Proceedings of the Annual Meeting of the Cognitive Science Society

Voice recognition depends on listeners' ability to retain and identify voices despite variability in the speech signal. Here, we examined the relative impact of changes in the speaker's language and changes in their emotional state on voice recognition. Participants were exposed to a speaker and later asked to identify that speaker from a lineup of voices. Across trials, exposure and test were either matched or mismatched in language and emotion, introducing variability in the speaker's phonological and acoustic characteristics. This design yielded four conditions: no-mismatch, language-mismatch, emotion-mismatch, and complete-mismatch. Overall, switching emotion impaired voice recognition more than switching languages, especially when the exposure took place in an unfamiliar language and a neutral tone of voice, demonstrating that the magnitude of these effects is context dependent. A tentative explanation of these results is proposed in the discussion.