Skip to main content

The Mechanical Turk: Artificial Artificial Intelligence

One of the earliest video games made for computers was chess. It’s not hard to see why chess was chosen: the rules are pretty simple, and the game is widely played. Artificial intelligence mastered chess early on, and programmers have long been able to set chess programs to play at different levels of difficulty. The first person to suggest that a computer might play chess was the celebrated computer scientist Alan Turing. Turing started talking about this in the 1940s, and in 1950, he wrote the first computer chess program. Turing himself was a weak chess player, but he started something, and a lot of others agreed. It was a common belief that by 1970, the world chess champion would probably be a computer. This never came to pass, of course, probably because human beings still got to decide who could enter chess tournaments in 1970 (and they haven’t given up that privilege yet), and humans never let computers in. With the arguable exception of IBM’s Watson’s appearance on Jeopardy! in 2011, humans would never go down that road. In fact, they already had.

Related image
Watson on Jeopardy!, doing rather well.
Or maybe not. The first artificial intelligence purported to have played chess actually appeared in 1770. This was the creation of Hungarian inventor Kempelen Farkas. Kempelen called his invention the Automaton Chess Player, or sometimes the Mechanical Turk. It was a remarkable machine. The Turk was a large wooden box with a chessboard on top. On one side was the carved wooden figure of a Turkish man, dressed in a turban and Ottoman robes. The machine was designed to play a strong game of chess against any human who sat opposite the “Turk”. Remarkably, it did just that. By all accounts, the “Turk” was very good at what it what it was designed to do. From the beginning, Kempelen set his sights pretty high. His “Turk” was brought to the court of Empress Maria Theresa of Austria, and she was impressed. The “Turk” could perform some of the most complicated chess moves, including the knight’s tour, a move that even the best chess players will execute with difficulty. The Mechanical Turk quickly developed a reputation and toured Europe, matching wits with intellectuals and heads of state from all over. The “Turk” went on to defeat the likes of Benjamin Franklin and Napoleon Bonaparte. Kempelen himself died in 1804, but the Mechanical Turk went on defeating excellent chess players in very public exhibitions until 1854, when it was destroyed in a fire. So: how did it work? Many observers wanted to know this, and Kempelen made sure they had something to look at. The side panel of the “Turk” could open easily, and it would reveal its very complicated clockwork-like mechanism, full of gears and belts and so forth, just like you might expect to see inside a clock. It was an impressive device. But the real question, the one that we know enough to ask today, is: how can you create artificial intelligence with clockwork? The answer is actually pretty simple: you can’t. Or, at least, it remains to be seen if one day someone will, because Kempelen Farkas didn’t do it, all appearances to the contrary. The inspiration for the “Turk” came not long before he introduced it to the world. Kempelen saw a performance at Maria Theresa’s palace of the illusionist François Pelletier, and after speaking with him after the show, vowed to return with an act that would top anything Pelletier had done that evening. In short order, the Mechanical Turk appeared, and started beating chess players left and right. But it’s not quite accurate to call the “Turk” “artificial intelligence”, because there was nothing artificial about it. Inside that large box under the chessboard was artificial artificial intelligence: a human being. Yes, the whole thing was a hoax. It was a pretty good hoax, too, but that’s what happens when you don’t skimp on your resources. Kempelen didn’t put just any human being inside that box. Instead, he managed to get the greatest chess players of his day to do it.

Image result for Mechanical Turk
Artist’s illustration of the Mechanical Turk without its most important component.
The “Turk” was an immediate sensation, starting its tour of Europe soon after its debut in front of the empress. Friends and colleagues naturally wanted to see the inner workings of the machine, but Kempelen of course didn’t show them, because all they’d see was an empty box festooned with ornamental gears. They’d also be able to figure out just how it worked. The “Turk’s” chessboard was made of a thin sheet of metal, and at the bottom of the white chessmen (which the “Turk” always played) were powerful magnets. Inside the box, the human chess player had corresponding pieces, also with magnets, and he’d use them to move the chessmen around the board above. There was even a horn set up in the “Turk’s” mouth, allowing the player inside the box to speak. The only word he would speak was “Échec,” which is French for “Check”, an essential word in any chess game. Despite the popularity of the “Turk”, Kempelen was not pleased with it. He was an inventor, after all, and would rather have spent his time on his real projects: steam engines and devices to replicate human speech. The success of the hoax did raise Kempelen’s reputation, as well as line his pockets, but he wanted to get away from what he saw as a distraction. The public, and his government, wouldn’t have it, though. Even though Kempelen dismantled his device and announced its permanent retirement, Emperor Joseph II of Austria ordered him to put it back together in 1781 for an exhibition before Russia’s Grand Duke Paul. Kempelen followed orders, though he wasn’t happy about it, and the “Turk” started another European tour, one that kept on going. He tried to sell the “Turk”, but never managed to find a buyer. Following his death in 1804, Kempelen’s son sold it to Bavarian musician Johann Nepomuk Mälzel for 10,000 francs—half the price that Kempelen had been asking. Mälzel was much more of a showman than the “Turk’s” inventor was, and gladly paraded it around Europe. In 1826 he brought the “Turk” to the United States, where it also caused a sensation. Mälzel carted the “Turk” to and from Europe, traveling as far west as the Mississippi River, into Canada, and as far south as Cuba. In 1838, on his last tour of North America, Mälzel’s companion, chess master William Schlumberger, contracted yellow fever and died. Mälzel was suddenly left without the essential mechanism that makes artificial artificial intelligence work. Without a “functioning” device, Mälzel packed up the show and set sail back to Europe. On that voyage, Mälzel died at sea, leaving the “Turk” and all his possessions with the ship’s captain, who saw to it that they made their way to Mälzel’s friend, American businessman John Ohl. Ohl sold the “Turk” at auction in Philadelphia in 1840, when it was purchased by John Kearsley Mitchell, the personal physician of Edgar Allen Poe, and a great admirer of the “Turk”. Mitchell was an enthusiast of the “Turk”, perhaps inspired by Poe’s celebrated essay “Maelzel’s Chess Player”, which (incorrectly) speculated that a mechanical chess player must always win. Mitchell gave some performances of the “Turk”, but interest in the novelty waned, and he eventually donated it to the Chinese Museum in Philadelphia where it spent its time gathering dust. The “Turk” never performed again. On July 5, 1854, a fire in the Chinese Museum destroyed the “Turk”. After the fire, Mitchell said, “Through the struggling flames… the last words of our departed friend, the sternly whispered, oft repeated syllables, ‘échec! échec!” The hoax took a while to be revealed. An 1857 article in The Chess Monthly magazine explained it, stating that “no secret was ever kept as the Turk’s has been”. Following the exposure of the hoax, there were imitators who created similar chess-playing automatons. Despite the “Turk’s” secret being known, spectators were still taken in, including President Grover Cleveland in 1885. In 1984, a replica of the “Turk” was constructed for a magicians’ exhibition in Los Angeles, but by then, the public was more skeptical. Few, if any, were taken in by the spectacle. In 1985, IBM announced it was developing a chess-playing program that could challenge the world’s great chess champions. In 1989, the program Deep Thought was pitted against chess master Garry Kasparov. Kasparov beat the program, so IBM went back to the drawing board and came up with a new one called Deep Blue. Deep Blue and Kasparov met up in 1996, and Kasparov won again. But in 1997, Deep Blue beat Kasparov in a six-game match, becoming the first computer ever to defeat a reigning world chess champion. How did the programmers at IBM do it? Did they finally work out an algorithm that could challenge the greatest minds of chess? Or did they give up and just pit an empty box with another chess master inside to pit against Kasparov? They’ll never tell.

Image result for deep blue
IBM’s Deep Blue: no room to hide anyone inside that thing.


Popular posts from this blog

Kick the Football, Charlie Brown

What's the lesson here? For nearly the entire run of Charles Schulz's Peanuts  comic strip, one running gag has been the football gag.  The gag is simple: Lucy Van Pelt kneels down on the grass, holding a football in place, and tells Charlie Brown to kick it.  Charlie Brown gets a good running start, ready to give it a good, solid kick, but at the last minute, Lucy pulls it away.  The final panel usually has a miserable Charlie Brown laying on the ground while Lucy looks over him, holding the football, telling him in one way or another that he obviously shouldn't have trusted her. The gag first appeared on November 14, 1951, when the strip was just over a year old.  In the first occurrence, the football was not held by Lucy but by Violet Gray, another little girl in Charlie Brown’s neighborhood.  (Violet would later become a minor character in the strip, and Lucy would become a major one.   Lucy wouldn’t appear in the strip until the following year.)  The f

43-Man Squamish: An Innovation in Athletics

For some people, one of the most tantalizing challenges is being told, explicitly or implicitly, that you can’t do something.  In 1965, MAD magazine writer Tom Koch laid down one such challenge.  He wrote an article laying out the rules of a sport he invented called 43-man squamish.  The article was illustrated by artist George Woodbridge, and judging by the mail MAD received from its readers, it was a huge hit.  Of course, Koch didn’t really intend the article to b e a challenge.  His idea was to invent a sport that was complex, convoluted, absurd, and ultimately unplayable.  It featured the kind of text readers of MAD, not athletes, would expect.  It’s an uncommon sport that has instructions like, “The offensive team, upon receiving the Pritz, receives five Snivels in which to advance to the enemy goal.  If they do it on the ground, it’s a Woomik and counts as 17 points.  If they hit it across with their Frullips it’s a Dermish which only counts points.  Only the offensive Nibling

Synanon: Self-Help Through Shame and Berating

In 1958, a recovering alcoholic named Chuck Dietrich discovered he had a talent for public speaking.  He was always a big hit at his Alcoholics Anonymous meetings, so he figured he’d take his talents and his $33 monthly unemployment check and try to give back to society.  Dietrich found he’d benefited greatly from A.A., but he was concerned about drug addicts, who weren’t admitted to the organization, because, as A.A. says, drug addiction is fundamentally different from alcohol addiction, and thus would require wholly different kinds of treatment.  Dietrich set out to help drug addicts and anyone else who needed support and organization in their lives.  That’s why he founded a two-year program called Synanon. The idea behind Synanon was to hold nothing back, because your chemical dependency was probably a symptom of your repressed emotions.  Synanon’s main activity was something Dietrich called The Game, which was designed to release these emotions.  To play The Game, all you did