Empirical Motivation

Abstract

A verbal particle in Swedish consists of a single word. A full phrase which includes a modifier and/or a complement can therefore never appear in the particle position. This fundamental fact motivates the proposal that particles are non-projecting words. The basic analysis of Swedish particles was sketched in chapter 1. The present chapter motivates and explains the proposal further (section 1). This chapter also discusses and rejects three alternative approaches. First, I consider an analysis where the particles project full phrases (section 2). This analysis runs into several problems. I show that the problems are not readily solved with a simple movement analysis, as it is necessary to posit mandatory movement in some cases, optional movement in some cases, and in yet other cases we need to obligatorily prevent movement. Second, I consider a morphological analysis, where the verb and the particle are assumed to be combined in the lexicon (section 3). This hypothesis is rejected since the verb and the particle do not need to be adjacent, and since the verb-particle ordering goes against the rules of compounding in Swedish, which dictate that compounds are head-final. Third, I consider the possibility of analyzing particles as clitics (section 4). This analysis is shown to be untenable as well, as Swedish verbal particles are not phonologically weak, a characteristic that is normally taken to be a defining feature of ‘clitichood’.