Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Processing Definiteness Ambiguity: Sustained ERP Response May Reflect Computations Specific to Semantic Definiteness
Poster Session B, Friday, September 12, 4:30 - 6:00 pm, Field House
Xiaoyu Yang1, Ellen Lau1; 1University of Maryland, College Park
Previous ERP research on the processing of noun phrases with a definite marker reveals that they pose a givenness and uniqueness demand on their referents within a discourse (Anderson & Holcomb 2005; Schumacher, 2009; Kirsten et al., 2014). However, semantic definiteness is not necessarily tied to the presence of definite markers. For example, the subject of ‘The dinosaurs are extinct’ does not refer to a group of specific dinosaurs even though it has the definite marker, whereas the subject of ‘Handouts are on the table’ does even though it is a bare noun phrase. It thus becomes an interesting question whether phrases with and without a definite marker still show any neural correlates of semantic definiteness/indefiniteness when presented without a disambiguating context. Here we report a study aimed at taking a preliminary step towards addressing this question by comparing ERP responses to coordinated vs. uncoordinated definite and bare plural noun phrases. Our results are consistent with the view that a sustained ERP response reflects computations specific to semantic definiteness. In a prior ERP study (Lau & Liao, 2018), a sustained anterior negativity (SAN) response was observed in coordinated bare plural noun phrases compared to their uncoordinated counterparts (e.g. sunlit ponds and green umbrellas vs. sunlit ponds ### green umbrellas) for natural language but not jabberwocky materials, providing potential evidence that the sustained response reflects the semantic consequences of bare plural coordination. However, their exact nature is indeterminate in this study, as bare plurals are ambiguous between a definite (an existing group of entities) and an indefinite (a conceptual kind) reading. One possibility is that coordination biases the bare plurals towards a definite reading. Here we compared coordinated and uncoordinated bare plurals with definite plural noun phrases (e.g. the scented gardens and the ardent dancers vs. the scented gardens ### the ardent dancers). If the previous SAN effects reflect differential computations specific to the definite vs. indefinite bare plural reading, we expect to see a different ERP response to the unambiguous definite plurals. We conducted one experiment presenting participants with only the definite sequences, and a subsequent within-subject experiment presenting both definite and bare plural sequences. In both experiments (n=18, 22 respectively), text materials were presented using RSVP, 600/ms per word. Participants’ task was to respond to an occasional recognition memory probe following some of the trials. We examined epochs of -200:2000ms time-locked to the coordinator/placeholder. In both experiments, we found no evidence of the SAN for definite plural coordination. However, at the second NP we observed a numerical posterior positivity for coordinated vs. uncoordinated definite plurals. Unexpectedly, when bare plurals were intermixed with definite plurals in the second experiment, they no longer demonstrated a SAN for coordination. We hypothesize that intermixing definite plurals and bare plurals might make their semantic contrast more salient, biasing the bare plural coordination towards the indefinite reading.
Topic Areas: Meaning: Lexical Semantics,