Tokenizing the Primes
Dennis Bouvard (@dennisbouvard)
September 2, 2024
Reworking the relations between currency, succession and the Stack provides a frame for returning to Anna Wierzbicka’s primes and giving them some real work to do. Wierzbicka’s discovery of a set of words which exist in all languages and that can be identified by their not being definable other than through (at some point) the use of the word itself should be far better known and I assume it’s not because it would make many discussions in which there are great stakes irrelevant. All of philosophy, for example, which finds its justification in hosting conversations over the meaning of words (and therefore comes into existence where the circulation of money and spread of literacy reach a certain threshold, along with the emergence of the problem of “tyranny”) should be put out of business: prime words simply mean what they mean, and all other words are articulations of primes which can be provided in the kinds of explication Wierzbicka has modeled extensively. Even ethical and moral questions can be addressed by exploring relations between the primes, even if Wierzbicka never seems to have explored this, perhaps because it would take her outside of the discipline of linguistics. (In general, even while invoking Leibniz’s project to develop a universal “character” as an inspiration, to my knowledge Wierzbicka has never undertaken the kind of “combinatorial” work I’m initiating here.) Constructing moral and ethical (and, for that matter, ontological and epistemological) discourses out of the primes would entail treating them as comprising a miniature language of its own with words bounded by other words in a way we could determine in ways that allow for useful conversations. What is the boundary and relation between, for example, “think” and “know”? We preface a statement with “I think” so as to tell our interlocutor that we are not claiming to “know,” which also means that saying “I know” means a certain train of thinking, presumably retrievable, has come to an end. (I’ll note here that Helen Bromhead points to important differences between the meaning of “I think” in different historical periods—so we need to keep in mind that Wierzbicka’s theory takes into account words in their collocations and not just as separate items.) Where the line gets drawn here is a source of data regarding historical change, traditions and the constitution of specific communities. Where do I draw the line between what I say has “happened” and what I say someone has “done”? Questions of morality and knowledge are implicated here. What are the protocols for saying something is “like” or the “same” as another, or that something is a “kind” of something? The vast vocabularies built over these primes, involving the permutations of these words in their interoperations, can be mapped as a series of historically weighted acts. Philosophy can be collapsed back into its origin: asking what someone means when they say something and, for that matter, reclaiming meaning as the highest purpose.
Part of my reason for returning to the primes is my growing conviction that the latter half of England’s 17th century is one of the most consequential periods in human history and by far the most consequential in modern Western history: the prosecution and replacement of the king and, albeit temporary, rejection of monarchy itself; the creation of the national bank and debt financing; the emergence of a primitive form of the “two party system”; the establishment of the preliminary form of modern science and a modern scientific community (the formation of the British Royal Society)—all of the elements of the Anglo-izing of the world that we have come to call “modernity.” Along with all this there is a fundamental transformation of the English language, which is obvious to anyone who reads something written, say, in 1620 alongside something written, say, in 1670, and which T.S. Eliot interpreted as a dissociation of sensibility, i.e, a kind of “fall.” (Augustan poet John Dryden referred someplace to his poetic predecessors as writing “before the Flood.”) Wierzbicka has herself dealt extensively with this question, with an eye at least as critical as Eliot’s, especially in Imprisoned in English and Experience, Evidence and Sense: The Hidden Cultural Legacy of English, and a student of hers, Helen Bromhead, has continued this inquiry in The Reign of Truth and Faith: Epistemic Expressions in 16th and 17th Century English. Bromhead narrows the transformation down to one from a language of “certainty” to a language of “doubt,” and shows how various words either disappeared or took on dramatically transformed uses and meanings in the process. This is an extremely helpful distinction and can be very readily mapped across all the other transformations in 17th century Britain mentioned above—and, I am proposing, can in fact enable us to anchor all those transformations in language. It’s not that the changes in language caused all those other changes—clearly, if we had to choose, the causation would be found to go in the other direction—but that language serves as the measurement of those changes, and measurement is more important than and our only path to whatever causality we’re interested in. One more thing: Wierzbicka very convincingly found the linguistic transformations she is identifying in the works of John Locke, and from a more quantitative perspective that may be true (I’m anyway in no position to challenge Wierzbicka on this) but I think it makes more sense, if we think about this linguistic shift as a paradigm shift on the model of Kuhnian scientific revolutions, that we should date its origin with Thomas Hobbes, in particular Leviathan, published, indicatively, in 1651, almost exactly mid-century and right in the middle of the English Civil War. It is Hobbes’s revolutionary proposal to use mathematical and scientific style definitions and reasoning to study human orders, along with the reduction of human action to “units” of self-preservation, self-interest and self-aggrandizement that seemed dictated by that proposal that would have massively leveraged the language in this new direction.
We can treat the primes as an idiom of its own in which we can describe the meaning of each word in its relation to other primes—that is, the primes exhaustively cover the possible semantic space and the boundaries on each word are marked by the other, proximate words. So, “think” is proximate to “know,” but also to “feel” and “say”—but, really any word can be proximate to any other word, because proximity is determined through sentences—Wierzbicka’s model presupposes that any meaning can be stated in declarative terms, demonstrating a faith in the declarative that I don’t so much share as think should be pushed to its limits. So, we say “I think..” when we cannot or do not want to say “I know…” This tells us a great deal about how we use the word “think,” and allows us to dispense with much philosophical and psychological reflection on “thinking”: thinking is not so much an internal “cognitive” process following certain rules or operating according to certain mechanisms as the marker of a hesitation and the opening of a space between “feeling” and “knowing” when “saying” something. We can bring “feel” back into play, thereby providing a link between the primes and Peirce, who insisted that the meaning of all signs ultimately registers as feeling. Feeling is firstness for Peirce, which would be the root word of any NSM explication of “presence” I would attempt. Presence is saying with some other that something is “the same” (to remain within the primes). We don’t have to say that we can say that something is the same when we also have the same feeling, or that we can ever know that such is the case but, rather, that we can only say that we have the same feeling when we are able to say that something else is the same. This is what I take to be the lesson of Wittgenstein’s refutation of the possibility of a private language, and I take the further implication that sustaining the joint attention implicit in “this is the same” would require establishing sufficient reference points whose interconnection we can say is the feeling the same thing. So, when someone says “‘I feel…’ it is like they are saying “I know we could say ‘this is the same’ many times about many things.” Given that there is no time or space limit required for the construction of presence, those many times and many things could have existed long ago or come into being in the far future and be anywhere in a particular construction of presence. Note that we are working horizontally in placing primes in sequences of declarative sentences. I have elsewhere, for example, come close to saying that knowing how doing and happening are the same—in less prime-like language, the more you can see (there is much to do with “see,” hear,” and “touch”) the world as saturated with intentions the more you can see the world as extending beyond any intentions, and vice versa. We’re working on ways of reading and writing here.
Now, the vertical dimension is how we can think about the generation of all the far more complex words and idioms that construct them along various horizontal registers by working our way from the ostensive through the imperative and interrogative and up to the declarative and back again. Wierzbicka’s explications are very scenic: they always involve some “someone” who can or wants to say that we can know or feel something if or when something is the case. I make no causal claims here, and have always been clear that there’s no reason to assume that the only words to be found in every language were the first words in any language (in fact, I consider this highly unlikely, since the primes include no word for God and, of course, no names), but we can nevertheless take any word or “chunk” in any language to be a compression of a particular array of sentenced primes and hence stacks of possible scenes operationalized by a particular word or phrase. We could not articulate this entire stack (or very much of it) in engaging with specific utterances or samples, but we can know it is there and target what we want to present as the most pertinent bit of it. We enter these compressed stacks by maximizing the horizontal relation between the primes comprising them—as suggested above, there is a moral and ethical component (“I want to say this is good”) to articulating the primes. There is always an optimal relation between what one says one feels, thinks and knows, and since all mental act verbs (and the adjectives and nominals into which they are converted) derive from these we can use the relation between these primes to inflect the words compressing their possible relations. This is a way of thinking: first, determine how a particular word or expression derives from the primes and then consider (how does “consider” derive from the primes?) how to further optimize the relations between the primes relative to some other proximal or possible utterance or sample. If, for example, “I think” is on the boundary with “I know,” then you want to direct what you can know to encroach upon what you merely think and disturb what you know with the new frontiers opened by knowledge of that regarding which we can thus far only say “I think.” And you can bring in “I feel” by testing what, at each point, it would mean to say “I feel” given what you have already said you think and know. This is a way of entering the disciplines, in their most compressed and metalinguistic forms, and converting utterances within the discipline into their scenic terms.
Now, Helen Bromhead’s identification of the transformation of the language of certainty to the language of doubt becomes important because, however we might say we dislike “modernity” it is only this transformation that gives us the means to think any of this or, for that matter, enable Wierzbicka to imagine her project. Think about (the “think” in “think about…” places “think” on the boundary with “do”) what happens when, instead of saying something like “X cannot do Y” you say something like “if X were to Y that would mean…” and you plug in there some implication of the unexpected, even impossible (but do we ever really know, and not just think, what is impossible?) then you open up a whole realm of thinking, or hypothesizing (what is hypothesizing if not thinking like you are knowing, a kind of scene of as if knowing?), of imagining possible events that would never have occurred without that linguistic shift: in the realm of certainty, turning out to be wrong (to think something that is not true—to discover that knowing overturns your thinking) must be s result of some betrayal; that is, one must saturate the space with doing, leaving no space for happening—a remnant of what Marshall Sahlins calls the “enchanted universe”—and therefore for playing thinking off against knowing or, for that matter, against the interplay of doing and happening.
We cannot restore the enchanted universe of the immanence of metapersonal command chains—no one is positioned so as to say something like “I feel that tree is saying…” in any other than a private capacity—the shared reference points, which would ultimately be grounded in ritual and sacrifice, are not there. We have to leverage the idiom of doubt cybernetically, so as to make what we say interoperable through whatever stacks of scenes with all the transactions of doings and happenings. This is what I have in mind (“have in mind”= I think+I can think this because I have said I think other things+I think this because if I think this I and others can think other things) when I speak of “tokenizing” the primes. (To hypothesize, then is to follow the command: think like you know when you know you don’t know: this is how “hypothesis” is a compression of “think” in its relation to “know.” But, of course, other explications will be possible—ultimately, all these compressions of the horizontal filed of the prime will become a space of inquiry drawing heavily upon specially trained LLMs.) I’ll return this in other posts, but let’s think about it a bit here. (Maybe Hobbes was even trying to do something like this, and we are retrieving a thread from him here.) We’re thinking here about how to make all language performative all the time—which is to say, the equivalent of ritual and magical, but predicated on doubt, or following the imperative to never be completely sure but never stop wanting to be. Doubt replaces certainty when vouching, honor and oaths are insufficient to make it possible to say “this is true,” and the only thing that can provide similar or even better support are disciplinary teams trained on trains of succession: the devotion to truth unmoved by appeals to interest relies upon “faith” in the possibility of improving succession choices by the occupant of the center; this is what all disciplinary activity comes down to.. We are already trained to read the media as producing tokens of succession—a given report on the president is a particular branch of the intelligence community trying to operationalize a particular policy or personnel shift, etc.; optimizing the horizontal spread of primes through the vertical stacks aims at saturating the field with such effects. Bring thinking closer to knowing, knowing to doing, prolong doing into happenings so as to bring happenings into proximity to thinking, test the boundary between feeling and wanting—operationalize the language as compressions of all these imperatives and in order to generate more of them.
The idea of tethering movements of different actors and programming them to reciprocally control one another goes back to the earliest days of cybernetics and practitioners like Gray Walter and Ross Ashby and is currently a major field of aesthetic inquiry: The New Normal (2021), edited by Benjamin Bratton and others, is filled with brilliant examples. The problem is to ground this practice in language so as to replace the break-up Hobbes initiated with a universe that will not so much be enchanted as fully idiomized. The meaning of any text I produce is to be the constraints it places upon text production, which is the presence it constructs; reciprocally, the constraints the text I produce imposes have further optimized the linguistic field by tethering the horizontal semantic field of the primes to ostensive>imperative>interrogative>declarative sequences. I’ll put the question this way: how can something someone says now anchor, through repetition and reference in various registers, actions to be taken in some indeterminate future? That is what is necessary for idioms to be tokenized and become currency, and to replace such guarantees as oaths as verification and authentification. It would kind of be like passwords that are fully public but only work when deployed by someone rightly positioned and knowing that and how they are rightly positioned. This would require the constant generation of idioms out of the primes and translated into other idioms: for example, thinking like you know becomes knowing that anyone can think like that: an idiom or series of idioms regarding hypotheses and disciplines and inventing ways of configuring actors so as to have new ways of saying “this is the same” emerges from these “primitive” idioms. Institutions, command structures and supply chains can be built around such idioms, like they are now built around slogans, mission statements, etc., but in ways that require constant learning. The primitive idioms meet the self-canceling idioms on the boundaries of the ritual, the juridical and the disciplinary, idioms of asymptotic convergence of debt enforcement and debt forgiveness. These are all ways of staging deferral, and placing the consummation of the want deferred on the horizon which, of course, we can never approach while turning the universe into a stack of signs of desire.