Voicing judgements by chinchillas trained with a reward paradigm

Abstract

Experiments were performed to replicate and extend previous findings of similar categorization of voiced/voiceless consonant–vowel (CV) syllables by humans and chinchillas. A reward paradigm was applied to the question of how stimulus range affects the voice-onset-time (VOT) corresponding to the voiced/voiceless category boundary. Each of four adult chinchillas and four human subjects identified synthetic CV syllables as voiced (/ba/, /da/, /ga/) or voiceless (/pa/, /ta/, /ka/) using voiceless standards of either 80 or 120 ms. In both humans and animals, extending the VOT range from 80 to 120 ms shifted category boundaries to longer VOTs, but to a different extent across listeners. Control experiments suggested that listeners were attending to different phonetic cues in a manner that depended on the listener, rather than on species. The results are interpreted in terms of similar contextual effects and use of multiple phonetic cues to voicing in humans and animals.