日時：9月30日（土）　13:00

会場：名古屋大学人間情報学研究科棟講義室1

発表者、発表題目及び要旨

1.　13:00-13:50
　　西田瑞生・出嶋真由美・藤森千博（名古屋大学人間情報学研究科）
　　"True A'-Movement in Japanese"

　これまで、日本語は英語などに比べて「島」の制約が非常に弱い言語であるという証拠がいくつか提出されており、それを基にした日英語の構造の違いを論じた研究が数多くなされてきた。
　本発表では、しかしながら、日本語の「NP + の」の移動が、従来日本語において観察されてきた移動現象（例；かき混ぜ）とは異なり、主語条件や wh-島の制約に従ったり、付加詞条件に関しても強い非文法性を示したりすることを指摘する。つまり、日本語の 「NP + の」の移動は英語等に見られる A'-移動とほとんど同じ性質を示すことになる。この経験的事実を基に、日本語の 「NP + の」の移動は主題句（Topic Phrase）の指定部へ移動すると分析される英語の主題化と同等の移動であると仮定する。
　もし本発表が主張する仮定及び分析が正しいものであるならば、日本語の主語や付加詞やwh-句などが英語等のそれらと同等の地位を持っていると結論付けられ、また、日本語にも英語等と同様の A'- 移動が存在することが示される。

2.　13:50-14:40
　　藤井友比呂（大阪大学大学院）
　　"Two Types of Multiple Nominative Constructions in Japanese"

  The following deals with multiple nominative constructions in Japanese. I will argue that constructions with more than one nominative NP are classified into at least two groups: the major subject construction (MSC), on one hand, and the external possession (or possessor raising) construction(EPC) (and the nominative object construction (NOC)), on the other hand.Relativization serves as a diagnostic, which shows that major subjects, but not other nominative NPs, block realization of thematic arguments (the major subject island effect). Based on the observation, I hypothesize that the nominative NPs in the EPC/NOC are, like regular VP-internal subjects,
assigned Case by T/Infl, whereas only major subjects, generated higher than T, are licensed by C. If the analysis is empirically correct, it implies that the cross-linguistic generalization that external possession is associated with theme but not with agent (e.g., Baker, 1988) should be maintained in Japanese as well (cf. Ura, 1996). Finally, I will suggest an explanation for the major subject island effect by appealing to the notion of phase, proposed by Chomsky (1998).
 

休憩（14:40-15:00）
 

3.　15:00-17:00
　　浦啓之（関西学院大学）
　　"Towards Eliminating Constituent Structures"

  It has long been alleged that constituent structures (or, phrase markers) are subsistent in the theory of grammar.  It has also been widely assumed that constituent structures (or phrase markers) can be identified with trees in graph theory, and trees are formally defined as a collocation of nodes and labels.  Under the Minimalist guideline, however, it is imperative for us to scrutinize whether each of these notions is a truly necessary substance in syntactic theory.  Then, it may come as a surprise to find that there is very scarce evidence in favor of them on conceptual grounds. Moreover, if Collins (1999) is right in claiming that no labels are needed in syntactic theory, then it is no longer possible to define trees by utilizing labels.
  In this talk, I will aim at showing that trees and constituent structures are not necessary in narrow syntax. More specifically, I will claim that there is no operation or condition that makes reference to constituent structures. As an immediate consequence of this claim, not only tree diagrams (and their subsidiary notions such as nodes, branches, etc.) and (labeled) brackets, but also sisterhood, c-commanding, and even domination cannot be defined in narrow syntax.  It will be demonstrated, however, that these notions are fully and safely abandoned under the system in which
formal features and the syntactic relations that are constructed from those features by syntactic operations in a manner constrained by the local economy are considered to be the only legitimate objects in narrow syntax. This is a natural extension of Collins (1999), which is a refinement of
Chomsky's (1994) conception of "Bare Phrase Structure" on the basis of more rigid scrutiny from the minimalist viewpoint.