The meta-fictional experiments of the latter half of the twentieth century emerged from a widely-felt sense that realistic literature had reached an impasse. John Barth spoke of a "literature of exhaustion," declaring that old-fashioned narrative prose concerning characters and events had hit a dead end, exhausted its potential, and otherwise—as he put it—"just about shot its bolt[.]"
One gets the sense that this conviction had something to do with the emergence of new technologies of ever-increasing complexity. What was to be the status, writers feared, of the "work of art in the age of mechanical reproduction"?—to use the resonant title of Walter Benjamin's essay. Italo Calvino, an exemplar of the new metafictional techniques, pictures in one scene (of his If on a winter's night a traveler...) a reader strapped to a series of electrodes monitoring her neurological reactions to each word and passage of the book she is consuming. Why? So that a computer can generate the perfectly-pitched novel designed to achieve maximum responsiveness from the human mind.
One is reminded as well of an experiment that Komar and Melamid ran in which they created the "ideal painting" by means of a statistical survey of mass preferences. The result? They depicted George Washington standing upon a green sward in front of a lake.
Calvino's scene and the Russian émigrés' project alike are plainly a satire upon the idea of mass-produced art. Yet they no longer seem like the comedic or parodic exaggerations of contemporary trends that they must have at the time. In the years since, after all, life has imitated art. With the rise of AI-powered chatbots, computers do now actually have the ability to create original, if lowest-common-denominator, works of art by mining heaps of human-generated data. These programs can now write plausible narrative fiction that has never appeared before, by combining the most preferred elements of the fiction that was written previously, much like Calvino's hypothetical machine & the Russian artists' experiment.
If people were already worried in the 1960s that literature might be displaced by technology, then—what are we to say about our contemporary era? Declaring the death of the novel in 1967 seems, from our current perspective, premature. Many fine realistic novels have been written in the half-century since, and some of them have aged better and retained more of their interest than Calvino's or Barth's experiments have. But as increasingly-sophisticated AI programs learn to infinitely reproduce work that is ever-less distinguishable from human-generated prose, we really might have to worry about what role the human arts could continue to play in the future of our civilization. Ours is surely becoming the real "age of mechanical reproduction."
I of course have argued before that the claims made for AI's transformative potential are somewhat exaggerated, and I stand by it. My fear, though, is that even if many parts of the economy remain relatively insulated from change, and even if automation of certain cognitive tasks in the workplace ends up creating more jobs than it eliminates, in the long run, the parts of our economic lives that will be displaced may be the ones we value and enjoy the most—in particular, things that involve creative craftsmanship of some kind. Even if I could hold a white-collar job in the age of AI, that is to say, perhaps a lot more of it would be devoted to attending pointless meetings than to writing something, even though I would infinitely prefer to spend my time on the latter.
Yet, perhaps declaring the death of the arts in 2023 will ultimately appear as premature to future generations as ringing the death-knell back in 1967 looks to us now. After all, John Barth's essay, it is worth recalling, is not an elegy for the art that was; it is a prospectus for a new literature and a celebration of the kinds of new art forms that he thinks will emerge and are emerging from the state of apparent impasse. In particular, he cites works by Borges that, instead of avoiding the subject of literature's exhaustion, confront it head-on, and then use this as the basis to create new art. Could something similar be achieved in our day?
Barth's essay includes an extended discussion of "Pierre Menard, Author of the Quixote," for instance—a Borges tale in which an avant-garde writer re-creates all of Cervantes' novel word-for-word. By making the story itself a study of the "mechanical reproduction" of art, Borges—in Barth's words—"confronts an intellectual dead end and employs it against itself to accomplish new human work." Elsewhere, Barth expands upon the point: Borges, he declares, is illustrative of "how an artist may paradoxically turn the felt ultimacies of our time into material and means for his work—paradoxically, because by doing so he transcends what had appeared to be his refutation, in the same way that the mystic who transcends finitude is said to be enabled to live, spiritually and physically, in the finite world."
Is something similar possible in our present moment? Can art threatened with extinction by AI manage to use this very fact to somehow transcend its own annihilation and create new works? One can imagine—least inspiringly—some avant-gardist using AI chatbots to generate works of publishable fiction, but doing so "ironically." Barth in his worst moments seems to approve of such would-be "droll" maneuvers; and in fairness they must have seemed more clever and original in 1967 than they would today, when we have all grown weary of postmodern irony, unless it is deployed ironically. More promising is a Borgesian literature in which a human author's prose interacts with and responds to the unexpected productions of a machine.
Perhaps the greatest scope for human creativity to transcend the limits of the age of AI, however, will be something else entirely. I submit it may be found in the domain of pataphysics (sometimes spelled with an easy-to-miss initial apostrophe, 'pataphysics).
This elaborate non-science, the product of the brain of experimental playwright and author Alfred Jarry (1873-1907), has been defined many ways. Jarry's own canonical account in his Exploits and Opinions of Dr. Faustroll, Pataphysician, includes such un-elucidating definitions as "the science of that which is superinduced upon metaphysics" (Taylor trans. throughout). In its later elaborations by subsequent authorities, however, 'pataphysics has come to be described as the "science of the useless." It incorporates such practices as designing and marketing unworkable machines and unhelpful robots. The Japanese art of "Chindogu," though it emerged through parallel evolution, nevertheless partakes of the same spirit.
If AI bots are going to get ever-more sophisticated at delivering what the largest number of people want from their arts, I am suggesting—then perhaps the answer is for human creatives to deliver what people don't want. If the domain of the useful and monetizable is going to be colonized by machines, then perhaps the new frontier for human creativity is to explore to its furthest reaches the realm of the useless. Is this how we might—as Barth advises—use the very extinction we confront as a basis for transcending it? Here among the useless, after all, would be scope to celebrate the joy of creation in its purest form. It would be creation as an utterly intrinsic good, creation for its own sake, stripped of any accretion of instrumental utility. Creation because we can, not because we have to.
Jarry describes the joy of 'pataphysics, after all, as stemming from its concern with "exceptions"—what modern philosophers of science might call "anomalies," or what Forteans would dub "the damned." These are the data points that don't fit into any relevant theoretical scheme: the total outliers. Of course, mainstream science—or, perhaps we should more accurately say, real science—can only concern itself most of the time with the opposite of exceptions. It looks for patterns. This is the only defensible way to approach a science or technology that has any claim to utility, since only patterns have any predictive force, therefore they are the only things that have pragmatic value in the effort to create inductive science.
If we free ourselves of the need to be useful, however, then we can discover the greater joys of dwelling in the realm of the unrepeatable exception—the useless datum or creation that serves nothing beyond itself. Are these not, after all, much more fun and interesting? Jarry's novel, to be sure, does not contest the utility of patterns; he merely suggests that they are boring. Normal science, he laments, is concerned by definition only with "unexceptional exceptions"—that is to say, individual data points that can be replicated under controlled conditions—and these, he declares, "possess no longer even the virtue of originality." Why not devote ourselves to that which is utterly sui generis, by contrast, and therefore possessed of greater intrinsic interest? Why not create what will never be re-created or mass produced, precisely because it serves no function?
Barth would probably say that, instead of talking about this, I should actually do it. He says he prefers the "virtuosity" that completes an avant-gardist experiment to the arm-chair theorist who merely says that such an experiment ought to be done. Perhaps I can only retort that this blog post is an example of me already fulfilling my own stated project. After all, will this blog post do any good? Will anyone other than its author even read it? Could it have been done equally well or better by an AI, if I had just plugged in key terms like "meta-fiction," "pataphysics," "Italo Calvino," etc.? Perhaps so. This, then, is already a pataphysical exercise. I have chosen to write it for myself not because it is useful, but simply because I wanted to; I enjoy it; and I wanted it to exist. You have been reading a 'pataphysical text.
No comments:
Post a Comment