Prosodic Compounding in Japanese and Korean

Abstract

Prosodic compounding in Japanese and Korean provides an argument for the theory that lexical prosodic constituents are formed on the basis of morphological structures (Cohn 1989, Inkelas 1989, Zsiga 1992). In this paper, it is proposed that two lexical prosodic categories-the prosodic root and the prosodic word-must be recognized to account for two distinct types of compounds in both languages. It is argued that such a distinction makes it possible to give a proper characterization for various phonological phenomena— the p/h alternation and the V/ø alternation in Japanese, and n-Insertion in Korean. This distinction also provides a straightforward analysis on the difference between the Seoul dialect and the Kyungsang dialect of Korean with regard to n-Insertion; the dialectal difference is derivable from the choice of prosodic domains. Apparent counterexamples—compounds that appear to consist of prosodic constituents of different types—turn out to belong to either of the two types. In Korean, these compounds are split into two classes depending on the prosodic type of the left constituent while in Japanese, all these compounds pattern with the prosodic-word compounds. This divergence between Japanese and Korean is attributed to the parametric difference in the recursiveness of the prosodic root; the prosodic root is recursive in Korean but not in Japanese.