No abstract

No abstract

The term “residual Verb Second ” is a misnomer for English, because V2 is, in fact, still productive in the language. Evidence for this comes from a previously undescribed negative inversion phenomenon innovated very recently in varieties of English. An analysis is proposed for how such a restrictive V2 system could, nevertheless, be productive, appealing to learner-driven models of language change in which novel structures can arise as artefacts of the acquisition procedure. Specifically, it is argued that innovative V2 arises when acquirers postulate a novel clause type characterized by a left-edge operator, which they analyse as a V2 environment by analogy with other non-declarative clause types involving such structures (e.g. interrogatives). This finds support from other cases of innovative V2 in English, Scots, and Afrikaans. Overall, we are left with a clearer picture of the status of V2 in English, and what it takes to innovate new V2 environments cross-linguistically.

This paper attempts to pinpoint the derivational timing of ellipsis by examining how it interacts with other phenomena whose timing is better understood. It is demonstrated that for morphophonological processes such as external sandhi and allomorphy, material adjacent to an ellipsis site cannot “see” ellipsis internal material. Under the mainstream [E]-based approach to ellipsis, this result indicates that the silence characterizing ellipsis arises early at PF, consistent with non-application of Vocabulary Insertion (and inconsistent with phonological deletion). However, it would also follow straightforwardly if ellipsis sites were sent to the interfaces alone, the moment they are licensed in the syntax (Aelbrecht 2010). This approach involving the Segregated Transfer of ellipsis sites is pursued here: beyond the mounting syntactic evidence supporting it, it is shown to be preferable on these post-syntactic grounds as well, requiring fewer assumptions to capture these novel findings.

In British English (BrE), a subset of pronominal objects of prepositions in have/with possessives may be optionally realised asprepositional object gaps(POGs). In this short paper, we introduce three core properties of this previously unreported phenomenon, and then outline a preliminary syntactic analysis to straightforwardly capture them. These properties are: POGs are only observed in BrE, POGs are only observed in have/with possessives, and POGs are only observed in structurally simplex complements of possessivehaveandwith. We show that these properties are straightforwardly captured by an analysis that treats POGs as arising from A-movement of the possessor. We claim that the locus of variation between dialects that permit POGs and dialects that do not is the feature specification of a single syntactic head, which either induces or precludes A-movement. This proposal accords with current Minimalist approaches to microparametric variation, in which all variation stems from the lexicon.