Speaker: Linmin Zhang (NYU Shanghai)
Title: Post-suppositions and wh-questions: Intervention effects revisited
Time: Friday, February 4th, 6:00pm – 7:30pm
Abstract: In this talk, I investigate a novel, post-supposition-based account for intervention effects of wh-questions, i.e., the degradedness or uninterpretability of certain wh-questions containing focus items (e.g., only) or quantifiers (e.g., exactly 3 girls), as illustrated in (1) (see also Beck 2006, Li and Law 2016) as well as (2) and (3).
(1) Focus intervention in Chinese wh-questions that contain only (zh?y?u)
a. zh?y?u [Zh?ng-S?n]F dú-le y?yìxué sh?.
only Zh?ng-S?n read-PFV semantics book
‘Only Zh?ng S?n read a book (or books) on semantics.’
[With only (zh?y?u)]
b. * zh?y?u [Zh?ng-S?n]F dú-le shénme sh??
only Zh?ng-S?n read-PFV what book
Intended: what book(s) did only Zh?ng S?n read?
[only
wh ? degraded]
c. shénme sh? zh?y?u [Zh?ng-S?n]F dú-le?
what book only Zh?ng-S?n read-PFV
‘What book(s) did only Zh?ng S?n read?’
[wh
only ? natural]
d. Zh?ng-S?n dú-le shénme sh??
Zh?ng-S?n read-PFV what book
‘What book(s) did Zh?ng S?n read?’
[Without only: Chinese is a wh-in-situ language]
(2) a. Exactly 3 girls are above 6 feet tall.
b. *How tall are exactly 3 girls?
(3) a. Exactly 3 girls read exactly 5 books.
[? distributive, ? cumulative]
b. How many girls read exactly 5 books?
[? distributive, # cumulative]
c. *How many books did exactly 3 girls read?
To account for these degraded wh-questions, I start with the interpretation of focus items and quantifiers in declarative sentences, examining how they affect the process of semantic composition. I argue that all these items that create intervention effects in wh-questions are triggers of post-suppositions in declarative sentences (see Brasoveanu 2013), bringing delayed evaluations. On the other hand, I propose that wh-questions (e.g., who did she kiss) are parallel to definite descriptions (e.g., the one she kissed) and also work in a post-suppositional fashion. Specifically, the semantic contribution of wh-words (e.g., who) is twofold: introducing a discourse referent, and then imposing tests of maximality as delayed evaluations. Intervention effects are considered due to an order conflict in semantic derivation, resulting from two distinct sources of post-suppositions.