Compression

A single sample includes the whole. This sentence is advocacy for a wild form of thinking, one involving abduction and an endless range of constantly re-examined (even self-re-examining) hypotheses. Even when you have lots of samples, you’re always hypothesizing from a single one of them insofar as there is always a privileged sample through which you compute all the rest as iterations. There’s a method of composition implicit here as well, in which each successive sentence would hypothesize from the previous one as a sample. Such a mode of composition would create a kind of mapping of the reality that is simultaneously being generated. And all this implies compression, which is a product and hinge point of disciplinarity and what I would point to as a sign pointing to the center and therefore as the meaning of transcendence. Peli Grietzer, in his “A Theory of Vibe” (easy to find), examines the kind of compression manifested in the easy to feel but hard to define “vibe” as “autoencoding,” i.e., rules for enabling the reconstruction of some phenomenon through the disassembly and reassembly of its components. Creating such rules is a moving target and hence subject to approximation and improvement, drawing upon a flexible “canon” (that’s Grietzer) of texts and objects to create a certain aesthetic and historical effect and affect. Even trying to describe a particular vibe involves compressed language: e.g., a certain film gives of a certain mid-90s urban vibe poised between the “vulgarity” of the “Reagan” 80s and the “return of history” post 9/11, effected through the foregrounding of some 90s-specific technology, landscapes, event, etc. “A vibe is therefore… an abstractum that cannot be separated from its concreta” (Grietzer, again). So, the practice of generating vibes would be to simultaneously get more abstract and more concrete, more extractive from your materials so as to generate samples that enable a wide range of hypothetical selvings to hang together and find themselves to be part of the sampling.

Compression, of course, has resonances (which Grietzer draws upon) with contemporary technological developments regarding AI and machine learning, but I think it also helps us to understand what makes for a successful and enduring ritual. The initial ritual on the originary scene would have had to compress a series of iterated gestures into a shared gesture that would enable each participant on the scene to look through the scene at the other participants and to further check and refine his own gesture. So, compression is scenic design; meaning is compression. If we go through the sequence of speech forms developed in Gans’s The Origin of Language we see a series of compressions as well: the imperative recovers from scenic implosion by introducing time into the scene, thereby compressing scattered ostensive gestures into deferring to the other; the interrogative does the same by compressing opposing imperatives by prolonging the initiating one so as to allow for a new mode of deferral to a new world of imperatives; and the declarative waves the demanded object away while preserving the possibility of its restoration, but in a new form that will reconfigure the scene. The declarative compresses the world into a range of possible ostensives. At the end of each compression is an exponential increase in the number of possible scene invokable or reconstructable out of the sign, and this increase is the accretion of deferrality so that activating one of the possible scenes leaves all the others in what we can now call, as we replace vocabulary drawn from older media with that drawn from emergent media, “latent space.” We could decompose any utterance or sample into so many articulated and articulable, within a shifting hierarchy of probabilities, possible ostensives, imperatives, interrogatives and declaratives. This is the kind of thing we do, in a commonsensical way, when someone explains something and someone else asks, well, what do you want to do about it—which is essentially a prolonged imperative demanding that an imperative be extracted from some declarative order.

This kind of decomposition or disassembly within the declarative itself is enabled by Anna Wierzbicka’s natural semantic primes, as evidenced by her own explications, which take a single word and break it down into a series of utterances that are invariably highly scenic. Every utterance is a compression of someone thinking about what someone says about what someone sees someone else doing when… Every word beyond the primes is therefore a compression, and is therefore always proto-disciplinary, requiring some shared scene upon which an ostensive might always lie at the end of however long a series of exchanges. This intensification of compression as the disciplinary space becomes more specialized is the reason for what is seen as the “jargon” of academic and scientific spaces (or, really, any kind of expertise). The reason the use of a strange word or phrase becomes strictly enforced is that the installation of that expression resulted from the trial and error of realizing that received vocabularies didn’t allow for the shared attention needed to generate new questions. And once someone leaves the discipline, even for a short time, upon his return its conceptual order will be alien, in part because one has been stripped away from that shared attention but also because such highly artificial (artificial in the sense of being explicitly designed and redesignable) vocabularies are always subject to change and the accretion of new nuances and implicit distinctions. The greater propensity of academics to write in long noun-phrases, often heading sentences, and thereby breaking with the almost compulsory subject-verb opening (even if with a brief introductory phrase), and which seems arbitrary and obnoxious to others, is really first of all a way of saving time (writing fewer words) but also a way of “name checking” a set of reference points all need to share for the discourse to proceed. This is why I am always working on developing a specialized vocabulary, or idiom, for center studies, even if it impedes popularization (never much of a possibility, anyway)—it enhances compressibility, and hence the exponentiality of the growth of name-able scenes.

I’ve said a few times that, if pressed on what I “believe,” my answer will ultimately be “language,” which seems to me far more miraculous than anything believed in by adherents to other faiths. And, of course, it’s also far more commonplace—a miracle occurring trillions of times every day. Believing in language is continually testing its compressibility, which provides me with a new way of returning to my privileging of the present, as in the present tense. I’ve made this argument quite a few times and do so completely seriously—one way I’ve put it, fairly recently, is to say that the past tense is too mythological and the future too pseudo-prophetical and therefore pompous. What you want to say about the past or future can be said in the present tense, if only you compress sufficiently. Doing so imposes a kind of discipline that sharpens the mind because you are constantly reminded that whatever you know of the past is only what you are receiving, organizing, enhancing and interpreting right now as you bring others’ attention to it and that whatever you can say of the future is redistributing the proportions of evidencies in the present which, sufficiently redistributed, does issue in novelty (there is no creation from nothing, at least not for us mortals)—and when you force that upon your attention you can train your attention on the repositories of the past, the archives, you have been drawing upon, how they have been transmitted, how you would like to be transmitting them now and which organs of perception and thought might single out otherwise neglected features of the evidencies for some shift in resolution and articulation. This seems to me a worthy heir to ritual, as it involves the continual offering of scenes around which anyone and everyone might gather.

This also means that what people often like to speak of in terms of “dialogue” are really cases of inter-language or transfer translations, which is to say some reciprocal modification of idioms, which in the end becomes compression. Memes, of course, are compressions. Resistance to compression is a condition of sustained and extensive intelligibility—if tens or hundreds of millions of people are to share a “culture” (i.e., a pedagogical and accrediting network) for decades or even more such that, for example, a hit movie still “strikes” audiences more or less the same, or at least trackable, way, fifty years later, a certain hostility to compression is probably necessary. In discursive terms, resistance to compression takes the form of examples and reformulations. You get to generalism, universalism and humanism through enough examples and enough of putting things “in other words” so as to spread sameness over compressions. In other wording unwinds compression by allowing for sentences resisting differentiation to be reiterated over multiple occasions. You might say this is the linguistic form taken by the High/Low vs. the Middle strategy of power centralization. Of course, that in itself is a kind of in other wording, which is hard to avoid when discussing resistance to compression, but if aligning the high, middle and low, or center intermediaries and peripherals (to use Chris Bond’s more recent terms), something that it is not clear has ever been done in a sustained way, is the goal, compression must be the way because this articulation can only be effected pedagogically and pedagogy is compressive if it recreates an idiom. My own energies are of course devoted to making center studies/anthropomorphics the one big compression; indeed, how could one assert that we are heading to an idiomatic future without also laying claim to the idiom of idioms, the idiom compressing idiom?

Here is Grietzer again:

One reason the mathematical-cognitive trope of autoencoding matters, I would argue, is that it describes the bare, first act of treating a collection of objects or phenomena as a set of states of a system rather than a bare collection of objects or phenomena—the minimal, ambient systematization that raises stuff to the level of things, raises things to the level of world, raises one-thing-after-another to the level of experience. (And, equally, the minimal, ambient systematization that erases nonconforming stuff on the authority of things, marginalizes nonconforming things to make a world, degenerates experience into false consciousness.)

So, a vibe involves the habit of treating an object as part of a collection of objects (all of which are “like” each other in some way), which is to say as a single sample that includes the whole, or a system, to which the addition of each object, including the addition of existing objects as new kinds of objects within the system modifies the system in ways that still resonate with the constitutive vibe. This is non-totalizable in the sense that there is no position or rule outside of this “world” that would give us the world—only continual training on reality, or things out there outside of this world that might find points of likeness within it can count as agency within the system. Grietzer’s parenthetical observation, furthermore, gives us a way of identifying “moves” that violate what we could call the rules of the “language game,” as per Wittgenstein—“erasing” stuff “on the authority of things,” rather than the authority of a continually retrained system.

I would say this means continuing to double down on compression against in other wording, and, even if in other wording and exemplifying have their place that place should be one that, rather than reducing compressed material to an existing field of commonplaces, brings the other words or examples into the system as things in a defamiliarized way. All this concerns our styles of thinking and making, first of all, but I think that makes it directly relevant to scenic design with machine learning and algorithmic governance insofar as something as simple but increasingly difficult as correct labeling is likely to become the primary human vocation under such machinery. Let’s say you’re researching (and, now, everything is researching) some newly coined medical term in 19^th century psychiatry and want to follow its operations across languages and disciplines—how often it appears, where it is approved of or criticized, how it spread and later declined, differently in different places, various implications of its translations, how it licensed certain medical diagnoses and practices, contributed to the formation of institutions and was implicated in the transformation of pedagogy, etc. You’d be doing document searches (imagine LLMs but with the ability to trace any outputs back to their original inputs) with increasingly precise questions aiming at locating, not only that term under specific institutional, historical and conceptual conditions, but terms “touched” by or “vibrating” (vibe-ing) with that term in various states of the system. You get to the point where, as with any real research, you are finding things you didn’t know you were looking for and which therefore take you well beyond the search terms you started with to search terms you keep remaking with each search and the returns of each search, as you learn along with the machine. As a result, every medical term you encounter now will vibe for you in a similar way, as long as you “keep faith” with compression and the integrity or selving of the idiom, and refuse to marginalize (in other words or exemplify) new things entering the system and thereby exclude them from the auto-encoding process. We do have an alternative to in other wording and exemplifying when it comes to unwinding or unfolding the system either for explicatory or initiatory purposes, and that is translation into the primes, which continues to raise the intellectual demands of entering the system while making it accessible to anyone wishing to enter. The system is both totally other and yet the same in each iteration—not coincidentally, this is itself an iteration of the problem of succession—and maintaining the functionality of such a system means being able to show how it is absolutely and necessarily even more the same while being even more wildly and disorientingly other at each point along with way. And, methodologically, this does come down to writing sentences in the present tense that have to “reformat” other possible sentences as noun, adjectival, adverbial and prepositional phrases modifying each other with increasing precision. That is, a writing class, dedicated to learning how to write the single sentence of the world. I don’t see what other than that one could want in terms of “transcendence.”