Notes on Inferential Role Semantics

If semantics is the study of meanings, and the meaning of an expression e is understood as what speakers of the language must know use e competently, then a natural first guess is that the correct semantics of names will involve their referents. One might, initially, make the further supposition that the meaning of a name is just its referent. However, this sort of ‘Fido’-Fido view immediately runs into the difficulties posed by Frege: although the referent(s) of the names in ‘Hesperus is Hesperus’ is (are) identical to the referent(s) of the names in ‘Hesperus is Phosphorus’ (assuming the latter is true), it is easy to imagine that the latter is informative, while it is difficult to imagine the same about the former. If we suppose, plausibly, that the difference in informativeness between the two comes from some aspect of the meaning of the names in the two sentences, then it is clear that a semantic theory that appeals only to reference lacks the resources to accommodate the difference Frege found. Frege’s wellknown solution to this puzzle was to introduce the notion of sense into his semantics: even though ‘Hesperus’ and ‘Phosphorus’ share a reference, their failure to share a sense will explain how the true identity statement ‘Hesperus is Phosphorus’ can be informative. For present purposes, it will be sufficient to note the following moral of Frege’s proposal: to cope with Frege’s problem, semanticists could no longer understand meaning solely in terms of reference, but now had to invoke a psychological component of meaning to explain how expressions figure in our cognitive economies. IRS theories are proposals for how this cognitive component of meaning should be understood. Roughly, according to such theories, the cognitive meaning of an expression is identified as the role that that expression plays in a lan-