A Theory of Catbgory Projection and Its Applications

This thesis proposes a new system of category pr0jection where Lexical categories and Nonlexical (or "Functional") categories project in different ways, which is crucially differerlt from the standard views in Wklich all categories project in the same fashion. In Chapte r 1, I in t roduce some of ttle ba sic not ions of Government-Binding Theory withiIl which all of the discussion in this thesis takes place. The aim of Chapter 2 is to show the fundamental difference between Lexical categories and Functional categories. That is, Lexical categories have Lexical Conceptual Structures (LCS) in the sense of Hale and Keyser (1985), whereas Functional categories do not have r,exical Conceptual Struct:ures comparable to the ones Lexical categories have, and the latter type of categories only have the function of "connf!cting" two syntactic units via some sort of "binding" and nYgreement." Based on this fundamental difference, a new projection system is introduced, in whictl Lexical categories project up to a single-bar level, allowing free recursion at that level, while Functional categories can project up to a double-bar level, taking a unique specifier and a unique complement. Chap t e r 3 e x p lor e s v a rio usc 0 n seq ue nee S 0 f t 11 e projection system introduced in Chapter 2. One important consequence is that the proposed projection system, combined with a "bottom-up" a-marking mechanism, predicts that the so-called n~xternal argument" appears within the projection of a Lexical head at D-structure, receiving the external a-role in that position, and then moves outside the Lexical projection to its S-structure position, for Case reasons. This move makes possible the explicit syntactic representation of what has been called the "impl ici t argument" both in noun phrases and in clauses (in the case