Poster F74, Tuesday, March 28, 8:00 – 10:00 am, Pacific Concourse
Low expectations: An ERP investigation of cue-based anticipatory processing in low constraint sentences
Kailen Shantz1, Darren Tanner1; 1University of Illinois at Urbana-Champaign
There is robust and growing evidence which strongly suggests that language comprehension employs predictive mechanisms, yet the conditions under which prediction does and does not occur remain unclear (cf. Huettig, 2015). To address this, we combined event-related potentials with a modified response cueing paradigm to investigate the time course of prediction in low constraint sentences where number marking or phonological information on verbs and determiners provide cues to the identity of impending nouns. We further examine how the reliability of cue-response mapping impacts prediction. We recorded EEG while native English speakers monitored low constraint sentences for nouns, and made responses with their left and right hands when a particular noun occurred in a sentence. We systematically varied within-subjects whether target nouns were preceded by an informative cue about the identity of the noun (e.g. the vs a/an or this/these). Between subjects, we also manipulated whether cues reliably mapped onto the same response hand. Preliminary inspection of the lateralized readiness potentials (LRPs) time-locked to the cue words suggest an LRP onsetting at approximately 250-300 ms post-stimulus onset when indefinite determiners provide an informative cue about the upcoming noun, but only for participants with a reliable cue-response mapping. LRP activity is not apparent in any other condition prior to the onset of the target words. These results suggest that not all possible cues may be used to generate expectations during language comprehension, and that prediction may be restricted to situations where it is easy to do.
Topic Area: LANGUAGE: Syntax