In particular the last paragraph and the discussion of what might be involved in programming a computer to write novels (a very engaging question which I have decided, for now at least, to keep very minimal), but a few other things as well:
Pressure on the category of the “human” comes from two directions, what we might call the “analytic” and the “synthetic.” What I mean by the “analytic” dismantling of the human is our capability and propensity to break the human down into a set of probabilities, with ever more subtle gradations: physically, we are all aggregations of DNA manifesting itself through interactions with an environment whose effects we are rapidly acquiring knowledge over—it will soon be, if it isn’t already, to treat a single individual as the predictable result of a history of activity, diet, genetic predispositions, places of residence, etc., and as a body whose future is equally predictable, given the known, and to some extent chosen, variables. This analytical dismantling is analogous to the replacement, in the digital humanities, of the individual text or art object with the database search as the object of inquiry: a particular text, or, for that matter, a particular sentence, is nothing more than the winnowing out of all the other possibilities generated by the preceding history of all cultural texts and practices. By the “synthetic” “remantling” of the human I mean our growing ability to engineer the body, first of all prosthetically, through the replacement of natural limbs and organs that have been damaged or lost, but, more consequentially, through the creation of new capabilities, through drugs, hormones, surgeries and, eventually, the prompting, manipulating and stimulating of the body’s own natural processes.
These two processes converge, as greater analytic prowess opens up new synthetic possibilities, while new synthetic inventions pose new analytic questions. Also, our deconstructive inheritance enables us to see that such developments reveal, not some unprecedented encroachment on a human domain that was once known and secure, but a more originary prosthetic being that constitutes us as human—with the first prosthesis being the human sign. Still, it may be that the “always already” deconstructive gesture serves as a kind of narcotic, comforting us with the assurance that what seems unprecedented has always already been with us. Maybe this time really is different, and something fundamentally human is at stake. At the very least, we have a definition of the human that allows us to test the hypothesis: does the advent of “transhumanism,” or the commitment to a fundamentally improved humanity (the ability to run a two minute mile, to live 300 years, to score 350 on an IQ test, etc.) through technological and medical advances render obsolete the originary understanding of the human as that species that poses a greater threat to itself than is posed to it by any external danger? Does a posthumanism theoretical perspective, which works to undermine the traditional conceptual boundaries separating individual from society, human from technology, and culture from nature, obviate the need for that understanding?
Answering these questions will simultaneously enable us to bring the concepts of technology and nature into originary thinking more thoroughly than we have so far. I would suggest the following hypothesis as an initial approach: the originary gesture aims at separating us from “Nature” and simultaneously takes “Nature” as a model for doing so. The separation from Nature is the creation of non-instinctual desire and of a community that transcends the animal pecking order. This we are all familiar with. The imitation of Nature is less explored, and more tacit. I don’t refer mainly to tribal rituals associating the members of the tribe with a totem animal, but to the much more pervasive adoption of natural constraints in the construction of tools, dwellings, weapons and, although this would require further exploration, the sounds and rhythms of human speech and the structures of human gesture and posture. Indeed, where else could early humans have derived models for these activities if not the natural world around them?
At the same time, such a concept of “Nature” could only have emerged much later, under civilized conditions. Like Jewish monotheism, or Western metaphysics (of which it is a part), the concept of “Nature” implies a thoroughly “declarative” culture, that is, a culture in which the declarative sentence is taken as the primary linguistic form. The history of the concept of “Nature” is obviously extraordinarily complicated, and I am not qualified to provide an authoritative survey—I will just venture some thoughts of some of its uses most pertinent to the question of the “Human.” It seems to me that a very productive way of understanding the valence of “Nature” as a cultural concept is as denoting a realm, in which humans participate, that is free of desire (that’s why imitating Nature would be a way of transcending desire). The separation of “Nature” as a non-desiring sphere of inquiry was obviously necessary for the emergence of the physical sciences, but just as importance is the notion of “Nature” in Stoic thinking, or the pastoral tradition, as a simple field of human activity enveloped by custom, tradition, and simple virtues, embedded in the natural world, and free of the “artificial” desires created by civilization. Even the “nastier” concept of Nature introduced by Hobbes is really more the basis of a calculus of social order, and hence a pacifying abstraction from actual social desires and rivalries.
So, Nature is something we seek to return to and restore; but Nature can also be the means of doing so, as the laws of nature discovered through science, mediated through the natural order of liberty established by the free market, leads to the technological transformations in nature that make prosperity possible. Up until fairly recently, it could be argued that all this was in accord with human nature: human nature as rational, as desiring, as toolmaking, as trading. The fully civilized order would then be the fully natural one as well. But it has proven impossible to keep the unnatural out. The primal victimary critique is of the notion of human nature, and for a very good reason: once you define human beings in terms of what they are or have (rationality, free choice, what have you), you provide a basis for excluding or conditioning the belonging of vast swathes of perhaps merely apparent humans from or to the class: those who are less rational, slavishly bound up with traditions or more instinctual desires, and so on. Once one accepts the perspective of the ontologically conditional human, it is the very positing of the human that becomes an unnatural, partisan, coup–d’etat—but against what, if not a more natural nature, posited as a truer equality in a more tight-knit community, in which rights and obligations are indistinguishable. In that case, those who have seized the category of human nature torment and dispossess those living genuinely human lives, closer to nature.
Let’s begin with where transhumanism and posthumanism agree—that this vein of thinking, predicated upon a fixed and definable human nature, supported by a metaphysics that sought to turn agreement about nature into a way of deflecting disagreements among humans, has run its course, with the further implication that the category of the “human” is, indeed, historical, contingent and contested. That still leaves open the question of whether the category can be abolished, though—historical, contingent and contested institutions as well as categories can persist precisely by virtue of being those things; and there must be some reason why the category changes in meaning and cultural positioning, and is struggled over, rather than simply dropped. I would return us to our originary definition of the human as the species that poses a danger greater to itself than any external threat, which is tied to our understanding of the human as the being that defers violence through representation. I think this definition survives posthuman questioning, and is in fact better suited to a more “emergent” sense of the human as never quite completed (a human “condition,” as Arendt had it, rather than a “nature”) than to a traditional notion of the human as a being with a fully installed nature, outfitted with a presumably impenetrable armor of rights. The danger that we face is never entirely predictable, the forms of representation needed to defer it never certain, and our capacity to find the means to discover the needed signs is always in doubt. What we are as humans is never “there”—all our signs could be rendered inoperative in one moment in the face of some unprecedented humanly generated threat. Of course, if that were to happen, all the boundary questions regarding the human and its others would also cease.
Transhumanism sees itself as extending the Enlightenment tradition. Once illegitimate, unnatural forms of domination have been eliminated, human beings can get to work addressing the externally imposed dangers and ills of humanity: producing sufficient food, fortifying shelter, curing diseases, and so on. All of these activities presuppose a human nature to be protected and enhanced: it is interesting, for example, that while we have for some time had the capacity to enhance eyesight to a remarkable extent, that capacity has been reserved for the laboratory and the observatory—no one seems to have considered offering, or demanding, eyeglasses or contact lenses providing even 20-5 vision for everyday use. But there’s no reason to assume that such things won’t be offered and demanded, and ultimately to be implanted in the eye, perhaps at an early age. There is no fixed line between the “natural” desire to remove an evident deformity like a cleft lip and the “artificial” one to make one’s lips look more like an admired movie star. Or between extending life in the sense of enabling us to live much longer with heart disease or curing cancer, and slowing or even reversing the molecular process of aging. Or developing a cure for Alzheimer’s, whether through neurologic advances or brain prosthetics, and providing the average person the ability to memorize Homer’s poems in one reading. Is there a certain point at which we would no longer be human? Only, we could say as originary thinkers, if such developments were to remove desire and resentment from our relations to each other, and it seems easy enough to imagine all kinds of ways in which the effects might be just the opposite—not only because the satisfaction of these desires would proceed at uneven pace, or because the desires might vary so widely, or even because even the most perfectly satisfied desires will be riddled with unanticipated side effects, but above all because the meaning of such transformations and their implications for questions of reciprocity will never be settled.
Similar questions are raised by the issue of Artificial Intelligence, addressed by Eric Gans in his recent Chronicle of Love & Resentment on the Digital Humanities. There is very little we can’t imagine programming computers to do: playing chess, of course, but why not writing novels or poetry, about any subject and in any form or genre one chooses? Certainly there will be computers capable of passing various versions of the Turing test. In the process, as Jaron Lanier, Jean-Pierre Dupuy and others have pointed out, we are likely to come to view our own mental processes and even emotional states in terms of computation, as the most advanced models in cognitive science already propose. Indeed, isn’t any declarative sentence a little algorithm, generating a finite series of differentially probable utterances by excluding (through negation and replacement) an infinity of other sentences that might, in a hierarchy of probabilities, have been uttered at that moment? At the very least, one can see the theoretical gains in seeing things this way.
But a funny thing happens on the way to the singularity. If we could imagine computers writing great poetry, we should consider it a matter of course that they could do something much simpler, something you really only need to be able to speak two languages to accomplish: translation. And yet they can’t do this, as anyone who witnessed the earlier hilarious attempts of google translate can testify. Or, at least, they couldn’t, when working according to a strictly computational model—that is, if you replace all the words, in the source language, one by one with words from the target language. Even allowing for grammar correct, which a computer can, of course, do, we end up with a mess—there is no way of accounting for the enormous variety and unpredictability of idioms in every language. But now it can be done. The way to solve the problem of computer translation is to amass an enormous database of existing translations (say, English to Italian)—then, for every phrase or sentence in the source text, the computer can search all the ways it has been translated previously, and then be programmed to choose from among various (presumably not that many) alternatives.
It seems to me that programming a computer to write novels, at least, would follow a similar, albeit far more complex, process: one would have to treat novels as “translations”—but of what? Other novels, for one thing. All forms of social texts, for another. A translation into what? Other novels would be translated into a different field of discourse: perhaps a historical field (translating a 19th century novel into 21st century social discourse), but perhaps a different field within the totality of contemporaneous discourse (a novel set in a psychiatric institution into one set on a beachfront resort); while social discourse would be translated into novelistic discourse, or “novelish”—or a dialect of novelish, of which there are many. Protocols would have to be established for translating (say) some segment of 19th century social discourse into some segment of 21st century, and a version of 19th century novelish into a version of 21st. This would involve determining rules for establishing equivalencies between these different fields of discourse: a chunk of 19th century political journalism into a chunk of 21st century (if we even choose to assume that these categories are equivalent, just because they both have the phrase “political journalism” in them). “Novelish” would have to be defined in terms of vocabulary and grammar, but also in terms of plot and character “codes,” perhaps involving the creation of templates that could serve as “attractors” for chunks of discourse. And the two different levels of translation would have to be articulated according to some rule.
This densely layered process, however it would ultimately be worked out, would require that multiple operations be applied recursively to the discourses drawn upon, meaning that every problem we solve along the way generates new problems, each decision branches off into multiple other decisions—the composer of such a text, computer or human, would need to articulate, through a series of constraints, a hierarchy of algorithms that would furthermore, be constantly “learning” and hence rearranging the entire series. At each point along the way we would find decisions that would necessarily be made more randomly if made by the computer, which would mean that the human composer would, by definition be “better”—unless, that is, we prefer the random decisions because they are more distant from the commonplaces and clichés that even the best writers cannot completely free themselves of—but in that case, it would be better for some reader, with a certain understanding and experience of the relation between familiarity and novelty in language and narrative.
What this means is that the googlization of the world, were it to reduce all objects and texts to search prompts to discover what that object or text was “translating” or transforming would not be in the slightest bit dehumanizing in any way that matters to our originary understanding of the human. There are infinite ways to derive target texts from source texts through a search process, and that process in turn will keep revising itself recursively. Each search decision is deeply embedded in body, history, biography, discipline and community. And any search will still be a discovery procedure for the signs that will defer, in however mediated a way, that potentially cataclysmic violence that will still be at the origin of our species.
Still, this notion of the sign as a prompt for a database search has ramifications for our understanding of the originary sign, which we would see now more as a constraint on future iterations, bearing the marks of the asymmetries and symmetries of its production than a single meaning apprehended by all: an emergent event rather than a completed state. What I mean can be illustrated by Johanna Drucker’s account of the sign understood materially:
Material conditions provide an inscriptional base, a score, a point of departure, a provocation, from which a work is produced as an event. The materiality of the system, no matter how stable, bears only a probabilistic relation to the event of production, which always occurs only in real time and is distinct in each instance.
… In each case, the performance constructs meaning as a result of engagement, the text is performed, rather than received. Materiality provokes the performance, and this is true whether we are talking about the workings of distributed systems in which resistance, voltage, and allocation of resources perform in accord with other processes and decisions, or whether we are referring to the reading of a poem.
I take the materiality of the sign, in Drucker’s sense, to suggest that we think of the originary gesture as a mark, mnemotechnic and prosthetic, on the natural world, including humans as part of that world. The effect of this mark on nature is to begin the process of transforming it into the materials for signs, tools, ritual objects, arenas and meals. Each new sign is constrained by the aggregation of previous markings, while inflecting the existing constraints, only thereby making it possible to see them as constraints. A constraint is a mapping of Nature, somewhere on the continuum between almost completely arbitrary imposition on one side and a nearly perfect iteration of a constraint imposed by Nature, on the other. Constraints, then, are an originary engagement with Nature, from which we are never quite extricated and in which we are never quite immersed, much like a preliminary search term, yet to undergo refinement.