An acoustic cue reliance pattern in second language speech perception

Abstract

The present study employed acoustic analysis to examine which acoustic cues, spectral (formant) or temporal (duration), native Japanese speakers rely on to identify four mid- and low back vowels in English as a second language. It is generally agreed that Japanese possesses five vowels, whereas English has 15 including diphthongs. This discrepancy is considered to cause confusion to the Japanese learners when they identify English vowels. Furthermore, Japanese possesses a long–short vowel contrast, which may help or cause confusion to the Japanese learners in the English vowel identification. In the present study, a perception test, which was a forced-choice identification task, was administered to Japanese learners of English, and the target back vowels used in the test were later examined acoustically. The results indicated that the learners confused the target back vowels that are temporally similar while they are spectrally different. The acoustic results, together with the patterns of confusion in identifying English back vowels, suggest that these Japanese learners of English are relying primarily on temporal cues rather than spectral cues.