Shui Mauser | 2026 I.S. Symposium
狈补尘别:听Shui Mauser
罢颈迟濒别:听A Bayesian Pragmatics Account of Common Ground: Behavioral and EEG/ERP Evidence
惭补箩辞谤蝉:听Neuroscience; Mathematics
础诲惫颈蝉辞谤蝉:听Grit Herzmann; Michael Bush
Perspective-taking is critical for communication. A central component of this is common ground integration, the process by which listeners use shared knowledge to interpret what a speaker means. However, the precise time course of this integration remains unclear. Some studies report early effects of common ground on language processing, while others find that such information is incorporated later. The present study investigated when and how listeners use common ground by combining a director鈥檚 task, in which participants must select objects based on instructions from a speaker whose visual perspective differs from their own, with a Bayesian computational model of perspective-taking. Participants completed the task while accuracy, reaction times, and brain electrical activity (event-related potentials, or ERPs) were recorded. Neither behavioral measures nor ERP components associated with referential processing (Nref and P600) showed reliable differences across conditions, suggesting that listeners incorporated the speaker’s perspective without measurable processing costs. Regression analyses also revealed no significant relationship between model-derived measures of uncertainty and brain responses. These findings suggest that listeners integrate speaker perspective rapidly and efficiently, consistent with accounts favoring early integration. The absence of significant condition effects may reflect the specific task demands or limited referential ambiguity in the paradigm, which may not have strongly engaged perspective-taking mechanisms. Future research should explore how varying task demands, multi-modal cues, and richer computational models can better characterize the perspective-taking process.
Posted in Symposium 2026 on May 1, 2026.
