In the Japanese cleft construction, multiple noun phrases (NPs) can occupy the focus position even if they do not form a constituent in the mainstream generative grammar. However, a single NP with the nominative case marker ga cannot. Kubota and Smith (2006, 2007) analyze the cleft construction with Combinatory Categorial Grammar (CCG, Steedman, 1996, 2000; Baldridge, 2002), but their analysis overgenerates a ga-marked NP in the focus position. Indeed, they recognized the obligatory omission of the nominative case marker in that position, and they assumed that some independently motivated principles should explain the distribution. However, it would be better if the distribution could be explained within the grammar formalism. This study aims to address this issue by partially incorporating the idea of constructivist analysis of argument structure from the mainstream generative grammar Kratzer (1996) into the CCG framework. Furthermore, I will show that this revision correctly predicts two syntactic phenomena where the ga-marked NP behaves differently from other case marked NPs.
日本語のcleft（分裂文・断裂文）において、focus（焦点）位置に絶対にガ格名詞句を置くことができない（たとえば、 花子を呼んだのは太郎がだ は非文。）が、他の格ならOK（花子が呼んだのは太郎をだ はOK）、という事実を、CCGの統語論で解決することを提案した論文です。 具体的に、Kratzer (1996) によるconstructivistのアプローチを参考にした格付与のシステムをCCGで形式化しています。 これにより、日本語においてガ格だけが他の格と絶対的に異なる挙動をするという事実（ガ格名詞句は長距離かき混ぜできない、ガ格名詞句はsmall clause内におけない）も説明できることを示しています。