Spectral and temporal implementation of Japanese speakers’ English vowel categories: A corpus-based study

Abstract

This study investigates the predictions of second language (L2) speech acquisition models — SLM(-r), PAM(-L2), and L2LP — on how native (L1) Japanese speakers implement the spectral and temporal aspects of L2 American English vowel categories. Data were obtained from 102L1 Japanese speakers in the J-AESOP corpus, which also includes nativelikeness judgments by trained phoneticians. Spectrally, speakers judged to be non-nativelike showed a strong influence from L1 categories, except L2 /ʌ/ which could be deflected away from L1 /a/ according to SLM(-r)and L2 /ɑː/ which seemed orthographically assimilated to L1 /o/ according to PAM(-L2). More nativelike speakers showed vowel spectra similar to those of native English speakers across all vowels, in accordance with L2LP. Temporally, although speakers tended to equate the phonetic length of English vowels with Japanese phonemic length distinctions, segment-level L1-L2 category similarity was not a significant predictor of the speakers’ nativelikeness. Instead, the implementation of prosodic-level factors such as stress and phrase-final lengthening were better predictors. The results highlight the importance of suprasegmental factors in successful category learning and also reveal a weakness in current models of L2 speech acquisition, which focus primarily on the segmental level. Theoretical and pedagogical implications are discussed. 

Keywords

• L2 vowel categories
• L2LP
• J-AESOP corpus
• SLM(-r)
• PAM(-L2)
• Japanese