Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Testing the social N400: Dissociating partner knowledge from co-presence
Poster Session E, Friday, October 2, 11:00 am - 1:00 pm, Wangari Maathai
Odysseas Myresiotis Alivertis1, Jasna Martinovic1, Vicky Chondrogianni1, Hugh Rabagliati1; 1University of Edinburgh
Recent studies have reported a “social N400” effect, whereby comprehenders show enhanced N400-like responses to sentence continuations that are predictable/plausible for themselves but expected to be implausible for a partner lacking access to critical supportive context (Rueschemeyer et al., 2015; Jouravlev et al., 2019). This pattern has been interpreted as reflecting sensitivity to a partner’s knowledge state during language comprehension. However, prior studies confounded the partner’s knowledge state with their mere co-presence, leaving open whether the effect reflects partner modelling or more general social-contextual influences on semantic processing. Separate work suggests that another person’s mere presence may modulate N400 amplitudes across conditions (e.g., Forgács et al., 2022), although evidence remains mixed (Sinha et al., 2025). To dissociate these accounts, we independently manipulated partner presence and shared knowledge in a preregistered EEG study, extending Rueschemeyer and colleagues’ joint sentence comprehension paradigm. We recorded 64-channel EEG as participants (N=39, M age=20, range=18-29) listened to auditory context sentences and judged the plausibility of critical sentences across three conditions (99 stimulus triplets): Plausible (expected/plausible regardless of auditory context; e.g. “My car has shiny new wheels now”); Implausible (unexpected/implausible regardless of context; e.g. “My boat has shiny new wheels now”); and Context-dependent, in which the same Implausible continuations were rendered expected and plausible by the preceding context (e.g. “I converted my boat into a car”). Participants completed three blocks: Alone; Hearing Partner (with a co-present partner who also heard the auditory context); and Non-hearing Partner (partner lacked access to context). In both joint blocks, participants judged whether the critical sentence would make sense to the partner, holding co-presence and judgements constant, varying only the partner's knowledge state. We pre-registered mixed effects regressions fit to single-trial N400 amplitudes (350-550ms, centro-parietal ROI). We observed robust canonical lexico-semantic N400 effects (Implausible >Plausible, βs =-.86 to -1.37, all p<.02) and Context-dependent targets patterned similarly to Plausible targets in the Alone condition as predicted. By contrast, we found no evidence for a general mere-presence effect (Hearing vs. Alone: β=0.16, p=.72), nor for the predicted “social N400”: the Context-dependent–Plausible contrast did not increase reliably in Non-hearing Partner relative to Alone, when the EEG participant was aware that the partner lacked the critical context (interaction: β=- 0.44, p=.39). A direct comparison of the two joint blocks revealed a difference in the Context-dependent–Plausible contrast (interaction: β=−1.06, p=.042); however, follow-up analyses suggested that this was driven primarily by an unexpected positive-going shift beginning in the N400 window for Context-dependent trials in the Hearing Partner condition, rather than reliably increased negativity in Non-hearing Partner. To summarise, recent work has suggested that the “social N400” indexes sensitivity to a co-present partner’s differing knowledge state. We sought to replicate this effect while evaluating whether it could instead be explained by the partner’s mere presence. Unexpectedly, we found no clear evidence for such an effect, suggesting that prior findings may be less robust than previously assumed. However, as our design necessarily differed from earlier studies we are currently conducting a direct replication of the original social N400 paradigm.
Topic Areas: Meaning: Discourse and Pragmatics, Meaning: Lexical Semantics