*Well, that's quite interesting, although I've got a couple of problems with it. *First, why allege that "probabilistic and connectionist approaches" are any closer to human cognitive activity than ...