Quite well-put, actually. I agree that it's a mistake to think of the brain as a "neural network in your head". Biological systems are extremely complicated, an

Quite well-put, actually. I agree that it's a mistake to think of the brain as a "neural network in your head". Biological systems are extremely complicated, and the brain is heavily influenced by the soup it swims in. Having said that, I think we might as well discuss it even if we don't understand it fully. If we're rejecting theories of cognition because they're probably not always right, then we're going to have to throw out the entire field of cognitive science. I do believe that a reductionist approach is critical for understanding cognition -- you could call that a religious belief if you want; I would prefer to call it an expectation based on the unprecedented progress that reductionistic approaches have made in understanding all kinds of things. We certainly have disciplines that examine emergent behavior of cognition -- psychology, I guess, is the prime example. But I think it's also important to try to understand the brain as a computational system, not just as a black box that produces behavior. I think an emergent, behavioral, holistic approach may be effective for characterizing cognition until we understand it better, but that a reductionist, symbolic approach is a superior method for knowing, manipulating or replicating cognition in all kinds of environments, not just a bunch of fatty cells inside a chemical soup. Also, after reading your response, I think that I misunderstood your original comment, which to me seemed vaguely derogatory. So, sorry for that.