Yun Jae-Sung. 1996. On Bracketing Paradox. Studies in Modern Grammatical Theories 8: 47-70. This thesis deals with bracketing paradoxes, i.e. constructions in which it seems necessary to assign two distinct structures to a word. We traces the recent history of this set of problems and some of the proposal for tackling them. Bracketing paradoxes that arise in the current framework of Lexical phonology give rise to some major complications. A main point made by Kiparsky is that morphological rules can only have access to information found on the level where they operate. In particular, bracketing paradoxes arise when attachment of level 1 presupposes the presence of level 2 affixes, or when level 2 affixation needs to see the internal bracketing of a word which is only available at level 1. We review two different theories that involve `rebracketing,`(Kiparsky, 1983ㆍPesetsky, 1985) and one different theory which rely on analogical word-formation(Spencer 1988, 1991). We also look into the prosodically based account that simply denies the morphological relevance of bracketings. Their approaches, it is claimed, can at best be regared as only a partial solution to the problem. We deal with an analysis diametrically opposed to previous studies, namely, the position of Robert Beard who argues for a separation of lexicon and morphology, a claim known as `Lexeme/Morpheme-Base Morphology.` In this model, the grammatical process of derivation and inflection are affected by L-rules in the lexicon, whereas processes of affixation are M-rules in the morphology. We show this thesis with a case study of the paradigm cases of the paradoxes which suggest that some of the paradoxes at least provide evidence for a `Lexeme/Morpheme-Base Morphology` approach. In conclusion, the idea of separating lexicon from morphology has an explanatory adequacy in accounting for bracketing paradoxes. This thesis speaks in favor of an autonomous morphological component.

• 윤재성