Is there something wrong in principle with the major direction of cognitive psychology these days? A blindness which comes from a too great confidence in its rationalist and mechanist assumptions?

Take our everyday performance, like catching a ball, or carrying on a conversation. The current mainstream in cognitive psychology sees as its task to explain these by some underlying process which resembles a computation. When we reflect, we are struck by the skill we exhibit in these performances: knowing just where to reach to intercept the ball, knowing just where and how to stand, what tone to adopt, what nuance of phrasing to use, to respond appropriately to what our interlocutor has said. To explain the performance would then be to give an account of how we compute these responses, how we take in the data, process them, and work out what moves to make, given our goals.

To reach an answer by computation is to work it out in a series of explicit steps. The problem is defined, if necessary broken up into subproblems, and then resolved by applying procedures which are justified by the definition. We resort to computation sometimes when we cannot get the answer we want any other way; and sometimes when we want to show that this is the right answer. Explicit procedures can be crucial to a justification of our result.