k-Ø And what phonology can do

Abstract

This article looks at the Turkish k-Ø alternation and questions whether it can be counted as a phonological process. After a discussion of what it means to be phonological, the empirical facts are weighed against the theoretical expectations. The alternation has several quirks, and any account treating it as phonological must allow for complex machinery to deal with those, thus weakening the predictive power of the theory employed. This is true of earlier rule-based accounts, but also of accounts that have been presented within Government Phonology, the framework that also this article is couched in. After careful revision of several problems that previous accounts have run into and that seem insurmountable, we must conclude that the k-Ø alternation does not qualify as phonological.