음운현상의 수의성과 음장

초록

]Korean has been generally regarded as a typical language displaying compensatory lengthening(CL, hereafter), since the pioneering work of P-G Lee (1978) in Korean phonology. The analysis of general characteristics of Korean vowel lengthening through hiatus resolution in the previous research has been generally focused on the fundamental characteristics of the mechanism of vowel lengthening proposed for typical CL-languages. This study, however, approaches vowel lengthening in Korean from a totally different view. The basic claim of the present study is that Korean is not a genuine CL-language, even though it exhibits some similar patterns of vowel lengthening as in typical CL-languages. This somewhat contradictory assumption indicates that the analysis of vowel lengthening in Korean needs a more careful investigation. Previous rule- or representation-based accounts have not been successful, especially in the sense that they have not provided a phonologically-convincing motivation of vowel lengthening in Korean. In the present study, vowel lengthening in Korean is analyzed as a foot-based phonological process, mainly relying on the assumption that vowel lengthening can be metrical rather than moraic at least in certain languages. In this respect, the preservation of timing units or moraic structure, which has been generally proposed in most autosegmental studies on CL, cannot provide sufficient motivation for vowel lengthening at least in Korean. Specifically, it is proposed that Korean vowel lengthening results from an Output-Output Correspondence constraint on metrical (foot) identity between phonological variants in hiatus resolution. Such an analysis captures the intriguing fact that "compensatory" vowel lengthening only occurs in cases where hiatus resolution is optional, but never in cases of obligatory hiatus resolution.

키워드