"Beliefs and Subdoxactic States" Stich (1978)

DRAFT (to be proofread)

So far I thought that, according to functionalism, the key property of intentional mental states was, apart from intentionality itself, their causal role. However, in the Stich paper causality appears connected to subdoxactic states, while the relation between intentional states is mainly conceived in terms of inference[A1] . Moreover, Stich suggests certain link between inference and consciousness (what implies a link between intentional states, esenctialy beliefs, and consciousness), whereas in the standard cognitive-scientific view the link traditionally underlined has been between intentional states and unconscious processes (For instance, Fodor 1987).

Although the link between inference and beliefs is in general clear, sometimes the position of the author is rather confusing, for instance, he admits that “inference relation includes other states in addition to beliefs” (514). It is true that he assume the existence of “specialized and limited” (507) inferential contact between beliefs and subdoxactic states, but in other occasions he emphasizes the opposite view: the two features which distinguishes beliefs from subdoxactic states are the acces to consciousness and inferential integration.

Why is important the distinction between beliefs and sybdoxactic states? Apparently the reason, according to Stich, would be the acceptance of intuitive psychology theory (which is coherent with this distinction, unlike the theory posited by Harman). But it is a sufficient reason (or ever the only one)? (Stich also suggests the importance to identify a particular psychological state -500, also 514-515-).

In general, the article shows that the concept of inference is problematic and possibly ambiguous (see 512 and 513 –perceptual inference, computational inference, etc.). Stich claims that the use of the concept in this context is an empirical question (513-514), but a matter of definition, I think, rarely is just an empirical question.   

Another interesting point is whether beliefs have to have the form of linguistic representation. In this were the case, then linguistic representation would be a criterion for beliefs (516-517).

Additional question: It is not very clear the relation between simulation and inference: the reason for neglecting the distinction, the author suggests, could be that “like Harman, many of those concerned with cognitive simulation have been so captivated with the promise of inferential accounts of the mechanisms underlying perception and thought that they have failed to note the rather special and largely isolated nature of the inferential processes between beliefs and subdoxactic states” (517). 

ANOTHER QUESTION: THE NON-INFERENTIAL RELATIONS BETWEENS BELIEFS: How are interpreted in this context contradictory beliefs, inconsistency, etc.?

---

 [A1]Could be worhty to note that Fodor posited a kind of nondemostrative inference: "the inference from like effects to like causes" (Fodor 1968, p. 627).