On Quantifying Into Epistemic Contexts

According to Quine and Hintikka (1) is sometimes used in such a way that in conjunction with (2) it implies (8). (1) is used on other occasions in such a way that this implication fails.' Let us say, following Quine, that 'knows' occurs transparently in (1) when such implication holds and opaquely when it fails.2 One thesis central to Quine's discussion of verbs of propositional attitude is that only when (1) is taken transparently may 'Cicero' be replaced by a variable which, in turn, is bound to a quantifier preceding the verb of propositional attitudes With respect to ordinary discourse Quine puts what appears to be an equivalent thesis thusly: "Where transparency matters in relation to indefinite singular terms is that there must not be a pronominal cross-reference from inside an opaque construction to an indefinite singular term outside the construction."4