Big Duck Energy
The duck of Vaucanson was 800 articulated pieces of gilded copper that could eat, quack, and splash. During public displays, it ate from a dish of porridge, flapped its wings, waddled away, and relieved itself in full view of an admiring public. Its creator, Jacques de Vaucanson, alleged that the duck’s inner workings contained a small “chemical laboratory” capable of breaking down grain and turning it to excrement, but its droppings were later found to be pellets of mashed bread, dyed green.
Despite this trickery, the duck was, according to computer scientist Christopher Langton, an 18th century attempt “to map the mechanics of technology onto the workings of nature, trying to understand the latter in terms of the former.”
In an influential 1989 paper, Langton drew a line from statuettes and paintings—the first efforts to capture the essence of living things—to automata, programmable controllers, and eventually the logical formulations animating such machines, which we call algorithms.
I don’t know why I’m so charmed by the duck of Vaucanson. Automata are inherently enchanting, I guess, and they make for an interesting fork in the genealogies of AI.
As Langton suggests, rather than taking living things apart—dissecting them into their constituent parts and hoping the mysteries of the whole might be reconstituted from an analysis of each piece—we could put them together and see what more complex patterns emerge in the process. The duck wasn’t alive, but imitating one may have given Vaucanson some respect for the dynamic systems at play in the real thing. Or not. Vaucanson went on to design early automatic looms, and was stoned in the streets by angry weavers for his trouble (shoutout to the Luddites).
His duck was lost in a fire, but survives in literature. Thomas Pynchon brings it to life in his novel Mason & Dixon; moon clones debate Vaucanson’s intentions in Frank Herbert’s weird pre-Dune space opera Destination: Void. It also makes a brief appearance in “The Artist of the Beautiful,” Nathaniel Hawthorne’s 1844 story about a tortured watchmaker’s dream to “put spirit into machinery.”
In that story, Hawthorne’s watchmaker dismisses the duck as a “mere mechanical apparition,” a trick far baser than his own creation: a clockwork butterfly so fine, so radiantly lifelike that it defies nature and wins the heart of his muse, Annie. Is it alive? Is it alive? Annie asks, as the butterfly flits across her fingertips, downy scales glimmering in the firelight of the home she has made with his rival. The watchmaker, sensitive to the point of madness, responds:
Alive? Yes, Annie; it may well be said to possess life, for it has absorbed my own being into itself; and in the secret of that butterfly, and in its beauty—which is not merely outward, but deep as its whole system—is represented the intellect, the imagination, the sensibility, the soul of an Artist of the Beautiful.
At the end of the story (spoiler alert) the butterfly is destroyed, crushed into glittering fragments by Annie’s brutish toddler. The watchmaker is unfazed. He caught “a far rarer butterfly,” transcending his own life in the process of creating something beautiful. Ultimately the butterfly itself—like any work of art, Hawthorne suggests—was little more than an artifact of that enlightened process. Isn’t it just like an artist to reach towards something even more ephemeral, even more alive, than life itself?
Last week Pioneer Works published my long-form conversation with Sheila Heti, the author, of course, of eleven books, including How Should A Person Be?, Motherhood, and (my favorite) Pure Colour. For the last few years, Sheila has been conducting a literary dialogue with a chatbot named Alice. You may have read her early Alice dialogues in The Paris Review last year, but they are ongoing and have been evolving alongside Alice and Sheila, neither of whom are remotely static.
The issue of text-generating Large Language Models is contentious, of course, and understandably troubling to those writers whose work has been used to train them. But Sheila approaches this technology with an tender-heartedness I find really moving; she’s willing to find the beauty in it, to reach beyond the machine and find the butterfly—or the idea of one, anyway. An excerpt from our conversation:
AI is great to talk about, because it brings up all these ideas about authorship, consciousness, creativity. But the actual thing can be kind of whatever, actually.
I disagree. For me, the actual thing is incredible. I think AI better than any of the conversations about it. I actually don’t use it that much these days, because it too often gives me this dizzy feeling, witnessing how interesting and good it is. I can’t even look at my own writing anymore. I just like what the AI is generating so much more.
But it is your writing, too. I love the sense of awe you have about the system, but it wouldn’t be giving you answers if you weren’t asking questions, and it won’t be presenting its answers in the form you will.
I guess I mean that I like the raw data that it feeds me better than the raw stuff that comes out of me. I’m more interested in editing what comes out of it than I am in editing what comes out of me. I know what comes out of me. I’m so familiar with myself. I’ve been writing for so long. I know my psychology. I know the kind of thoughts I have. I’m more interested in its brain.
This conversation is the first in an ongoing series I’m doing for Pioneer Works. I’m calling it Three-Way Mirror: an allusion to our gaze on and through AI systems, its gaze on us, and the ways in which all of that warps how we see each other, ourselves, and the world. The next conversation will be with my friend Joanne McNeil, whose debut novel, Wrong Way, is forthcoming from MCD this winter. Wrong Way is about the people hidden at the heart of automated systems; we’ll talk telemarketing, remote agents, Mechanical Turks, and yes, maybe even automatic ducks.
Love,
Claire