Well, if I knew that tying in a pop song to a post was the way to get lots of you guys annoyed I'd have done it more often. I defend my choice to do so because Betty Everett and Alan Turing were proposing ways of doing the same thing, trying to figure out what's going on where it can't be seen. And I don't say that for humorous effect, it's true in every sense, the fact that one was a famous scientist and one was a R&B-soul singer who had one big pop hit makes no difference in that.
It is quite obvious that Everett's thinking in the matter was, actually, more sophisticated. But, then, she was dealing with actual minds of which she and her audience had enough experience to know that just asking the boy, "Do you love me" carried the real possibility of getting a false answer that she would believe for any number of reasons. She knew just getting an answer to questions wouldn't tell her what was really going on in the boys mind, she was looking for other methods to verify that.
Unfortunately, there is no way to do that by testing, you have to choose to believe or not believe it but you're never going to have absolute proof of it. But, clearly the singer realized that just wanting it to be true or her erotic desires made her susceptible to fooling herself, she warned against trusting "his charms" and "his arms" in the song. It's not only in love that what we want to be true carries a powerful incentive for suspending critical judgement and coming to conclusions that should be suspended for lack of reliable evidence.
Turing's proposal assumed that the person getting responses to questions posed to a machine in his test would have a level of sophistication in judging what was going on inside it by just asking questions that is not warranted. For all he knew the machine could be made that, though not thinking would sucker people into believing it was. Which was exactly what happened with Joseph Weizenbaum's ELIZA bot program about fifteen years after he proposed his test. And it wasn't just unsophisticated, uneducated people who really thought that the machine understood the exchange the person had with it, it was college students, grad students, psychologists - who sold themselves as such experts on minds - and scientists such as Carl Sagan who would attain a cult like following on his pop books instructing people on how to spot "baloney" in claims. Sagan's "skeptical" acolytes are all over the web, I suspect a lot of them are among those who got into a lather at my dissing Turing's follies yesterday. What a bunch of suckers those guys are, they don't even understand what it is they're buying. Clearly, in terms of reality testing, Betty Everett was miles ahead of them.
As to the anger over my saying that Turing was an idiot when he picked the boy he didn't know up but an even bigger one for walking into to a police station and, in effect, confessing to the cops for what was then a sex crime, one which the cops could certainly be guessed would be only too eager to imprison him for, there are no apologies required on my part. It was as stupid a thing as any idiot of any identity has ever done and Turing was no young fool when he did it. Gay bashing by cops was and is common enough for a gay man in early middle age to have known about it. Turing was old enough, smart enough and aware of public events in Britain where men being arrested for consensual gay sex was covered in the trashy British press so that he should have been able to guess how the police were going to think about the situation he not only volunteered but called their attention to. Maybe you have to be a gay man who lived when gay sex was illegal to understand how incredibly clueless Turing was being. Perhaps a gay man who reached that age who had been in an hermetically sealed box with no input from the real world his entire life would have been unaware of what would happen, that wasn't Turing. I'm sure Betty Everett wasn't that clueless about what was likely going on in other peoples' minds.
If Turing didn't know that the results would be he'd be arrested for sex crimes and that he'd provided them with a virtual confession of that, it only proves my case that his method of discerning if a computer is thinking by getting output to questions was totally and obviously inadequate.
Turing, a genius able to guess at the thinking of the Nazis in their encryption technology on the basis of logical analysis and mathematical reasoning, among the most impressive acts of logical mind reading for all times, may have been misled by that into thinking that kind of mental activity was a bigger part of intelligence than it is. I suspect, especially after reading Weizenbaum's reaction to the reaction of psychologists WHO ARE SUPPOSED TO BE SCIENTIFIC EXPERTS IN MINDS got suckered by his ELIZA program into mistaking what it was doing as thinking. And Weizenbaum was not intending to deceive people with it, he was merely demonstrating the possibility of using natural language with computers, he was entirely upfront about it, you didn't need to decrypt his intentions.
My point yesterday is that people are always taking for granted the human abilities that the Turing test depends on when everything about people makes it known that we don't possess those abilities to that extent. His test can't do what it's supposed to do for that reason.
I haven't seen the movie they made about him, though I've heard the biographer whose account of his life the movie was supposed to be based on was pissed off that they made stuff up about Turing, especially around the role of a woman Turing proposed marriage to in the early 1940s. Well, you know what corners they have to cut and what liberties with reality they make when they're trying to reduce a person into a two-hour movie. I'm not entirely certain Turing knew his own thinking all that reliably that he could have admitted it even to himself, which might also be a hint as to how incredibly badly he read the cops ten years later. And he was in his own mind, it wasn't in some other black box. I don't think he understood how big the thing he was trying to reproduce was and how much of what our minds do can't possibly be simulated by calculations. The habit of practicing the reductionism science requires seems to take over quite a lot, just like the habits of people in other areas of life. Like love.
Yet people believe computers will be able to contain the totality of a human mind and that "we" will attain immortality in that way. It was a lesser but still quite impressive bunch of computer guys who came up with that one. My guess is that their hair-brained idea is not based in different assumptions than the ones Turing made. I wish I could ask Betty Everett what she thought of that idea.
No comments:
Post a Comment