VP ellipsis in English and Swedish: a comparative study

Abstract

This comparative study analyses the verb phrase ellipsis phenomenon in English and in Swedish, which exhibit both similarities and differences. In this analysis, VP ellipsis –an elided verb phrase whose meaning is recovered by that of another, semantically identical verb phrase in its surrounding– is treated as a full-fledged syntactic structure that is omitted in the Phonological Form as a result of a [E] feature. This syntactic structure is the same in English and Swedish in contexts of VP ellipsis, which occur when the ellipsis licensor head is a modal or an auxiliary in Swedish, adding to the case of English infinitival to and pleonastic do, both of which are potential licensors of English VP ellipsis. The most striking dissimilarity between these two languages is due to the nature of the support verbs. English do and Swedish göra are different in both semantic and syntactic nature and do not merge in the same position. Göra is not a licensor of VP ellipsis in Swedish, as do is in English, but a licensor of VP pronominalization. When göra is present the verb phrase is pronominalized instead of elided. Moreover, the Swedish support verb can coexist with auxiliaries and modals since it may be nonfinite. This is because it behaves similarly to a main verb. In the case of English, do is completely grammaticalized and need to be finite which makes it incompatible with auxiliaries and modals. The study also looks at the V2 effect in verb second languages that has consequences in the extraction of arguments from the ellipsis site. In English both subjects and objects can move out from the elliptical structure simultaneously. In Swedish, which is a V2 language, this is impossible in cases of both VP ellipsis and VP pronominalization due to competition for the landing site when a sentence is projected.