Adult and infant perception of two English phones.

Previous research has shown that young infants easily discriminate both native and non-native consonant contrasts, but by 10-12 months of age infants perform like adults and easily discriminate only native consonant contrasts. The present study was designed to determine what kind of experience is required to maintain discrimination of native consonants. To address this question, English listeners of three ages (6-8 months, 10-12 months, and adults) were presented with the phonetic difference, [da] vs [t = a]. This distinction occurs in English but is not phonemic: [t = a] occurs when it follows an [s] (as in /sta/). If passive exposure is sufficient to maintain discrimination, all age groups should discriminate [da] vs [t = a]. However, if phonological status plays a role, then older infants and adults should fail. In experiment 1, English adults judged [da] and [t = a] to be equally good instances of the same phonemic category /da/. In an AX procedure in experiment 2, English adults discriminated [da] vs [t = a] better than chance but worse than native phonemic levels. In the Conditioned Head Turn procedure in experiment 3, adults and 6- to 8-month-old infants discriminated [da] vs [t = a], but 10- to 12-month-old infants did not. Taken together, these results are most consistent with the hypothesis that phonological status plays a role in maintaining discrimination of phonetic information.