Presentation
Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Lightning Talks
Neural signatures of conversion and stress alternation in English noun-verb recognition
There is a Poster PDF for this presentation, but you must be a current member or registered to attend SNL 2023 to view it. Please go to your Account Home page to register.
Poster A32 in Poster Session A, Tuesday, October 24, 10:15 am - 12:00 pm CEST, Espace Vieux-Port
Charles Redmon1,3, Anna Gupta2, Aditi Lahiri1, Frans Plank1,2, Carsten Eulitz1; 1University of Oxford, 2University of Konstanz, 3University of Essex
INTRODUCTION. Word stress in English is governed in part by word class, with stress predominantly initial on nouns and non-initial on verbs. Further, English has many noun-verb pairs where stress is the only indicator of word class: e.g., cónvert vs. convért (cf. Sherman 1975, Kelly 1988). Previous research has examined how monosyllabic cases of conversion (noun and verb identical in pronunciation) are processed, where no stress alternation is possible (Pliatsikas et al. 2014, Wheeldon et al. 2018). These studies show increases in processing difficulty for words of greater complexity (boating [[[boat]N]Ving]V as compared with [[walk]Ving]V) despite having identical surface forms. The present study aims to fill this gap by examining the interaction between conversion and lexical stress. METHODS. Participants completed a cross-modal priming lexical decision task where they heard complex words over headphones followed by visually presented stems. Prime-target pairs were structured according to four conditions: (A) stress-alternation (constrúcting→cónstruct[N] / constrúct[V]), (B) conversion from noun base with initial stress (píloting→pílot [N/V]), (C) conversion from verb base with final stress (desígning→desígn [N/V]), and (D) verb-only with final stress (fóllowing→fóllow [V-only; no N counterpart]). An equal number of nonword targets were presented following real-word primes so that the lexical status of the prime could not bias the decision to the visual target. Finally, EEG was recorded while participants completed this task and subsequently pre-processed and averaged into ERPs by condition, priming (related vs. control), and participant. RESULTS. Based on preliminary results from 15 participants (32 will be presented at the conference) three of the four conditions showed significant priming as reflected in N400 reduction over the interval of 300-500 ms following stimulus onset. Within this set, the initially stressed noun-base forms (B) showed the greatest priming (~1.4 μV) followed by stress-alternating and finally stressed verb-base conditions (A/C, ~1 μV each), though this interaction was not significant. The verb-only cases (D) showed a smaller N400 reduction (~0.8 μV) which was not significant. These results are distinct from those from response times in a parallel behavioural experiment (N=32) wherein the verb-only condition (4) showed significantly greater priming than the other three, though all conditions primed; however, this priming effect was due primarily to slower baseline reaction times to verb-only words that words that are multiply represented as nouns and verbs. DISCUSSION. This pattern of results indicates a complex interaction between word class, derivational structure, and stress. When the stress on the prime is initial, participants may expect a noun reading, which is available and metrically consistent in B (píloting→pílot), but not in D (fóllowing→fóllow), yielding priming in B but no priming in D. When stress is non-initial, such as in conditions A and C, both prime-target pairs have the same metrical structure and derivational structures (both are verb-basic), and thus their priming effects are equivalent. Finally, when the base stress is held constant (e.g., initially stressed noun in A/B; cónstruct/pílot), there is greater facilitation from primes with consistent stress (B) than when the stress shifts between verb and noun (A).
Topic Areas: Morphology, Prosody