Language categorization by adults is based on sensitivity to durational cues, not rhythm class

Studies of listeners’ ability to distinguish languages when segmental information is eliminated have been taken as evidence for categorical rhythmic distinctions between language groups (“rhythm classes”). Furthermore, it has been suggested that sensitivity to rhythm class is present at birth and that infants must establish the rhythm class of their native language as a precursor to language acquisition. We tested the hypothesis that adult listeners’ ability to distinguish between languages is better predicted by differences in specific durational cues than by putative rhythm classes. We examined the categorization of language pairs using utterances in which only durational characteristics were preserved. We found that English listeners could distinguish between not only English and Spanish (from different rhythm classes), but also between different accents of British English. Furthermore, patterns of categorization between and within languages highlighted the contribution of speech rate, durational contrast and utterance-final lengthening.

► Language discrimination studies have supported distinct rhythm classes.
► We converted natural English and Spanish utterances into flat sasasa speech.
► Listeners distinguished Spanish and English, and also different English accents.
► They used speech rate and localized durational cues, such as final lengthening.
► The results indicate listeners are sensitive to timing rather than to rhythm class.