Currently, there are two major ``paradigms'' of computational cognitive science. To lead up to these, several dichotomies (albeit overly simplified ones) can be made:
(1) Researchers who study (human) cognitive behavior either (a) believe that there are mental states and processes that mediate input stimuli and output responses (this position may be called ``cognitivism'') or else (b) believe that there are no such mediating states or processes (or that it is unscientific to talk about any unobservable such states or processes--the position of behaviorism).
(2) Cognitivists believe either (a) that all mental states and processes are computational in nature (and here there is a further dichotomy between (i) the weak and (ii) the strong computational views) or else (b) that at least some (and perhaps all) such processes are not computational. Position (2b) is held by a number of scholars who believe that there are inherent limitations on the ability of computers to simulate or produce mental phenomena (e.g., Dreyfus 1979, Searle 1980; cf. Johnson-Laird 1988: 26). It is certainly a position that provides many of the most interesting and hardest challenges to the computational cognitivists. One such challenge is the problem of the nature of consciousness. Another is the problem of subjective qualitative experiences--e.g., what kind of computational theory can account for our experience of pain or of the color green? But (2b) is also a position that is often ridiculed as ``mysticism'' or as a contemporary version of vitalism.
(3) The dichotomy between the two major paradigms is between (a) those computational cognitivists who believe that cognitive computations are ``symbolic'' and (b) those who believe that they are, rather, ``connectionist''.