Locality and clause size matter: Semantic selection of nominalized interrogatives

In this talk, I investigate s(emantic)-selection of interrogatives with a particular focus on nominalized clauses. Comparing empirical patterns from Korean, Japanese and Turkish, I provide following observations. First, what seems to be non-local s-selection of interrogatives in these languages is subject to the weak version of Phase Impenetrability Condition (PIC2; Chomsky 2001). Second, these languages opt for different strategies of encoding the interrogative information in nominalized contexts. Based on these observations, I argue that s-selection of interrogatives is satisfied by a type of agreement dependencies. This syntactic dependency requires valuation of a selection feature, and fails if it crosses more than one phase boundary (due to PIC2). Different strategies of encoding interrogatives in nominalized clauses are attributed to different size of nominalization and also to their locality profile. My claim suggests that s-selection of certain clause types is better captured as a syntactic dependency rather than a lexical constraint (in line with Safir 2021). Variations in size of nominalization further provide a support to the view that finiteness comes in different size of clauses (Adger 2007; Wurmbrand & Lohninger 2023).