Wataru Uegaki
Project Introduction + Constraints on the lexical semantics of question-embedding predicates
In recent years, several authors have argued that responsive predicates—the clause-embedding predicates that can embed either a declarative or interrogative complement—semantically select for questions rather than propositions, based on a number of empirical observations, e.g., false-answer sensitivity of factive predicates (Theiler 2014, Theiler, Roelofsen & Aloni 2018; cf. George 2011), interpretation of nominal complements (Uegaki 2016), behavior of Predicates of Relevance (Elliott, Klinedinst, Sudo & Uegaki 2017), and the interpretation of the Estonian verb mõtlema (Roberts 2018).
However, there is also a possibility that the question-oriented semantics for responsive predicates is too powerful, i.e., too unrestricted. That is, by itself, the theory allows us to define a cross-linguistically non-existent predicate. As a candidate for such a non-existent predicate, Spector and Egré (2015) discuss a fictitious predicate *shknow, which means ‘know’ when it embeds a declarative clause but means ‘wonder’ when it embeds an interrogative clause.
In this talk, based on proposals in the recent literature (Theiler et al. 2018; Uegaki 2018), I will discuss possible ways to constrain the question-oriented semantics for responsive predicates so that a non-existing predicate like *shknow can be ruled out, without abandoning the empirical advantage of the theory. These constraints constitute possible working hypotheses for the lexical-semantic universals in the domain of responsive predicates.
In the beginning of the talk, I will also introduce the general agenda of the ModUni project.