Acoustic analysis of perceived accentedness in Mandarin speakers’ second language production of Japanese

Abstract

Many second language (L2) learners, particularly adult learners, retain foreign accent on their L2 production. What acoustic sources give rise to the perception of foreign accent? This study examines beginning and intermediate Chinese learners’ production of Japanese in terms of segmental and suprasegmental features and investigates the relationship between acoustic characteristics of the L2 production and accentedness ratings provided by native Japanese listeners. Results of acoustic examination indicated that learners’ production varied considerably from that of native speakers in terms of durational features of stops, spectral features of some vowels, pitch (F0) peak alignment, and F0 contour. Multiple regression analysis identified the second formant of /u/, F0 peak alignment and contour as the strong predictors of perceived accent, accounting for nearly 90% of variance. These findings confirmed Flege’s Speech Learning Model hypothesis—L2 sounds that are similar to L1 sounds, while subphonemically distinct, seem to pose greater difficulty for acquisition than dissimilar sounds. Moreover, longer classroom experience was found to show limited effects in reducing perceived accent, with slightly greater effects on segmental than suprasegmental variables.