부산대학교 도서관

The purpose of this paper is to account for the asymmetry between a suffixed word and a prefixed word or a compound in Korean Coda Neutralization within the framework of Correspondence Theory. The asymmetry is shown to result from difference in a constraint ranking. Specially, I propose that the relative position of Base-Identity constraint, which requires an identity relation between the source word base and its output form, is responsible for the asymmetry between a suffixed word and a prefixed word or a compound. In a suffixed word, the markedness constraints and IDENT-IO constraint, which requires an identity relation between input and output, dominate Base-Identity constraint (IDENT-IO≫Base-Identity). On the contrary, the markedness constraints and Base-Identity constraint dominate IDENT-IO constraint (Base-Identity≫IDENT-IO) in a prefixed word or a compound. In addition, I consider n-Deletion in English between level 1 and level 2 suffixes, which shows the similar phonological asymmetry with coda neutralization in Korean, and show that the phonological difference resulted from English affixation can also be accounted for by the same line with coda neutralization in Korean.