The seminal work such as Abney (1987) and Szabolcsi (1983, 1994) suggests that noun phrases have sentence-like structures. More specifically, noun phrases have functional categories analogous to I or C. The views raise the following question. How parallel are structures of clauses and noun phrases? In order to answer the question, we focus on interesting asymmetries which occur between nominal structures and clausal structures. More specifically, this paper explores restrictions on non-occurrence of so-called genitive markers previously discussed by An (2009, 2012, and 2014) and impossibility of movement inside the nominal domain. We also discuss further asymmetries related to quantifiers and adverbials which are not properly analyzed in the previous literature. We show that non-parallel behaviors are explained under the assumption that positions related to discourse are not available inside nominal domains.

In possessor agreement constructions in Korean, where a possessor and its possessum show case agreement, the possessum seems to resist extraction. Interestingly, such restriction disappears in elliptical contexts and right dislocation. This paper focuses on the issue, presence or absence of the restriction possessum movement shows. Cho (2007) suggests that the movement asymmetries observed in the possessor agreement construction are accounted for as a consequence of cyclic linearization requirement and shows that the contrast between elliptical context and non-elliptical contexts is correctly predicted under the cyclic linearization-based analysis. Following this analysis, I suggest that the contrast between leftward movement and right dislocation contexts is similar to the contrast between non-elliptical and elliptical contexts.

This paper explores the absence or presence of CP ellipsis in Japanese and Korean. Saito (2007) argues that in Japanese and Korean, arguments such as DPs and CPs can undergo ellipsis unlike in English since agreement is optional in these languages. He further puts forward an LF copying analysis of argument ellipsis. A couple of puzzles, however, need to be resolved. First, no extraction out of CP ellipsis should be explained. Second, the fact that CP ellipsis is sensitive to selection of matrix verb should be explained. We suggest that apparent DP and CP ellipsis in Korean are all instances of a null pronoun, the so-called pro. We reanalyze the apparent instances of DP and CP ellipsis discussed in Saito (2007), and propose that they indeed involve deep anaphora pro but not surface anaphora ellipsis.