초록 열기/닫기 버튼

The purpose of this paper is to show that opaque outputs cannot be completely explained within the framework of Generalized Sympathy(GS) proposed by Jun (1999). Although GS is successful in explaining some of opaque outputs, it fails to account for opaque outputs of affixed words like criticize in which Velar Softening (VS) gratuitously applies. Furthermore, a constraint ranking required for analyzing medicate in which VS underapplies conflicts with one for afiixed words like electricity. In order to account for underapplication of VS in morphologically independent words like kick as well as normal application of VS in electricity, a little revision of GS is necessary in an analysis of affixed words. However, opacity occurring in medicate is not explained by the constraint ranking required for affixed words within the refined version of GS. The bottom line is that opacity, expecially occurring in affixed words, is a challenge to GS.