Inner vs. Outer A-not-A Questions

Mercredi 11 Février 2015, 15:00 to 17:00
Invité: 
Dylan Wei-Tien Tsai (National Tsing Hua University)
Organisation: 
UFR Linguistique LLF UFR LCAO
Lieu: 

Salle 681C
Université Paris-Diderot
Bâtiment Grands Moulins, aile C
5 rue Thomas Mann
75013 Paris

This paper distinguishes two types of A-not-A questions in Mandarin according to their distributions and interpretations: One is triggered by an outer A-not-A morpheme hosted by an assertion projection in the left periphery; the other is licensed by an inner A-not-A morpheme situated on the edge of vP. This distinction is of particular interest in the context of the cartographic approach advocated by Rizzi (1997) and Cinque (1999), under which we propose a novel way to separate the two constructions in terms of their relation to a variety of adverbials: For instance, inner A-not-A is typically blocked by frequency/manner adverbials, while its outer counterpart, being much higher and discourse-oriented, is subject to no such blocking. More specifically, we analyze outer A-not-A as the head of AstP (assertion phrase), which is sandwiched between IntP (interrogative phrase) and EviP (evidential phrase), whereas inner A-not-A is taken to be part of the vP periphery in the spirit of Ernst (1994) and Law (2006). Under this analysis, A-not-A questions are triggered by a cluster of a strong uninterpretable V feature and an interpretable Q feature: inner A-not-A is associated with the head of vP, whereas outer A-not-A is associated with the head of AstP in the left periphery. More specifically, the strong [uV] feature is checked off by attracting the closest verbal head, thereby assuming the A-not-A form. At LF, the [iQ] feature is further adjoined to Int head to check off the weak uninterpretable Q feature in Int (Chomsky 2000, among others). One piece of evidence for this minimalist/cartographic analysis comes from the fact that light verbs and certain manner adverbs may optionally undergo inner A-not-A construals. We have built a fine-grained cartographic analysis of the two types of A-not-A questions based on their correspondences with modals, adverbials and light verbs, as well as their height of interpretation. This move in turn accounts for the fact that only outer A-not-A is speaker-oriented and infelicitous in an out-of-blue context.