Search Abstracts | Symposia | Slide Sessions | Poster Sessions
When A Man Says He Is Pregnant: The Neural Correlates of Speaker-contextualized Language Processing
Poster Session C, Friday, October 25, 4:30 - 6:00 pm, Great Hall 3 and 4
Hanlin Wu1, Zhenguang Cai1; 1The Chinese University of Hong Kong
#Introduction# One fundamental aspect of spoken language is that it carries both linguistic content such as the explicit message and extra-linguistic contexts such as the speaker’s identity. Over the past two decades, researchers have used ERPs to investigate the neural correlates of speaker-contextualized language processing by exposing participants to implausible utterances where the message mismatches the speaker’s identity. However, the findings were mixed, with some studies showing N400 effects (e.g., Van Berkum et al., 2008) while others showing P600 effects (e.g., Foucart et al., 2015). We hypothesize that this inconsistency of findings relates to how listeners deal with different types of implausibilities. When the utterance violates people’s social stereotypes (socially implausible utterances) such as “I’m going to have a *manicure* this weekend” spoken by a man compared to a woman, listeners tend to integrate the utterance with their social stereotypes, resulting in an N400 effect reflecting an effortful integration of the message, the speaker’s identity, and listeners’ social stereotypes. In contrast, when the utterance violates people’s biological knowledge (biologically implausible utterances) such as “The first time I got *pregnant* I had a hard time” spoken by a man compared to a woman, listeners tend to revise their perception of the utterance to reconcile this extremely low plausibility, as predicted by rational accounts of language processing (e.g., Cai et al., 2022; Gibson et al., 2013), resulting in a P600 effect reflecting an effort to revise the utterance for a possible interpretation. #Methods And Results# To test our hypothesis, we recorded 80 socially plausible/implausible Mandarin utterances and 80 biologically plausible/implausible utterances. Plausibility ratings (30 participants: 15 males, 15 females) showed that socially implausible utterances were significantly more plausible than biologically implausible utterances. ERPs (another 64 participants: 32 males, 32 females) showed that socially implausible utterances triggered an N400 effect, 300-900 ms after the critical word onset. In contrast, biologically implausible utterances triggered a P600 effect, 600-1000 ms after the critical word onset. Additionally, social implausibility effects decreased as a function of the listeners’ personality traits of openness (measured by BFI-2, Zhang et al., 2022) for both ratings and ERPs (N400 magnitudes), while such observation was absent for biological implausibility effects. #Discussion And Conclusion# These findings suggest a new framework for speaker-contextualized language processing, where different types of utterances go through distinct neural-cognitive processes, depending on how an interpretation can be achieved. Specifically, listeners arrive at an interpretation (posterior) based on their world knowledge (prior) and the utterance (evidence). For socially implausible utterances, listeners integrate the utterance with their social stereotypes to arrive at an interpretation, resulting in an N400 effect. For biologically implausible utterances, as they deviate too much from the listener’s biological knowledge, an error revision process is triggered for the perception of either the message or the speaker’s identity, resulting in a P600 effect. Listeners’ openness traits modulate their social stereotypes but not biological knowledge. Our proposed account explains the empirical inconsistencies in speaker-contextualized language processing and sheds light on how the human brain processes language in a wider context.
Topic Areas: Meaning: Discourse and Pragmatics, Meaning: Lexical Semantics