You’ve probably never heard of Edmund Cartwright, but if you’ve heard of the machine he’s credited with inventing, the power loom, it’s thanks to a recent surge of interest in the Luddites, a people’s movement in England in the early c19 that protested, and physically attacked, the mechanization of the weaving industry and the way that human life and labor was reconfigured and exploited to serve machines and the capital that owned them. Maybe you’ve encountered Gavin Mueller’s Breaking Things at Work or Brian Merchant’s Blood in the Machine or this great comic that ran on the Nib (RIP). When the Luddite movement was finally crushed, their name was quickly redefined by the state and capital to describe someone backwards, opposed to technology out of ignorance or fear.
There’s a greater chance you’ve heard of the Mechanical Turk (the late c18 automaton), or of Mechanical Turk (the Amazon service). The latter provides data processing and other digital piecework services to clients who want the convenience (and cost) of automation for tasks that can’t actually be automated yet, by sourcing digital gig workers to do things like tag data for AI training, clean up messy datasets, or test interfaces. The service is named, with a cynicism that nearly parallels that of Palantir, for a famous mechanical hoax of late c19 Europe, a humanoid machine that purported to play chess.
The Mechanical Turk was the top half of a human figure that emerged from a box which appeared to only contain conventional clockwork. The articulated limbs of the Turk would, when the machine was wound up, play games of chess against a human opponent, apparently pretty well. It was styled as a “Turk” because Europeans at this time thought the game of chess came from Turkey (it probably came from India by way of Turkey), and because such extravagantly complex automata that served only to entertain and amaze were associated with the supposed decadence and wealth of “the east”.
The Turk was a fraud. There was a guy in there. The opening of the box to reveal clockwork was a magic trick, an optical illusion. There was just a guy in there, playing chess using the robot’s arm. This is, needless, to say, a painfully obvious allegory for the modern craze for “AI”-powered digital services, which are all one of two things: either directly powered by human labor, but sold as AI, or powered by machine learning tools that have been trained on and by human output and labor. There’s always just a guy in there. Just a lot of little guys in there.
The Turk and Cartwright’s power loom sit on either sides of a peculiar interface — that between automata and machines as devices of demonstration, entertainment, and wonder, and as tools of reconfiguring human labor, the processes we call mechanization and, in social historical terms, industrialization.
In the wikipedia article on the Turk, you will find the following claim:
The Turk was visited in London by Rev. Edmund Cartwright in 1784. He was so intrigued by the Turk that he would later question whether "it is more difficult to construct a machine that shall weave than one which shall make all the variety of moves required in that complicated game".
The Turk was touring in London in 1783-1784, but there is no evidence that Cartwright saw it in action. An 1824 Supplementum to the Encyclopedia Britannica quotes a letter from Cartwright narrating a conversation he had in 1784. The patent on a spinning machine was due to expire, which would mean an explosion in spinning mills, and Cartwright was of the opinion that the next step of textile work in need of automation was that of weaving. His interlocutors thought it would be impossible to automate weaving, but Cartwright had been working on this for some time.
I controverted, however, the impracticability of the thing, by remarking that there had lately been exhibited in London an automaton figure which played at chess. Now, you will not assert, gentlemen, said, I, that it is more difficult to construct a machine that shall weave, than one which shall make all the variety of moves which are required in that complicated game.
In a 2018 podcast discussion of the Turk, the writer Tom Standage argues that Cartwright’s inspiration by the Turk was a “good thing”, and therefore that the deception or trickery of the Turk was a good thing, because it spurred “progress” and technological development. When I first listened to this discussion the other day, I was so startled by Standage’s assertion that I googled for confirmation of the connection and found the wikipedia quote above. I was quite sure Standage had claimed that Cartwright saw the Turk, as wikipedia does, and I even posted about it on Bluesky, in my startlement. It was simply too good to be true - that these two technological moments on either side of the industrial revolution could be so intimately and personally linked.
But really — it’s too good to be true and bore a closer look. Did Cartwright see the Turk play chess, or did he simply hear about it? I went back and listened again, and found Standage only said Cartwright heard about the Turk. I tracked down what little recent scholarship there is on the Turk. The historian Simon Schaffer (who helped me access the primary sources quoted here) has argued that the connection between the loom’s invention and the Turk, and particularly the story that Cartwright was inspired by the Turk, is not only apocryphal, but an ex post facto attempt to explain how the age of the wondrous automaton gave way to the age of the mill (and the Luddites). The story of the connection does not enter circulation until 1820, decades after Cartwright’s patent, when the forces unleashed by the power loom and similar tools were in full swing.
The Turk was touring in London in 1783-1784. It seems Cartwright might have learned about it by reading a newspaper. A notice in the Morning Chronicle of Oct 21, 1783, explains the way the machine plays chess, including indicating check and refusing incorrect moves, then performs a kind of parlor trick with a knight, and then — and this really gets me —
Finally, this Automaton gives answers to questions proposed by the spectators. This is accomplished by shewing with its fingers on an alphabet-table the letters, which compose the words necessary for that purpose.
That’s right, the Turk had a chat interface, using a ouija board. Gilding the lily imo, and pushing your luck to boot.
Indeed, there’s another way of understanding the story of the Turk, and that’s through the copious and voluminous attempts, throughout its touring life, to debunk and explain it. Some people tried to identify hidden means of controlling it from offstage, or hypothesized a small person inside. These accounts were published in newspapers and as standalone pamphlets. No published account seems to have quite gotten it right (just a normal guy in there), but in another sense every attempted debunking was right: there was simply no way for a mechanical automaton to be doing the things the Turk apparently did.
One story about the Turk is about the intricacy and efficacy of its fraud. A related story is that of Cartwright and the power loom, that the invented capabilities of the Turk inspired inventors to dream bigger about what a new machine might be capable of. But I’d like to see another story written about it, one about all the people who correctly observed that it had to be a fraud — they might not have sussed out the exact trick, but anyone who knew anything about mechanisms had to have known that it was a trick. Automata in 1783 simply did have the capacity to see or listen as the Turk purported to do, never mind the question of automating the process of playing chess.
Sometimes things are too good to be true and sometimes skeptics are right. Sometimes, a little common sense reflection about how technology actually does work can immediately make it clear when someone is lying to you about what it is doing or can do. The more I think about the story that Cartwright was inspired by the Turk, or reread his account of citing the Turk as evidence that machinery could achieve greater levels of sophistication, the harder it is for me to escape an impression of him as the Turk’s most famous mark. In fact, I’m not entirely convinced that Cartwright was not being tongue-in-cheek when he challenged his friends naysaying the power loom with the example of the Turk.
Unless I miss something, the stakes of the original fraud of the Turk were essentially nil — except for the price of admission, no one was selling anything with the Turk, or making any greater claims on the basis of its feats. For the skeptics who sought to debunk it, automata like this were a kind of entertainment, a puzzle to be solved on the stage of public opinion (recall that Turing’s test has its origins in a parlor game). Indeed without the story of Cartwright and the Turk, the story might be one of wonder and magic. But the stakes of the fraud are applied retroactively by the claim that the Turk “inspired” Cartwright. For Standage (quoted above) those stakes are nothing less than the Industrial Revolution. I don’t see why they shouldn’t be essentially the opposite: that Cartwright was so foolish as to not realize there had to be a guy in there, and as a result invented a new mechanical technology that would be applied with little concern for its human cost. The lack of necessary skepticism and common sense required to be taken in by the Turk is transmuted, in this tale, into the myopia necessary to build capital the tools it needed to advance the march of immiseration and extraction.
But — at the risk of being too pat — the story of Cartwright and the Turk is itself something of a Turk. The passing of the torch from the era of the clockwork marvel to the era of the mill was never so near and tidy; history is not a parlor trick. They’ll open the door and show you the gears and close it again, but it’s always worth checking again, because there’s probably just a guy in there.
