ABSTRACT
Rizzi (1997) presents what has been termed a ‘split analysis’ of the complementizer phrase (CP). This is sketched in(1)below. The proposal is very broad, and it is claimed to explain a variety of facts about the English complementizer system. In this chapter, I will not take on many of the large aspects of the proposal. Here, I wish to deal with two phenomena, each of which raises some degree of doubt about the extent to which a split CP analysis is explanatory of the CP architecture of English. This in turn raises questions about the universality and perhaps the efficacy of split CP. The central facts to be discussed here concern echo questions (EQs) and their correspondence to CP structure under various assumptions about what CP structure is like. As I will try to show, a description of the patterning of possible and impossible EQs in English appears to be fairly straightforward under what I will call the ‘classic’ CP analysis. In contrast, the description of EQs becomes more complicated under the split CP analysis. Following this is a brief consideration of negative inversion (NI) constructions and a phenomenon that I will term ‘echo negation.’ If NI constructions parallel questions in their involvement with CP structure, as has been claimed (e.g., Haegeman 2000), then one might expect their ‘echo negation’ forms to have properties similar to EQs; however, as we will see, they do not. Thus, these findings constitute a small bit of evidence that the classic analysis of CP, one consisting of Spec, CP, and C, may be more appropriate for English. This result is compatible with work on other syntactic phenomena in English that also appear not to require a split CP analysis, including negative inversion (Sobin 2003) and the Comp-trace effect with and without adverbials (Sobin 2002). These findings are compatible with those of Newmeyer (this volume), who also argues that a split CP analysis is not needed for, and in fact induces complications in, the analysis of elements dislocated to the left periphery.
