Abstract

This dissertation work is premised upon the observation that semantic information is required in order to group phonetically distinct word-tokens into phonemically equivalent word-types. For philosophers, like W. V. Quine, who have a dim view of meaning, this claim regarding the semantic basis of natural language phonology, if true, is problematic. This is why in a series of publications, Quine has attempted to avoid any appeal to semantics in his efforts to reconstruct phonemic word-type equivalence. Consistent with his rejection of meaning-based analytic equivalence, Quine rejects a meaning-based account of basic phonological equivalence. However, given the empirical failures to "reduce the phoneme" to acoustic, phonetic, and/or distributional data within phonology---combined with the philosophical short-comings of a behavioral reduction---a consistent disavowal of meaning prohibits Quine from acknowledging the legitimacy of basic phonological equivalence anymore than he wishes to acknowledge the legitimacy of analytic equivalence. In this way, a consistent Quinean rejection of meaning entails not only a skeptical dismissal of analytic equivalence, but also a dismissal of basic phonological equivalence. ;Yet the observation that semantic information is required in order to group word-tokens into phonemic word-types also has a Fregean application. This is because the semantic information required to reconstruct basic phonology is not entirely intensional in character. In addition to typing word-tokens of various predicates in terms of properties, word-tokens of names must also be phonologically typed. So whereas Quineans reject synonymous substitutions because they are based upon meaning, Fregeans claim reference is insufficient to guarantee the substitution of co-referring names salva veritate. But if basic phonological substitutions among word-tokens are premised upon co-reference , a consistent rejection of co-naming substitutions within belief contexts also threatens the substitution of allophonically related word-tokens in belief contexts. The semantic basis of natural language phonemic categorization has a number of implications vis-a-vis contemporary philosophy of language