Dutch Nominalization and the Architecture of Grammar

Abstract

This thesis focuses on the question of whether an active lexicon, or, more broadly, a pre-syntactic computational module, is necessary. The phenomenon of nominalization is of special significance in this debate, as it involves categorization, argument structure, and transformation, which are all issues which in the multiple generative engine conception are, in whole or partly, associated with the lexicon or non-compositional cycle. I focus on the empirical issue of nominalization in order to explore the theoretical issue of the architecture of grammar. My empirical domain is Dutch. I find that the chosen theory of nominalization, rooted in Distributed Morphology, being the most prominent and consequential single generative engine framework, is unable to account for Dutch nominalization. Furthermore, abstracting away from specific theories to the single generative engine framework itself, I find that the Dutch nominalization data poses fundamental problems for a single generative engine framework, making it empirically implausible.