Chaos Bewitched Moby-Dick and AI
In his Truth and Method (1960), the German philosopher Hans-Georg Gadamer insisted affectingly that human understanding is an essentially historical process. The “meaning” of texts (novels, poems), the significance of past eras or events (WWII, the Renaissance) — none of this can ever be, somehow, “settled” or “put to rest”. On the contrary, because we are historical beings, because our consciousness is formed in and of time, understanding will always itself be a function of intersecting temporalities, what he called a “fusion of horizons”: the perspective from one moment must meet the perspective of another; the widening of view, the shift in what can be seen, this is the experience of coming to understand. There are no cheat codes. And it cannot be mechanized. It is the forever dialog of now with then, through which we become what we are (and the past does too). In the piece that follows, Eigil zu Tage-Ravn experiments with some digital-analog horizon-fusing, asking a GTP-3-driven AI system for help in the interpretation of a key scene in Moby-Dick (1851). Do androids dream of electric whales?
— D. Graham Burnett, Series Editor
March 21, 2023
It is among the most memorable moments in American literature. At the start of Chapter Three of his masterwork, Moby-Dick (1851), Herman Melville has his protagonist, the existential castaway Ishmael, newly embarked on his fateful commitment to go a-whaling (is it a kind of suicide? a mere lark? both?), push open the door of a port-side rough-house for boozing mariners: the so-called “Spouter-Inn”. It is dark. It is dank. Savage weaponry from cannibal isles spikes the walls. And those who hunker at the bar are renegades and isolatoes, gruff men of the sea. In the half-light, Ishmael immediately discerns, in the entryway itself, like a warning over the threshold, “a very large oil-painting”.
Ishmael pauses. He puzzles. He peers. He even goes so far as to pry open a little window in the vestibule, to shed some light on a most inscrutable, a most intractable, a most maddening canvas: “Such unaccountable masses of shades and shadows, that at first you almost thought some ambitious young artist, in the time of the New England hags, had endeavored to delineate chaos bewitched.”
But what most puzzled and confounded you was a long, limber, portentous, black mass of something hovering in the centre of the picture over three blue, dim, perpendicular lines floating in a nameless yeast. A boggy, soggy, squitchy picture truly, enough to drive a nervous man distracted. Yet was there a sort of indefinite, half-attained, unimaginable sublimity about it that fairly froze you to it, till you involuntarily took an oath with yourself to find out what that marvellous painting meant. Ever and anon a bright, but, alas, deceptive idea would dart you through.—It’s the Black Sea in a midnight gale.—It’s the unnatural combat of the four primal elements.—It’s a blasted heath.—It’s a Hyperborean winter scene.—It’s the breaking-up of the icebound stream of Time. But at last all these fancies yielded to that one portentous something in the picture’s midst. That once found out, and all the rest were plain. But stop; does it not bear a faint resemblance to a gigantic fish? even the great leviathan himself?
Few images of pure formlessness have attained to the canonicity of this darkly imagined canvas, which has been a touchstone for generations of Melville critics. The grizzled Robert H. Zoellner, for instance, writing in the early 1970s, discerned in Ishmael’s panic before the painting nothing less than “a glimpse of th[e] anteperceptual world” — itself a function of “philosophical or religious doubt, impinging corrosively upon man’s Adamic acceptance of the phenomenal datum.” Hence, Ishmael, standing before the beckoning disorder of a smoky vortex, feels “the consequence of faith’s receding wave” in “the dissolution of those hard, sharp lines, those incisive shapes, which characterize a thoughtless acceptance of appearances.”1 By these lights, the murky canvas of Chapter Three sets up “the ontological vacancy beneath appearances” that, for Zoellner, lies at the heart of the novel as a whole — and which, of course, the ghost-leviathan himself makes furiously and fatally real.
Zoellner wrote that of all the works of literature he had encountered, only Moby-Dick actually frightened him. So much so, in fact, that he called his critical efforts “sheer self-defense”. But if this context of existential combat perhaps purpled his prose, it did not push his interpretation of Chapter Three into outlier territory. On the contrary, his Melville-scholar contemporary, James W. Nechass, writing a few years later, offered a very similar interpretation of the “boggy, soggy, squitchy picture”. Focusing on a close reading of “negative affix” in the (non-) description of the (unseeable) painting, Nechass argued that a swarm of privative and negatory terms (“nameless”, “indefinite”, “unimaginable”, “involuntary”, “unnatural”, etc.) all bespoke the apophatic central theme of the book as a whole: “the ungraspable mystery of life” and “the unknowable truth of the universe”. If, ultimately, it seems that a whale might be at the center of the elusive canvas, that merely confirms the conflation of malevolent cetacean and satanic vacancy from which Moby-Dick moves.2
And a series of more recent readings have continued to shed lurid blacklight on Ishmael’s dark canvas. For Manfred Pütz, the scene is carefully constructed to conflate the position of Ishmael with that of the person holding Moby-Dick in hand, since Melville’s own novelistic achievement is itself a dark storm of murk and metaphysics, before which we as readers stand as befuddled as Ishmael confronting the “black mass” of “nameless yeast”.3 Ought we read that key phrase “black mass” in its fully diabolical, rites-of-Beelzebub sense? Jonathan Cook, writing in Inscrutable Malice: Theodicy, Eschatology, and the Biblical Sources of Moby-Dick (2012), suggests yes. Since he sees the scene in the Spouter-Inn as a Bunyan-esque set-piece, where the pilgrim walks directly into the mouth of hell, there to encounter a menace of Burkean sublimity. The painting is then not only a foretaste of the destruction to come, but also a visual warning over the gate: abandon all hope, ye who enter here. What about the possibility that Melville was playing, in his depiction of painterly formlessness, with a reference to the contemporaneous work of J. M. W. Turner, whose abstracted seascapes were themselves mocked as “pictures of nothing, and very like”? Several scholars have made the case.4
But is the painting actually a painting of nothing? Not really. Or, anyway, not exactly. There is that “one portentous something in the picture’s midst”. And in a final paragraph on the encounter, Melville depicts Ishmael as persisting in the work of discernment — even going so far as to consult with the old-timers at the bar, soliciting their views, and creating thereby a kind of extemporized hermeneutic community. With their help, he eventually makes out a tale of woe in the whirl of darkness, a tale that is itself a premonitory figure for the novel as a whole: a demasted ship in the gale, and perhaps a breaching whale, in its death-throes hurling itself upon the ruin.
For the art historian Darcy Grimaldo Grigsby this scene of specifically collaborative interpretation/recovery is revelatory: begrimed by time, smoke, and grease, the original painting has in fact become illegible. It is history that has rendered it opaque. Its layers of patina disclose its venerable antiquity even as they disguise its image and smear the picture to a mummy-brown stew. In the end, Ishmael is only able to make out its form with the help of those who remember — who are, of course, also those directly implicated in its acquired opacity. “Ishmael’s dark painting is almost lost, but it remains just barely legible on the walls of the Spouter-Inn, surrounded by aged whalers who spit and smoke and drink and talk and see themselves in it.”5 They have, across time, puffing on their pipes, made it nearly impossible to see. And they are the ones on whom its interpretation must now rely.
Grigsby thus succeeds in bringing the question of history itself to the fore — history as interpretation, and interpretation as historical. She sees the fateful painting of Chapter Three as not merely a nebulous icon for the blind force of the unknown, but as cameo for the historicity of consciousness itself, a parable for the dynamic processes by which surface becomes depth — and the equally dynamic processes by which that sedimentation is reversed. Meaning happens in both directions.
Open AI released the initial “DALL- E” image-generating software in January of 2021, updating to “DALL-E 2” the following year. Both platforms make use of the GPT-3 (Generative Pre-trained Transformer) model, a staggeringly powerful artificial intelligence tool that has “digested” a vast amount of human-generated online content, and developed the capacity to make original and context-responsive outputs that are often difficult to distinguish from new human-generated material. By kitting out DALL-E with GPT-3 models tuned to image analysis (and trained up on billions of image-caption conjunctions) the programmers of Open AI created an essentially unprecedented graphical tool: a system that creates novel visual images from text “prompts” provided by the user. It brings strange new eddies of coherence out of the unformed darkmatter of our interwebbed existence. Compelling eddies. Chaos bewitched.
Offered free to the public, DALL-E and DALL-E 2 precipitated an explosion of online play and artistic experimentation — along with a great deal of head-slapping delight, some ethical handwringing, a bit of (well deserved) Silicon Valley braggadocio, and a smattering of gob-smacked Op-Eds. Through the portal of the software, it is possible to feel one is looking into the nascent dream-world of the androids, presently perched on the very cusp of the singularity. At the same time, however, they are very definitely dreaming in our language, since (for now) all the machines know is stuff they get from us. This means each new series of pictures that sift up out of the mind of the GPT-3 engine can also be understood as “ours” in a real way: DALL-E 2 knows the visual unconscious of humanity much better than we do; it shows us new things, which are, simultaneously, algorithmic recombinations of our collective imaginary.
The two on the left had a ghostly air. But the Melvillian mood did not feel honored by the graphical idiom on which the system had fixed. I thought it over, and realized the machine had no way of knowing that we were talking, here, about a painting. It seemed quite fair to tip it off about that. I modified the prompt with a prefatory “an oil painting of . . .” and ran it again.
Nothing not to like. Although I found myself at this point wanting to go back to the start, and take a different tack. After all, I was here working with simply one random sentence in Melville’s total evocation of the mystery canvas. Why that sentence over the others? Why not simply drop the whole paragraph in as the prompt, and see what happens? I left the “nineteenth-century nautical oil painting” preamble, and cut and pasted the whole ’graph right out of Project Gutenberg. I am not sure how much of the text the machine actually processed, but its results were quite new:
Each of these seemed to me “a boggy, soggy, squitchy picture truly”. And indeed of any of them I might be tempted to cry out something along the lines of, “It’s the Black Sea in a midnight gale!” or, “It’s the unnatural combat of the four primal elements!” Or perhaps even, “It’s a Hyperborean winter scene! It’s the breaking-up of the icebound stream of Time!”
The drama’s done. All true, too. Every word of it. I pushed away from the screen with a woozy air of one who could not quite believe what he had seen. Could flukes — very proper flukes! — have lifted their dark sinews to the moonlight out of that black roiling water? And then did I see the cloud-tending bird, white of wing, bend its lofting arc toward that heaven from which Ahab hid his eyes? What force, then, did twist with redeeming will those scimitar blades into a true and glowing heart, that figure of tenderness and love? Had I not just seen it rise, with the quiet grace of a cold dawn out of the inky depths of that primeval maelstrom?
Eigil zu Tage-Ravn is an independent scholar and collector. For many years he served as the keeper of a distinguished private library in Copenhagen, where he also ran a small contemporary gallery. He has contributed essays and criticism to various publications, including Cabinet (Brooklyn and Berlin) and Parkett (Zurich). He chaired the ESTAR(SER) “Committee on Publications” from 2006–2015, and has a chapter in the recent In Search of the Third Bird: Exemplary Essays from the Proceedings of ESTAR(SER), 2001–2021 (London: Strange Attractor, 2021), edited by D. Graham Burnett, Catherine L. Hansen, and Justin E. H. Smith. Presently retired, he lives on a houseboat in the Netherlands with his wife, Lucienne, and their Broholmer, Søpapegøje.