| Phi 109-02 Fall 2013 |
|
(Site navigation is not working.) |
Alan Turing (1912-1954) was a mathematician who had done theoretical work on computation in the 1930s that was an influence on the idea of a “stored-program computer” as that developed in the 1940s. More practically, he was involved in project during WWII that used electronic devices for code-breaking and, after the war, with the operation of one of the first real computers. That was the Manchester Mark 1, which he refers to here as the “Manchester machine.” It (and its competitors for the title of being the first) began operation in the two years before this article was written, so Turing is speaking of something that was then very new. These first sections of Turing’s paper are introductory. He describes and begins discussion of the “imitation game” that is now called the “Turing test” and then introduces the idea of a digital computer.
• Think about the details of the game and its rationale described in §§1-2, and also think about other set-ups that might test conversational ability. Why do you think Turing has chosen the set-up he has?
• The restriction to digital computers in §3 defines the concept of artificial intelligence as it is usually understood, but it is worth asking yourself whether there might be other directions in which one would look for intelligent machines or intelligent artifacts.
• §§4-5 are Turing’s account of what a digital computer is. When he says “infinite capacity computers” have “special theoretical interest” (p. 439), I think he has in mind the general description of systems of computation in his theoretical work of the 1930s, something people now refer to as “Turing machines.” The “discrete state machines” he speaks of in §5 are a more general idea. (It’s worth paying some attention to the table displayed on p. 440: although it won’t play much role in our discussion of Turing, an analogous table will be discussed in the article we will discuss on Thursday.)
• Turing’s prediction of storage capacity in the introduction paragraphs of §6 were substantial underestimates of the state of things a dozen years ago. Although there has been substantial interest in various forms of a “Turing test,” I don’t know that anyone attempted to compile statistics for the “imitation game” exactly as he describes it.
I have assigned only two of the objections that Turing goes on to consider but, of course, don’t hesitate to look at others if you have time.
• Some take Turing’s article to be a response to the lecture he quotes at the beginning of objection (4). (Jefferson was based in Manchester, and his lecture was prompted in part by discussion of the new computer there; if you are curious about what he said, you can find the full lecture on line.) I’ve assigned Turing’s response as much as anything for his example (the term “viva voce” is used for oral exams and also for oral testimony, and Turing seems to have both in mind).
• Ada Lovelace (1815-1852), who Turing quotes for objection (6), was the poet Byron’s daughter; she had broach mathematical interests and was a friend and supporter of Charles Babbage (1791-1871), who developed a number of mechanical computing devices and planned what was effectively a programmable computer (for which Lady Lovelace is believed to have contributed a program). Note that Turing’s response to this objection is effectively continued in his final section.
• The discussion of “learning machines” presented in §7 is heart of his reply to objection (6), but it is only one of a number of interesting ideas in this section. Notice, in addition to his main discussion on pp. 455-459, the arguments he offers before turning to learning machines as well as the issue about how to employ them that he points to in his next-to-last paragraph.