Tom Standage is science correspondent of the Economist. He is the author of The Victorian Internet and has written for Wired, the Guardian, the Independent and the Daily Telegraph. He is married and lives in Greenwich.
In The Mechanical Turk, Tom Standage, Science Correspondent for The Economist, has brought the full, incredible story of the famous eighteenth-century chess-playing automaton, the Turk, to new light. Raising increasingly relevant questions about artificial intelligence and consciousness, the story of the Turk is now firmly established as a landmark in the development of ‘thinking machine’ technology. Here, he brings this enigmatic character to life.
When did you first hear about the Turk?
I can't remember exactly, but it was probably when I was a teenager, when I first got interested in artificial intelligence. I had a book called The Policeman's Beard is Half-Constructed, which contained poems and prose written by a computer program called Racter. I spent several months working on my own software to do the same sorts of things: hold conversations, construct sentences, write poems, and so on. I also wrote programs that could make logical deductions based on simple statements and generate horoscopes. I read around the subject of machine intelligence generally, and it was probably then that I first came across the Turk.
What was it about the Turk's story that particularly interested you?
What appealed to me was the way in which this automaton prompted a debate, in the late eighteenth century, about whether a machine could think or not. I loved the idea that people were debating the possibility of thinking machines over one hundred and fifty years before the first digital computers were built. We like to think that the ‘artificial intelligence’ debate is a modern phenomenon, but it's not. I'm rather fond of collecting examples of this kind of thing - historical precursors of modern scientific and technological breakthroughs. My first book, The Victorian Internet, looked at the parallels between the telegraph networks of the nineteenth century and the modern Internet. My second book, The Neptune File, was about how the planet Neptune was detected in 1845 by mathematical analysis of its effects on other bodies - which is how astronomers are now detecting planets around other stars.
The Turk is a detective story as well as a book about the history of
technology. How did you piece together the Turk's somewhat mysterious
The biggest problem was distinguishing fact from fiction. There are lots of myths and legends surrounding the Turk, many of which are routinely reported as fact. Even the Encyclopaedia Britannica from 1911, my favourite snapshot of the Victorian world-view, has a completely erroneous explanation of its secret based on a story put about by the magician Robert-Houdin in the nineteenth century. Wolfgang von Kempelen, the Turk's creator, is almost universally referred to as a baron, but he wasn't. This kind of thing happens repeatedly, so I had to go back to the original reports and work through them in chronological order to see what could be trusted. It was then possible to see mistakes and fabrications propagate from one account to another.
I went back to old journals and sources in English, French, and German, and communicated with a researcher in Budapest, who translated excerpts from Hungarian sources into German for me. I also visited the Library Company of Philadelphia, which has a huge archive of Turk-related material. And I talked to the members of what I call the ‘Turk Mafia’ - a group of magicians, chess experts and academics, most of whom communicate with each other, and all of whom are passionately interested in the Turk and its story. I've even ended up bidding against members of the Turk Mafia when Turk-related items come up for sale on eBay.
Why were so many people prepared to believe that the Turk was genuine?
There seem to have been a number of reasons. The Turk's first visit to Paris, for example, coincided with the first public demonstration of a hot-air balloon by the Montgolfier brothers. If flying machines, which were supposedly impossible, could in fact be built, then why not a thinking machine? Mechanical technology was advancing quickly, the industrial revolution was getting started, and displays of mechanical toys of amazing complexity were very popular. The way in which the Turk was presented made a big difference too. John Gaughan, a Los Angeles magician, has reconstructed the Turk. And when you see it playing, even if you know the secret, it's really convincing. It seems to tap into a really fundamental human compulsion to believe that it's real.
Do you believe a ‘thinking machine’ will be possible any time soon?
It all depends how you define ‘thinking’. I go along with Alan Turing, the great British mathematician, who sidestepped this philosophical question with his ‘Turing test’: he defined a thinking machine as one that can convince someone that it is a human in a written question-and-answer session. In other words, for practical purposes, a machine that appears to be intelligent-that can answer questions, or drive a car in response to spoken directions, or whatever-is as good as a machine that is ‘really’ intelligent. The philosophers can go off and argue about whether or not it's really thinking, or has a mind, or whatever, but from an engineering point of view it's what the machine can do that counts. I expect we'll see machines like this in the next few years: speech-driven PCs, personal assistants, that sort of thing. They will appear to be thinking. Will they be HAL-like artificial minds? No. But they'll still be useful.