Divide et Impera---Separating Operators from their Variables

Date of Completion

January 2010


Language, Linguistics




This dissertation investigates the role of lexical decomposition in several syntactic and semantic issues. Two phenomena in particular are discussed: Wh in situ in Wh-Movement languages and the behavior of even in English and Brazilian Portuguese. For the first phenomenon, I propose that separating Wh-phrases into two parts that can move independently from one another can explain crosslinguistic differences between, among others, French and Brazilian Portuguese with respect to the syntactic contexts that allow for Wh in situ. I also investigate different strategies for establishing syntactic covert relations and argue that Agree and Unselective Binding are the two strategies that are needed to explain Wh in situ in Brazilian Portuguese. Next, I look at the same phenomenon from a semantic/pragmatic point of view and conclude that, in a Wh-Movement language like Brazilian Portuguese, Wh in situ is only possible when the speaker assumes that the presupposition of the question is part of the Common Ground. I also discuss some crosslinguistic variation in this respect. I then move to a semantic phenomenon, i.e. the behavior of even in English and Brazilian Portuguese. More specifically, I investigate the behavior of even under a predicate like glad. I show that in this case a lexical-decomposition analysis is not necessary or enough to explain the facts adequately. On the other hand, I propose that the distribution of stressed any can be accounted for if we assume that it can be decomposed into any itself and a silent even. ^