In this thesis I examine the question that was left unanswered in Tulling (2014). Why can some SFP1s, like the Dutch hè, attach to [-Q]-questions, while most SFP1s are unable to attach to questions...Show moreIn this thesis I examine the question that was left unanswered in Tulling (2014). Why can some SFP1s, like the Dutch hè, attach to [-Q]-questions, while most SFP1s are unable to attach to questions? This paper focuses on rhetorical wh-questions. First the different theories on the structure and meaning of rhetorical questions are discussed. Then I define what I consider a rhetorical question, and provide tests to distinguish rhetorical questions from ordinary questions in Dutch. I argue against the proposal that ordinary questions and rhetorical questions are semantically and structurally the same (Copanigro & Sprouse 2007), and argue in favor of the idea that RQs and OQs are actually semantically and syntactically different from each other. Combining earlier theories on the semantics and structure of rhetorical questions (Han 2002, Obenauer & Poletto 2000) and generalizations on the structure of SFPs using the split-CP hypothesis (Law 2004, Sybesma & Li 2007), I propose the following to account for the main question of this paper: In contrary to wh¬-words in ordinary questions, which are positioned in [Spec, ForceP], the wh-words in rhetorical questions are positioned in [Spec, EpistP]. Following Sybesma & Li (2007) SFPs are base generated across the different layers of CP. The SFP1 hè is base generated in Epist0. SFP1s have an inherent feature [+SFP] which requires to be checked by all the relevant sentence information. In ordinary statements, the entire sentence moves to [Spec, EpistP] to check this SFP feature in sentences with hè. In RQs the wh-word carries all the relevant epistemic information of the sentence (Han 2002), and the SFP-feature can be checked. In ordinary questions there is no wh-word in EpistP, and the [+SFP]-feature cannot be checked; the sentence is ungrammatical. Show less
