Something to Say

 I have continued my reading of John Barth's short fiction this week, moving on from Lost in the Funhouse to his next collection of interwoven novellas: the wonderfully inventive and occasionally quite moving Chimera, originally published in 1972. One could say that here Barth continues to explore his great theme of the angst the modern writer feels at having nothing to say. Chimera is in part, like its predecessor, a book about not having anything to write a book about. 

I should clarify that I in no way mean this pejoratively. Barth himself is entirely self-conscious about this fear; he himself—as I quoted last time—admits that one of his recurring obsessions is that of the specter of creative "impotency" (to which he draws the obvious sexual analogy in both the earlier collection and in Chimera). And he is committed to responding to this situation, not by retreating into silence and despair, but by forging something out of the very impasse in which he finds himself: he seeks, as he wrote in Lost in the Funhouse—echoing the themes of his earlier essay on the "Literature of Exhaustion"—"to turn ultimacy, exhaustion, paralyzing self-consciousness [...] against itself to make something new and valid, the essence whereof would be the impossibility of making something new." 

For writing so much about the impossibility of writing, Barth was accused at the time of course of a sterile academicism and self-absorption. Some critics, as I noted in the previous post, attributed the involuted productions of the postmodern writers to the unique conditions of their life: they were alleged to be a generation of university instructors who had never left the academy, and who therefore had deprived themselves of anything serious to write about not because of the times in which they lived, but because of their own aversion to and insulation from ordinary human experience. 

Writers like Gore Vidal or Tom Wolfe therefore judged that the postmoderns had found nothing to say about life, not because modernity had deprived literature of value, or exhausted its potential to convey meaning in new ways, but because the postmodern authors in America's universities have never lived. They had nothing to say because of who they were, and the specific way that they had lived (or failed to live), not because of "the way we live now" in some broader sense. Such, at least, was the Wolfe diagnosis—the world, he said, in so many words, is still interesting—it has never been more interesting!; and there is still much to write about within it, if you would only be willing to step into it. 

As much as critics at the time might have been inclined to sneer at the plight of the academic author who can find nothing to write about other than the problem of finding nothing to write about, though, one now feels that Barth and his contemporaries were simply ahead of their time. Perhaps the judgment that literature had exhausted the potential of its traditional forms may have seemed premature in the 1960s and '70s (Wolfe and Vidal seemed to be saying, in response to Barth's plaint, "that sounds like a you problem, not a me problem). But even if that is true, one feels Barth's angst is even more timely and justified in the present, in the age of chatbots and other forms of generative AI. And it is worth noting that Chimera even precisely foresees this concern. Part of Barth's anxiety about the author's obsolescence has to do with the specter of literally being replaced by machines. 

 Most of Chimera is not about this, of course. It first explores the challenges of the contemporary author to find something new to write about, or some new way to say it, in other guises. In the first of the book's three tales, a present-day Barth steps back through time for a charming confrontation with his literary hero, Scheherazade, the archetype of the storyteller, who heroically uses the power of story not only to prevent mass murder, but is also compelled by her circumstances to narrate and maintain suspense at the peril of her own life. Barth confides to the Scheherazade of his rendition that he has suffered a sense of exhaustion at the potential for literary forms; he touchingly admits to the great mythical storyteller that his best works will never be the equal of "her least." And through this metafictional dialogue, Barth manages—just as he promised he would, according to the "ultimacy" program quoted above—make something moving and new—genuine storytelling—out of the very fact of being unable to tell a story. Bravo!

The other two stories in the collection are perhaps even more unwieldy in their concatenation of mythological, autobiographical, and metafictional elements—but I would in no way call them failures for that. And both find still further ways to generate new literature out of the very topic of the exhaustion of literary forms, just as Barth set out to do. The first of these latter two stories, centered on Perseus, shows the Greek hero viewing panels telling the tale of his own life, confronts ambiguities and discrepancies in the different versions of the character's fate, and ends with him properly ensconced in the heavens as Medusa—now transformed into a constellation herself and become his lover—comments ironically on the literary defects in his rendition of the tale. The third and final story in the series, the "Bellerophoniad," describes the efforts of the eponymous hero to conform his life to archetypes of the hero's journey as described in the great works of comparative mythology, even as circumstances militate against him.

Strung throughout these stories are intertextual references to the other novellas, and to Barth's earlier works. An all-female society, for instance, created by Shahryar in order to dodge the strict letter of his murderous vow to his brother, Shah Zaman, in Barth's "Dunyaziad," turns out later in the volume to be the origin of the tribe of Amazons in Greek Myth. So too, the anonymous Greek minstrel who appears in Lost in the Funhouse is briefly mentioned in this collection, where he is described as the progenitor of writing. And an anonymous editor who goes by the initials J.B. in Barth's campus satire Giles Goat-Boy (which I confess I've never finished, it being a great heavy slab of a book that will require a serious commitment of time when I get around to it) shows up in the "Bellerophoniad" on one occasion, as the author of a document—one of several into which Bellerophon's shape-shifting tutor Polyeidus transforms himself in the course of the narrative. 

One would have thought that J.B. stood for "John Barth," of course, but here the author playfully reveals that the character's true identity is as the descendent of one "Jerome Bonaparte," whose own surname was "Bray," and that it was merely a coincidence of their two sets of initials that had allowed Barth to pass off the prior work as his own. This digression within the "Bellerophoniad" depicts Bray moving into the tidewater home of a contrite Barth who supports him temporarily as he seeks a grant from a foundation to finance his next great authorial project—an attempt to create "scientific fiction" (as opposed to "science fiction") generated by a computer program fed with all the world's literature. Giles Goat-Boy apparently involves a similar conceit of computer-generated prose, as the novel itself is purported (as a fictional device) to have been constructed by a campus-wide super-computer. 

Here is where Barth seems most prescient with regard to our present age of chatbots and AI avatars (when online pranksters are preparing imminently, for instance, to launch digital versions of all the 2024 presidential candidates, trained on countless hours of online video of all these individuals, and when many of the generative AI programs can already respond convincingly to prompts instructing them to "write in the style of..." a major author (one wonders with a chill, for instance, if they would not be able to write a Barth-style metafictional experiment if you asked them to)). Tell me honestly, after all, if Mr. Bray's proposal from 1972 does not sound eerily reminiscent of what modern machine learning has actually been able to approximate half a century later: his "scientific fiction," the narrator tells us, would be composed using a computer "that, once a number of works by a particular author were fed into it [...] could compose hypothetical new works in that author's manner." 

Ultimately, Mr. Bray's proposal for how to train these computers is quite similar to how the generative AI models are actually built: by feeding in enough literary data, he surmises, the machine could learn to detect the statistical patterns involved between all the literary elements, and use it to create something like the pure Platonic form of narrative art. The only difference is that Mr. Bray didn't have the internet at his disposal in 1972, and its endless terabytes of data—so he imagines feeding libraries of reference materials into the machine instead. "[H]e began programming his machine," says the narrator, "to compose, not hypothetical fictions, but the 'Complete,' the 'Final Fiction.' Into its maw [...] he fed all the 50,000-odd entries in 'Thompson's Motif-Index of Folk Literature [etc....] Thus equipped, the machine was to analyze the corpus of existing fiction [...] and reduce the ideal to a mathematical model," thereby eventually generating the "quintessential fiction." 

This is what I mean when I say that Barth's angst about the obsolescence of the literary author, far from being overstated, was merely ahead of its time. What appears in this novel as a kind of sci-fi conceit, to illustrate the larger problem of the exhaustion of literary forms through their endless repetition and the excessive self-consciousness of the modern, has become, in the third decade of the twenty-first century, a kind of literal threat. If Barth felt that literary form was already at risk of being used up in 1972, how much more then must we share his anxieties today—especially those of us who work in any kind of writing or communication field—when it is actually possible to train machines on the corpus of an author—or on videos of their movements and expression—and produce convincing machine-generated imitations and likenesses. The thought of a Mr. Bray using a chatbot to create parody versions of Barth novels, or sequels to Barth's works that the true author never penned, is closer now to reality than ever before. 

But if that really is the extreme situation that we confront—the long-heralded Death of the Author, except not through annihilation, but rather through redundancy—we are nonetheless no more entitled to despair now in 2023 than Barth was at the time he wrote his great collections of metafiction. His advice then has merely become more timely and urgent. Now as then, the only way out is through. We cannot unwind the progress of technology or of the development of literary forms. We cannot rediscover a naive and unselfconscious position vis-a-vis literary artifice any more than we can un-design technological innovations that have already been unleashed on the world. We therefore have no choice but to work with these developments, and to do so within the epoch of history we have found ourselves. We are going to have to make of "ultimacy" itself something new—to turn the Death of the Author, paradoxically, into the Author's next great act. 

How to do that? Plainly, we will need a novel whose great theme is the very difficulty of using human subjectivity to find something to say, that needs saying, when everything sayable can seemingly be said by machines. Perhaps one will be able to find material for this theme through, as Barth describes his own methods in this collection, "go[ing] forward by going back, to the very roots and springs of story." I cannot more than echo Barth's hints at the direction such a literature might take. But it seems clear to me that we need a Barthian metafiction for the age of AI. And that is why Chimera and the author's other classic works are worth rediscovering today.