Search

In this paper, I address a recent proposal by Yu-Cho and Sells (1995) to explicate the constraints on the morphosyntax of inflectional affixation in Korean based on the attribute TYPE, which in large measure coincides with the bound vs. free distinction but also diverges from it in crucial cases. The theory of TYPE offers an insightful account of a class of affixes-the cross-categorial affixes-that attach either to nominals or verbals. I suggest, however, that when a fuller range of relevant data is taken into account, constraints on inflectional affixation in Korean are better handled by the distinction between bound verbal roots vs. free nominal roots on the one hand, and that between affixes which attach to bound forms and those that combine with free forms. The alternative that is proposed overgenerates. To deal with this problem, non-morphological means of filtering overgeneration are suggested and argued to be necessary.