#clausal embedding

Our speaker will be Tanya Bondarenko from Harvard University and Patrick D. Elliott from the University of Düsseldorf

Abstract: This goal of this presentation is to provide an account of an observation made by Sharvit (2023, henceforth Sharvit’s puzzle): NPIs can be licensed in complement clauses that occur inside of definite descriptions, but not relative clauses. We show that all existing accounts of clausal embedding fail to account for this fact, including Kratzer’s content-based semantics (Kratzer 2006, 2016) which is based on Hintikka’s (1969) treatment of attitude verbs as universal modals, and recent refinements which equate the content of an attitudinal eventuality with the proposition denoted by the embedded clause (Moulton 2009, Elliott 2017, Bondarenko 2022). Independently, the latter equality-based approaches are thought to be problematic since they render attitude reports non-monotonic, thus failing to capture a class of entailments which are straightforward from a Hintikkan perspective. In this talk, we propose that monotonicity of attitude reports is best modelled with the notion of incrementality (Krifka 1998). We show that once equality semantics is supplemented with the idea that monotonic attitudes have incremental propositional content, we not only fix the bad predictions of this approach about entailment, but also solve a puzzle about NPI licensing (Sharvit 2023) that other approaches cannot account for.