A language effect in voice identification

Bilingual students recorded messages in English, Spanish, and English with a strong Spanish accent. Monolingual English‐speaking subjects heard a single message and attempted to identify the voice in a six‐person line‐up 1 week later. The line‐up message was delivered in the same language and accent as the initial message. Voices were identified best when speaking English and worst when speaking Spanish. Identification accuracy was intermediate for the accent condition. There were no reliable differences among conditions in false alarms when the target voice was absent from the line‐up. The effect of language was replicated using a 30‐min retention interval. Familiarity with the language and language constraints on voice characteristics were discussed as possible explanations of the language effect.