Effects of consonantal voicing on English diphthongs: a comparison of L1 and L2 production

An acoustic comparison was made between native and non-native English diphthongs. We focused on the extent to which diphthong duration varied due to post-vocalic voicing. Talkers recorded three types of diphthongs /aΙ, eΙ, oΥ/ in /CVC/, /CVnC/ and /CVC↔/ words in which the second C contrasted in voicing. The major difference between the talker groups was found for the diphthongs in /CVC/ and /CVnC/ words for which the voicing effect was larger in native than in non-native English. Two groups also differed clearly in the duration of the nasal in /CVnC/. While native talkers lengthened both diphthongs and nasals before /d/, non-native talkers lengthened only diphthongs in this position. In /CVC↔/, there was a greater similarity between the two groups for both the mean and variation of diphthong duration. The data in this study indicate that transferring the L1 Japanese durational patterns to English would result in a positive transfer for /CVC↔/, but a negative one for /CVC/. L2 learners need to acquire the knowledge and skills to pronounce the diphthong with distinct production strategies according to the syllable type in order to approximate to the phonetic norms of the native English speaker.