Abstract

To infer emotions in speech, listeners can use the way people speak (prosody) or what people say (semantics). We hypothesized that bilinguals and musicians would rely more on prosody than on semantics. In two online experiments, we collected data on 1041 young adults, who listened to sentences with either matching or mismatching semantic and prosodic cues to emotions. Participants then identified the emotion enacted by the speaker’s prosody (ignoring semantics; Experiment 1) or in the semantics (ignoring prosody; Experiment 2). In both experiments, performance suffered when prosody and semantics conflicted. Musicians were better at resisting the interference among bilinguals, but not among monolinguals. Thus, the musician advantage may not be due to a difference in weighting prosody over semantics, rather an overall better ability to inhibit irrelevant information. As for the hypothesized bilingual advantage, the findings warn that musicianship is critical to control for.