Tuesday, March 14, 2006

Myths About Turing

David Leavitt, The Man Who Knew Too Much, Norton, 2006.

Alan Turing is an iconic figure. He was one of the main founders of the mathematical theory of computability. He proved that first-order logic is undecidable. He was one of the main brains behind the allies' successful effort to break Nazi secret codes during WWII. He designed one of the earliest general-purpose, program-controlled computers. He pioneered artificial intelligence. He was persecuted by British authorities for being homosexual. And much more.

But there are also a number of myths about Turing.

Myth #1: he invented the notion of stored-program computer. Truth: Turing's universal machines are not properly called stored-program; roughly speaking, this is because they have no memory component separate from input and output devices.

Myth #2: he created the computational theory of mind, according to which mental capacities are explained by neural computations. Truth: the computational theory of mind is due to Warren McCulloch, who published it in 1943 in collaboration with Walter Pitts.

Myth #3: Turing committed suicide because of being persecuted for his homosexuality. Truth: well, myth #3 might be true, for all we know. But there is very little evidence for it. Turing might have committed suicide for other reasons, or he might have died because of an accident. Turing's archives in Cambridge contain a lot of testimonials by people who knew and saw Turing during the last weeks of his life, many of whom express doubts that he committed suicide. To my knowledge, no historian has done justice to those documents in print. At a minimum, it's fair to say that no one knows why he died.

In his new book, author David Leavitt misses a perfectly good opportunity to correct some of these myths about Alan Turing. In addition, the book contains a lot of technical mistakes. For example, Leavitt mistakenly states that Kurt Gödel showed that Bertrand Russell and Alfred North Whitehead’s system in “Principia Mathematica” was inconsistent. (what Gödel actually showed is that if the system is consistent, then it is incomplete.)

If you wish to read more serious work on Turing, you should read articles by Wilfried Sieg and Jack Copeland. References may be found in my own article on Turing, or on their websites. (However, be aware that Copeland subscribes to myth #1, and as I argue in that same article, Sieg is not always right either.)

0 Comments:

Post a Comment

<< Home