@article{oai:nagoya.repo.nii.ac.jp:00027878, author = {前田, 満 and Maeda, Mitsuru}, journal = {IVY}, month = {Oct}, note = {It has been claimed that the extractability of an element from wh-islands depends on its intrinsic properties. Wh-movement of an object out of wh-island is marginally permitted, while that of an adjunct is generally impossible. In this sense, wh-islands may be referred to as weak islands, as distinct from strong islands, such as subjects or adjunct clauses. It has been widely claimed that the weak island effect of embedded questions should be explained by Bounding Theory. Chomsky (1986) follows these lines, and proposes an account for the effect in terms of the notion of barrier. In particular, embedded questions constitute certain barriers. Recently, however, an alternative account is put forth by Cinque (1990), who argues that the weak island effect is entirely derivable from the notion of Relativized Minimality without recourse to Bounding Theory. The aim of this paper is to defend the latter position. To achieve this, we cast light on the behavior of simply preposed adjuncts with respect to the extraction from within wh-islands, where the term "simple preposing" is intended to mean Ā movement of non-operators: movement of such adjuncts is typically exempt from the weak island effect of embedded questions. This fact is unpredictable under Chomsky’s (1986) account, since it is predicted under the view that adjuncts cannot be extracted from wh-islands, whether the movement involved is wh- movement or simple preposing. Thus, it forms an argument against Chomsky’s account. However, although this fact motivates our claim in this paper, it also constitutes a potential problem at the same time: Relativized Minimality also excludes extraction of an adjunct from wh-islands, as it stands. We attempt to solve this problem by arguing that simply preposed adjuncts, but not those that are wh-moved, may be reconstructed (or undone) into its D-structure position at LF. This assumption is based on the observation that simple preposing of an adjunct, unlike wh-movement of an adjunct, does not necessarily establish a semantically significant operator-variable relation. This amounts to saying that simple preposing of an adjunct may be merely "stylistic" in the relevant sense. If simply preposed adjuncts are reconstructed into its base position at LF, they do not have any significant semantic import.}, pages = {127--144}, title = {付加詞の前置とwh島の効果}, volume = {25}, year = {1992} }