The identity of phasal boundaries has mostly been considered in light of minimal CP-TP-vP-VP structures. The question this paper addresses is where the clause internal phase boundary lies in light of more complex structures in which aspectual projections intervene between TP and vP. I claim progressive aspect to be unique amongst aspectual forms in English in that it is part of the clause internal phase, whilst perfect aspect and all higher functional items are contained within the CP/TP phase. This claim accounts for many peculiar quirks of progressive aspect in English, namely in VP ellipsis, fronting phenomena, idioms and existential constructions.On the theoretical front I argue that this division in the aspectual hierarchy is best understood through a variable approach to phases in which the highest projection within a sub-numeration acts as the phase, irrespective of what that projection is. This denies vP of its exclusivity as the clause-internal phase, and allows the progressive layer to project the phase when present. This approach generally sits in line with the move towards a dynamic understanding of phases, as per Bobaljik & Wurmbrand (2005), Wurmbrand (2012, to appear) and Bošković (to appear a, b).
IntroductionIt has long been observed that certain domains appear to exist in natural language which exhibit syntactic, phonological and semantic independence from the rest of the structure surrounding them. An ongoing issue for generative grammarians has been how these facts should best be captured within a syntactic framework. Various proposals have been made to explain these seemingly opaque domains, such as Chomsky's Barriers (1986) model. The Minimalist Program (Chomsky 1995), however, abandoned such earlier proposals in favour of phase theory (Chomsky 2000(Chomsky , 2001. Under phase theory, the syntactic derivation is essentially built up in a series of discrete chunks of structure, known as phases, rather than forming the entire derivation in one go. Once complete, each phase is sent off independently to PF and LF for pronunciation and interpretation, thereby establishing the apparent independence of certain domains. With respect to the main clausal spine, Chomsky (2000, 2001) assumes this to be bifurcated into two discrete phases; the lower phase demarcated by vP, and the higher phase demarcated by CP. The focus of this paper will be on the identity of the lower, clause internal phase.For the most part, the phasal boundary of the clause internal phase is only ever considered in the context of a minimal CP>TP>vP>VP structure.1 But what happens if we look at phases in the light of more articulated structures? Consider, for instance, the sentence below:(1) Betsy must have been being paid to keep quiet about the crime.
2In this example there is clearly a more detailed structure intervening between TP and vP involving a number of aspectual projections. Phases are rarely explored in the context of these more elaborate structures, and on the few occasions that they have been considered, ther...