t-구개음화와 신·구 유표성 제약

T-palatalization in Korean and New·Old Markedness Constraint

Previous studies have found t-palatalization to show neutralizing and phonemic change, applying only in derived environments, while s-, n-, l-palatalization shows non-neutralizing and allophonic change, applying in both derived and non-derived environments. And also it has been shown that Korean has two levels as far as palatalization is concerned. Two levels are phonemic/lexical level and allophonic/postlexical level which has different rankings of constraints. More than one level of constraint hierachy leads to revise the strict parallelistic view of OT in order to have any descriptive adequacy in Korean palatlization.. The ranking under question is that of Pal and Ident(ant), which is reversed depending on the different levels: Ident(ant)≫Pal at the lexical level and Pal≫Ident(ant) at the postlexical level. However if the t-palatalization in Korean was supposed to affricate /t/, level-ordered OT may not be argued. In this paper, it has been shown that the t-palatalization is occurred owing to the constraint *[tj(i)]σ . Under the comparative markedness theory, this constraint is divided into the old one and new one. The opacity of t-palatalization which was referred to the derived environment effects can now be explained by the ranking of N*[tj(i)]σ≫Ident[t]≫O*[tj(i)]σ.