On features and the MLC

This thesis investigates the proper characterization of triggers and locality conditions governing the displacement property of human language. Following the introductory statement of issues and background assumptions in Chapter 1, Chapter 2 presents the currently predominant view of displacement on which it involves a matching relation between uninterpretable features of the target/probe and interpretable features of the goal. It is shown that such a view faces serious problems in both undergenerating grammatical constructions and allowing ungrammatical ones, and in incorporating complex and vague notions in the computational system such as the mechanism of ‘equidistance’, the suppressive role of inherent Case, the ‘activating’ role of uninterpretable features, etc. An alternative on which only uninterpretable features are subject to the operation Attract/Move is presented in Chapter 3, and it is shown that such an analysis makes correct empirical predictions and avoids conceptual problems inherent in the opposite view. Chapter 4 proposes a principle of Late Expletive Insertion on which the ‘superraising’ and several other construction are ruled out without the need for the mechanism of ‘defective intervention’. Chapter 5 extends the proposed analysis to the core cases of A\u27-movement, showing that it provides a straightforward way to capture the effects of Pesetsky\u27s Path Containment Condition in the Minimalist framework. I show that the proposal can be reconciled with evidence for the successive cyclic character of movement under a version of Chomsky\u27s Form Chain condition. I also show that a view on which structural Case is but a reflex of agreement relation faces serious empirical and conceptual problems. Chapter 6 recapitulates the main results of the thesis.