Mechanisms of chain formation

The theoretical domain of the present investigation is the mechanisms of chain formation. The empirical domain is the nature of resumption. I provide compelling arguments in favor of a movement-based analysis of resumptive chains. However, unlike more traditional analyses, I do not take a resumptive pronoun to be a (minimal) copy of its antecedent. Instead, I argue that resumptive elements and their antecedents are distinct syntactic entities, which form a constituent with their antecedents upon First Merge. Resumptive chains are the result of stranding (subextraction) under A-bar movement. My proposal makes correct predictions in various domains pertaining to the interpretive consequences of resumption, the relation between resumption and clitic doubling, and cases of agreement mismatch between the resumptive pronoun and its antecedent, which turn out to be crucial in defining the nature of resumption. I define as precisely as possible how resumptive chains are formed, which necessitates a theory of extraction. The answer I suggest is strongly reminiscent of Ross\u27s (1967). For Ross, movement was unbounded. Crossing an island in and of itself did not suffice to yield a deviant output. Rather, only certain types of rules were sensitive to islands. I revise Ross\u27s taxonomy in such a way as to make agreement processes island-sensitive. Movement triggered in the absence of agreement can be island-insensitive. By stranding resumptive pronouns, antecedents are able to undergo Move without Agree, and thereby void islandhood. A careful examination of the properties of resumptive pronouns is shown to predict when the latter will be island-sensitive. The final chapter of this work expands the data base by examining more marked instances of resumption, and shows how these can be accounted for at no cost. In particular, cases of mixed chains, resumptive pronoun fronting, clitic left dislocation, and interacting A-bar dependencies are analyzed. Instances of so-called intrusive pronouns (resumption restricted to island contexts) are examined, and formally distinguished from cases of genuine resumption. The chapter ends with a discussion of some implications of the present analysis of resumption for domains like Weak Crossover, parasitic gaps, reconstruction, and asymmetries between interrogative and relative clauses.