실재적 기저음운

Abstract

This study is concerned with the problem of psychological reality in setting up under lying segments for Korean stops, affricates and liquids. Those Korean stops are particularly interesting because their phonemic/allophonic behavìor is very different from that of the major Indo-European languages we are familiar with. For instance, voicing is not distinctive in the obstruents, and /l/ and /r/ are allophonic variants of the only Korean liquid. The problem is that people are still posìting (and usìng) voiced or voiceless obstruents as underlying segments for stops and affricates, and either /l/ or /r/ for the liquid. This is in part due to the lack of symbols to represent the obstruents which are neither voiced nor voiceless and the liquid which is neither /1/ nor /r/, but more important is Korean linguists' attitude toward underlying segments and the rules for deriving surface phonetic forms. One of the most important points of this paper is that the psychologically real, sub-stantive underlying segments should have no distinction or indication as to voicing (for obstruents) and no l/r distinction (for liquid). For this, it is necessary that the phonetic/phonemic transcription systems which are currently being used be modified and enlarged, because there are no adequate symbols for the Korean consonants mentioned above. Another point is that speakers do not really derive surface forms from unique underlying representations; rather, they know the distributional patterns of the surface forms and the relationship between those forms in ' the pattern in a sort of interpretive (as compared with generative) manner. This is a very important idea that needs to be further developed.