The true genius of automata
- May 26, 2023
- Gaby Wood
An automaton chess player bewitched the public of the eighteenth and nineteenth centuries - but its real value was all-too-human.
In 1769, an eminent Hungarian mechanician built a machine to amuse the Habsburg empress, Maria Theresa. It was an automaton dressed in Turkish clothes, seated behind a cabinet to which a large chessboard was affixed. When its opponent made a move, the Turk – as it became known – would lift its arm, grasp a piece of its own, and set it down on another square. The mechanics were impressive. But the Turk’s game was even more so. The Automaton Chess Player challenged some of the best players in Europe, and later toured America. Over the course of its long life it played against Philidor, Benjamin Franklin, Catherine the Great and Napoleon – and rarely lost a match.
Thirty years earlier in France, Jacques de Vaucanson had caused a stir with an automaton that could play the flute, and Wolfgang von Kempelen’s new automaton chess player was said to be ‘for the mind […] what the flute player of M. de Vaucanson is for the ear’. It was, in other words, an early attempt at artificial intelligence. Could a machine think? Could it play chess? Audiences responded to the automaton with wonder – so much so that its popularity worried its inventor. Soon after it was first shown, he dismantled it.
Kempelen never in fact said his automaton could think. He claimed only that it was a toy, a trifle made to amuse the empress. ‘Philosophical toys’ were much-exhibited during the Enlightenment; not merely marvels but vessels for ideas. When Vaucanson came up with a small companion for his flute player – a duck that could defecate – it didn’t seem to matter that it was (almost obviously) a fake. The point was to display a proposition: if a machine could shit, would that mean it was alive?
Kempelen designed water pumps and steam engines; he made a typewriter for a friend of Mozart who was a blind pianist; he built the fountains at Schönbrunn palace, the Habsburgs’ summer residence. This machine was one he described, overtly, as an ‘illusion’. As Vaucanson and Kempelen knew, in order to consider the possibility that artificial life could replace human intelligence, it is not necessary for it to exist. It’s only necessary for it to be imagined. Conversations around AI today are not far from the reactions to the Turk in the eighteenth century. And we are well past the point where a machine can beat a man at chess.
Twelve years after he’d dismantled it, Kempelen was asked to resurrect the automaton. The request came from Joseph II, Maria Theresa’s son and successor, who was engaged in an anti-Turkish diplomatic campaign. Kempelen was by now a privy councillor. The politics are not irrelevant: whatever the original reason for the automaton’s dress, there was some orientalism involved – forces associated with darkness were also associated with the East. As entertaining and clever as the chess automaton was, it did Joseph II no harm to advertise the public’s mysterious, uncanny opponent as Turkish. Kempelen was granted a two-year leave of absence from his post at court, and took the automaton on tour.
As soon as they embarked – travelling from Vienna to Dresden, Paris and London – attempts to ‘expose and detect’ the automaton’s secret began. By 1790 over half a dozen pamphlets in this vein had been published, and more were to come. ‘It cannot usurp and exercise the faculties of the human mind,’ wrote one commentator. ‘Exercise’ would have been enough; but ‘usurp’? As if human beings had a natural right to their abilities, rather than simply being able to do certain things. Something about the automaton’s apparent capacity for reason made people who prided themselves on ratiocination reveal their irrational fears through language. It’s not conscious thought that makes us human, these reactions suggest, but unconscious ones. (We’ll return to this idea.)
In exhibition rooms across Europe people crossed themselves on entering, and it was said that ladies who attended the exhibition sometimes fainted from fear. Some suggested that the machine must be manned by a demon of some kind. What Kempelen called an ‘illusion’ in the magician’s sense of the word was often taken to be a spectral one – a hallucination, or a communication with the dead.
An apocryphal story is told of the seventeenth-century French philosopher and mathematician René Descartes: that on the last journey of his life he travelled by ship with a daughter named Francine, who turned out to be an automaton. On discovering her, the sailors, believing this dark invention to be responsible for a storm that threatened their lives, threw her overboard.
Descartes – the man behind the idea that reason is what makes us human (‘I think, therefore I am’) – was indeed interested in mechanical life. We know that he had plans to make a dancing man and a flying pigeon. He did have a daughter called Francine. She died of scarlet fever at the age of five, and her death was said to be the greatest sorrow of his life. But this Francine is a fiction. Why was she invented? The story gets at some of the anxieties people had about automata, and at their perceived connections with the dead.
In 1814 E.T.A. Hoffman, whose short story ‘The Sandman’ became the basis for Sigmund Freud’s examination of ‘the uncanny’, rehearsed his explorations of that idea in another story, ‘The Automata’. In this one, a machine clearly based on Kempelen’s is said to speak, and is given psychic powers. When asked a question, it ‘seems to read the very depths of the questioner’s soul’. One character describes such machines as ‘mere images of living death’. He imagines a scene that would become central to ‘The Sandman’: a living man dancing, unwittingly, with a lifeless mechanical partner. ‘Could you look at such a sight, for an instant, without horror?’ he asks.
In the story about Descartes, Francine is a consolation. Driven by grief, the philosopher resurrects his daughter. In Hoffman’s account the Turk and his fellow machines are akin to revenants, bringing frightening news from a land beyond that of the living. In both cases, death is part of the story.
When we respond to artificial intelligence – the notion of it or, now, the fact of it – with fear, we may rationally argue that our concern is what makes us human. But we are also asking: what keeps us alive? And what makes that worth it? Right from the start, artificial life has presented itself to us as a kind of haunting: a reflection of ourselves as unliving, or unnecessary. This, it seems to me, is not the machine’s problem – nor even, fundamentally, a problem introduced by the machine. It is one to be examined in ourselves. When the chess automaton was due to play against Philidor, the greatest player of the eighteenth century, Kempelen approached Philidor in advance and asked if he’d be gracious enough to lose against the automaton. Philidor agreed, but only if the automaton played well enough for Philidor not to lose his reputation. It didn’t. That was one of the few games the automaton lost. But many years later Philidor told his son that no other game of chess had left him so tired.
The chess player’s cabinet was four foot long by two foot deep and three foot high. It had three doors at the front, a single drawer at the bottom and two small doors at the back, one of which was hidden in the figure’s spine, and revealed only when the Turk’s fur-trimmed coat was lifted. The figure’s right hand rested on top of the chest (it played right-handed) and in its left it held a long, thin pipe. Its head (clad in a turban) moved on its neck, its eyes moved in their sockets. It surveyed the board before making a move. It rapped to indicate check or check-mate.
Wherever it was exhibited, the showman’s routine was the same: it involved an elaborate display of machinery, much examined by would-be unveilers. Kempelen would open the door on the viewer’s left to reveal a compartment lined with dark-coloured cloth and filled with cogs, cylinders, wheels – it looked like the inside of a clock. He would then move to the back of the cabinet and, opening the door directly behind the first, shine a candle through the opening, so that the absence of anything but machinery was fully visible. He then locked the door at the back, came around to the front again and opened the drawer at the bottom. From this he would take the chess pieces, and a cushion for the figure’s elbow (another philosophically brilliant effect: the machine must be comfortable). The drawer would remain open as Kempelen revealed the interior of the main compartment by opening the two doors on the viewer’s right. This space was lined, like the adjoining one, with dark fabric. It was mostly empty, except for two quadrant-shaped pieces of metal near the top, and what looked like wires connecting them. At this point, all the drawers and doors were (or appeared to be) open. Kempelen would then turn the machine around so that its back was facing the audience and lift the back of the figure’s jacket to reveal two more compartments: one containing a mechanism like a piano pedal, and another one containing wires, in the figure’s thigh. The audience could now see that the figure was seated on a decoratively carved wooden stool, attached to the cabinet. Bunches of cloth made up its Turkish pantaloons, and its feet were crossed in a side-saddle position. The idea was to give the audience the impression that they were seeing the whole interior at once – or, as one commentator put it, ‘the automaton stripped naked’.
From the outset, it was assumed that there must be a person inside the machine. Those seeking to expose it were concerned only with where the person could be hidden, and what sort of person could fit into the space. Was it a child? One writer calculated that the cabinet could fit a ten or twelve year-old. Did the human hide somewhere in the box, or slip into the figure and see the game through its clothes? And what of the level of chess? Was Kempelen directing the moves remotely? He seemed to have a magnet in his pocket…
When I first wrote about this machine twenty years ago I was gripped by the pamphleteers and their furious attempts to expose the Turk. Now other scenes strike me as rather more poignant. The Turk when seen from behind feels more alive than it does from the front: because of the invasion of privacy, perhaps, in removing its clothes. There is an intimacy in this, a connection, and a sense of responsibility. Kempelen’s friend Gottfried von Windisch described visiting him at home, where you would walk through the house Kempelen shared with his wife and children to reach his workshop and study – rooms filled, respectively, with tools and marvels. There were large oak dressers containing books and antiques, the walls were covered in prints made by Kempelen himself. Kempelen had children, I think to myself now as I picture his house, This changes everything. It doesn’t, of course (the facts are the same) but these two images – the Turk stripped bare and Kempelen’s human progeny, not all of whom lived to adulthood – seem to me to give the story a place in the world of human feeling, and not merely that of rational thought.
Kempelen dismantled his machine for a second time in 1789, 20 years after he built it. In retrospect this appears to have been a response to a pamphlet by Joseph Freiherr zu Racknitz, which exposed the secret with some errors but a large degree of accuracy. That was never said at the time, though, and after Kempelen’s death in 1804 the automaton passed into the hands of Johann Nepomuk Maelzel, a man credited with the invention of the metronome. Maelzel – undeterred by the apparent exposure – gave the Turk a second life, touring not only Europe but the United States and parts of South America in the first half of the nineteenth century. The machine ‘died’ in 1854 in a fire at a museum in Philadelphia, almost two decades after Maelzel’s death. In that time, as one expert later put it, the automaton’s success had relied on members of the public who had ‘refused to know a secret which had been exposed and published a dozen times’. They wanted to believe it. Or wanted not to believe it. Belief had to continue to be in play.
As one of the operators finally revealed first-hand in 1834, there was a human player inside the machine – not any old player but a good one. In order for the automaton to continue to win its matches, members of the chess-playing circle all over the world had to be initiated into its secret and recruited to play inside the box.
Conditions were extremely cramped. The ‘director’, as the human player was called, would slide back and forth and pull up partitions as the exhibitor showed the interior to the audience. There were breathing holes in the cornicing around the lid of the box, but it was still hot and airless inside. The player saw by the light of a candle, and there were candelabras on top of the chest to mask the smell of wax. Each time the opponent made a move, the chess piece would stimulate a magnet on the underside of the board, and the director would replicate this on a small travelling chess board so he could keep track of the game, before operating the real wonder of the machine: a pantographic device that allowed the false figure to move the chess pieces visibly. Maelzel added sounds to the machine – he not only had it say ‘échec’ but added mechanical noises that could be set off by the director from inside, in case he needed to mask a sneeze.
Why did they do it? For the chess coterie, the question was not whether there was someone inside the box, but who. They attempted to recognise their fellow men by the machine’s style of play. The guessing game, in other words, was based on character and idiosyncrasy – exactly what it was thought the automaton couldn’t have. And it didn’t; what had character was the player within. Yet it was by hiding inside a false self that they each proposed to reveal their character most clearly: can you guess who I am, without seeing me?
There was another reason to attempt the operation of the automaton. Many of the best players in the world have been fond of blindfold chess – the moves are all retained in the player’s mind. Playing inside the box must have been akin to that: physical discomfort and near-darkness made them concentrate all the more. It accentuated their powers. For the artificial being it may have been an impossibility, but for the players who operated it, playing inside the Turk was a game of pure reason.
Was the machine acting in the service of the man, or was it the other way round? The memoirs of those who played inside the box, and their descriptions of others, leave one thinking about its aftereffects. Some players were small – one, Charles Schmidt, was a 16 year-old student, and Jacques François Mouret, who gave away the secret, was said to be ‘practically a dwarf’. But a maths professor named Anderson and the more famous Wilhelm Schlumberger were over six foot tall. Schlumberger travelled around America with Maelzel and nearly gave the game away when, on a hot day in Baltimore, Maelzel opened the windows as Schlumberger was emerging from the box backstage. They were seen by two boys in a neighbouring building. Schlumberger developed a pronounced stoop and died of yellow fever while touring with the automaton in Havana. Mouret fought with Maelzel (‘Tell the King the automaton has a sore throat!’ he threatened, when Maelzel wouldn’t pay him before they performed at court in Amsterdam). He eventually became paralysed, sold the secret to a magazine in Paris and died destitute, of drink.
The automaton chess player was a puzzle of pure reason. But is unreason, eventually, what makes us human? Many writers, from Stefan Zweig to Vladimir Nabokov, have written of chess as leading to insanity. Is that something we should strive to protect in ourselves, or has it only brought us trouble? In the case of these men, their symbiotic lives with the Turk appear to have driven them to the final, fallible, most human qualities. A machine could not, at the time, play chess. But it could not disintegrate or go mad either.
No one noticed that Kempelen was trying to show them the future. The pantographic mechanism within the Turk wasn’t just there as a decoy, to hide a man – it was the true subject of the exhibition, a feat of mechanical ingenuity that would go on, many years later, to contribute to the development of artificial limbs. In the late eighteenth century his audiences were too busy being frightened of the mysterious figure, its uncanny affect and the threat it represented. Yet Kempelen had invented something that could help mankind rather than replace it. No one imagined that, or no one was looking.