>To sum it up: Is there some way to combine the fuzzy quality >that intelligence >relies on with the rigid quality of not making a single mistake? >Is >generalisation a limit to intelligence? Hofstader mentions that in GEB. Here's the way I see it: When you generalize, you chunk information, jump out of the system and plays with the generalized information(but loose some control over the sub-details of it). Like in chess... You first learn the rules, then a few traps, then go to tactics. Afer tactics becomes obvious, you go to a higher level , you think strategically. At this point, you might filter the board SO MUCH to the point that you might miss the simple fact that a bishop is attacking a rook. (something amateurs often won't do). The way I play is something like: Check for some success in strategical level. Check for success in tactical level. Check for success in avoiding traps. Check for success at the rawest level(is there a threat going on now). Since the "brain only pays atention to one thing at the time" (sorry I don't know a neurological, low level way of putting this) not making a mistakes will occour if you operate at each level of abstraction at the time, or find ways to leave subneural nets operating at the lower levels by themselves. Rafael Anschau ______________________________________________ FREE Personalized Email at Mail.com Sign up at http://www.mail.com/?sr=signup