Freddie deBoer's recent post (Great Writers "Tell" All the Time) rolls out a few usual suspects (Jane Austen, Fyodor Dostoevsky) to demonstrate that the "show don't tell" bromide has been one of the more harmful pedagogical cliches of what Mark McGurl describe in his book The Program Era (not that deBoer name-checks that book anywhere in the post). The other axioms are "write what you know" and "find your voice".
I mention deBoer simply in passing despite his inspiring this post because I have also been reading Alan Jacobs and Ted Gioia on artificial intelligence and how it has been used in the arts. Jacobs was the one who proposed that AI slop has become what it was because the Iowa Writers Workshop formulas made it so easy for Ai to replicate those formulas. I discussed that over here.
Ted Gioia has been warning that AI slop is utterly inhuman and even if I can agree on paper I think the grand failure of Ted Gioia has been the failure to explain how and why AI slop is what it is. How is it that the AI project, which has been utterly human from start to finish, become abruptly inhuman only at the content generated from the AI prompt? As Ethan Hein so eloquently put it, AI slop predates AI.
deBoer's proposal is that rules that make sense for screenwriting and playwriting migrated into prose where "show don't tell" is not automatically (or even usually) the wisest way to write. It is here that I can't resist shifting to an essay Leonard Meyer wrote in the last century wherein he contended that one of the foundational mistakes of the integral serialists was presuming that all levels of cognitive engagement worked in the same way with music. Meyer's counterproposal was that just as a molecule doesn't behave the way a cell does; a cell doesn't behave the way an organ does; an organ doesn't behave the same way a system does; and a biological system is not the same thing as organism in a similar way the behavior or patterns of musical organization at the level of the phrase are not the same as those for a larger musical chunk. As Ben Johnston put it in the writings that made it into Maximum Clarity, twelve-tone exhausted its own set of viable options faster than all tonal systems that came before it.
AI has been creating creepy slop because we taught it a bunch of lies about what good art does, if not necessarily about what we consider good art. I respectfully differ from Bryan Townsend that AI slop generates cliches. Love is the biggest cliche there is but we need love. We also can't all be super-Benthamites who abjure all the arts.
Engaging all the arts on the basis of a few simple top-down formulas has been what we've had going on in the arts in the United States for generations. It's one thing to lament the identikit nature of post Joseph Campbell Star Wars everywhere monomythic monotony. That's to complain about the output. The problem with the input filter is that Joseph Campbell's whole approach was garbage. It doesn't necessarily improve if the variation comes via Northrop Frye. These moves have probably had the same effect on narrative as Anton Reicha's description of grand binary form had on early 19th century piano sonatas, leading to a bunch of interchangeable and forgettable piano sonatas from that period nobody knows about because those sonatas were written by people who didn't end up getting canonized. Ethan Hein is utterly write that AI slop predates AI but I am willing to grant that Reicha did not intend for his description of what has come to be called sonata form as a mere formula or recipe.
Ted Gioia claimed in his history of music that through the power of music we can all be wizards. That's a lie. I've made this point in the past but the big theme running across Peter Beagle's novel The Last Unicorn is the contrast between the magic of seeming and the magic of doing. The magic of seeming can seem powerful and amazing but it creates appearances and impressions for those who are able to be won over by it. The magic of doing is a lot more powerful but also more difficult to control and fewer people have the ability to master that kind of magic. It can be read very easily as a parable about two kinds of art. As big and scary as The Red Bull is he's blind and, as the magician points out in the story's end, the monster conquers but never fights. When the unicorn turns, stands her ground and prepares to fight The Red Bull he relents and the unicorns trapped in the sea are released. The old king Haggard meets his death with a strange sense of relief as the bull he'd captured is leaving him.
King Haggard's domain can be read in the early twenty-first century as a kingdom of AI slop with a king who wants to see, if just once in his life, someone who has the magic of doing and not the magic of seeming. Beagle, tellingly, once described himself as not seeing the king as a villain because he saw himself in the character. A particularly acid literary scholarly sort could quip that The Last Unicorn is literature in the magic of seeming and not doing a la whoever-else-that's-been-canonized gets assigned to you in a college literature course. But the academic world of liberal arts education has to have been administered by a bunch of King Haggards for AI slop to be what it is here in 2026.
Apropos of nothing I'm reading James Crenshaw's book on Ecclesiastes. Crenshaw takes the book to have been a fabrication passed off as the work of Solomon so there are ideas Crenshaw entertains I find less and more persuasive but an idea of interest to me is this--Crenshaw pointed out that in the wisdom literature of the Hebrew Bible there is no redeemer. Job hopes, briefly, that he can have a redeemer who delivers him from his misery but Qoholet does not conceive of God as a redeemer and the Book of Proverbs gives you rules to live by because people who don't screw up their own lives have done at least that much. The Wisdom Books are books with no memory of the Exodus and even no recollections of the Law, let alone the Prophets.
The proverbial is wise but also formulaic and yet we have generations of liberal arts educators who swear up and down in theory and practice that the liberal arts are supposed to be prophetic and powerful and shake things up and call down and call up chthonic spirits blah blah blah. But what gets taught is literally proverbial and formulaic. The problem isn't that proverbs are cliches. Cliches are often true and even often wise but Proverbs warns that a fool wielding a proverb is still a fool. Proverbs also warns that if you see a man who is wise in his own estimation there is more hope for a fool than for him. Qoholet warned that even if a wise man claims to have figured everything out that happens under the sun don't believe him.
In his commentary on Job 1-21 C L Seow noted that many biblical scholars have groused that Elihu seems to add nothing to the debate. That's obviously not true. Seow pointed out that Elihu's new contribution is that God reveals some things to some and not others and that this is so is mysterious. Job's friends rely on what may be good theology as far as it goes but the shoe-horn Job's suffering into their established ideas and double down on the necessity of being right.
The crisis of AI slop in the arts seems to me as though it's art generated by Job's friends rather than Job. It's the art that relies on the wise formulas of tenured sages that has no prophetic element because ... when's the last time you heard of a North American tenured academic who thinks prophets are actually prophets and not scribes with political interests whose hegemonic agendas can be deconstructed from the strictures of language and analyzed with respect to systems of control?
We live in such a time that people can publish books in which they not only assume Q was a real document but that the political agendas of the scribal community that produced Q (upon which the Synoptic Gospels are assumed to have relied) can be nailed down. Look, I get that Q is useful as a scholarly shorthand in New Testament studies for all the material in common across the Synoptic Gospels. That doesn't oblige me to think some Q document ever actually existed--that's a scholarly super-fad of the last century or so. Let's think of AI slop in the arts as what you can get if you think the Gospel of Q was ever actually written and not the fabrication of American scholars who think they're smarter than the people who wrote the Gospels. Their scholarly magic can be described as a magic of seeming and not doing. It's the committee made reductionist take on what some self-selected scholars decide the "real" teachings of Jesus were once all the mythic stuff got removed.
And Ted Gioia can wonder how AI slop got to us? No, I think it's pretty patently clear what kind of technocratic self-assured American educational norms helped get us to the age of AI slop.
In postmodern biblical studies we can see that people have realized that you can speculate on the subdivisions of J, E, D and P all you want but that will not get you closer to interpreting the actual Torah as a narrative. If you spend your time separating P and E materials and don't actually read the narrative of Genesis 9 and wonder whether Noah's curse was simply unjust then you're splitting hairs within the realms of academic fads rather than directly wrestling with how unjust Noah's curse is within the Genesis narrative if you stop and think it over. I am not kidding about scholars being so busy separating strands of the Genesis narrative into source critical ruminations they don't deal with the narrative but I'll let Justin Michael Reed's book do the work he already did in it.
You can take that tale of source critical hair-splitting that doesn't engage the narrative itself as a parable and a tale of how we arrived at AI slop. The mystery is not that we got here to the point where AI slop art is with us in a new form, it's that people who taught the way they taught the arts for the better part of a century, devoted to what Jacques Ellul described as the dogma of technique, didn't step back to think that this outcome was on the way.
No comments:
Post a Comment