Title: Cross-linguistic patterns in the selectional restrictions of non-veridical preferential predicates

Speaker:  Ciyang Qing (青慈阳) (University of Edinburgh)

Time: 16:00 ~ 18:00  (May 10)

Abstract:  This talk concerns the selectional patterns of clause-embedding predicates. While some predicates such as know and be certain are compatible with both declarative and interrogative complements, e.g., Alex knows/is certain [that/whether it is raining], other predicates such as believe and hope are (or at least canonically reported to be) compatible only with declarative complements, e.g., Alex believes/hopes [that/*whether it is raining]. There are various proposals in the literature that attempt to explain such selectional restrictions of clause-embedding predicates from their semantic properties. However, most of the discussions so far focus on English and a handful of other languages, and therefore it is unclear whether existing proposals would lead to robust cross-linguistic generalizations. In this talk, I will use Uegaki & Sudo's (2019) generalization about Non-Veridical Preferential (NVP) predicates such as hope/wish/fear/worry as a case study and report interim results from our ongoing cross-linguistic investigation of clause-embedding predicates. Uegaki & Sudo report that NVPs in English are incompatible with interrogative complements and provide a formal analysis that aims to explain this restriction. However, recently White (2021) provides corpus examples challenging their generalization/analysis. By examining White's counterexamples, I suggest that while Uegaki & Sudo's empirical generalization needs to be revised, their formal account provides a useful baseline to analyze the various types of counterexamples. I will then show how the new analysis leads to more robust crosslinguistic generalizations.