Beliefs vary in degree of accessibility to consciousness. Current theory in cognitive science maps out a spectrum of accessibility in which beliefs accessible to consciousness (such as conscious ideas about occurrences) could be located at one extreme, with beliefs that are inaccessible to consciousness (such as tacit knowledge of complex syntactic rules governing language use) at the other.

Gross proposes a study of the nature of beliefs that are in principle inaccessible to consciousness and their relation to beliefs that are accessible to consciousness, particularly in light of the puzzlingly contradictory conclusions that can be drawn on both counts. Offering the case of linguistic meaning to illustrate these difficulties, Gross demonstrates how competing considerations can be adduced that would require beliefs about linguistic meaning to be both accessible and inaccessible to consciousness. On the one hand, the simple fact that language use is typically an instance of intentional action (e.g., "I'll pick up the tomatoes") requires that beliefs about linguistic meaning be accessible to consciousness, the propositions of current linguistic theory. On the other hand, it requires that beliefs about linguistic meaning be viewed as inaccessible to consciousness (e.g., the proposition that speakers have a tacit-but unconscious-knowledge of set theory in producing a sentence such as, "Some cats have no tails"). Empirical theory about language thus leads in two contradictory directions. Gross will explore parallel questions of the nature of unconscious belief and its problematic relation to conscious belief, such as the relation of religious belief to unconscious mental mechanisms posited by contemporary evolutionary psychologists.