Warning: This column includes sexually explicit language, literary theory, conventional pronouns, and other potentially offensive material.
In one of her more recent Substack entries, Leah Eichler reflects on a New York Times book review of an AI novella, Death of an Author, lamenting the precedent set by The Times’s publishing of this review. Playing off the NYT reviewer’s observation that the book left him feeling hollow, Eichler writes:
Similarly, the review left me hollow. Not because of how it was written. . . But the fact that the review exists. It doesn’t take a prophetic writer to see how quickly both books and their reviews — maybe entire newspapers — can be ultimately scripted by artificial intelligence. Will we be out of jobs? Some of us, certainly . . . .
I don’t want to dismiss the threat AI poses to the writing profession, but what I find more interesting right now is how the whole phenomenon is an academic fever dream come true—though no one seems to have noticed yet.
What do I mean?
Back in 1968, the French literary critic Roland Barthes wrote his famous essay, “The Death of the Author,” in which he argues there’s no meaningful relationship between a literary artifact and the biological entity that brought it into the world. “Authors” aren’t creators like God creating ex nihilo. On the contrary, they’re essentially typists through whom cultural currents express themselves, no more responsible for meaning than antennae on top of a television could be said to have produced TV shows.1
Barthes himself was channeling an emerging cultural discourse that rose out of the so-called “New Criticism,” a 1950s school of literary criticism that advanced “the intentional fallacy”—the notion that the author’s intended meaning is irrelevant to any interpretation because a. there’s no way of knowing for sure what that intention was (even if the author tells us, he could be misleading us or himself) and b. because “who cares?” Once an author has released his work to the public, it’s no longer his. Meaning is not made of what he or she intends but rather through what the literary critic Stanley Fish would later call “interpretive communities.”
I’ll give you an example of how it works. I once heard a teaching assistant complain at a literary conference that students resisted her understanding of a scene in Romeo and Juliet. As the play opens, two Capulets notice some Montagues, sworn enemies, approaching. One Capulet says to the other, “Draw thy tool!” to which his friend replies, “My naked weapon is out; quarrel, I will back thee.” To this budding professor, “Draw thy sword, and I will back thee” obviously meant, “Take out your phallus, and I’ll perform anal sex on you.” She found it frustrating that her students resisted such a reading.
Of course, Shakespeare certainly had bawdy sense of humor. Nonetheless, it seems improbable he intended these lines to be read that way. But if a student were to object that “Shakespeare didn’t mean that,” the proper academic response would be, “You’re committing the intentional fallacy. You think you can know or that it matters what Shakespeare intended. It doesn’t. What matters is that a queer audience2 might see these lines as constituting a homoerotic double entendre.”
That was back in the 1990s. This line of thinking is so ingrained today that any self-respecting sophomore English major would know better than to raise the issue of intention in discussing a play.
So here’s the thing: if intention doesn’t matter, if authors are just complex antennae that channel lines of cultural discourse, then there’s no meaningful difference between a human and a machine intelligence “authoring” a novel.
For that matter, there’s no such thing as a novel, because not only did 1960s and 70s literary critics do away with authors, they also exploded the concept of art. It’s not a novel or novella. It’s not a poem. Or a short story. Or a song. It’s a “text.” Words like “novel” privilege a text the same way that words like “author” privilege the biological entity through which the text arrived. We don’t “write novels.” We generate texts. So, academia has suggested for the past half-century.3
From a literary theory perspective, the only difference between James Joyce and ChatGPT is that “James Joyce” is the name associated with a more advanced biologically generated language-based chatbox. He may have “thought” he authored Ulysses, but really, all he did was reassemble old texts and combine them with currents of discourse prevailing in his own time.4
So, if AI gets more proficient, if it eventually generates more readable, more entertaining, less “hollow” texts, literary theorists ought to celebrate. They’ll have been proven right. You and I, rubes that we are, thought there was something like a soul trying to communicate with us when we were reading, say, Mrs. Dalloway. But really, it was just the product of a clever algorithm.
Of course, the celebration may be brief, seeing as there won’t be much need for English Professors anymore. At such a point, it would make more sense to farm out literary analysis to the Engineering schools who have a better understanding of complex algorithms. Or better yet, farm out such interpretations to other AIs. ChatGPT or the like could write the next post-human masterpiece, and then one of its electronic colleagues could explain its significance. And all of us retired English professors could watch in admiration.
It’s funny, you know. There’s a lot of hand-wringing nowadays about declining enrollment in English Departments. When you consider that literary scholars long ago replaced authorship with algorithms, it’s not so hard to figure out.
And what does all this have to do with being a perplexed Jew?
Well, part of it hearkens back to my dark night of the soul observations. But it also has to do with why I left “The Church of Art.”
There was a time when William Blake was my prophet, Matthew Arnold my priest. When I got to Vassar in the 1980s, my religious outlook took a serious hit from the readings I was doing in English and History. But, with the help and encouragement of my professors, I turned to literature for meaning and purpose. Poetry became scripture and scripture became poetry. Blake’s The Marriage of Heaven and Hell was my bible; Arnold’s Culture and Anarchy my Talmud.
But then I got to graduate school. At the University of Chicago, professors laughed at the idea of taking a poet at face value or even trying to unearth deeper levels of meaning from the poet’s perspective. “No, you can’t do a Blakean reading of Blake’s poetry,” I was told. The point was not to try to understand what a writer was “trying to say” but to expose the ways in which the text reflected its culture of origin and functioned either to shore up an oppressive status quo or to “subvert it.”5
After seven years or so of that sort of thing, I kind of lost my enthusiasm—not because I believed any of this, but because the community I sought, the worshipers of literature, didn’t exist anymore. The people who ought to have been its pastors had demolished6 literature in much the same way that parishioners demolish their church in Jamie Quatro’s wonderfully disturbing story “Demolition.”
Or, to use another metaphor, I, a one-time disciple, arrived at the Holy Temple of Art, only to find that hostile scholars and theoreticians had stormed it and were sacrificing to alien gods its most sacred works. I paid obeisance as required to earn my place in the new priesthood, but in my heart, I was in exile, in galus. Until I found Orthodox Judaism . . . .
That’s how TVs used to work in the 1960s, with these weird metal sticks that picked up broadcasts ;-). The metaphor is mine, not Barthes.
I.e., one of an almost infinite number of interpretative communities.
Having the privilege, for the moment, of teaching creative writing, I not only don’t embrace such jargon, I strongly discourage it. “There is no text in my class,” I tell my students, not because I agree with Stanley Fish that there is no such thing as an authoritative interpretation but because I disagree with the profession’s reduction of art to artifact. We don’t use the T-word in my classes. We call a short story a short story.
Well, to be honest, Finnegan’s Wake really does read like it was written that way.
"Subversion” was all the rage in the 1990s. If you loved a text, you tried to prove it was subversive so as to save it from attacks as being “reactionary.”
They called it “deconstructing.”
I really like the closing metaphor in this essay, which demonstrates that iconoclasm profanes institutions that have great value.
You are actually giving me flashbacks to grad school. I chose to write my (never completed) dissertation on Defoe, because I didn’t really like his books much and knew it would be much easier to “deconstruct” and subvert them. It would have been too depressing to do that to Henry James, or Jane Austen, or Samuel Johnson, or one of my other beloved authors. But isn’t it revealing that grad school encouraged me to think this way?
And isn’t it limiting too? To take your opening example, from Romeo and Juliet, I have no doubt that Shakespeare intended the double entendre. (I know, I know, authorial intention.) His plays are full of these double entendres, which enrich their meanings. These are young men, after all, and they are trying to prove their masculinity by perpetuating this ridiculous and deadly feud. Their swords are a symbol of that masculinity. But how boring to say that the only acceptable way to read the scene is to say it’s about gay sex. Ok, now what? And so what? The theory professors who wanted us to subvert every text were themselves imposing a new orthodoxy, and one that is much less interesting than what came before.