Generalized inversion and theory of agree

Abstract

In this thesis I examine some of the fundamental questions surrounding inversion structures. I first provide an analysis of Locative Inversion. I show that the mixed A- and A- syntactic behavior of the fronted PP in English could be derived once we understand how the featural composition of locative phrases influences on the Probe-Goal relation between C and the postverbal DP. In particular, I argue that there is a correlation between syntactic categories of locative phrases and typological differences in the syntactic patterns in Locative Inversion: in Mandarin Chinese, Chichewa, Kinande and Gungbe, locatives are (or can be) represented by nominal categories (i.e. equipped with complete cp-features) and these locatives exhibit pure A-properties in Locative Inversion; in English and Sesotho, however, they are characteristically represented by non-nominal categories and the locative phrases are thus forced to undergo two-step movement from an Aposition to an A-position as avoidance of intervention effects in the Agree system.I also discuss a variety of (generalized) inversion constructions, including English Quotative Inversion, Sentential Subject and French Stylistic Inversion. In these constructions I show that since a cp-deficient constituent moves to [Spec, TP], additional operations (such as topicalization) have to take place so as to destroy the potentially offending structure created by the fronted defective elements.

(cont.) Specifically, I suggest that these are related constructions because they all display a mixture of A- and A- properties.Finally I focus on the generalization concerning the placement restrictions of arguments by Spell-Out, in particular the principles that force argument externalization from the vP and VP. I argue that argument externalization is motivated by Case-related concerns.