Abstract

Jerry Fodor (1994) proposes a solution to Quine's inscrutability–of–reference problem for certain naturalized semantic theories, thereby defending such theories from charges that they cannot discriminate meanings finely enough. His proposal, combining elements of informational and inferential–role semantics, is to eliminate non–standard interpretations by testing predicate compatibility relations. I argue that Fodor's proposal, understood as primarily aimed at Mentalese, withstands Ray's (1997) and Gates's (1996) objections but nonetheless fails because of unwarranted assumptions about ontological homogeneity of target language predicates, and problems with Fodor's reliance on predicate conjunction to resolve ambiguity. Naturalized semantics thus remains without an answer to the inscrutability objection.