More precisely: The Descartes (-Turing) Test Is Stupid
The “Turing Test” is a big deal in Artificial Intelligence and logic, for reasons that are assuredly not flattering. As the “Test” is obviously flawed. The Test confuses conversation and imagination, while identifying both to intelligence.
The fundamental error of this Descartes Test is mathematical (ironical, as Descartes was one of the greatest mathematicians, ever: he invented algebraic geometry, the foundation of all modern science and technology).
The Descartes Test overlooks the fact that the set of all possible conversations is not just countable, but even, certainly, in practice, finite. Thus a machine could plausibly, encompass all possible conversations, as it means interlinking a finite set with logical chains.
(To excuse Descartes, the notion of countability had not yet been clearly defined in his time; it leads, in turn, to the finiteness of speech, modulo my finite mood.)
Other objections to the Descartes Test show up in the essay reblogged below (which I wanted to write long ago, and have alluded to, here and there).
The “Turing” Test certainly ought to be called the Descartes Test, in light of the quote given in the attached essay. To know that the Turing test was actually invented by Descartes is of no small consequence.
Who “invented” what is not just a question of justice. And no just a question of the history of the systems of thought. It’s also a question of logic: knowing an idea appeared early on is a hint that it ought to be obvious, for example.
This process of associating the correct labels ought to be extended to all fields on inquiry. For example, Johanus Buridanus formulated clearly the law of inertia, circa 1320. That’s more than three centuries before the Anglo-Saxon gentleman generally celebrated as its author, was born.
This is a testimony that the Church was incredibly efficient, in the late Fifteenth Century, in its repression of advanced thinking. Buridan was put to the “Index”… Except in Cracow, where Copernicus studied, and, when he was dying, the latter re-published Buridan heliocentric proposition.
This ought to be a warning to the pseudo-scientist attitude about the Multiverse and Strings: too much craziness could lead to an anti-science backlash, on the ground of common sense.
(Fortunately, there is biology, which is much more scientific than physics, these days!)
The Turing Test pretends that intelligence is all about conversation, a finite process. It’s not. It’s about imagination (a much larger process).
Patrice Ayme’
by Massimo Pigliucci
You probably heard the news: a supercomputer has become sentient and has passed the Turing test (i.e., has managed to fool a human being into thinking he was talking to another human being [1,2])! Surely the Singularity is around the corner and humanity is either doomed or will soon become god-like.
Except, of course, that little of the above is true, and it matters even less. First, let’s get the facts straight: what actually happened [3] was that a chatterbot (i.e., a computer script), not a computer, has passed the Turing test at a competition organized at the Royal Society in London. Second, there is no reason whatsoever to think that the chatterbot in question, named “Eugene Goostman” and designed by Vladimir Veselov, is sentient, or even particularly intelligent. It’s little more than a (clever) parlor trick. Third, this was actually the second time that a chatterbot passed…
View original post 2,075 more words
Tags: conversation, Descartes, Imagination, Turing Test
Patrice Ayme's Thoughts 
June 19, 2014 at 3:44 am |
I never thought about it at the time, but in one of my scifi ebooks, I showed how a schoolboy with computer skills could defeat the Turing test, and what really amused me is that on reading the example where the computer won, it did much the same as what I wrote, except it was somewhat more sophisticated. The important point is that the computer takes over the conversation, once it gets past the initial stages, and it knows what the initial stages are going to be so it can be prepared.
In my novel, I had the machine entered into a science competition, and the presentation was pretty bad, so the first comment from the judges was obvious – “What a poor display!” That gave the machine a clear run to take over, and since it had a function (dispensing fruit juice) it could take command of the conversation by asking whether anyone wanted a drink. The more sophisticated responses actually did something very similar, and steered the questioner along the expected path. It proved nothing, other than the gullibility of those who wish to believe.
LikeLike
June 19, 2014 at 5:55 am |
Wow Ian, you are pretty busy! A novel too? It’s true that this (meta)trick you expose, to steer the conversation into a particular universe would totally work. It reduces the set of intelligent answers enormously.
In any case, even without crafty reduction such as the one you propose, the problem is that the set of all possible intelligent conversations, although large, is finite.
Even without using my theory of the Largest Number! Indeed there are only so many words a human being living 1,000 years can form…
PA
LikeLike
June 19, 2014 at 6:40 am |
Turing Test Is Stupid, the original title you used, is funnier, hence better.
LikeLike
June 19, 2014 at 1:48 pm |
Well, agreed. Yet I had a reason, involving the observation of how search engines search…
LikeLike
June 21, 2014 at 10:20 am |
It amazes me that so many skilled people can take seriously Turing test, which its author meant to be a joke. To my opinion, John Searle’s Chinese Room should refute to everyone the validity of this test. Yet many people still insist to claim, that the human mental phenomena is nothing more than input, some kind of compilation process and output. To me it seems all these people with such a claim can’t mean it seriously that they are nothing else but some kind of supercomputer, unless they are a supercomputer 🙂
LikeLike
June 21, 2014 at 4:17 pm |
Agreed. I don’t like Searle: he forces his students to buy his hyper expensive small books that say nothing. Anyway, it’s the Descartes Test. What I read from Turing about the Test he stole from Descartes was not amused. Nor I am.
This said, I agree with what you said. My argument that you can make a perfect Descartes Test and still have no intelligence whatsoever. Or, rather, mechanical intelligence (in the sense of classical mechanics) is what the Descartes Test defines. Mechanical intelligence is extremely useful, say, to drive a car. But ain’t human intelligence.
PA
LikeLike