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

 
Poster C58, Thursday, November 9, 10:00 – 11:15 am, Harborview and Loch Raven Ballrooms

Tracking the neurophysiological correlates during the computation of agreement dependencies: the access of grammatical feature and associative representations in spoken language

Jane Aristia1, Angèle Brunellière1;1University of Lille, CNRS, UMR 9193 - SCALab - Sciences Cognitives et Sciences Affectives, F-59000 Lille, France

Most of event-related potential (ERP) studies (for a review, Molinaro, Barber & Carreiras, 2011) investigating the computation of agreement dependencies revealed three neurophysiological correlates related to grammatical violations: left anterior negativity, N400, and P600. Nonetheless, little is known in terms of how the representations are accessed during the agreement processing. Some prior studies suggested the using of abstract representations in terms of grammatical features to compute agreement dependencies (Harley & Ritter, 2002; Pearlmutter, 2000) between words, whereas others proposed the access to associative lexical representations to code the probability between words (Truswell & Tanenhaus, 1994; Seidenberg & MacDonald, 1999). Moreover, the neurophysiological correlates associated with grammatical violations have been mainly examined in reading task. Therefore, the present study aims to explore on how the representations are accessed during the agreement processing in spoken language. To this end, we conducted an ERP experiment, in French pronominal subject-verb agreement wherein we manipulated the abstract features involved in agreement violations and the associative frequency between the pronominal subject and the verbal inflection. First, four experimental conditions are created as followed (1) congruent, (2) person violation, (3) person & number violation, (4) more number than person violation (i.e., orthographically it is a number violation but in spoken language, it can be recognized as number and person violation) to test the access of abstract features. The incongruent stimuli were twenty percents of the total stimuli and participants were asked to make a response if they hear a nonword. The results based on the neurophysiological correlates associated with the grammatical violations are in line with the previous studies suggesting the access of grammatical features to compute agreement dependencies. More precisely, the person violations elicited a posterior P600 wave at around 800 ms from the onset of verbs in comparison with the congruent conditions whereas the number and person violations triggered an anterior negativity at around 600 ms and a P600 at around 900 ms. Additionally, more number than person violations elicited a late negativity at around 700 ms over the anterior sites. To probe the activation of associative representations, we analyzed separately the pronouns depending on how frequent they are followed by a particular verbal inflection (e.g., nous proceeded by –ons inflection is more frequent than je by –ai). We found that low associative frequency between the pronominal subject and the verbal inflection induced a stronger negative response in comparison with high associative frequency. Altogether, beyond on the access of abstract grammatical features, this study highlights the role of associative representations during the computation of agreement dependencies.

Topic Area: Grammar: Morphology

Back to Poster Schedule