I wrote:
> The thing is, Watson would feel no contraction between this response and > the one he gave a minute before. He would not recognize that if the second > response is true, it invalidates the first one. That is too high a level of > abstraction for Watson, and for many people. > As far as I know, Watson wouldn't. That is my impression reading about Watson. I could be wrong. I have just been reading generalized descriptions of Watson, and I read about previous language parsing and internal representation. As I said, computers may eventually be able to do this. When I say "Watson understands" in this context I do not mean he understands the way humans do. It is a different thing. Far more limited. The mechanism is different. The mechanics are different, which is like saying that an airplane flies even though it does not flap its wings. In my opinion it is reasonable to say that a bee can think and communicate by dancing. Brain tissue is brain tissue, and the information definitely does pass from one bee to another. The thinking is rudimentary and bound entirely by instinct, and the communication is totally different from human language, yet they fall into the same general category as human cogitation and language. If you agree that a bee can think, it is reasonable to say that Watson can understand human language. Actually, it is more accurate to say the entire bee hive "thinks" as one entity, and plans its future. The notion of an entire group of animals "thinking" as if they were one organism is alien to us. If something that alien can be said to "think" it is not such a stretch to imagine a massively parallel processor computer like Watson "thinking." Watson's biggest weakness compared to people is its inability to grasp (parse; guess; model) context in any depth. I doubt that computers will think using the kinds of mechanisms human brain tissue does, although there used to be some interesting programs that simulated neurons and synapses. I do not think Watson is anything like that. - Jed
