屈折素 分解 假說과 障壁 槪念의 定義 = The Split INFL Hypothesis and the Definition of Barrier

다국어 초록 (Multilingual Abstract)

The split INFL hypothesis has been exercising a great influence on current studies of syntactic phenomena in various languages ever since it was first proposed by Pollock (1989). The curx of the hypothesis is that the functional head INFL must split up into two new functional heads, T(ense) and Agr(eement). It is assumed by Pollock(1989) that the two newly-created functional heads have their own projections, and in consequence, the maximal projection of the functional head I, i.e., IP, must now split up into two new maximal projections, i.e., TP and AgrP.
If we accept the split INFL hypothesis, therefore, it means that the number of logically possible barriers increases since maximal projections are potential barriers in the sense of Chomsky(1986). The purpose of this paper is to see if the standard definition of the notion of barrier suggested by Chomsky (1986) can successfully handle the phenomenon of head movement on the fine structure of canonical clause entailed from the split INFL hypothesis. It turns out not to be true that Chomskys (1986) original formulation of the notion of barrier can treat head movement on the structure assumed in the split INFL hypothesis. We suggest that various notions and assumptions used for the definition of barrier in Chomsky (1986) be modified in order for the definition to deal with the phenomenon of head movement on the new structure. Our discussion in this paper boils down to the following:
(ⅰ) The notion of lexical category used in the definition of barrier must be defined in such a way that lexical categories are those non-affixal elements which have the value of [+/-N, +/-V] features and phonetic matrices as well. As a consequence, such auxiliary verbs as aspectuals must be included in the set of lexical categories.
(ⅱ) The notion of L-marking must be defined without reference to the notion of -marking since it would be very difficult to maintain that functional categories like T, Neg, and Agr -mark complements if we accept the standard notion of -marking. If we were to assume -marking as a necessary condition for L-marking every functional head should -mark its complement since otherwise the maximal projection of the head of that complement category would become a barrier. In this respect, we claim that -marking should not be a necessary condition for L-marking.
(ⅲ) A lexical category can L-mark wherever it occurs. In other words, a moved lexical category α may maintain its L-marking properties even when it has already L-marked its rightful complement in its pre-movement position. In this case, the value of L-marking may not change even if the L-marking head has moved to a higher position. This phenomenon is analogous to that of γ-marking suggested by Lasnik and Saito (1984; 1992).
(ⅳ) The notion of minimality barrier as formulated by Chomsky (1986) may not apply to the fine structure assumed in the split INFL hypothesis. If it were possible for intermediate projections to function as minimality barriers as in Chomskys rigid minimality framework, it would be impossible to present any correct account of the phenomenon of head movement since there would appear various intermediate projections of functional categories. We therefore abandon Chomskys rigid minimality and adopt Rizzis (1990) relativized minimality instead. In accordance with the latter theory, we do not accept the barrierhood of intermediate projections but rather the obstaclehood of their head with respect to head movement. The problem with Neg can be solved if we adopt Pollocks (1989) assumption that Neg is intrinsically inert for government and that heads intrinsically inert for government do not count as potential intervening head governors for the minimality principle.
The main results of the present research above have been derived chiefly from the interaction between barriers theory and head movement. Some modifications to the results might be needed if such movement phenomena as NP-movement, wh-movement, and quantifier movement are taken into consideration. We reserve this for future research.

목차 (Table of Contents)

• Ⅰ. 序論
• Ⅱ. Chomsky(1986)의 障壁 槪念의 核心
• Ⅲ. 屈折素 分解 假說
• Ⅳ. 屈折素 分解 假說과 障壁 槪念
• Ⅴ. 結論
• 參考文獻
• 영문초록