Issues in the Syntax of Movement: Cross-Clausal Dependencies, Reconstruction, and Movement Typology

Abstract

This dissertation is a study of the syntax of phrasal movement and its effects on interpretation. Specifically, it argues that the reconstruction patterns in Mongolian scrambling and raising (ECM), which cannot be easily subsumed under the standard A/A'-distinction, emerge from independent properties of grammar. On the basis of a detailed empirical study of Mongolian phrase structure and movement, the current work also explores a series of interrelated issues including free word order, movement typology, Binding Condition C reconstruction, radical reconstruction, case assignment, and feature valuation. Chapters 2 and 3 present an overview of Mongolian syntax, with a special focus on the free word order phenomenon in this language. I demonstrate that short scrambling in Mongolian shows consistent A-properties, but intermediate and long distance scrambling show mixed A/A'-properties. Based on the empirical findings in Chapters 2 and 3, Chapter 4 focuses on Condition C reconstruction effects in scrambling. Mongolian scrambling displays paradoxical patterns in terms of Condition C reconstruction which cannot be easily subsumed under the standard A/A'-distinction. Building on Takahashi and Hulsey (2009), I argue that Condition C reconstruction effects in Mongolian do not track movement types, but are instead directly tied to the language's case system, which controls the applicability of Wholesale Late Merger (WLM). In particular, accusative case in Mongolian is assigned as a dependent case, and nominative is assigned via Agree (cf. Baker & Vinokurova 2010). I suggest that this case assignment mechanism in Mongolian has nontrivial consequences for how Condition C is interpreted in scrambling chains. The novel data presented in this chapter strongly suggest that a WLM-based account of reconstruction effects requires a fine-grained view of the case mechanism of the language in question. Further, the direction pursued in this chapter contributes a new perspective to the study of reconstruction effects in different types of movement, especially long distance scrambling. Chapter 5 builds on the results obtained in the preceding chapters and discusses some well-known issues concerning the A/A'-distinction and the syntax of exceptional case marking (ECM). I present additional novel data on the interaction between ECM and wh/topic licensing in Mongolian, which cannot be easily related to the standard A/A'-distinction. I propose an Agree-based analysis for Mongolian ECM, which derives the relevant movement properties from the features involved in Agree relations, without directly appealing to the A- or A'-status of syntax positions. In addition, I also investigate the differences among raising (ECM), scrambling, and other types of movement in terms of Condition C reconstruction and radical reconstruction. I propose that the relevant contrasts can be accounted for based on case assignment and Agree relationships involved in different types of phrasal movement.