You are viewing the SNL 2017 Archive Website. For the latest information, see the Current Website.

Poster C52, Thursday, November 9, 10:00 – 11:15 am, Harborview and Loch Raven Ballrooms

ANY ERP effects

Hongchen Wu1, Jun Lyu1, Aydogan Yanilmaz1, John E. Drury1;1Stony Brook University

[INTRODUCTION] English ANY lives a double-life. It can serve as a Negative Polarity Item (NPI) (e.g., "There wasn't ANYONE in the room") requiring the presence of a licensor like negation (cf. "There was *ANYONE in the room"). But ANY also can be a Free Choice Item (FCI) (e.g., "ANYONE can do that"). Within linguistic theory, the relationship between the NPI/FCI-variants of ANY is controversial. For example, accounts of FCI-ANY conflict, alternatively proposing that it introduces universal quantificational force (like EVERY; Dayal 1998), or that it is rather a type of indefinite (Kadmon & Landmann 1993; Horn 2000; Giannakidou 2001). [PRESENT STUDY] We examined ERP responses to EVERY/ANY in four embedded contexts, varying the presence/absence of a main clause subject licensor ((1a)/(2a) vs (1b)/(2b)) and whether the Target occurred in the embedded subject position (1a/b) or the "pivot" position of an existential construction(2a/b). Universals/definites are deviant in pivots positions (EVERY in (2a/b) is unacceptable; McNally 2011). (1a) {LICENSED/SUBJECT} None of us thought that EVERYONE/ANYONE was in the room. (1b) {NON-LICENSED/SUBJECT} All of us thought that EVERYONE/*ANYONE was in the room. (2a) {LICENSED/PIVOT} None of us thought there was *EVERYONE/ANYONE in the room. (2b) {NON-LICENSED/PIVOT} All of us thought there was *EVERYONE/*ANYONE in the room. Previous ERP findings examining these types of logical semantic/pragmatic deviance (Drury & Steinhauer 2009) led us to predict that the ungrammatical cases (*) in (2) would elicit a P600 effect relative to licensed ANY in (2a). Our interest in the (1) cases was to observe whether ANY might (not) differ from (universal) EVERY. [METHODS] In a reading/judgment ERP study, adult native English speakers (N=18) viewed critical sentences RSVP (500 ms), providing acceptability judgments after every trial. EEG was recorded continuously (32 channels), and ERP mean amplitudes were examined for 100 ms time-windows for 1200 ms epochs, with -200 to 0 ms serving as the baseline. [RESULTS] EVERY/ANY subjects (1) revealed an N400 effect (300-500 ms) in Licensing contexts (ANY>EVERY in (1a)). Strikingly, there were no EVERY/ANY differences in the Non-licensing subject contexts (1b), although unlicensed ANY subjects were consistently rejected sentence-finally. As pivots (2), ANY yielded a larger N400 than EVERY in both Licensing and Non-licensing conditions (main effect of Target word only). Subsequently, however, we found Context x Target word interactions due to the presence of relative positivities for both EVERY conditions and for ANY in Non-licensing contexts relative to ANY when it was licensed. [CONCLUSION] Additional processing occurs whenever ANY is accessed/retrieved as an NPI (in the scope of a licensor in subject position, or in existential constructions), indexed by larger N400s. The absence of (Non-) Licensing differences for ANY subjects suggests that whether FCI-ANY is treated as an indefinite or a universal, in subject contexts it incurs no additional processing effort relative to universal EVERY (either in terms of retrieval or integration/composition operations). We discuss these patterns both in the context of current debates on NPI/FCI-ANY and in connection with theories/models of the etiology of the N400 and P600 ERP components.

Topic Area: Meaning: Lexical Semantics

Back to Poster Schedule