Speaker: Johanna Alstott (MIT)
Title: On two types of aspectual coercion and before-/after-clauses: Evidence from processing
Time: Thursday, October 2412:30 PM
Location: 32-D461
Abstract: As originally observed by Anscombe (1964) and Heinamäki (1974), certain sentences with before and after are prima facie ambiguous between a strong reading and a weak reading. There is an entailment relation between the two purported readings, and as such two types of theories of these sentences have been proposed. Deflationary theories (Heinamäki 1974, Beaver & Condoravdi 2003, Krifka 2010) posit that the sentences in question are not actually ambiguous: their only LF corresponds to the weak reading, and the strong reading falls out as a subcase. Coercion-based ambiguity theories (Condoravdi 2010; Rett 2020), by contrast, claim that the LF for one reading of the sentences in question has a coercion operator that the other lacks.
This work-in-progress tests an online processing prediction that Retts (2020) coercion-based theory makes and deflationary theories do not. Retts (2020) theory, which relies on two coercion mechanisms (inchoative coercion and completive coercion), makes processing predictions because coercion has a distinctive psycholinguistic profile (Piñango et al. 1999; Brennan & Pylkkänen 2008, etc.): parsers show processing slowdowns in a coercion sentence upon realizing that insertion of a coercion operator is necessary for its grammaticality or felicity. Past work has tested coercion-based theories of various data via processing experiments, arguing for coercion if the claimed coercion cases have an online cost; however, no past work has tested Retts (2020) theory in this way. We do just that, presenting two self-paced reading experiments and pilot results for a third that probe for the existence of Retts inchoative and completive coercion operators. We find initial evidence for Retts completive coercion operator but no comparable evidence for inchoative coercion (cf. Brennan & Pylkkänen 2010).