ABSTRACT
This book is an exercise in what might be called ‘philosophical archaeology’: a logical excavation of the foundations of AI. The guiding principle here is the Wittgensteinian theme that, in order to understand the nature of AIwhat kind of theory it is and what kind of problems it raises-and in order to understand the reasons why the cognitive revolution was so quickly usurped by the post-computational mechanist revolution, we need to examine not just the mathematical and psychological origins of AI but, also, its philosophical roots. As we shall see in this chapter, this last point is especially relevant to the AI view of concepts, which has a distinctly Kantian ring to it. It must be noted from the start, however, that merely establishing the Kantian orientation of the AI view of concepts does not ipso facto constitute an indictment of AI. After all, Kant welcomed the ‘confirmation or refutation by experiment’ of his theory of cognition; and Piaget was clearly not disturbed by being labelled a Kantian (see Piaget 1965). So why
should AI not claim for itself the honour of fulfilling Kant’s vision of advancing the ‘science of metaphysics’?
