Search Abstracts | Symposia | Slide Sessions | Poster Sessions
ERPs during compound picture-naming to dissociate morphosyntax and phonology
Poster Session B, Friday, September 12, 4:30 - 6:00 pm, Field House
This poster is part of the Sandbox Series.
Ellen Lau1, Sophie McCarthy1; 1University of Maryland
Novel compound formation as in 'lemon basket' or 'cat seat' necessarily involves the combination of existing phonological (/kaet/, /sit/), syntactic (Noun1, Noun2), and meaning ({cat}, {seat}) units. But what happens in the case of known compound nouns with fixed meanings, like 'airbag' and 'strawberry', is a more interesting question with broader implications: do mental representations and computations continue to make reference to smaller units, even when it seems not to be strictly necessary for successful production and comprehension? Previous research on the production of known compound nouns in error patterns of patients with aphasia (e.g. Blanken, 2000; Badecker, 2001; Chiarelli et al. 2007) and using behavioral and ERP measures (e.g. Koester & Schiller 2008; Lorenz et al. 2021) has provided considerable evidence that production of known compounds continues to make reference to smaller linguistic units. However, whether those smaller units are morphosyntactic, phonological, or both, is still a matter of some debate. Here we use an auditory priming manipulation during a compound picture naming ERP study to provide further evidence. A recent ERP study by Cuonzo et al. (2024) found that compound constituent priming of naming ('bag' -> 'airbag') resulted in an ERP effect with a qualitatively different polarity and distribution from phonological priming ('bat' --> 'airbag'). However, in their materials, the amount of phonological overlap was smaller in the phonological prime condition (onset and nucleus) than in the compound constituent prime condition (full syllable). In the current ERP study, we use phonological wordforms containing embedded wordforms ('acorn') to create a non-compound phonological priming condition ('corn' --> 'acorn') that contains full syllable overlap. Participants hear the spoken prime word (e.g. 'corn'), followed by a target picture (an acorn) which they must name. Phonological priming effects are assessed by comparing the response to picture naming when preceded by a phonological prime vs. an unrelated control (e.g. 'truck' --> 'acorn'). Compound constituent priming effects are assessed by comparing the response to compound picture naming when preceded by the compound head ('bag' --> 'airbag') vs. an unrelated control (e.g. 'tree' --> 'airbag'). Each participant sees 30 items from each of the 4 conditions, with condition-item pairings counterbalanced across two lists, intermixed with 134 monomorphemic picture naming fillers. Participants are familiarized with the list of picture names prior to the experiment but not with the pictures themselves, to minimize the contribution of episodic memory to the task. We will measure production accuracy, reaction time, and ERPs from the onset of the picture, during the production planning period from 0-700ms (any trials with faster naming RTs will be excluded to prevent motor contamination of ERPs). If compound constituent priming in picture naming is due to shared morphosyntactic units, we predict qualitatively different ERP effects from phonological priming. On the other hand, observing qualitatively similar priming effects across conditions will suggest that these effects largely reflect facilitation due to shared phonological form. Data collection is currently underway.
Topic Areas: Language Production, Morphology