and more to the point, the computer doesn't even know what chess is. It's just adding, subtracting, fetching instructions from memory, etc. It's kind of like how a guy in a box doesn't really understand chinese, or how none of your brain cells actually know what slashdot is.
Careful. You are (as you probably know) repeating Searle's argument. This sounds like an obvious truth, but it is not. Does a computer know what chess is? You would say no, because you look at the program and observe it juggling bits and bytes. And since the juggling of bits and bytes is not understanding, you conclude that a computer does not understand chess. But can we conclude that YOU do not understand chess because to play chess your neurons are firing, and obviously neurons do not understand chess? No, we cannot. "Understanding" happens at another level of consciousness.
Now, the problem with the level of consciousness of a computer such as Deep Blue, is that the ONLY thing that it has any knowledge about, is chess. Thus it cannot talk about chess, because it has no knowledge of language. It cannot assess the cultural value of chess, because it knows nothing about culture. The only thing it knows anything about, is a field of 64 squares, on which 32 pieces move according to predetermined patterns. However, it knows that particular field VERY well, much better than most humans.
So, the question is whether Deep Blue has a kind of "chess consciousness", in which it really "understands" chess. (Trouble here is, of course, that the terms consciousness and understanding are not well-defined, but let's assume that they mean what is commonly taken as their meaning). And it is very hard to argue that it does not.
Arguments for why Deep Blue has no chess consciousness are usually along the lines of "it does not understand chess because it always makes the same move in the same situation." Not true, as the match of Deep Blue and Kasparov showed. Or, "it does not understand chess because it cannot learn new chess behaviour." Not true, learning algorithms are pretty common nowadays. Or, "it does not understand chess because it cannot explain its moves." Not true, usually a chess computer is perfectly capable of explaining its moves, albeit in a special-purpose language. Or, "it does not understand chess because it uses brute force calculations exclusively." Not true, if that would be the case Deep Blue would need about 10,000 years to make one move. Or, "it does not understand chess because it does not care about the game." Well, that is probably true, but we are now talking of assigning Deep Blue a consciousness of a higher level than just "chess," and I would never argue that it possesses that.
Personally, I believe that Deep Blue has a chess consciousness. True, that consciousness finds its basis in programming (and probably has been automatically configured by Deep Blue itself), but that does not invalidate its quality.
