The Representation And Derivation Of English Schwa

This thesis presents a nonlinear analysis of the vowel schwa [a] in English with the application of Underspecification Theory, Feature Geometry Theory, Syllable Structure Theory and Lexical Phonology Theory, which provides insights into the underlying inventory of English vowels. The close examination of schwa phonetic properties, phonological manifestations and a review of previous research on schwa representation and derivation reveal that this vowel has more than one underlying correspondence, and undergoes different phonological surface processes. Therefore, the idea is adopted here that the surface representation of schwa vowels has three underlying sources. It is assumed here that there is an empty vowel, which is contrastive with other vowels in terms of its absence of place node. Then, three different underlying representations and derivation processes can be accounted for by virtue of the empty vowel. Although schwa manifests in three ways, it is given a uniform account as the default specification for an empty vowel, no matter it is underlyingly present, or derived, or inserted. The non-alternating schwa is the vowel specification for an underlyingly empty vowel; the alternating schwa is the vowel specification for a derived empty vowel which is derived in turn by delinking the place node from full vowels in unstressed syllables, due to the weak position assigned by the stress assignment rule; while the epenthetic schwa is the vowel specification of an epenthetic empty vowel which is required by the syllable structure. Therefore, schwa is excluded from the underlying inventory of English vowels, but instead, a placeless empty vowel is introduced. Furthermore, the underlying representation and derivation of [(?):] and other related phonological phenomena can also be accounted for by use of the empty vowel.