Speaker Demographics Modulate Listeners' Neural Correlates of Spoken Word Processing

Abstract

Abstract In language comprehension, listeners expect a speaker to be consistent in their word choice for labeling the same object. For instance, if a speaker previously refers to a piece of furniture as a “couch,” in subsequent references, listeners would expect the speaker to repeat this label instead of switching to an alternative label such as “sofa.” Moreover, it has been found that speakers' demographic backgrounds, often inferred from their voice, influence how listeners process their language. The question in focus, therefore, is whether speaker demographics influence how listeners expect the speaker to repeat or switch labels. In this study, we used ERPs to investigate whether listeners expect a child speaker to be less likely to switch labels compared to an adult speaker, given the common belief that children are less flexible in language use. In the experiment, we used 80 pictures with alternative labels in Mandarin Chinese (e.g., yi1sheng1 vs. dai4fu, “doctor”). Each picture was presented twice over two experimental phases: In the establishment phase, participants listened to an adult or a child naming a picture with one of the labels and decided whether the label matched the picture they saw; in the test phase, participants listened to the same speaker naming the same picture by either repeating the original label or switching to an alternative label and, again, decided whether the label matched the picture they saw. ERP results in the test phase revealed that, compared to repeated labels, switched labels elicited an N400 effect (300–600 msec after label onset) and a P600 effect (600–1000 msec after label onset). Critically, the N400 effect was larger when listeners were exposed to the child speaker than to the adult speaker, suggesting that listeners found a switched label harder to comprehend when it was produced by a child speaker than an adult speaker. Our study shows that the perceived speaker demographic backgrounds influence listeners' neural responses to spoken words, particularly in relation to their expectations regarding the speaker's label switching behavior. This finding contributes to a broader understanding of the relationship between social cognition and language processing.