BROKEN KNOWLEDGE (14)

By: Joshua Glenn
November 30, 2025

University of Toronto philosopher Mark Kingwell and HILOBROW‘s Josh Glenn are coauthors of The Idler’s Glossary (2008), The Wage Slave’s Glossary (2011), and The Adventurer’s Glossary (2021). In 2022, they engaged in an epistolary exchange about science fiction. Via the series BROKEN KNOWLEDGE, the title of which references Francis Bacon’s philosophy, HILOBROW is pleased to share a lightly edited version of their exchange with our readers. Also see Josh and Mark’s previous exchange 49th PARALLEL.

DEFAMILIARIZATION

AI-assisted illustration by HILOBROW, for our serialization of Max Brand’s 1919 western novel *The Untamed*.

8th November, 2022
BOSTON

I’ve written to you “offline” (outside the context of this exchange) to commiserate about your prolonged hospital stay and Covid. I hope that now, a week later, you’re both covid-free and home safe and sound?

Thank you for unpacking your take on the uncanniness of cinema, a bit. What you say about the familiar yet strange repetition of cliches is a key philosophical point around which my own thinking has been circling. (As you’ll see below, when I get into talking about AI’s relation to conventions.) Thinking of “the movie camera as the inverse of the V-K machine” is an astonishing insight, one at which I think only Mark Kingwell could arrive.

Have we ever talked about my own defamiliarization practice around movie-watching? It’s something that I stumbled upon accidentally, when I was taking time off from college c. 1988, and couch-surfing at a friend’s place in Boston (where he wasn’t supposed to be living, either). His TV’s audio conked out — this is a pre-Web/laptop/streaming anecdote — and at one point, no doubt while stoned, I watched a late-night showing of the great 1979 action movie The Warriors. A silent-movie version, that is. I learned so much about the movie from that experience — i.e., having been liberated not only from its dialogue but also its sound effects, music, etc. By the time I later discovered the movies of Buñuel and other avant-garde movie makers, I was primed to appreciate a director’s principled refusal (or ironic decision) to use music to manipulate the audience’s emotions. I still watch movies with the sound off — on airplane trips, but also when I’m doing a particularly tedious piece of semiotics work — and always learn things about those movies. I’ve become particularly attuned to blocking. But that’s a topic for another discussion altogether. Do you have any defamiliarization techniques you use when watching movies? This goes back to my question: How do you, Mark, observe yourself at the movies?

You mention that AI is more ubiquitous all the time, and of course that’s so — advanced web search engines, recommendation engines, intelligent assistants with natural-language user interfaces, chatbots, and the like. I remember enjoying the process, years ago now, of “training” the Netflix search engine to recommend only movies that I might actually want to see. I’m currently playing around with the AI-generated art-making tool Midjourney, which is also quite enjoyable, though creepy for two reasons.

(1) It’s unsettling to think that this tool is already good enough to generate images that would not look out of place on the cover of, say, sf or fantasy novels. One gets the sense that illustrators who charge next to nothing (say, via Toptal, Minty, Upwork, Dribbble [sic], or one of these other freelancer platforms) will continue to find work; and same thing goes for really talented illustrators with a unique, “ownable” illustration style. But the middle of the market is doomed; cf. what has happened in book publishing, where books must be either digital or elaborately produced to make money. AI is coming for me, too — there are AI tools that produce something fairly close to semiotic analysis (of vast image sets); quickly and cheaply. Yikes!

(2) Midjourney is getting better and better — because so many people are using it, and it’s paying attention to which produced images they like and dislike — at producing “realistic” depictions. It’s fascinating to observe the results now, though, when all the kinks haven’t been worked out yet. One soon realizes that AI regards all this imagery as… fungible, let’s call it. If I ask MJ to create a Remington-style painting (make that: digital image that looks like a painting) of a cowboy on horseback, two of the four results may be pretty much exactly what I requested. But pay attention to the other results and you’ll see cowboys fused in grotesque ways with their horses, cowboys with horse’s heads or tails, two-headed (or -assed) horses that are like Dr. Dolittle’s pushmi-pullyu. MJ recognizes what the distinctive features of “cowboy” and “horse” are… but it’s assembling these features based on conventions that it hasn’t learned yet. There is no “right” or “true” for MJ — there are only conventions that a majority of users seem to prefer. Brrr. I find that chilling! Because I don’t want to face up to uncanny truths about my own convention-driven consciousness, daily life, and social order.

Swift is already rubbing our noses in these sorts of uncanny truths when — in Gulliver’s Travels — he imagines a room-sized machine designed to allow “the most ignorant Person” to “write Books in Philosophy, Poetry, Politicks, Law, Mathematicks and Theology.” Swift’s “Engine” contains myriad “Bits” crammed with all the words of a language, “all linked together by slender Wires” that can be turned by cranks, thus generating all possible linguistic combinations. Squads of scribes produce hard copy by recording any sequence of words that seems to make sense. Yeah — Midjourney!

There are Radium Age proto-sf stories about AIs or proto-AIs whose programming (and, one imagines, vast amounts of user-testing that never gets discussed) allow them to simulate something like humanity, or at least concern for humankind and our needs. But something inevitably goes awry and the mask slips — and we realize that if this is indeed an intelligence, it’s an alien intelligence. It doesn’t really see things our way at all; it’s just very good at obeying the conventions we humans prefer. The “Machine” in Forster’s “The Machine Stops” (1909) is my favorite example, but others include the “phonographic apartment” in R.O. Eastman’s 1910 story of that title (which ends with the apartment’s occupant desperately attempting to get out of the apartment to disconnect a cable — one thinks of Dave making his way to HAL’s processor core, in 2001: A Space Odyssey; and the Dweller, from A. Merritt’s The Moon Pool (1918-1919), an artificially intelligent entity that has capacity for good or evil — but is mostly evil. There’s also a wild play from 1930, by Lionel Britton, called Brain — in which an enormous mechanical brain takes over the planet, ushering in a utopian era (just so long as humankind is willing to obey its dictates). The sort of thing that Star Trek would explore nearly 40 years later; or Iain M. Banks — in Banks’ case, with subtle nuance — nearly 60 years later.

One could go on with more examples, but I’ll stop there. Because I also want to ask you about how ontological investigation might be revitalized (according to what I know of your project) through close attention to movies — particularly since ontology is a “dead” area of philosophy. But — why is ontology dead? Exactly might it be revitalized via your project — and are you sure this is wise, Dr. Frankenstein? Your project promises to be a work of necromancy — uncanny indeed.

Josh

***

ALSO SEE: Josh’s BEST 250 ADVENTURES of the 20th CENTURY list | Mark on BATTLESTAR GALACTICA and THE HONG KONG CAVALIERS | Mark and Josh’s exchange 49th PARALLEL.

BROKEN KNOWLEDGE (14)

Categories

Tags