Alan Turing

1912 - 1954

One of the Twentieth Century's mathematical titans, Alan is largely credited as the founder of computer science. Although that may be unfair to other early mathematicians - after all, you can certainly go back to Charles Babbage in the early Nineteenth Century and even back to Blaise Pascal almost two hundred years earlier. But if you want to pick one person who was responsible for making hand mathematical calculations an increasingly lost art, Alan's as good a pick as any.

Although fairly well known today and achieving iconic status among the mathematically inclined, for a long time Alan remained little known to the great unwashed (such as people who put up cartoon websites). But in 1968 Arthur C. Clark made a reference to Alan in his novel 2001: A Space Odyssey. There we learned that Alan has "settled" the question of whether the spaceship's heuristically programmed algorithmic computer - HAL, of course - could "think". After all, we're told, HAL could "pass the Turing Test with ease".

Alan didn't call it the Turing Test. He called it the Imitation Game. Let someone sit down and ask a computer questions - whether verbally or by keyboard is immaterial. Any questions would be permitted. They could even be framed to deliberately discover if it was a computer or a person at the other end. If the computer could then carry on a prolonged conversation so the human could not tell if he or she were talking with a computer, then the computer must be thinking by any reasonable definition of the word.

Alan believed eventually computers would be constructed that could pass the Imitation Game and so, he thought, we would eventually have computers that could think. It wouldn't happen overnight, of course. But Alan thought by the end of the Twentieth Century, computers could fool people 70 % of the time after five minutes of conversation.

The computers we have today are much more computationally powerful, faster, and smaller than anything Alan imagined we'd have. It was said for decades computers would not be able to beat chess grandmasters. But that has happened. Shoot, they not only beat grandmasters at chess, but they even play Jeopardy better than humans (although CooperToons thinks there was a bit of an advantage given to the computer regarding response time). Unfortunately (or fortunately depending on your point of view), no computer has ever been built that can pass the Turing Test to Alan 's specification. Are they coming close? Well, recent tests have shown a group of people can be fooled about 60 % of the time that a computer is a human. That seems pretty good. But what throws a monkey wrench into the circuits is that in the same time about the same amount of people thought a person was a computer.

So this brings up a point that really isn't considered too much. The Turing Test may be more of a test on how gullible people are, rather than if the machine is actually thinking. Other people don't even think it's a fair question, and today, where it has been officially ruled in the United States that corporations are people with all the rights thereof, we certainly don't want to be unfair to computers.

A major problem here is that the word "thinking" has multiple meanings. If we mean that thinking is "doing calculations like computers", the computers are indeed thinking. But if we are asking whether computers have consciousness, are aware of their own existence, stuff like that, then the answer is, in an Official CooperToons Humble Opinion, no better known today than it was in the 1820's when Charles Babbage invented his "difference engine". In other words, does a computer any more conscious than - that is does it "think" any more than - a light circuit in your home?

Or another question. Do insects, like ants, think as humans do are they organic automatons responding reflexively to stimuli? The latter is certainly possible. Robotic "insects" have been created whose actions and behavior appears to all intents and purposes identical to live insects, yet they are controlled only by simple stimulus-response algorithms. In one sense these creations satisfy the Turing Test, not regarding thinking, but in appearing to be alive. But almost certainly these robots do not "think" any more than an automobile.

There are certainly various theories as to what causes and how to detect consciousness. One theory - and one that may very well be correct - is that consciousness arises when the number of the synaptical connections in the brain reaches a certain value and have certain arrangements. Or perhaps consciousness is a continuum which gradually forms as the connections increase. If these theories are correct, then it is feasible that computer can become conscious, much like the computer on the well-done but fairly spooky movie The Forbin Project. The computer is programmed to control nuclear weapons and stop war. It is so large it has a consciousness and decides the only way to stop war is to become a supreme dictator.

On the other hand, Alan's paper on the "Halting Problem" has given us a bit more definitive conclusion on what computers can and cannot do. Alan had posed the question on whether you could program a computer to automatically solve all problems, or if the problem was not solvable, to say so. For what it's worth, Alan demonstrated you could not (for a Most Merry and Illustrated, well, Illustration of this surprisingly easily understood proof just click here.)

Alan was far from a wimpy egghead. He was a world class distance runner, and in 1947, he ran a marathon in 2 hours and 46 minutes. Pretty impressive since the World Record at the time was 2 hours 26 minutes. But we have to remember this was the day where the Olympic Teams were filled from the ranks of the amateur athletic clubs (rather than picked from groups of professional multi-millionaires). After all, it was in Alan's time that a young Oxford medical student named Roger Bannister broke the Four Minute Mile.

The tragic outcome of Alan's short life is well known and we need not dwell on it at length. It's probably sufficient to point out that his personal preferences seem to be little different than those of a number of American politicians and religious leaders, who when not indulging in the lifestyle themselves, spend a lot of their time and legislating and preaching against it.

References

Alan Turing: The Enigma, Andrew Hodges, Simon and Schuster (1984). Alan's biography, volumous, well written and informative. But ... there is an awful lot of detail that might be a bit overwhelming for those who are looking for some light reading.

"Computing Machinery and Intelligence", Mind Volume 59, pp. 433-460. Alan's original paper. Easy to read and understand.

Official Alan Turing Website.

2001: A Space Oddessey" , Arthur C. Clarke, New American Library (1968). Arthur has the computer in the book speaking in an idiomatic conversational style with the astronauts. So yes, HAL does pass "the Turing Test with ease". But we know that HAL does think like a human being, is self-aware, and even has insticts for self-preservation. So strictly speaking, though, a computer like HAL would not prove the Turing Test is valid. All it means is if an electronic entity does have thought processees like a human, then it would pass the Turing Test. The reverse does not follow.

Oh, yes. MIT's real life computer and cognitive scientist Marvin Minsky also gets a plug.

A CooperToons Opinion: 2001 is one of the best science fiction novels written and holds up well even nearly half a century later. But although kudos are often thrown out for the pioneering special effects of the film, particularly of the astronaut speeding through the nebulae, it's very easy to see how it was done (optical printing of arial photography). And so it's kind of like watching a magician do a trick you can figure out.

But more to the point, unless one knows the story by reading the novel, the movie is disjointed, incomprehensible, and quite frankly, at its best the acting is merely adequate. Viewers walk out wondering what the hell was going on. In that sense, the film is a failure when taken independently from the book.

One last tidbit. Marvin Gardner, the writer who penned the Scientific American "Mathematical Games" columns (which were never the same after he left), hypothesized that the name "HAL" was formed simply by moving the letters IBM one letter back in the alphabet. It turns out, though, that was pure coincidence. Clarke later told Marvin that this derivation was a complete surprise to him.

"Alan Turing predicted that computers would be thinking by now. Was he correct?", Christian Science Monitor, June 23, 2012. An update on the Turing Test.