf('()')
: isSymbol(atom) ? runicListOf(atom)
: atomicPrintOf('!?')
// how to print runic lists
, quotationMark = runeOf("'")
, runicListPrintOf = runicList =>
prependedListOf(quotationMark
, prependedListOf(runicList,quotationMark))
// how to print proper lists
, openParenthesis = runeOf('(')
, spaceRune = runeOf(' ')
, closeParenthesis = runeOf(')')
, isParenthesis = x => isIdentical(x,openParenthesis) || isIdentical(x,closeParenthesis)
, properListPrintOf = properList =>
prependedListOf(openParenthesis
, prependedListOf(properListPrintEachOf(properList)
,prependedListOf(spaceRune,closeParenthesis)))
, properListPrintEachOf = properList =>
isNil(properList) ? nil
: prependedListOf(runeOf(' ')
, prependedListOf(printOf(carOf(properList))
, properListPrintEachOf(cdrOf(properList))))
// how to print dotted lists
, dottedListPrintOf = dottedList =>
prependedListOf(openParenthesis
, prependedListOf(dottedListPrintEachOf(dottedList)
,prependedListOf(spaceRune,closeParenthesis)))
, dotRune = runeOf('.')
, dottedListPrintEachOf = dottedList =>
isAtom(dottedList) ? prependedListOf(
listOf(spaceRune,dotRune,spaceRune,nil),atomicPrintOf(dottedList))
: prependedListOf(spaceRune
, prependedListOf(printOf(carOf(dottedList))
, dottedListPrintEachOf(cdrOf(dottedList))));
This continues my read of Will Durant’s “The Story of Philosophy” from 202504142055.
The first chapter on Plato is divided into ten sections:
The first section “The Context of Plato” begins with a geological description:
“If you look at a map of europe you will observe that Greece is a skeletonlike hand stretching its crooked fingers out into the Mediterranean Sea.” [pg. 5 Durant “The Story of Philosophy”]
Geology selects geography which plays a large part in the survival of species, be they human or not. Durant begins with a geological description that has not changed much since the time of ancient Greece from 1100 BCE to 146 BCE. We are connected to the shape of greece on which Socrates and Plato lived out their entire lives.
He then describes the surrounding geographies and narrows his view quickly upon the “city-states”:
“Greece was broken into isolated fragments by these natural barriers of sea and soil; travel and communication were far more difficult and dangerous then than now; every valley therefore developed its own self-sufficient economic life, its own sovereign government, its own institutions and dialect and religion and culture. In each case one or two cities, and around them, stretching up the mountainslopes, an agricultural hinterland: such were the “city-states” of Euboea, and Locris, and Aetolia, and Phocis, and Boeotia, and Achaea, and Argolis, and Elis, and Arcadia, and Messenia, and Laconia— with its Sparta, and Attica— with its Athens.”[pg. 5 Durant “The Story of Philosophy”]
Having arrived at Sparta and Athens, the Greco-Persian Wars enter the scene: “Sparta provided the army and Athens the navy.”[pg.6] Athen’s navy became a merchant fleet and new cultural practices flowed through the city’s ports. Finally, Durant has made his way to philosophy:
The Sophists were
“traveling teachers of wisdom, who looked within upon their own thought and nature, rather than out upon the world of things.”
“They asked questions about anything; they stood unafraid in teh presence of religious or political taboos”
“In politics they divided into two schools. One, like Rousseau, agrued that nature is good, and civilization bad; that by nature all men are equal, becoming unequal only by class-made institutions: and the law is an invention of the strong to chain and rule the weak. Another school, like Nietzsche, claimed that nature is beyond good and evil; that by nature all men are unequal; that morality is an invention of the weak to limit and deter the strong; that power is the supreme virtue and the supreme desire of man; and that of all forms of government the wisest and most natural is aristocracy.”[pg. 6-7]
Next, the Peloponnesian war (430-400 BCE) begot the oligarichal control of Athens by Critias and his “Thirty Tyrants”. So it is that Critias “was a pupil of Socrates, and an uncle of Plato”[pg. 7]
Here are some notes that I told my self to make:
“Even the concepts of “places”, “paths”, “action”, “distance” (and for that matter, “two”) are ultimately just mathematical abstractions.
The fact that we privilege some abstractions as more “real” than others is a byproduct of familiarity, not fundamentality.” 3:27 PM - 4/19/25 @getjonwithit
“My personal definition of mathematics (i.e. “the set of all things about which one can reasonably prove theorems”) is pretty broad. If an abstraction can be made precise, then it’s mathematical.” 5:21 PM - 4/19/25 @getjonwithit
“early use of abstraction was paeolithic people creating different sets of symbols for different animal/bird species which were based around the creatures footprints. It comes up a lot in cave paintings” [1:11 AM - 4/20/25]
“other nice bit of symbolism is that arches in churches exist because early churches were tree glades and when they started to build dedicated spaces for ceremony/worship, they wanted to mimic the tree arches.
the domes are symbolic of the sun, as well”[1:49 AM - 4/20/25]
This continues the work on my little lisp from 202504181456.
Runic lists are the foreign companion to javascript’s native strings. Just as strings can be concatenated together, so too can proper lists. One proper list is prepended to an other by an appropriate sequence of pairings by the function designated by ‘consOf’
let prependedProperListOf = (properList1, properList2) =>
isNil(properList1) ? properList2
: consOf(carOf(properList1)
,prependedProperListOf(cdrOf(properList1), properList2));
An example:
isIdenticalString(concatenationOf('this is ','a test')
,stringOf(prependedProperListOf(runicListOf('this is ')
,runicListOf('a test'))))
true
A method of prepending proper lists, dotted lists, and atoms follows from transforming a dotted list into its associated proper list (which then also takes care of the atomic case):
let singletonListOf = x => consOf(x,nil)
, properListOf = x =>
isNil(x) ? nil
: isAtom(x) ? singletonListOf(x)
: consOf(carOf(x), properListOf(cdrOf(x)))
, prependedListOf = (x,y) =>
prependedProperListOf(properListOf(x),properListOf(y));
A note on speed and efficiency: they are the result of shortcuts which do not mutilate the logical consequences of the definitions given. Before such definitions are given, there are no comparisons to be made, and hence no speed or efficiency of which to speak.
Appending mirrors prepending.
let appendedListOf = (x,y) => prependedListOf(y,x);
Finally, the printer prints a list or an atom as a runic list that can be sent to the native language of javascript as a string which is then displayed. There is first the problem of transforming a symbol (which is just a native javascript string in my little lisp) into a runic list; but, surprise, we already implemented that as the function designated by ‘runicListOf’. There is only one catch, if the atom being printed is designated by ‘nil’ then print a runic list that gets displayed as ‘()’ so as not to forget what part it now plays:
let isSymbol = x => isString(x)
, atomicPrintOf = atom =>
isNil(atom) ? atomicPrintOf('()')
: isSymbol(atom) ? runicListOf(atom)
: atomicPrintOf('!?');
The readings in the fourth edition of Perry, Bratman, and Fischer’s “Introduction to Philosophy” are divided into six parts:
Each part is filled with selections from primary sources. One of my aims is to cover the breadth of these sources and to supplement that breadth with the depth of the texts from which these selections were made.
There must be no doubt as to where I am coming from when looking over the data of philosophy: they are records of responses whose origins are through reinforcing practices of verbal communities whose cultures evolved by selection from variations on cultures past. Such reinforcing practices were and are mediated by bits of behaving biology: organizations mediated by organisms mediated by organelles and so on.
There are doubts as to what specific chains of contingencies selected these organisms, these behaviors, and these cultures. The boundary between what is known and what is unknown is often, and perhaps unavoidably, blurry. Science and logic have done better than the rest of human’s practices to uncover the shape of our strongest doubts. There appears to be no better reason for them than that.
What others have to say about the world, or the worlds, enters into a mix. The soup of social behaviors is seasoned by each of us. Our favorite flavors are contingent upon the ingredients with which we are endowed.
Philosophy is a part of science. The behavior of the philosopher is behavior and is, hence, part of the science of behavior. Psychology as something other than the science of behavior is part of the speculation of philosophers. Such psychologies are a part of science as much as any pseudoscience is part of what is to be explained by a complete account of human behavior.
Just as the molecular biologist can spend their life pouring over the peculiairites of abberant corners of the ocean, so too can the experimental analyst probe the behaviors and practices of prescientific practitioners. They are, after all, the progenitors of modern science, which, if human practices are to be taken as a going concern, is the progenitor of future sciences largely unknown or, perhaps, even the progenitors of some more effective or, goodness forbid, less effective practices beyond science itself.
To read philosophy is to triangulate: where is what you have to say located with respect to what others have to say? How far can speaking and listening take each of us when dealing with our shared world? These are among the questions I ask, and which others have asked and answered well before me. What can be learned from listening to my self as I do those others? As an other to my self, what can be said of my speech and its relation to the speech of others better spoken than mine?
With that I turn to the texts, and I shall return here to tease out more questions, and, if I am so lucky, to find the threads of wisdom that often pose as our best tentative answers to them.
A new mistake I make with markdown to add to the first I identified in 202504171525:
I put the link inside parenthesis and then put the display text inside square brackets AFTER the parenthesis!
This continues the work on my little lisp from 202504171613.
I apologize for these fragments of progress on this particular project: I’ve decided to just get things done as they occur to me rather than go out of my way to first make them more easily explainable up front. Everything will still be completely explained, but I am certain there are better ways of writing this all out than what I have done thus far.
Runes are special atoms that are designated by javascript strings that start with the “runeMark”:
let runeMark = '^'
, isRune = x => ('string'== typeof x ) && runeMark == x[0];
Examples:
isRune('^test')
true
isRune('test')
false
A list is runic if it is proper and each of its left parts is a rune:
let isRunic = list => isNil(list) || (isRune(carOf(list)) && isRunic(cdrOf(list)));
Runic lists shall paly the part of strings in my little lisp (for now because I dont’ know what consequences may come to select a new design). This follows the convention of Paul Graham’s Bel where he spoke of chars I speak of runes: for this project chars are always native javascript. The strange distinction must be made to preserve the distinction between symbols in lisp, runic lists, and javascript strings: this is something that has already tricked me and may very likely trick you as well!
I’ll write some examples for the function designated by ‘isRunic’ after I introduce some slightly edited string functions from Bit Strings and Binary Trees:
let emptyString=''
, isIdenticalString = (x,y) => x==y
, isEmptyString = string => isIdenticalString(string,emptyString)
, firstCharOf = string => isEmptyString(string) ? emptyString : string[0]
, restCharsOf = string => isEmptyString(string) ? emptyString : string.slice(1)
, concatenationOf = (...strings) =>
strings.length ? strings.shift() + concatenationOf(...strings) : emptyString
, isString = x => 'string' == typeof x
, isRuneMark = x => isIdenticalString(x,runeMark);
isRune = x => isString(x) && isRuneMark(firstCharOf(x));
This is also the first example of redefining a variable in javascript: ‘isRune’ comes to denote a function specified in language more like the idioms I’ve adopted so far with their ‘of’s and ‘is’s. The names are much longer than I’d write if I was writing just for my self, but I’m not doing that am I? Examples (I do not yet have a comprehensive method of testing, but have tested out a few different methods of testing and so far examples are good enough):
isRune(concatenationOf(runeMark,'test'))
true
isRune('test')
false
Next, the two functions that go from strings to runic lists and back again:
let runeOf = char => concatenationOf(runeMark,char)
, runicListOf = string =>
isEmptyString(string) ? nil
: consOf(runeOf(firstCharOf(string)), runicListOf(restCharsOf(string)))
, charOf = rune => restCharsOf(rune)
, stringOf = runicList =>
isNil(runicList) ? emptyString
: concatenationOf(charOf(carOf(runicList)), stringOf(cdrOf(runicList)));
Examples:
isRunic(runicListOf('this is a test'))
true
stringOf(runicListOf('this is a test'))
this is a test
Why start with strings and runic lists? Runic lists are the foreign companion to javascript’s native strings and they shall be the porthole through which the foreign items of my little lisp are displayed in our native tongue.
The first main part of my little lisp that must be made is the printer. It has to take a pair or an atom and transform it into a runic list that can then be read by our human eyes. Perhaps if we could see into the computer like superman with x-ray vision, then we wouldn’t need to worry about making a printer. But, making a printer also helps when making a reader which transforms key presses into foreign code.
For my little lisp there is only going to be one kind of atom: symbols. Runes are then special symbols: those that start with the rune mark as above. What was spoken of as ‘strings starting with the rune mark’ shall now be spoken of as ‘symbols starting with the rune mark’. While I’d rather not use the word ‘symbol’ because it leads to questions like “What does this symbol symbolize?” there is a long history of its use in lisp and I am not yet prepared to break that chain of consequences.
Here is all the code from this entry in one place:
let run=code=>{console.log(code,'\n',eval(code));}
let consOf = (car,cdr) => [car,cdr]
, carOf = cons => cons[0]
, cdrOf = cons => cons[1]
, nil = 'nil'
, isIdentical = (x,y) => x==y
, isNil = x => isIdentical(x,nil)
, isPair = x => Array.isArray(x)
, isAtom = x => !isPair(x);
run('isPair(nil)');
run('isAtom(nil)');
let runeMark = '^'
, isRune = x => ('string'== typeof x ) && runeMark == x[0];
run("isRune('^test')")
run("isRune('test')")
let isRunic = list => isNil(list) || (isRune(carOf(list)) && isRunic(cdrOf(list)));
let emptyString=''
, isIdenticalString = (x,y) => x==y
, isEmptyString = string => isIdenticalString(string,emptyString)
, firstCharOf = string => isEmptyString(string) ? emptyString : string[0]
, restCharsOf = string => isEmptyString(string) ? emptyString : string.slice(1)
, concatenationOf = (...strings) =>
strings.length ? strings.shift() + concatenationOf(...strings) : emptyString
, isString = x => 'string' == typeof x
, isRuneMark = x => isIdenticalString(x,runeMark);
isRune = x => isString(x) && isRuneMark(firstCharOf(x));
run("isRune(concatenationOf(runeMark,'test'))");
run("isRune('test')");
let runeOf = char => concatenationOf(runeMark,char)
, runicListOf = string =>
isEmptyString(string) ? nil
: consOf(runeOf(firstCharOf(string)), runicListOf(restCharsOf(string)))
, charOf = rune => restCharsOf(rune)
, stringOf = runicList =>
isNil(runicList) ? emptyString
: concatenationOf(charOf(carOf(runicList)), stringOf(cdrOf(runicList)));
run("isRunic(runicListOf('this is a test'))");
run("stringOf(runicListOf('this is a test'))");
The notation used to explain The Popr Programming Language is very much like the notation I devised for my own language, and is a clear indication that there is a convergence of notational practices that lend themselves to visualization near these concatenative methods.
It is sad that they are spoken of as “concatenative” when, as far as I can tell, this is just one way of elaborating on any principles of programming stack machines. Without the concrete contingencies of stack machines there is no concatenative programming and concatenative methods may not even end up being the most appropriate for stack machines. Again, as far as I can tell, the key to a specific method of programming is how you would write an interpreter for that programming language in that programming language.
This is not just some honorary show to LISP: it is the most accessible way to bring up the logic of the theory upon which a given method of programming is built. It forces the programmer to confront the difference between use and mention, logic and theory, and often reveals the language’s place on among models of computability.
A major conclusion from the substance of accumulating the links to papers on what is called “concatenative programming languages” is that modern writers do not know the difference between a theory and its logic. The inability to notice that the predicate functors of Quine are not combinators as in the more familiar combinatory calculi is a great sadness: both to our sciences and to our teachers.
Finally done accumulating the links that I wrote down. Now on to my little lisp.
All the links across the internet are rotting. Searches are no longer helpful: one system often says that there is nothing relevant while another quickly brings what you’re looking for to the top only to bury it later. My saddness knows no bounds. There are a few people that could break out of the environments which control them (and the world) in unhelpful ways, but they have no good reasons for doing so.
If your shit isn’t in plain text documents then it is more likely to rot than anything else. There are no easy ways to accept or deal with this: it is a kind of culture war both foreign and yet somehow familiar.
Some links to papers and presentations on concatenative programming languages and related things:
This is unlikely to be around for long: it will be too hard to find it.
There is a lot more to this one than can be got at right now. I really do hope that someone is out there (besides just the way back machine) keeping track of things like this before all is lost.
Ilia shares my unhappiness with broken links directly to Chris Barker in this one!
It is in large part because of the difficulty in assembling this list of links (and finding the sources mentioned by the documents pointed to from these linked sites) that I have largely committed to monolithic methods of writing. It is not what I would prefer, but it is also not so inconvenient when what I’m ultimately going for is an accurate record of the evolution of my verbal behavior.
This continues work on my little lisp from 2025 0415 1548.
Nil is a proper list and any pair whose right part is a pair whose right part is itself a proper list is a proper list; and a nonproper list is called a dotted list.
let isProperList = x => isNil(x) || (isPair(cdrOf(x)) && isProperList(cdrOf(x)))
, isDottedList = x => !isProperList = x;
Historically, a pair whose left part is designated by ‘x’ and whose right part is designated ‘y’ was designated by ‘(x . y)’ and ‘(x y)’ was short for ‘(x . (y . nil))’ so that, in general, ‘(x y …z)’ is schematically short for ‘(x . (y . (…z . nil)…))’. Thus, the only abbreviations which contained a dot were those whose right parts were nonnil atoms e.g. ‘(^a ^b . ^c)’ is short for ‘(^a . (^b . ^c))’ and ‘((^a ^b) . ^c)’ is short for ‘((^a . (^b . nil)) . ^c)’ where all atoms beginning with ‘^’ designate their self.
Some writing projects that I may never get around to:
Who am I kidding, I’m definitely going to write about that.
Side note. the markdown convention of using square brakets around the display text for a link which is given in curved brackets does not work: I constantly put square brackets where curved brackets go and vice versa. This seems like something that would have been quickly uncovered with a little “beta testing”. Dare I say that a solution may be found in some method of slashes?
Today is the first day that a problem I’ve had for some time was clearly revealed to me: I’m having trouble keeping up with my self. Hopefully tomorrow I can get out some more work that I did yesterday and today:
Oh! And the Durant reading was on Greece, Socrates, and Plato. To my great surprise, I just discovered that Whitman’s 1865 “O Captain! My Captain!” is specifically described as a “Ship of State” metaphor which is most famously mentioned in Plato’s “Republic” Book 6. This is something like an example of what I mean when I say, metaphorically, reading always makes the world seem larger and smaller than it once was.
I left a note for my self on twitter. This is me making the note I’d made a note for my self to note.
The account @ResonantPyre, a friend, wrote a few paragraphs on Wittgenstein’s 1935 ‘Lectures on Personal Experience’. Nonsense is under the microscope. ‘Sound and Sense’ was the title of a book on poetry that I was forced to buy to finish some requirement of my undergraduate education (little did I know what a wonderful friend I would find in that professor even though I was a fish out of water). I defaced my copy by prepending a ‘non’ at each place where ‘sense’ occurred.
Meaning, sense, semantic agreement, and all other spooks of sentences are cast out, reluctantly, with the full thrust of Ramsey’s proxy functions encapsulated by Quine’s “cosmic complement”. Indeterminancy of reference alone does not preclude propositions (the poltergeists of sentences). On occasions we fix our meaning of a sentence to outright intersubjective agreement: unus pro omnibus, omnes pro uno.
The problem with propositions— meanings made eternal— is their purported transcendence: everywhere forever, forever everywhere. A well paraphrased verbal response stands for now and does well enough without proposing propositions. We lose nothing by leaving those who are lost without them: they come back around eventually.
We go to dictionaries “to get the meaning of the word” and when we open to the right page we get on where we left off.
There’s obviously much more to this all than anything like the flob above. It’s okay to have some fun when you fill it out with fuller explanations sooner rather than later.
When I was young, and still very much under the control of Russell’s Principles and Principia, I found my way to Egner and Denonn’s “The Basic Writings of Bertrand Russell” and clung to it as an other might the primary text of their religion. Everywhere I found the methods of Principia applied to Russell’s world. That world was one I hoped to glimpse within my own.
It is only now, some fifteen years later, that I see less of Russell and more of logic in his memorabilia. It was always the logic of his works, and not their well written projections, that caught and kept me reading. The error of Russell’s ways were somewhat corrected by him. Clarity and exactness were logic disguised and upon confronting this perversion (he excised logic overtly in “An Inquiry into Meaning and Truth”) he left the fruits of philosophy for something more attractive to modern minds.
Logic looms throughout his works, but the mind left him open to mystical influence. He rested in a way somewhat more obscure than Descartes “I think therefore I am.” It is, for me, the difference between singular terms and Quine’s elimination of them by Russell’s own singular descriptions.
Expedience was never sufficient for Russell, there was always more to it than just that. Thus, even though he is so often worshipped as a controversial figure, it has gone unnoticed that public comprehension is a halmark of convention (whether such conventions are in vogue is another matter e.g. nothing sells quite like good versus good).
Russell was no real threat to the world. For all that was hoisted upon him through his personal affairs, he made a cuckold of Frege and logic got lost in the shuffle. Quine remediated what Russell ruffled. Carnap did too, but never got far enough from Russell’s outlook e.g. his philosophy of physics is limp.
What once of logic had exhausted Russell was energized by Quine: sprawling theories and unspooled speculations were cut off from their metaphysical wellsprings and reality pierced the veils of clarity and exactness. Rather than complexify the world, Quine made due with the complications of a worldclass watchmaker. But whither the warmth of love and the sensations of the soul amongst such machinations? Those who clutched for them betrayed their own insecurities.
Principia Mathematica was a report from the frontlines of the final war which we wage on to this day. It tormented him for the rest of his life: had he done more to help the world or to hurt it by having so vividly revealed the brutality of the battlefront? There seemed to be no peace nor no hope in the trenches. This would not keep the homefires burning: what else was he to do?
As is often the case, I have added some other books to read in parallel with Druant’s “The Story of Philosophy”. They are
I’m pretty sure those are the only books in my library that are directly aimed at giving a sorta story or a sorta history of philosophy as a whole. There is also the classic
that may come up in what may be a larger piece of the reading puzzle I have to solve in order to do all that I must before I die. As much as I recommend people go directly to primary texts, there is no good way to get started on the whole of a thing than by reading how someone else did it: there are some writers who can explain what they did as if it was done by an other and they are to be cherished. We’re all human, in the end, and that alone binds us into the most startling of enterprises.
It seems that people who were a part of history wish they had written down more of the history of which they were a part.
There is less to learn from what was forgotten than from what was remembered.
Time tramples all.
This occurred to me while reading John McCarthy’s 1979 “History of Lisp”.
First a summary of the work done on my little LISP from 202504112248
let consOf = (car,cdr) => [car,cdr]
, carOf = cons => cons[0]
, cdrOf = cons => cons[1]
, nil = []
, isIdentical = (x,y) => x==y
, isNil = x => isIdentical(x,nil)
, isPair = x => Array.isArray(x)
, isAtom = x => !isPair(x);
In English (that carefully distinguishes between quotations and their purported designations for reasons that shall be explained later),
== to the second (in the language of javascript) and returns the javascript item designated by ‘false’ elsewhere;== to the js item designated by ‘nil’, and returns the js item designated by ‘false’ elsewhere;From my work on Bit Strings and Binary Trees, the practice of designating a function by ending it in ‘Of’ and designating the functional representation of a predicate by beginning it with ‘is’ helps a lot when working out the logic of the programs under construction. This is also a partial explanation for all the ‘designated by’s and quotations that spell out the letters of what most people would call the names of the designated functions or objects.
I’m writing here now after having written the following section to say that it goes a little too far from talk about programming a LISP in javascript and if that’s all that you care about then you can just skip past this break to the next one. What you’re missing is a more detailed explanation of the general plan for dealing with syntax and semantics. It doesn’t get too deep, but does go even further out of bounds than I can presumably tolerate.
As much as there is the temptation to write phrases like ‘the function consOf’ or, better, ‘the function consOf’ which resorts to a different typeface in order to weakly emphasize that there is something different about the purported designatum of the so rendered phrase, this has only led to decades and centuries of mistakes and wasted efforts promulgated by untaught writers and readers.
This is especially the case when writing and reading programs e.g. the problems that so many people have with so called ‘pointer arithmetic’ is no such simple thing as a ‘bad design decision of a particular programming language’. The problems of meaning and reference, the life blood of semantics, crop up wherever languages are planted (which includes some of our most powerful social technologies). They are fundamental problems that can not be swept under the rug before the guests arrive.
When a writer leaves it to the reader to figure out what is supposed to be a quotation and what purports to be designated by such a quotation, there is only hell to pay. Even if the writer commands the reader to take on the responsibility of “keeping conventions in mind” this only works if there are other reasons for them to do so (I think of Kleene’s 1952 “Introduction to Metamathematics”).
This problem is not as unfamiliar nor as fussy as it seems e.g. classic problems of scoping in logic and in programming are fundamentally about not having resolved a concrete method of reference or meaning which deals well with syntactic and semantic agreement.
Rather than ignore the fundamental problems I’ll pick the strongest, perhaps tentative, methods that are already within my reach and take it from there: nothing is as final as finality claims to be.
I have already mentioned Quine multiple times throughout these notes and to anyone who is so unfortunate to have followed me on twitter. As much as people detest Quine’s obstinate rigor (a term that I only just discovered is already stated poetically as “Ostinato Rigore”), the articulation of his uncertainty is beyond anything else that I’ve seen.
The plan is simple: semantics breaks into reference and meaning, reference breaks into designation and denotation, and denotation
“is where the action is. It takes designation in stride, for a singular term can be recast as a predicate that happens to denote just one thing if any. The singular term ‘Boston’ designating Boston, can be reconstrued as a predicate ‘is Boston’, denoting only Boston. Anything said about Boston can be paraphrased using ‘is Boston’.”[pg. 59 of Quine’s “From Stimulus to Science”]
Quine shows just how denotation comes to be defined in consistent theories by first generalizing Tarski’s method of defining truth to denotation and then admitting a hierarchy of predicates of denotation, finite in number for a given theory of denotation. This settles, once and for all, the work of denotation in a theory which is rich enough to admit a method of semantic ascent e.g. by way of quotations.
While such a hierarchy of predicates of denotation are less than what most expect, e.g., from a comprehensive theory of truth (“Truth, one might risk being quoted as saying, is just a degenerate case of denotation.”[pg. 65 Quine’s “From Stimulus to Science”]) it is all that can be got without plunging any such theory into the classic antimonies.
Now, among the practices of programming there are those called “denotational semantics”. This is but one way of explaining what so many people speak of as the meaning of a given program. It is much easier to see how all such talk paraphrases into Quine’s methods by beginning with a programming language which drops variables from the beginning and hence eases the complexity of any subsequent definition of programmatic denotation as a projection from a logically consistent theory of denotation.
Sadly, now is not the time or place for me to complete such a project. As almost always seems to happen when working on interesting things, other interesting things come up, they are temporarily entertained for the sake of maintaining momentum, and then when they start to take over and drive at a new direction, there’s nothing left to do but drop it and return to where the enthusiasm last let off on the path you had original planned.
I won’t leave the other part of semantics hanging though: meaning. The simplest explanation of how I deal with it is to point again to Quine. This is not because I accept Quine without criticism or correction, but because there is no one who has yet combined the philosophies of Quine with the philosophy of Skinner’s radical behaviorism in the way that seems to be characteristic of my outlook and actions in regards to these matters.
In the classical story, sentences have meaning. It is what a sentence means that matters most when we speak with each other. If you don’t know what a sentence means, then the sentence is effectively worthless to you. Sentences are merely the aftereffects of meaning. In the metaphor of traditional signals processing, it is sentences that send the message but it is the meaning that is encoded or decoded from the sentence.
Skinner’s analysis of verbal behavior ends all speculation as to what potential philosophic value there can be to disputing the existence or nonexistence of meanings (or information, in the modern parlence, or propositions in more sacred realms). But, meanings die hard, and it is harder still for most to imagine a world without meanings much less to get along in one. While Skinner drops meanings outright and simply begins without them until he comes around to the verbal behavior that we otherwise said would not occur without the existence of meanings, Quine gives traditionalists a good shake.
He starts with “sameness of meaning” rather than with meaning. At least we can say when two sentences “have the same meaning” if we can not say what the meaning is, or if we might somehow get along without purporting anything more than a predicate “means the same as”. Just as in the case of denotation, I can not go through how “sameness of meaning” goes broke. The concluding outlook is this:
“I submit that intersubjective sameness of reference makes no sense, as applied to abstract objects, beyond what is reflected in successful dialogue.” [pg. 70 Quine’s “From Stimulus to Science”].
For more detail on any of these things I can point to the text quoted, oh so many times in this entry, Quine’s 1995 refinement of his 1990 talk “From Stimulus to Science”. It is a hard read, something which I have said many times before, but the value of each of its words and sentences is second only to the value of his “Methods of Logic 4th edition”.
That all being said…
Thankfully I wrote up all that thinking because, for some reason which is not clear to me, it let me see that ‘nil’ is to designate a symbol/atom and that should help me with the design of this little LISP!
Here is the code from above revised
let consOf = (car,cdr) => [car,cdr]
, carOf = cons => cons[0]
, cdrOf = cons => cons[1]
, nil = 'nil'
, isIdentical = (x,y) => x==y
, isNil = x => isIdentical(x,nil)
, isPair = x => Array.isArray(x)
, isAtom = x => !isPair(x);
This entirely avoids the problems that come with ‘nil’ designating a javascript array. So the key examples that show the critical change are:
isPair(nil)
false
isAtom(nil)
true
The ‘LISt’ in ‘LISt Processing’ is defined inductively as
Recursively we define proper lists as
let isProperList = x=> isNil(x) || (isPair(cdrOf(x)) && isProperList(cdr(x)))
, isDottedList = x => !isProperList(x);
As the second definition suggests, a list which is not proper is called dotted because of the historic practice of writing a pair whose left part is x and whose right part is y as ‘(x . y)’. This matters because when printing atoms and pairs there will be times when a dotted list must be constructed so that the person looking at the printed result can tell the difference between ‘(^a ^b)’ and ‘(^a . ^b)’.
This entry continues my read of Will Durant’s “The Story of Philosophy” from 202504132323.
The introduction hovers around a familiar analogy: science is to knowledge as philosophy is to wisdom. He says that science analyzes and philosophy synthesizes; that science takes apart what philosophy must put back together; that science gives us the power to do only what philosophy can tell us is worth doing; that science without philosophy has no value in that science without philosophy is like a fact without a feeling; that we must not forget that science is a descendant of philosophy; and that philosophy brought science into this world and can just as easily take it out of this world.
Next, he breaks apart philosophy in the traditional way, only to follow his own advice and focus on “the great men of philosophy” rather than the great fields of philosophy:
The chapters outline the path Will takes through the story of philosophy:
In the introduction to the second edition Will makes repeated apologies for what he has already apologized for in the main text: it is riddled with sentences that would have experts on the relevant topics fuming and it leaves out Eastern philosophers entirely (which he tried to correct in the first volume of “The Story of Civilization”).
There is no book on philosophy as a whole which has not failed to present that whole without holes.
Yesterday, I stumbled on https://www.game-cities.com/ by Konstantinos Dimopoulos who shared a link to Ultima and Worldbuilding in the Computer Role-Playing Game Carly A. Kocurek and Matthew Thomas Payne that can be read online for free.
It’s a short and quick read. I’m always left wanting more from the books I read on the history or philosohpy of various video games. In the end, the books that I want on my favorite games are basically literate programs that give an elaborate history of each design decision. That is not yet the kind of book that is easy to get your hands on.
I hope to come back to Dimopoulos’ site and read some of his articles on urban design in video games.
A way of eliminating variables from predicate logic which combines the methods of Quine’s predicate functors with the practices of programming in J and FORTH permits predicate letters which, initially, generalize the familiar two place predicate notation— e.g. ‘x is identical to y’, ‘x belongs to y’, and ‘x is the father of y’— by schema such as ‘xyzFstuvw’ so that each predicate has a number of left places— three in the example schema— and a number of right places— five in the example. Thus, the predicate ‘x pairs y with z’ is rendered perspicuously as ‘xPyz’. The left places are referred to as the pile of the predicate and the right places are the list of the predicate.
The predicate functors of recombination (which generalize the stack notation of FORTH to predicate logic) are
from which all others are defined e.g.
and so on. The remaining predicate functors (which generalize the tacit notation of the J programming language to predicate logic) are defined from
For more on predicate functors and the method of homogenization which permits translation back and forth between predicate functor logic and ‘predication’ logic see Quine’s “Methods of Logic 4th edition”.
Predication logic is the most appropriate name for what is unhappily called ‘first-order logic’ or otherwise called ‘predicate logic’. As I see it, with predicate functors elimination of variables, the only items in the lexicon of a logical theory are the predicates and hence ‘predicate logic’ is what might otherwise be called ‘predicate functor logic’.
There are two things forthcoming:
There are additional changes which permit a full generalization of APL and J’s functional operators that generalize e.g. matrix multiplication. These correspond to predicate functor functors: you can continue to go up and up the grammatical hierarchy and even contemplate predicate functor functor functors if you desire.
Though, it appears as if there is a limit to how high you can go and still get something helpful out of it: it seems to me that predicate functor functors are as far as you can go since after that everything else looks exactly like the first three levels of operations on the grammatical hierarchy of a logical language.
Today is a Nat King Cole day. His voice, his piano, and his trio have brought me hours of delight.
Some things I’d like to get done today:
I’ve read that Shirer’s account is ‘outdated’ or ‘narrow’ or ‘not historically accurate’. No book is historically accurate: we wouldn’t know how to check for historic accuracy even if our life dependend on it. That is a bit of an overexaggeration: it is largely the result of the metaphorical distance between our records of day to day life and works like “Schedules of Reinforcement” by Skinner and Ferster where the challenges of keeping an accurate record of behavior are uncovered in all their ugly details.
Whether Shirer’s book is adequate for an understanding of the third reich is moot: no book is going to give “the whole story” as if there was some big story in the sky that would truly enlighten us if we could just reach it. I see what I’ve read of Shirer so far as a starting point: it is so well written that you would be hard pressed NOT to finish it. Not all history books have that same luxury.
So far, I’ve read that Richard J. Evan’s “Third Reich Trilogy”:
is a much more accurate report on the contingencies that selected and ultimately extinguished the third reich.
I have so many books to read that it is hard to see how I’ll ever get to reading them all.
While skimming over John McCarthy’s 1979 “History of Lisp” two things caught my attention:
He also mentioned something about LISP having no effect on those working in recursion theory, but that now seems to be a historical hiccup.
My interest in (1) is much less than my interest in (2). I am reading the mentioned paper now: “First Order Programming Logic” by Cartwright and McCarthy in 1979.
I just finished reading it (202504140019) and am both happy and sad. My interest was primarly the result of being consumed by the methods of logic programming. But, alas, “programming logic” and “logic programming” though just one swap away from each other does not bring them together as one. That is the sadness: this paper does not not reduce pure LISP to the methods of logic programming.
What it does do though is still very important, and I will have to read this paper again before I can explain it and the many things that it brings up as they relate to logic programming: it calms the anxieties of those who continue to seek something beyond predicate logic when making proofs about programs.
Cartwright shows how to set up a theory (kind of like the traces in what Hoare first introduced as 1978 “Communicating Sequential Processes”) that transforms partial functions into total functions allowing for the introduction of an equivalence axiom that permits convenient proofs of properties of the original partial function e.g. an interpreter.
McCarthy’s contribution is a minimization schema for each partial function (this is very likely akin to the schematic method I devised for setting up a schematic theory of transitive closures of an unspecified predicate of the lexicon of the theory) that ends up being equivalent to Cartwright’s method.
Together these methods indicate clearly and exactly that no more than predicate logic is needed to reason about recursive programs, even those once as unfamiliar as interpreters/compilers.
These are happy things to know because it strengthens the general significance of my proposal that the practices of programmign are simply a subcollection of the practices of predicate logic.
The significance to this specific paper, and any others like it, is that such formalisms are no more a surprise than a formalism of predicate logic itself.
I got a copy of Will Durant’s “The Story of Philosophy” some months ago from a book fair. Though I read some of it back then, I do not recall any of it now. One of the joys of reading SO MANY fundamentally interesting things (sometimes over and over again) is that you can actually forget some of them!
Most of what I read is nonfiction, but I’ve started reading fiction recently e.g. Agatha Christe, le Carre, Chaucer, and Dante. Most of the fiction I’ve read in my life is from when I was less than fifteen years old e.g. Asimov, Pullman, Tolkein, Salinger, Wells, Orwell, and Shelly. Since I haven’t reread those books I’m uncertain whether I have forgotten them or not.
Anyway, I finished reading the introduction to Will Durant’s “The Story of Philosohpy” and had some things to say (apparently).
The success of Will’s book “The Story of Philosophy” allowed him and his wife to write “The Story of Civilization” which I have yet to finish reading (in fact, I still have yet to finish the first book in the series!). I am drawn to the Durants because they wished to see the world as a whole: to put Humpty Dumpty back together again.
Mine is the second edition and the eighth printing: copyrighted last in 1933 and published no later than 1953 by Simon and Schuster.
The introduction to this second edition echos my own outlook on our modern world: although we are technically more interconnected now than ever before, there is less wisdom among us than ever there has been. Much of our world was designed and built by a few of us. They have drained knowledge of all that makes it rich.
To be wise is to threaten the prevailing world order. The wise see beyond the shades of political esotericism through to the fundamental conflict between controllers and countercontrollers. Hope has not died, and it is still expected by a few of us that beyond the bounds of individualism and collectivism there is something like a peaceful culture: one that does the controlling and countercontrolling through us in only the most conspicuous of ways.
I finally finished my entry on Bit Strings and Binary Trees! It took longer than I expected, but I also uncovered some unexpected things along the way e.g. a quick and easy way to find shorter bit strings for a given binary tree.
Something I also learned: there is no faster way to catch bad writing than to publish it.
Something else I learned: if you don’t end a markdown list with two returns it thinks the trailing line of text is part of the last entry in the list.
A friend just introduced me to two great bits of music:
I was able to guess that Ginastera had learned from Copland. I described Ginastera as a “faster Copland”.
Seru Giran was too short: I wanted more after listening to it. It is a very comforting album that seemed to combine popular styles from both North America and England.
If I could just finish one thing before finishing another thing then I would be able to finish more things that people care about now than things they’ll care about later. This is a roundabout explanation of why I work the way that I do.
Warren Buffett wrote about selling and buying a jet in his letter to Berkshire sharehodlers in 1989. On more than one occasion I have had to explain to people that it is more than a luxury, though that is mostly what the shareholder letters show off as the big problem.
There is almost no limit to the potential value of being able to go from point A to point B around the world with tangable objects in tow. Apple’s three jets full of iPhones should help people to more easily grasp why Warren went from naming his jet the ‘indefensible’ to the ‘indispensable’.
Bytes of memory stored in a computer can be treated as strings of binary digits or as little binary trees. Let me explain what I mean while I give you a javascript program that decodes a binary tree into a bit string and encodes a bit string into a binary tree. (Most people would swap ‘encode’ and ‘decode’ in that sentence.)
Computers tend to use bits, bytes, and sometimes other more exotic words to do either place value arithmetic (binary in almost all cases, ternary in some rare cases) or to address other words of memory.
Technically they do place value arithemetic of remainders (sometimes with respect to the word size, and othertimes in less obvious ways that are otherwise more familiar as “logical operations”).
Storing addresses is an easy way to avoid having to copy words to new locations: go to where the relevant data is rather than waste the time and space to replicate it somewhere else.
Words also have another vital role to play in computers: they store commands that end up telling the computer what to do with all those addresses and numbers.
There is an alternative way of looking at words as lists of ones and zeros: each string of bits corresponds to a binary tree and each binary tree corresponds to a string of bits. It is easier to decode a binary tree into its binary string than it is to encode a binary string into its tree.
Since the binary trees contemplated here are all dirty— they are all built from pairs whose subcomponents are pairs or the nil pair— there is only one terminal condition which must be checked in any recursive definition: is this the nil tree? Before that, here is the code for building pairs from their left and right parts and for getting the left and right parts from a pair.
let pair=(leftPart,rightPart)=>({leftPart,rightPart})
, leftPartOf=pair=>pair.leftPart
, rightPartOf=pair=>pair.rightPart;
The nil pair is not to be confused with the empty pair: the empty pair has no parts, but the nil pair is a special pair (or what some might call an atom) that is identical to its left and right parts (see 2025 0411 2248 for more on the logic of atomic and empty items). The way we make an object like that in javascript is a bit esoteric:
let nil={get leftPart(){return this},get rightPart(){return this}};
Setting up getters (and setters) with javascript objects is just weird and I really only read the documentation for such things when I already really know what I want. Otherwise, you will get lost trying to make sense of the design choices that brought us the programming language of the internet. There is an alternate world where LISP was the language of the internet: that world would have been better than this one. An even better world is one where it was FORTH, but I digress.
The above definition of ‘nil’ makes it so that
nil == leftPartOf(nil) && nil == rightPartOf(nil)
true
which happens to be the defining feature of atoms in a more general setting:
let atom=x=> x==leftPartOf(x) && x==rightPartOf(x);
Sadly, we won’t be needing to talk any more about atoms. We just need to know whether a given tree is nil or not:
let isNil=x=> x==nil;
Oh, I should explain the difference between ‘dirty trees’ and ‘pure trees’ because most people are familiar enough with ‘pure trees’ and few people use the phrase ‘dirty trees’. Dirty trees are ones that include something other than pairs or the empty pair among their subcomponents. Pure trees are ones that don’t include anything other than pairs or the empty pair among their components. They are pure in more ways than one: they do not admit of talk of atoms, and if you go down one branch or another you eventually hit the empty tree (but it may take infintely long to get there in some cases, those also being cases that I would say are inappropriate).
So now we have all we need in order to talk about binary trees. Onto what we need to talk about bit strings:
let zero='.'
, one=','
, concatenate=(...strings)=> strings.length ? strings[0]+concatenate(...strings.slice(1)) : '';
There are two building blocks to bit strings: zero and one. They are to be distinct:
zero != one
true
and they can be combined by concatenation. The principles governing concatenation, like the principles governing binary trees, are actually a lot more elusive than it first seems: thankfully, if we just go with what javascript gives us there is no reason to get stuck in the theories of strings and trees, even though those can be fun places to be stuck.
The function designated by ‘concatenation’ takes any number of arguments that are strings and puts them together in the order they are taken (from left to right).
concatonate(zero,one,zero,one,one)
.,.,,
Why use ‘.’s for zeros and ‘,’s for ones? Because then the binary decoding of a tree is also the program for the stack machine that actually constructs the binary tree it decodes! But first, lets decode some trees:
let decode= tree => isNil(tree) ? zero
: concatenate(decode(leftPartOf(tree)), decode(rightPartOf(tree)), one);
In English, decode takes a tree and first checks if it is identical to nil. If it is then we already know how to decode that: it’s just a zero. If it isn’t nil then we have already made the assumption (quite a strong one really) that it must be a pair with a left part and a right part. So, all we need to know to write out the decoded tree is to
Once we have those two things we just write out the code for the left tree, followed by the code for the right tree, and cap it all off at the end with a one.
It will be shown that the decoded strings are almost always longer than they need to be: any initial zeros can be trimmed from the front of the decoded string without any loss to subsequent encoding.
Some examples
decode(nil)
.
decode(pair(nil,nil))
..,
decode(pair(nil,pair(pair(nil,nil),nil)))
...,.,,
The decode function is one-to-one (injective i.e. each string that comes out of it is only the result of decoding one and only one tree). There are some bit strings that will never be the result of such decoding e.g. ‘,’ and ‘.,’. Thus the method of decoding is not a one-to-one correspondence (bijection i.e. each tree matches to one and only one string and each string matches to one and only one tree). In general, there is a philosophic simplicity to establishing one-to-one correspondences from one universe of items to another. For more on the fruits of such labors see Quine’s 1946 “Concatenation as a Basis for Arithmetic”.
When treated as postfix notation (as in reverse Łukasiewicz notation) the ‘,’ comes to work as the function ‘pairUp’ of a stack machine. So many of the practices of programming are described by tiny stack machines that there is great value to being familiar with them perhaps without ever programming a real stack machine.
From the method of decoding it can be seen that the last ‘,’ pairs together the left and right parts of the tree. This is how the encoder works! It simulates a little stack machine and uses the bit string as the list of commands necessary to construct the decoded tree.
First, we can use a tree to simulate a stack. All we need is a way to push, pop, and maybe peek at the top of the stack. When I implement stacks with trees, I always take the top item to be the right part of the tree simulating the stack. Some people prefer to take the left part of the simulating tree as the top, and there’s nothing mathematically wrong with that (perhaps not even something logically wrong with that), but the programming language LISP has long since forced us to interpret such a setup as a list and not a stack i.e. the left part of a tree is the first item in a tree simulating a list.
let push= (stack,item) =>pair(stack,item)
, pop= stack => leftPartOf(stack)
, peek = stack => rightPartOf(stack);
If instead of allowing nil into our universe of pairs we had stuck ourselves with a pure theory of pairs, then ‘peek’ and ‘pop’ would have to be written differently e.g.
let error = message => console.log(message)
, emptyPair=[]
, isEmpty = pair => pair.length == 0
, purePop = stack => isEmpty(stack) ? emptyPair : leftPartOf(stack)
, purePeek= stack => isEmpty(stack) ? error('empty stack') : rightPartOf(stack);
Most people teach and prefer the implementation that includes error messages. They want the machine to tell them when something is wrong by anticipating how it could go wrong. My preference is for a crash based method of design: rather than go looking for errors leave as much of the design as possible out on the rim of “don’t cares”. Not only does this make your code a lot shorter (which makes it so there are fewer places where things could ever possibly go wrong), it also releases you from the anxiety of trying to catch your mistakes before you make them.
There are some standard techniques that we pick up over time for avoiding problems that have already happened again and again in the past. Some people are taught by others to avoid them and some are taught by the immediate consequences of having made such mistakes. The latter almost always teaches better than the former, but that is not the end of this general problem as to how to practice programming.
There is no universal way to detect logical errors in programs. No one is there to look over your shoulder and tell you “Hey, this isn’t actually what you wanted to make in the first place.” Though we have new machines that speak to us in such conversational tones— after having been forcefed some scrap of code— there is no part of such machines that prepares them for anything more than a world like the selecting past (but that goes a much longer way than most people once supposed, e.g., half a decade ago).
But, there is a way to entirely avoid logical errors: that is what logic itself is for! Thus, rather than make trouble for ourselves by starting with the logic of pure binary trees (where, e.g., the definition of ‘properPop’ deals clumsily with degenerate cases), we begin with a theory that admits no such road bumps even if, in degenerate cases, it gives us unfamiliar results (e.g. the definition of ‘pop’ I have adopted here gives nil when you pop nil as the simulation of the empty stack).
Back to encoding bit strings to trees! I already gave away the key to solving this problem: the bit string is a program that tells a tiny stack machine how to build the encoded tree. Thus, we go along the bits, from left to right, and when we hit a zero we push a nil onto the stack, and when we hit a one we pop the top two trees on the stack and pair them into a new one that we push back on the stack.
Unlike pairs, there are no standard ways of getting at the anatomy of a string (there is much more to say about that and about Quine’s protosyntax in general). Thus, we go with some traditional methods:
let emptyString=''
, isEmptyString = string => string==emptyString
, isZero = string => string == zero
, isOne = string => string == one
, firstOf = string => isEmptyString(string) ? emptyString: string[0]
, restOf = string => isEmptyString(string) ? emptyString : string.slice(1);
Some examples:
isEmptyString(emptyString)
true
isEmpty(concatenate(zero,emptyString,one,one))
false
isZero(firstOf(concatenate(zero,one,one,emptyString)))
true
isOne(firstOf(restOf(concatenate(zero, one, one, emptyString))))
true
concatenate(zero, one, zero) == restOf(concatenate(one, zero, one, zero))
true
Not all of those string functions are really needed in the encoder, but there’s nothing wrong with covering familiar ground when it doesn’t take you far from your path. There are two operations that our little simulation of a stack machine has to accommodate: push a nil and push a pair made from popping the top two items from the stack. The first is simple enough:
let pushNil=stack=>push(stack,nil);
It is the second that poses a few problems (but not anything too troublesome). Rather than explain it, I’ll just show the code and go from there:
let emptyStack = nil
, isEmptyStack = stack => stack==emptyStack
, topOf = stack => peek(stack)
, secondOf= stack => peek(pop(stack))
, pop2 = stack => pop(pop(stack))
, pairUp = stack => push(pop2(stack), pair(secondOf(stack), topOf(stack)));
For those familiar with stack operations in other circumstances, it may be strange to take a functional view of stacks: popping a stack returns a stack without its top, it doesn’t change the state of some stack accessible outside of the scope of the current executing function.
Now all that is left is to break the encoder into a function that gets the ball rolling with an empty stack and one that takes a string and a stack and gets to work:
let encodeHelperBeta = (string, stack) =>
isEmptyString(string) ? topOf(stack)
: isZero(firstOf(string)) ? encodeHelperBeta(restOf(string), pushNil(stack))
: isOne(firstOf(string)) ? encodeHelperBeta(restOf(string), pairUp(stack))
: encodeHelperBeta(restOf(string),stack)
, encodeBeta = string => encodeHelperBeta(string,emptyStack)
Lo, this is not the last version of ‘encode’ and that is why it is called ‘encodeBeta’. There are three things of note in the beta definition given:
Here are some examples, but because we have no way of seeing the encoded tree I have to put it through the decoder— this ends up being helpful because it should decode into a string that we could put back into the encoder to make the tree all over again.
decode(encodeBeta(concatenate(zero, zero, one)))
'..,'
decode(encodeBeta(decode(encodeBeta(concatenate(zero, zero, one)))))
'..,'
decode(encodeBeta(concatenate(zero,zero,one,zero,one)))
'..,.,'
decode(encodeBeta(concatenate(one,one)))
'...,,'
Alternatively, we could have built up the parenthetical notation familiar to most people for ordered pairs e.g. where ‘(x,y)’ is short for “the pair whose left part is x and whose right part is y”. It would also be helpful to know when two trees are equal so that we might be able to write out the construction of a tree and compare it to the encoding of its abbreviation as a bit string. While javascript comes with its own built in string functions, it doesn’t give us built in pairs or functions for working with pairs. Given what is known from Solomon Fefferman’s work on Finitary Inductively Presented Logics, it is probably a bad idea to design a language that doesn’t work with ordered pairs (and for those less theoretically minded, there’s always the conveniences of LISP to look at).
That last example is well worth looking at more closely: it took the concatenation of one with one (which is not a string that will ever come out of the decoder) and it encoded just fine and when we sent that tree through the decoder it returned the concatenation of zero, zero, zero, one, and one! So we have a much shorter way of decoding this particular tree: we can write it as ‘,,’ instead of ‘…,,’ when we have said how ‘encoderBeta’ works!
The following example shows why it is still called ‘encoderBeta’:
decode(encodeBeta(concatenate(zero, zero, zero, zero)))
'.'
Where did all those other zeros go that we concatenated together? How is it that we only got a single zero when we decoded the encoded tree?
After the encoder checks that the string is empty it peeks at the top of the stack and returns it as the result. Since there are no ones in that concatenation, there are no pairs made out of the nil trees on the stack (of which there are four before it shows us the top nil).
There’s nothing wrong with the beta encoder except that we can make a new one that doesn’t leave anything on the stack. Presumably, doing so will give us some new ways of abbreviating binary trees as bit strings! The simplest change is to pair up everything that’s left on the stack and then return that.
let pairUpEverything = stack =>
isEmptyStack(pop(stack)) ? topOf(stack)
: pairUpEverything(pairUp(stack))
, encoderHelper = (string, stack) =>
isEmptyString(string) ? pairUpEverything(stack)
: isZero(firstOf(string)) ? encodeHelper(restOf(string), pushNil(stack))
: isOne(firstOf(string)) ? encodeHelper(restOf(string), pairUp(stack))
: encodeHelper(restOf(string),stack)
, encode = string => encodeHelper(string,emptyStack);
Everything on the stack gets paired up by repeatedly applying ‘pairUp’ to the stack until there is only one item left on it (this is checked by seeing if popping the stack leaves only the empty stack behind). Some examples:
decode(encode(concatenate(zero, zero, one)))
'..,'
decode(encode(concatenate(zero, zero, zero, zero)))
'....,,,'
decode(encode(concatenate(one, zero, one, zero, zero, zero)))
'..,.,...,,,'
decode(encode(concatenate(zero, zero, one, zero, one, zero, zero, zero, one, one, one)))
'..,.,...,,,'
The last two examples suggest a simple way of getting the shorter decoded string from the longer decoded string of an encoded tree without having to guess and check encoding and decoding:
There are some more complicated javascript functions being used to define the operations of trimming and checking for occurrences of ones and zeros, but no more than what you can figure out for yourself from context clues or a quick search:
let occursIn = (string, item) => string.includes(item)
, trimLeftZeros = string => string.slice(string.indexOf(one))
, trimRightOnes = string => string.slice(0,1+string.lastIndexOf(zero))
, trim = string => trimLeftZeros(trimRightOnes(string))
, shorten = string =>
isEmptyString(string) ? '.'
: occursIn(string,zero) && occursIn(string,one) ? trim(string)
: string;
And, here are some examples:
shorten( concatenate(zero, zero, one, zero, one, zero, zero, zero, one, one, one))
',.,...'
decode( encode( shorten( concatenate(zero, zero, one, zero, one, zero, zero, zero, one, one, one))))
'..,.,...,,,'
There are a few ways that these shorter decoded bit strings help out e.g. they let us fit more trees within a single word of memory and, if we are silly enough to do so, fit multiple tiny trees within a single word of memory.
Future John here (from 202504272330): the shortener doesn’t work in the case where there is a one, a zero, and the first occurrence of a one is the last occurrence of a zero e.g. ‘00000111111’. You can fix this by checking for this exact degeneracy and then keep whichever block of identical digits is shorter.
Next I’ll come up with a way to showcase how each bit string or binary tree locates chuncks, down to words, of memory. After that it will probably be time for some bit string and binary tree arithmetic. No promises!
Here’s all the code I used to write this whole note:
let pair=(leftPart,rightPart)=>({leftPart,rightPart})
, leftPartOf=pair=>pair.leftPart
, rightPartOf=pair=>pair.rightPart;
let nil={get leftPart(){return this}, get rightPart(){return this}};
let run=code=>{console.log(code,'\n ',eval(code));}
run('nil == leftPartOf(nil) && nil == rightPartOf(nil)');
let atom=x=> x==leftPartOf(x) && x==rightPartOf(x);
run('atom(nil)');
let isNil=x=> x==nil;
let zero='.'
, one=','
, concatenate=(...strings)=> strings.length ? strings[0]+concatenate(...strings.slice(1)) : '';
run('zero != one');
run('concatenate(zero,one,zero,one,one)');
let decode= tree => isNil(tree) ? zero
: concatenate(decode(leftPartOf(tree)), decode(rightPartOf(tree)), one);
run('decode(nil)');
run('decode(pair(nil,nil))');
run('decode(pair(nil,pair(pair(nil,nil),nil)))');
let push= (stack,item) =>pair(stack,item)
, pop= stack => leftPartOf(stack)
, peek = stack => rightPartOf(stack);
let error = message => console.log(message)
, emptyPair=[]
, isEmpty = pair => pair.length == 0
, purePop = stack => isEmpty(stack) ? emptyPair : leftPartOf(stack)
, purePeek= stack => isEmpty(stack) ? error('empty stack') : rightPartOf(stack);
let emptyString=''
, isEmptyString = string => string==emptyString
, isZero = string => string == zero
, isOne = string => string == one
, firstOf = string => isEmptyString(string) ? emptyString: string[0]
, restOf = string => isEmptyString(string) ? emptyString : string.slice(1);
run("isEmptyString(emptyString)");
run("isEmpty(concatenate(zero, emptyString, one, one))");
run("isZero(firstOf(concatenate(zero, one, one, emptyString)))");
run("isOne(firstOf(restOf(concatenate(zero, one, one, emptyString))))")
run("concatenate(zero, one, zero) == restOf(concatenate(one, zero, one, zero))");
let pushNil=stack=>push(stack,nil);
let emptyStack = nil
, isEmptyStack = stack => stack==emptyStack
, topOf = stack => peek(stack)
, secondOf= stack => peek(pop(stack))
, pop2 = stack => pop(pop(stack))
, pairUp = stack => push(pop2(stack), pair(secondOf(stack), topOf(stack)));
let encodeHelperBeta = (string, stack) =>
isEmptyString(string) ? topOf(stack)
: isZero(firstOf(string)) ? encodeHelperBeta(restOf(string), pushNil(stack))
: isOne(firstOf(string)) ? encodeHelperBeta(restOf(string), pairUp(stack))
: encodeHelperBeta(restOf(string),stack)
, encodeBeta = string => encodeHelperBeta(string, emptyStack);
run("decode( encodeBeta( concatenate(zero, zero, one)))");
run("decode( encodeBeta( decode( encodeBeta( concatenate(zero, zero, one)))))")
run("decode( encodeBeta( concatenate(one, one)))")
run("decode( encodeBeta(concatenate(zero, zero, zero, zero)))");
let pairUpEverything = stack =>
isEmptyStack(pop(stack)) ? topOf(stack)
: pairUpEverything(pairUp(stack))
, encodeHelper = (string, stack) =>
isEmptyString(string) ? pairUpEverything(stack)
: isZero(firstOf(string)) ? encodeHelper(restOf(string), pushNil(stack))
: isOne(firstOf(string)) ? encodeHelper(restOf(string), pairUp(stack))
: encodeHelper(restOf(string),stack)
, encode = string => encodeHelper(string,emptyStack);
run("decode( encode( concatenate(zero, zero, one)))")
run("decode( encode( concatenate(zero, zero, zero, zero)))")
run("decode( encode( concatenate(one, zero, one, zero, zero, zero)))");
run("decode(encode(concatenate(zero, zero, one, zero, one, zero, zero, zero, one, one, one)))");
let occursIn = (string, item) => string.includes(item)
, trimLeftZeros = string => string.slice(string.indexOf(one))
, trimRightOnes = string => string.slice(0,1+string.lastIndexOf(zero))
, trim = string => trimLeftZeros(trimRightOnes(string))
, shorten = string =>
isEmptyString(string) ? '.'
: occursIn(string,zero) && occursIn(string,one) ? trim(string)
: string;
run("shorten(concatenate(zero, zero, one, zero, one, zero, zero, zero, one, one, one))");
run("decode(encode(shorten(concatenate(zero, zero, one, zero, one, zero, zero, zero, one, one, one))))");
It is late and it feels like there is a fat tire tightening around skull. My temples throb and my eyes hurt. Nothing comes from closing my eyes but the feeling of pains marching across my face. If I don’t tell myself to relax then I’ll soon be clenching my teeth until my ears start to ring. A few deep breaths may loosen my neck and shoulders, but my forehead remains scrunched.
Despite these unfavorable conditions, I have added some more to the 2025 0412 1422 Paper on Logic.
I’m making a note here because I am so happy to have a single text document that I can easily copy and paste onto a single monolithic website. This is the closest I have come to explaining myself to others in a way that is mostly accurate and mostly accessible.
For now there are only a list of timestamped entries to show you. Soon these shall evolve into more substantial blocks of templates, contemplations, and conclusions. As what I have to say reaches a critical mass, it bubbles up into a more solid and long standing brick in the foundation of my principles and practices. I am glad to share these steps in the evolution fo my behavior.
It has been a while since I’ve worked directly on my paper on logic. In the past I have had to shove what little I can into a thin thread of tweets. Now I am not stuck in a tight corner and can let my explanations relax into the more casual style that my friends and family are subjected to when I trap them in a “Johnversation”.
My outlook on logic is now almost entirely that of Quine’s (amended by Skinner’s outlook on verbal behavior). His “Methods of Logic 4th Edition” (and you need the 4th edition!) has almost split in half from how many times I’ve read it. It doesn’t help that Quine’s main method of proof for predicate logic is about half way through the book and that it is the one section I’ve probably read and reread the most.
Whereas Quine’s “Methods of Logic 4th Edition” narrowly focuses on the mechanics of logical practices, his “From Stimulus to Science” broadly applies those mechanics to, what Quine called, a “Breast Pocket Theory of the World”. As much as I disagree with the details of Quine’s theory, e.g. he mistakes something like Skinner’s experimental analysis of behavior for a dispositional stimulus-response theory, I agree with his methods.
Where Quine goes wrong, Skinner often goes right, and where Skinner goes wrong, Quine is there to clearly identify the problems if not also to provide compact solutions.
That is enough on where I’m coming from when I talk about logic. Now for parts of my paper on logic.
There are a few unique things about my methods of logic. The most important is that I aim to present predicate functor logic not as the elimination of variables from predicate logic, but as an autonomous system independent of predicate logic.
Whereas predicate logic is a consequence of grammar and truth, predicate functor logic is a consequence of grammar and denotation. Closed sentences are true, predicates denote, and closed sentences are a degenerate form of predicate: they are zero place predicates. Correspondingly, truth is a degenerate form of denotation. In predicate functor logic we can define truth of a zero place predicate as the universal (or existential closure) of the quotation of its padding denoting.
The phrases “existential closure”, “quotation”, “padding”, and “denoting” are all technical terms and I have used them as such with full knowledge that they are foreign to most people e.g. “padding” is most likely unknown to anyone who is not yet familiar with Quine’s predicate functors.
The first sentence of the paper, so far, is little more than the opening of a mystery:
Logic is the science of validity as a consequence of grammar and denotation.
It is a sad sentence, and I am unhappy with it. Simpler alternatives occur to me e.g.
Logic is a consequence of grammar and denotation.
or perhaps
Logic is that science which is the consequence of grammar and denotation.
but neither of those, nor the first, really clear the way for what is to come. Such introductory sentences tend to come much later in my writing of a paper. They are a summary of what is to come, and when what is to come has not yet come, there is nothing to concretely summarize.
The key components to the introductory sentence shall remain ‘grammar’ and ‘denotation’. They are an echo of Quine’s “Philosophy of Logic 2nd Edition” which clearly shows the ways in which logic is the sum of grammar and truth. He has a beautiful metaphor:
Logic chases truth up the tree of grammar. [pg. 35]
and
The logician talks of sentences only as a means of achieving generality along a dimension that he cannot sweep out by quantifying over objects. The truth predicate then preserves his contact with the world, where his heart is. [pg. 35]
Together these give a clear hint as to where Quine is coming from when he talks about the methods of logic. The problem of “achieving generality” by talking of sentences is a critical part of Quine’s method of semantic ascent:
The move is what I call semantic ascent: mentioning an expression by name instead of using it as a component clause. [pg. 92 of From Stimulus to Science]
In “From Stimulus to Science” Quine does a better job of showing how quotation is just one potential method of semantic ascent by showing how it can be extracted from ‘that’ clauses e.g.
[pg. 90 of From Stimulus to Science]. For most people familiar with Quine, semantic ascent is intertwined with his theory of propositional attitudes (as the examples amply indicate). But, there is nothing so grand in the method: it is a technical solution to a technical complication:
We can generalize on ‘Tom is mortal’, ‘Dick is mortal’, and so on, without talking of truth or of sentences; we can say ‘All men are mortal’. We can generalize similarly on ‘Tom is tom.’, ‘Dick is Dick’, ‘0 is 0’, and so on, saying ‘Everything is itself’. When on the other hand we want to generalize on ‘Tom is mortal or Tom is not mortal’, ‘Snow is white or snow is not white’, and so on, we ascend to talk fo truth and of sentences, saying “Every sentence of the form ‘p or not p’ is true’, or ‘Every alternation of a sentence with its negation is true’. What prompts semantic ascent is not that ‘Tom is mortal or Tom is not mortal’ is somehow about sentences while “Tom is mortal’ and ‘Tom is Tom’ are about Tom. All three are about Tom. We ascend only because of the oblique way in which the instances over which we are generalizing are related to one another. [pg. 11 Philosophy of Logic]
Quotations as a method of semantic ascent are haunted by philosophies of the past which are unable to distinguish between schematics and the items they purport to schematize. For example, “of the form ‘p or not p’” is usually written without the interior quotes, that is as “of the form p or not p”, and this is otherwise excused as an expedient to effective writing where the reader is expected to know where the invisible quotations are to be put. This has continued to confuse people who then take the occurrence of the letter pee in “of the form p or not p” as a pronoun referring to some otherwordly item called a proposition that somehow is the meaning of what a sentence like “Tom is Tom” says.
To be trapped down such dark allys of past mistakes is as silly as being stuck doing arithmetic with roman numerals. We can look back and marvel at our mistakes and laugh at the labors we now save by teaching better methods.
All of this is to say that the single sentence
Logic is the sciene of validity and validity is a consequence of grammar and denotation.
which is the first in my paper on logic is woefully inadequate as a summary of even what little I have said above and what little I have yet to say on the rest of logic.
Perhaps the proper “first sentence” is the second one
Compounding is (denotative) functional when, exclusively, each like compound denotes or each like compound does not denote, where and only where (waow), exclusively, each like component denotes or each like component does not denote. Chains are compounds compounded functionally.
The phrase “denotative functional” generalizes to predicates what is said as “truth functional” of sentences. The definition of denotational compound is clumsy with its “exclusively”s and “like”s, but the alternatives are woefully inaccurate and undermine the task at hand. Logic does not have the benefit of its fruits before they have ripened. As much as I would love to say that whether a compound denotes or not is a function of whether its components denote or not, there is no thing that can yet be spoken of as “a function”.
It is only from the methods of logic that we come to pin down the items of the world. While there are many who say that they are in touch with the world in some direct way, e.g. Russell repeatedly leaves room for singular terms so that they can be used to designate particulars of immediate experience, there is not yet a way that I have come to be in touch with the world as such others describe. It is only through their reports that I am even confronted with the problems of such seemingly personalized revelations.
There is another problem that sometimes occurs when making such definitions: do I not appear to be speaking of “exclusive alternations” when I have not said what it means to say an instance of the sentence schema “exclusively p or not p”? Is there not a circularity to my methods?
Quine was confronted with the same problem so often in the first publication of his early text “Mathematical Logic” that he had to add an appendix to help the anxious reader. There he explains the difference between “theorem” and “metatheorem” as well as “deduction” and “formal deduction”. There (and in his Methods of Logic) he is confined to philosophic methods and only indirectly mentions in Methods of Logic the scientific origins of logical practices.
I have gone out of my way to use the phrase ‘where and only where’ or ‘when and only when’ to distinguish the preformal/informal or preschematic methods from the schematic methods of logic. To explain the origin of verbal behavior with the forms “Exclusively Tom is wet or Tom is dry” is to leave the science of logic for the science of verbal behavior more generally. Just as humans spoke grammatically before speaking of grammar, they spoke logically before speaking of logic.
At this time I do not have a better explanation as to why I can use phrases that appear to have the same form as those said to be defined e.g.
Joint denials denote waow each of their components do not. Negations are self joint denials: they denote waow their component does not. Alternations are negations of joint denials: they denote waow some of their components do. Conjunctions are joint denials of negations: they denote waow each of their components do.
With all that said, I have only introduced the following sentences from my paper on logic
Logic is the science of validity and validity is a consequence of grammar and denotation.
Compounding is (denotative) functional when, exclusively, each like compound denotes or each like compound does not denote, where and only where (waow), exclusively, each like component denotes or each like component does not denote. Chains are compounds compounded functionally.
Joint denials denote waow each of their components do not. Negations are self joint denials: they denote waow their component does not. Alternations are negations of joint denials: they denote waow some of their components do. Conjunctions are joint denials of negations: they denote waow each of their components do.
Here is all that I have so far:
Logic is the science of validity and validity is a consequence of grammar and denotation.
Compounding is (denotative) functional when, exclusively, each like compound denotes or each like compound does not denote, where and only where (waow), exclusively, each like component denotes or each like component does not denote. Chains are compounds compounded functionally.
Joint denials denote waow each of their components do not. Negations are self joint denials: they denote waow their component does not. Alternations are negations of joint denials: they denote waow some of their components do. Conjunctions are joint denials of negations: they denote waow each of their components do.
Subcompounds of compounds are their self or those of their components. Substitutions of like compounds for like nonchain subcompounds are (denotative) functional. Functional substitutions of functional substitutions of compounds are functional substitutions of their self.
Compounds are (functionally)
Alternations of compounds with their negations are valid (they denote waow the compound does or its negation does, i.e. waow it does or does not, so, each functional substitution denotes). Conjunctions of compounds with their negations are nonconsistent (they denote waow their compound does and its negation does i.e. waow it does and does not, so, each functional substitution does not denote). Compounds are implied by and equivalent to their self.
Functional substitutions in
Interchanges of equivalents in a compound are equivalent to that compound (each functional substitution of a compound matches the same of its interchange, except perhaps for the same of the equivalents which otherwise denote in tandem, so each denotes waow the other does i.e. they are equivalent).
Interchanges of equivalents in
Compounds are equivalent to
I have found that it is easier to be wrong than it is to be right and that it is easier to be right after having been wrong. If there was a way to avoid being wrong then I would only act in that way. Logic helps but is slow. Sorting through the wrong bits takes more than what logic has to offer.
I see logic as part of science and see science as picking up what logic fumbles and drops. Logic works on the theoretical side of science and laboratories work on the experimental side. As I have said many times before “Libraries select theoretical practices and laboratories select experimental practices”. Science is compounded of theoretical and experimental practices.
One of the reasons I’ve allowed myself to write in this longer form is to make it easier for other people to see that I am often wrong, and that it is only from first being wrong that I get anywhere near being right. The definitions, the natures, or the essences of what is right and what is wrong are not easy to suss out.
Many people have spent their whole lives trying and failing to establish what is right and what is wrong. As much as I have a sort of optimism that if I work hard enough then I may one day uncover some piece of a complete theory of right and wrong, I am not so foolish to think that I shall do anything like what the best already have.
With that in mind, I offer these writings that are most certainly mistaken for no other reason than that they occurred to me.
I talk about behavior a lot. Does that imply that my behavior is under my careful control? No. It is one thing to know a science and another thing to have a technology in hand. The things of technology are the consequences of science, more or less. But, the knowledge of science lacks the concrete dimensions of levers, pullies, medicines, and computers. All that goes into scinece is often ephemeral: here for but a moment and gone forever, lost to our remote contact with the past through fragmentary records that soon fall apart.
Those who manufacture products, be they assemblies of machines, or chemicals, or organisms, or programs, or etc., work in ways often mistaken as scientific. They manufacture with the controls offered by technologies i.e. with the descendants of scientific practices.
The difference between science and such manufactureres is the difference between sensitivity and insensitivity. The manufacturer is deliberately insensitive to the strange variations responsible for science. This is not what most people think or say: they say that business is as much a science and an art as physics or biology. They insist that it is so in large part because the best businesses are almost always the result of some new technological consequence from science.
This does not always seem to be the case. Some businesses appear to survive despite science. Much that is “made for TV” objects to being called pseudoscience. It sells, and there is a science to sales!
These are all problems of control: can you make a person buy your product? Some say you can’t, others say you can. Some say that people only buy what is an intrinsically good product: you can’t trick your way into becoming a good business when your product is technologically sophistocated.
What’s with people who make a living as oracles? They make predictions, often asserted with the authority of a proclamation, and profit, one way or another, from betting that they’re right and ultimately being right. But, are they betting on what they think they are? Are they right when their bet pays off? There seems to be some difference between constructing bets that are verly likely to pay off and making predictions like those people assume come from our best sciences.
How do we know the future? When we speak of the future, as in predictions, what is it that we are speaking of? Can speach actually be said to be about anything? Certainly the speaking occurrs, but what good reasons do we have for saying that such speach is about this, that, or the future?
Matios Berhe suggests I read
I have now added these to my long list of “stuff to read”. Some day, sooner rather than later, I shall provide a complete list of texts in my library and the schedule on which I read them. There are two main reasons to do so:
I already have a copy of Kant’s Critique of Pure Reason that I read first when I was in high school. There are a few texts from that time which have survived long enough to find themselves on my shelves. One of the treasures from that time is Russell’s “Problems of Philosophy”. In no way did I learn all that Russell and Kant had to teach in their books when I first read them so very long ago. There is something special though about reading things that you do not yet understand.
The more I read the more I don’t at first understand, and yet, at the same time, the more I learn about the world as a whole. There is something about reading which makes the world bigger and smaller than it once seemed. When I’m not in such a metaphorical mood, I am quick to say the world is no bigger nor no smaller than it is: to say otherwise is to plunge such a theory into contradiction from which no logic prevents each and every imaginable conclusion.
People are usually scared by contradiction for the wrong reasons: this is somewhat corrected by phrases like “law of explosion”. The problem with contradiction is not really a problem: when you have a contradiction you find yourself without the benefits of a consistent theory. Some people are ok with that.
I finally have a place to put all my daft drafts without chopping them into little tweet size pieces. Why has it taken me so long to return to this long form method of writing? It’s mostly that strange feeling I get that this is like a message in a bottle sent out into an enormous sea of nothing.
Once I figured out that it wasn’t sending the message that mattered most, I warmed up to the practice of writing from my little island on the internet.
I’m implementing a little LISP. Like the LISPs of yore, there are two basic items: pairs and atoms.
Pairs split into left and right parts. Paul Graham wrote of “halves” when explaining pairs in Bel.
Whether you talk about the anatomy of a pair by “breaking it into pieces”, “splitting it into halves”, or “decomposing it into components”, there is one rule that governs the logic of pairs:
pairs are identical where and only where their left components are identical and their right components are identical.
This is otherwise know as “the law of extensionality of ordered pairs”. It is a fancy way of mentioning the precise conditions upon which to count two items in a universe of ordered pairs as identical. The single premise “each item is (a,b,c,x,y, and z such that a pairs b with c, x pairs y with z, and a is identical to x, if and only if b is identical to y and c is identical to z)” logically implies each conclusion of any theory of ordered pairs.
Now is not the time or place to get distracted by problems of logic, theories, and ordered pairs. As much as I would enjoy a full digression on the origins of logic, the extensionality of effective theories, and ordered pairs as a paradigm of philosophic investigation, those are not the problems that I am presented with here and that I have to solve now.
Back to my little LISP.
Pairs are identical where and only where their parallel components are. For now, “(x,y)” is short for “the pair whose left part is x and whose right part is y”. Thus, (x,y) is identical to (a,b) where and only where x is identical a and y is identical to b.
Since I’m coding up my little LISP in javascript I’m using the following definitions:
let cons=(x,y)=>[x,y]
, car=x=>x[0]
, cdr=x=>x[1]
These just say that ‘cons’ is the name of a function that takes two arguments and returns an array whose zeroth component is the first argument and whose first component is the second argument (note this confusing sentence is why some people prefer different methods of indexing into arrays of data). The function named ‘car’ takes one argument and returns the item at index zero of it.
If you do not already know how javascript arrays work, and how indexing into javascript arrays works, then these explanations are not helpful. Thankfully, if you do not already know how javascript arrays work, you get the joy of learning that first before coming back here.
I have not yet found a faster way to introduce programming than to present it after teaching logic up to and including Quine’s main method (of proof for predicate logic). Since more people are familiar with javascript than they are with Quine’s main method— or with the methods of logic— I shall continue to avoid what might otherwise appear as an entirely roundabout way of teaching programming from scratch.
The function named ‘cdr’ takes one argument and returns the item at index one of it.
Both ‘car’ and ‘cdr’ are names inherited from working with programmable machines (aka “computers”) during the late 1950s and early 1960s. If you want to learn more about where ‘car’ and ‘cdr’ come from, and perahsp even how to pronounce ‘cdr’ (which I say as “could-er”), then track down John McCarthy’s 1979 “History of Lisp” and give it a compassionate read.
With ‘cons’, ‘car’, and ‘cdr’ we can make pairs, get their left part, and their right part respectively. There are algebraic or equational ways of introducing these basic functions e.g.
car(cons(x,y))==x
cdr(cons(x,y))==y
'car(x)==car(y) and cdr(x)==cdr(y)' is equivalent to 'x==y'
but again this is just a distraction from the problem to solve (implementing a little lisp like language). Note, such purportedly algebraic or equational theories appear to follow entirely from pure logic when Quine’s method of schematic theories is extended to schematic theories of functional predicates (see Quine’s ‘Philosophy of Logic’ for a schematic theory of identity).
Following the conventions established in the wonderful book “The Little Schemer” by Friedman and Felleisen, cdr and car are only used in contexts where their arguments are pairs with left and right parts. There is one pair without any parts. It is unique (in more ways than one).
let nil=[]
, id=(x,y)=>x==y
, empty=x=>id(x,nil)
The item named ‘nil’ is an empty javascript array. Not all empty arrays are equal in javascript. Equality, identity, and indiscernibility are hard to deal with when you don’t start with logic. Surprisingly, ‘id’ works almost like it does in a language like Bel even though it is not obvious that it should. Remember the ‘law of extensionality of ordered pairs’? Well, it is violated by almost all programming languages.
There is a silent assumption that is among the many hidden away behind any explanation of a foreign programming language: they actually run on real machines. On most machines there are places and objects at those places. Objects in different places can’t be identical! (Think that over: identical things can’t differ in place apparently.) And yet, sometimes they are!
Pairs may have parallel components that are identical without being identical themselves. This happens when two pairs are silently assumed to be in different places but each of them places their components in the same places! There are lots of solutions to this problem of identity. Most of them are needlessly complex and overburdened with crusty philosophies. My solution is as sad as it is simple: just go with what javascript gives you.
My real solution is to go with logic: identity is indiscernibility in that any two predicates which fill out each instance of the premises of identity are coextensive and indiscernibiltiy is one such predicate (in a logical theory with a finite lexicon). For more on this see Quine’s “Philosophy of Logic”.
Atoms are not pairs (and pairs are not atoms).
let pair=x=>Array.isArray(x)
, atom=x=>!pair(x)
Is nil an atom? No, not here. Is nil an atom elsewhere? Sometimes. Early LISPs only had pairs and symbols. Symbols are often a large part of LISPs. They are the workhorses of reference: they end up sometimes playing the part of proper nouns and sometimes playing the part of pronouns. Here again, there is much to be learned from predicate logic. Here again, I must avoid dipping into such unfamiliar wells of simplicity.
The general problem of atomic items provides ample opportunity to make use of degenerate cases of common objects e.g. Quine provides for atoms in his “Set Theory and its Logic” by picking out atoms with the reflection of membership i.e. ‘x is atomic’ is short for ‘x belongs to x’. Most people avoid making convenient use of alarming degeneracies: they’re really missing out on some good fun.
Atomic items and empty items have a long and sordid history. Atoms are not supposed to have parts, and yet we are often forced to say that they do. Empty items always end up behaving like atoms in many respects— e.g. they are empty and presumably “don’t have parts”— but they almost always tend to be unique where as there tend to be many many atoms in any given theory.
The only place I have found clarity on these boundary problems is, you guessed it, logic. In a theory of ordered pairs, it is enough to provide for the existence of an ordered pair that neither has a left part nor has a right part. It is then unique since any two purportedly distinct empty pairs have identical parallel components (vacuously). Atoms are then accomodated by one of a few methods, each of which match up with Quine’s atomic methods in “Set Theory and its Logic”. An item of a theory of ordered pairs is atomic when it is identical to its components (or, just one of its components, in which case the other components allows us to distinguish between atoms with, e.g., identical left components: are these even atoms?).
Sadly neither of these principles are used in LISPs. Again, this is because there is that silent assumption working in the background of most explanations: the programming language runs on machines in our shared material world.
This is enough thinking on this for now.
I can not guarentee that there will be no spelling errors in what I write. Rarely do I spell well, and rarely do I take the time to check my spelling. Rather than check my spelling I would keep on thinking. Spell checking stops thinking except when looking up the word to be spelled spurs some better bits of writing. Definitions have the power to prompt better writing when they present short and sticky explanations. Those sensitive to spelling errors will be greatly disappointed (that is until I go back through and spell check this document, that is, if I ever get around to doing that).
I do not have much familiarity with using this software and can not be certain how what I’ve written here will appear. I am also not familiar with markdown. Will there be new lines in the final document if I put them here? The answer is no.
Does this start a new paragraph? The answer is yes.
Does this 20250411 automatically.
ok, rather than get lost in the details of editing I will just go on and start writing.