All You Need Is Language

When forced to answer questions like, what is the ultimate telos, what does it all mean, what allows us to reject nihilism, etc., I always fall back on the answer: language. Language is literally meaning, and if you think you can say there is no meaning that just means that language is transparent to you and you need to make it visible and audible. The fact that asking about meaning means something situates you within language, and if you are situated within language you have taken upon yourself the originary and never completed project of deferring violence and you might as well get on with it. There’s no justification for humanity and if you don’t care whether we go extinct or blow ourselves up or whatever, well, I can try to take you seriously but the resentment in such attitudes is so evident as to render the attitudes themselves comical. As Dostoevsky or Girard would say, it’s not that you don’t care about humanity; it’s that you want others to see you not caring about humanity. But the further implication of locating meaning in the most obvious place of all, language, is that the “good” (another one of those terms those undergoing political transformation in proximity to NRx and related movements has been confronted me with) is the maximization of language—sometimes clarifying and even purifying it, as the logical positivists tried to do (even if not quite in their way), but also enriching it as working to release its energies which are further, propulsive energies of deferral. It also means bringing all other vocabularies into linguistic or semiotic ones, in part as a way of appropriating the tendency, also a tendency of language, to expand other vocabularies—economic, technological, etc.—into those where explorations of language should prevail. And I’ll try that here by languaging “originary debt,” “singularized succession,” converting assets into data, and we’ll see what else.

How to describe everything in terms of language has been a longstanding question and quest for me and I’ve gone about it in some crude ways that may have been necessary preliminaries. This is the project of making all language performative, which means deferring the “tyranny” of the declarative, and the imperative has to be central to that. The whole world is held together by imperatives—declaratives couldn’t do it, and ostensives provide insufficient continuity—and all those imperatives must track back to a single one—a kind of spine of all human history. I’m still obeying the imperative to follow this through, but it must be done in terms of dialogues, between ostensives and imperatives and imperatives themselves. Imperatives must find their confirmation in ostensives—if “thy will” is to “be done,” it must be seen to be done, and this witnessing must be communicated to and confirmed by others. But “thy will be done” is itself an imperative, issued to whom is not entirely clear, at least to me: who is to ensure that His will will be done on earth, as it is in heaven? I see the petitioner as requesting that God ensure His will is done on earth but also extending the divine command to those fellow petitioners who would have to actualize that divine will. Like every genuine petition, it’s not for some benefit, but to help me know what you want me to do, and, therefore, how to do it. Prayers are always on the boundary between these dialogues among ostensives and imperatives on the one hand, and declaratives on the other. The declarative is necessary for reality, that which remains regardless of whether we know or recognize it, to interpose itself in those dialogues, and the question is how declaratives can do that generatively rather than by trying to nip imperatives in the bud by trying to derive all of them from declaratives—that’s the problem of “metaphysics.” None of this is new, but the question remains: how to make this discourse performative, part of a way of speaking, writing and thinking: an idiom rather than a “theory”?

The difficulty has always been that any utterance is context bound, and so it was always naïve to think a sentence could be broken down into an ostensive-imperative-interrogative network—one could always situate it just as plausibly within some other network. But LLMs may solve this problem for us. It will be possible, if it isn’t already, to ask an AI model to generate any number of possible contexts, or scenes, or stacks of scenes, within which a given utterance could have been made, and to rank those possible stacks of scenes in order of probability given whatever variables we wish to introduce. Each utterance is a compression of deferral, so each hypothetical stack of scenes would model a different hierarchy of deferrals—deferrals of mimetic violence the contours of which could be extracted from the utterance. We could focus more precisely on the “operator of negation,” that general, open-ended imperative to no longer do something that one could never be finished not doing (“stop smoking” is Gans’s example in The Origin of Language) and which I would see as an extension of the originary event of deferral, recalled in this crisis of the imperative. The “positive” imperative—first of all to supply something, then to do something—precedes the operator of negation, but is subordinated to it within the declarative. But the operator of negation must itself give way to an affirmative imperative—not doing something must give way to doing something. The declarative is the operator of this transition. If I modulate my anger at an offensive remark made by my host at dinner into an ironic comment that might be ignored, denied, or taken as a friendly warning, I am marking an immediate deferral in my utterance; but in doing so, I am also contributing to broader systems of deferral, those we call “neighborliness,” ‘lawfulness,” “civility,” and so on and, ultimately it’s a small drop in the bucket of “civilization”—but all this insofar my ironic remark, probably a declarative sentence, grammatically speaking, is on the edge of soliciting (soliciting is a kind of mediated imperative) some kind of more or less implicit apology or, perhaps, further provocation. We should now have the data and computational power to fully represent such “scenery” in hierarchically articulated possible outcomes all along the stack—it would be a kind of literature, or cinema.

Now, if money is credit, or a means of marking the relative power of creditors and debtors, then credit is ultimately located in the entire civilizational stack: you accept my 100$ today because you are confident that you will be able to spend it in turn tomorrow on goods that would have gone for 100$ today. This is really a lot to be confident about and it should not be taken at all for granted, because a whole network of institutions is necessary to make this confidence reasonable. This is why leftist vandalism, even in its seemingly harmless forms, should be rejected with the utmost prejudice—it is destructive of credit because it discredits the institutions credit relies upon. We must, then, be able to fold credit back into the imperative order—those imperative-ostensive conversations issuing in and flowing out of declaratives. “If you lend me 100$ today I will pay you back in a year”—we can’t say what this means, and if the person saying or hearing it can’t without know what $100 will “mean” in a year, and we can’t know that without knowing what our governing and juridical institutions will look like in a year, including the stack of scenes built to condition us to new trajectories of succession. And so, we have things like interest, which price in the inherent riskiness of assuming that the conditions under which money has been lent will approximate those under which it will be repaid or not repaid (and whether or not it is repaid is itself embedded into those conditions). All language is like that—all of our promises, and every utterance is to some extent a promise, at least regarding the ostensive conditions under which its truth will be affirmed. How well we can know what the conditions—general conditions but also those of the individual debtor and creditor—will be like when the debt comes due gives us the meaning of any utterance. And we can load this question up onto the deferral capacities or maybe “energies” of the local players and the institutions in which they have been invested and are maintained with greater or lesser care.

The question of pedagogy has come to solve these problems for me in large part because teaching and learning is always on that boundary between the ostensive/imperative and the declarative. Credit can then be expressed in terms of pedagogical futures: the further down the line you can expect learners to practice the modes of deferral that will issue in imperatives to build the institutions of deferral the more you can credit those willing to “trade” on those futures. But a further inscription must be made here: to fully convert assets into data, currency into learning curves the entire sacrificial order constituting our current transactional networks must be, I can’t say abolished but made to undergo another increment of deferral. Our language is permeated with sacrifice, ultimately, still, human sacrifice—some must die for the community to continue—because that has been the only social form of deferral we’ve ever had, and for clear reasons: that creature on the originary scene which created the community was also our prey, with whom we established exchange relations involving its willing self-sacrifice and our sacrifice of a portion of food and devotion in return. Language will only be worth currency when we convert it sufficiently in ways that defer sacrifice, even metaphorical sacrifice, even in the form of speech forms like irony. A lot of modern and postmodern literature and art has been trying to teach us that, even if I’m not sure much of it has known that’s what it’s doing. Can we stop thinking in terms of “punishment,” for example? Can “punishment” become archaic for us? Liberalism and leftism have long profited from the sense that we can no longer sacrifice in good faith, but in the process has only provided more primitive forms of sacrifice, regressing to vendettas and corrupting the law into a vehicle of vendetta. So, I’m not talking about letting murderers and rapist roam free. But I’m also not speaking about therapeutizing crime, turning criminals into patients—another attempt to transcend sacrifice. Or, for that matter, treating “illegitimate” violence as simply something to be prevented or to protect potential victims against, like storms and floods—yet another attempt at post-sacrifice. What I have in mind is the kind of thing I’ve spoken about before and which we only now are coming into possession of the technology to operationalize: what we might call tokenizing violence—providing everyone, in compact form, with information regarding the likely outcomes of any imminent or possible interaction, information formulated in a juridical vocabulary, the richest moral vocabulary we have, along with devices to activate so as to regulate those interactions based on that information (with varying degrees of automaticity, which is also to say, with vary degrees of openness to appeal). We don’t have to lock up murderers but no one would need to let a murderer into his establishment, no one would have to enter an establishment that lets murderers in, while all of social life, including public areas like streets and parks, would be treated as “establishments.” At the same time murderers could be dealt with at varying degrees of mediation, and there would also be establishments founded by communities for housing them. Criminals might token mine in an effort to lessen the degrees of mediation separating them from others, providing data of their improved behavior within their limited mobility, allowing others to give them the “credit” they think they deserve. Victims, former and potential, could contribute to the data pool as well. We could imagine scaled up versions of this approach for relations between nations. (There is a female comedian—I couldn’t find her name through a Google search so I’m working from memory—who has repudiated being funny on the grounds that all humor requires a “butt” of the joke and hence is implicitly violent. She has been ridiculed for this, but she possesses a true anthropological insight, and using the comedy “routine” as a way to reflect on what we desire in such a routine is an innovative and valuable form of art and—I don’t know since I’ve never watched her perform—might even turn out to be “funny” in new ways. Can there, in fact, be humor without a “butt”? We can see humor as a form of deferral—joking rather than lynching—but we confront the same problem: once we see it for what it is, can we continue to participate in good faith?)

Maybe money, and economic language more generally, is inseparable from sacrificiality, and further transference from money to learncoin means further distancing from absolute ostracism. Our language would become increasingly hypothetical, conditional, probabilistic and “measured,” always detecting degrees of deferral currently operative and potentially needed. This will always leave plenty of room for straightforwardness: the operator of negation needs its transitions into affirmative imperatives. But, saying we take Trump as a kind of exemplar of straightforwardness, everything he says could also be translated into learncoin without losing—maybe even with gaining—some of its perlocutionary force. In our interactions with AIs “all” that might be left is for humans to say “it’s the same picture,” or it’s not, or giving degrees of likeness, and then to request more samples (in the process providing more samples) of certain kinds. This will give us more rigorous and demanding work than we have ever done, and more rewarding too. What will provide confidence and credit will then be the likelihood that samenesses and likenesses will be detected and authenticated at ever greater degrees of resolution and with ever greater consistency, as more potential samenesses and likenesses are produced at the same time. In other words, greater degrees and further layers of joint attention, of people being able to answer the question, “what am I looking at here?” Registering these degrees of deferral is what maxes out the language, making deferral itself, in the form of the operator of negation generating affirmative commands the primary imperative. And we can speak and think this way and give credit to others who do so and who build scenic architecture enabling others to do so, and further scenic architectures to staff those scenic architectures, and so on. And those are people you could trust to transmit your wealth to future generations, until wealth no longer takes the form of legally protected property and assets but of the credit one has accrued through one’s contribution to singularized succession in perpetuity.

In thinking of language as deferral we think of every discourse, every text, even every utterance, as an archive of conversations and we can then think of the samples we provide as better to the extent that they archive more conversations, more varied conversations, more conversations likely to be lost. (Think of how Fustel de Coulanges reconstructed an entire social world out of texts written centuries later.) And we can now think of such archiving as imperative, as demand and petition, directed toward the databases we contribute to and draw upon, an imperative that requests further pathways along which the archived conversations might go. Asking for a more direct, affirmative command out of these ongoing conversations is tantamount to interrupting one of them midstream, which we are in fact free to do with any conversation and which Paul de Man once contended, cited Schelling (I think), is in fact constitutive of any conversation as free intercourse (for me, free intercourse=indefinite deferral). The learning machines we will share the scene, as props or addresses, with, will thereby hold anything worth calling “wealth,” as an ongoing central intelligence to which all are contributing but some more consequentially than others. Like with the endowments of some universities, you keep accumulating, while never touching the principal, only using it as collateral for loans for ongoing investments. That’s what the most supremely gathered and curated data will be once we no longer need money as a standard of value and unit of account.