Language, the Ultimate Source and Analysis of Data

Insofar as I have an epistemology it’s borrowed from Charles Sanders Peirce’s notion that truth is what will turn out to be the case in the long run by the community of inquirers, which in my understanding not only makes truth a kind of infinitely receding target but makes it an effect of seeing to the long-term coherence and integrity of the community of inquirers. In that case, the meaning, or eventual ostensive truth, of anything I happen to be saying now, is located in some hypothetical figure upon a hypothetical scene at some hypothesized time in the future—or, really, a sequence and distribution of such figures on such scenes at such times, to be further specified by continuations of the discourse initiated by that statement. This also means that part of the meaning of what you say is your own promise to do what you can to ensure the establishment of that scene and the preparation of the figures necessary to populate and act upon it in the way implicit in whatever you’ve said. You start to design a scene, however unlikely or distant, with every utterance even if you’re doing no more than scouting out the location, so to speak.

This also means that the meaning of any utterance depends upon the indirect cooperation of all the other people saying things, which is to say sketching out the design of scenes that will more or less overlap with the ones undergoing development in your own utterance, and what you say is more meaningful to the extent that you implicate (fold in with your own) those other scenes while inscribing an awareness of the limits of your awareness of all of them. A kind of economy is built into meaningfulness, then, because a severe selectivity is involved wherein you build into your statements criteria for the other disciplinary scenes you are willing to rely upon and how far you are willing to rely upon them. There’s no single rule here for determining which scenes to rely on—no one is obliged to restrict themselves to what you can prove right here and now beyond some arbitrarily determined standard of proof—sometimes the more speculative, hypothetical, even fantastical suggestions are the most meaningful. If there’s a rule it’s that you keep learning to discriminate between the ever proliferating, converging and diverging idioms, which refine the ostensive outcomes in ways that can only be mastered on the scenes, which is to say you engage in perpetual language learning and facilitate the same for anyone who might be downstream of your utterances.

All of this is stuff I’ve said, more than once, even if not so recently, but my own mostly followed rule in such matters is that I repeat myself in order to introduce a new degree of consistency between the various idioms I’m working on. And what I’d like to do here is bring this discussion of meaning and truth into convergence with my recent thinking on data and algorithms and, even more importantly, bring a more definitive resolution to my concept of singularized succession in perpetuity by treating it explicitly as the successor concept to the sovereign or central imaginary. This is all to formulate with increasing precision and effectiveness the notion of “listening to the center” or, for political scientists, the problem of political obedience. The sovereign (more recently: central) imaginary is what you arrive at when you utter some political desire (universal healthcare, no more immigration, etc.), imagine the kind of political authority that could satisfy that desire in such a way that you could recognize it to have been satisfied, further imagine the scenic conditions that could enable the stable continuation of that authority (because you don’t want you political desire met just for a single administration) and then consider whether that authority would in fact do what you want it to do because, after all, if it could do it, it wouldn’t have to and might have something very different in mind. If, like the hypothetical recipient of the saying of the eternal return by Nietzsche’s demon you can embrace whatever this authority would command by way of or instead of fulfilling your desire, then your political desire is meaningful.

How singularized succession in perpetuity catches this up is in the implication that to be ready to obey such an authority is to (to continue to speak a little Nietzschean) “will” its continuation on its own terms, which in turn means to participate in whatever way falls to you in ensuring that authority select its successor, and not just any successor but a successor that it most likely to choose its successor, and so on in perpetuity; and what is implicit here but what I have not yet made so explicit is that if singularized succession is the optimal condition what can this mean other than exercising singularized succession in perpetuity in the way appropriate to whatever disciplinary spaces each and every one of us is implicated in. (Can there be singularized succession in perpetuity at every level in the social order? What if a higher level disapproves of the way someone under his supervision selects his successor? This would be an indication of dysfunction to be addressed pedagogically.) This, then, is the purpose of scenic design—to set the stage as precisely or vaguely as necessary for the understudy to step in when needed and ready (to help the understudy be ready when needed). This, then, is technics: inventing, refining, spreading and repairing devices so that the figure proposed by any utterance, that successor however distant, will be in a position to enter and revise whatever idiom might pop up on that scene. Even the most seemingly utilitarian form of tool making, producing a physical object specifically suited to have controlled effects as part of a limited work process only make sense within a process of transmission and pedagogy in which the tool is refined and marked within an essentially ritualistic master-apprentice relationship (somebody showed you how to use a hammer).  So, technology is a set of assignments for distance language learning, because what else could be the point of technology other than a specifically distributed population having a new field of ostensives, which means a new realm of meaning, which they couldn’t have had without that technology? And discerning a sorting out and generating in the process a new field of ostensives is language learning.

This also means that the designing of such “assignments” follows the discovery, creation and revision of a new idiom, perhaps on the margin but transferred to the center (there’s an growing literature on the overlappings of the artistic avant-garde with intelligence and military agencies) which can advance some imperative of governance. Technology is always the transformation of existing forms of human cooperation and coordination, including the gestural, motion, normative and other ultimately linguistic capacities involved in coordination and cooperation. So, some problem of organization related to war, or colonization, or the creation of economies of scale, or continental transportation is taken up on some disciplinary scene which starts to design the scene that would solve that problem along with educating, training, and distributing those who will populate it. That technology is a set of assignments for distance language learning is another way of saying that technology is post-ritual governance, but it’s a better way of saying it because it enables us to highlight and then draw consequences from the fact that, post-ritual, the only source of the good is not some moral or ethical “principle” but language itself, or the ongoing commensuration of ostensives, imperatives, interrogatives and declaratives. Any transcendent principle or being you would want to posit would, anyway, be a set of technologies for commensurating the various levels of speech acts: believing in God, or in equality, or in fairness, or in human dignity, etc., insofar as it’s more than than a commonplace, involves saying and doing certain things under certain conditions, according to certain standards, enforced more or less formally in various ways, and so on—the authenticity of one’s beliefs or commitments will need to be assessed, so why not just say that we can trace the ostensives, imperatives and interrogatives that are addressed and in turn issued by a particular sentence or discourse and thereby determine the best way to respond so as to increase the commensurability across the levels. This approach precludes all “nihilism,” because we can’t run out of language, and we will always face an unimaginable and inexhaustible plenitude of possible ostensives.

Listening to the center therefore entails creating an architecture for housing the commands of the center, which means prolonging those commands by streaming them off into more localized commands to inquire into extending and fortifying the architecture. This means constructing the apparatus for testing out the central imaginary, because if you can imagine the mode of centrality that would satisfy, cancel and transcend your political desire you can imagine an intermediate arrangement that would be closer to the installation of that mode of centrality than other possible arrangements, and a social technology (all technologies are social technologies) that would approximate those arrangements, and be made more consistent with those technologies that would approximate a later version of those arrangement—and even if you don’t know how to fill in the gaps, others will, or they’ll discover that they can’t be filled thereby invalidating that form of that political desire. If we’re to argue about political ideologies, then, we can transfer those arguments into arguments for political machinery that will constrain the commands issued at various levels in such a way that the commands in their implementation would “look like” the field of ostensives your “ideology” or “philosophy” will have in the end been seeking. Listening to the center means building a world in which you and everyone else could choose your successors in perpetuity, but this also means learning how to choose in many different ways and to select for successors who may not be like, much less the same, as you in any obvious way. Who will be best suited to carry your work forward 70 years from now? You can only have a vague sense, but it’s still a sense and therefore one that can be improved upon, in part by improving your sense of who might carry the work forward 40 years from now, 20, and 2—doing better in the short term will reduce the vagueness of the long-term (this also makes reducing vagueness in the long run part of what “doing better” in the short run means).

What is good? Stringing together declaratives that exhaustively answer questions that have been framed so as to be exhaustively answerable, because they have followed the extension of imperatives aimed at preserving an ostensive that has been created by a string of declaratives that exhaustively answer... Ostensives, imperatives, interrogatives and declaratives all fit into each other in a way that makes a coherent act embedded in an order of which that act is a fitting piece. This is the linguistic formulation of singularized succession in perpetuity. We are now well positioned to technologize this formulation as well: algorithms are articulations of imperatives and declaratives (if... then sequences deriving instructions from the specifications of a statement being met), with the data being the articulated field of ostensives from which the interrogatives regarding whether conditions have been met are derived. If you think algorithmically upon data you will be thinking better and, far from reducing you to the terms of the technology, this form of practice or selving submits the technology to human ordering. When you’re designing an algorithm you’re trying to produce an operational, iterable answer to some question like, “how can we be safer,” “how can we be healthier,” “how can we create clearer images,” and so on, but what the words “safe,” “healthy” and “clear” mean in any of these cases can only be specified by their use within the idioms that have called for (or not) the algorithmic designing. Thinking algorithmically upon data is a form of reflection upon data, and this will become evident as we think algorithmically upon data well beyond the rather simplistic questions we pose now, and start to take all of language, which is to say all possibilities of meaning, as the source of such thinking. Look at any conversation, even the most trivial and casual, you’ve had lately, and take some unsettled question from that conversation and consider how you might design an algorithm that would reveal the meaning of that question, and you will see that it will be all the more your question, and not one any computer could have generated, as a result of such engagement with technology. (And such straining with language against language will always have something funny about it, because it involves a bit of yielding and deferral.)

Language is itself the greatest source of data without which we would have no other data. As you gather data you generate more data, first of all regarding the gathering of data, which refers to chains of preferences, priorities, necessities, traditions, and competencies. We will never be outside of this, and all the playing with self-referentiality of innovative writers over the centuries takes on its direct social and political meaning here. The best way to control for all the “posts” the data you gather, archive, curate and analyze will subsequently open is to be the end point, the singularized successor, of some predecessor, in some way he couldn’t have expected but would nevertheless recognize as his own. This comes down to developing a style, and a style that makes explicit and formalizing the learning of idioms coming home in whatever discourse you’re working through. Style is technology—if you imitate and translate another’s style you’ll be saying things you wouldn’t have thought of otherwise; idioms are shared styles, and creating styles of learning idioms is a way of proposing new idioms.

You enter through someone else’s stylistic singularization of another idiom—there’s no other entry point. A style is always at some distance from the “classical style,” which tries to create the illusion of reader and writer occupying the same scene, with the prose offering a transparent window onto some shared event—even if one is writing in the classical style, you will be at some distance from some other possible version of the classical style since the shared scene is constructed, not given. Everything in the prose marks or conceals some infrastructure making the text possible. You can restrict access to the scene in some way or open it beyond the parameters of classicism, exposing the infrastructural conditions of classical prose itself. Either way there’s some movement between those on the scene and those in the viewed event—someone we’ve been looking at gets called on to come look with us, or someone looking with this finds himself on the scene. Someone sneaks off to go behind the scene or someone from behind the scene is brought. out to center-stage.

All of these maneuvers depend on making the scene of writing more a scene and that involves linguistic involutions and conversions: words moved from context to context, so that they are marked by and mark the discourse itself, words taking on different shapes, nouns used as verbs, verbs and adjectives nominalized, etc. There will be a large role for language generation and translation programs in the science and technology of style, but one can always begin by carrying out operations on a text and generating a discourse simply by saying what seem to be the differences from the original created by the operations. This is how language becomes more data rich, data driven, data savvy and data converting. If your style is compelling and demanding enough, if it’s like a home constructed so as to incessantly test your perceptions, muscles and reflexes, then you’ve built a technology designed to attract potential successors and repel pretenders. And style is not only to be found in writing, but in posture, gesture and patterns of decision making. So, listening to the center ends up meaning maximizing the arrangement of all of your “parts” so as to provide a set of implicit instructions for others to do completely different things that couldn’t have been done without the model.

Think about what’s involved in learning an idiom: imitating the speech of others, semi-randomly, making a lot of embarrassing mistakes; twisting and stretching your own language, in ways you sometime realize after the fact and that often won’t fit; tracking the responses of others more inside than yourself and modifying your attempts accordingly; bluffing, at least on occasion, and figuring out what to do when you get called or the other folds; trying to put into new words thoughts that exceed your present linguistic capacities because they are been drawn out by the new idiom which you haven’t sufficiently mastered so as to express them in it. This is the case for learning a new slang as much as entering an academic discipline. Those points where you’re using new vocabulary in a familiar grammar, or familiar vocabulary in a new grammar, are the most interesting ones—that’s also what produces change within idioms, so that the idiom generates new ostensives. A new technology, even the most conventionally conceived one, designed to just make something easier for people, situates its users in this kind of position. But, even while we could examine, say, bridges, in these terms, this description clearly best fits information technologies, which always involve the creation of new discursive rules, new networks, and new observer posts (surveillance), known and unknown. Treating a new expression as coming as a result of the exhaustion of some attempt at extensive mimesis, when you realize that your attempt to be like everyone else, or like some model, has made you something completely different, and what that will be is yet to be determined, has to be useful. Propose some terms for determining what that new thing will be, and you will have the blueprint of a technology to help you sift through the data flow accumulated for solving your successor problem. And once you say, “when you’re learning a new idiom,” it’s easy to see that we are never doing anything other than learning new idioms and so it’s really just a question of keeping attention on ongoing problems of succession, which is to say, the continuity of order cognizant of maximal possibilities of disorder, that constitutes every use of language.