Acoustic variability in the speech of second language learners of American English as a function of accentedness

Abstract

The primary issue of interest in the present study concerned acoustic variability among L2 learners of English with different degrees of accentedness. Specifically, we were interested in determining whether L2 learners with stronger accents differ from L2 learners with weaker accents in terms of the amount of within-subject variability they manifest when producing English consonants and vowels. Twenty L2 English learners from nine different L1 backgrounds and a group of 20 native English control subjects produced a number of target sounds contained within CVC words that were embedded in a carrier phrase. Accent ratings for the twenty L2 talkers were obtained, and acoustic measurements were made of various consonants and vowels; coefficient of variation [(S.D. ÷ mean) × 100] was also computed for each of the acoustic measures. A number of temporal and spectral comparisons were made between L2 talkers with stronger versus weaker accents and with the native control subjects. Results indicated that although L2 subjects with stronger accents sometimes showed greater inter-subject (i.e., group) variability, they did not typically show more within-subject (i.e., token-to-token) variability than subjects with weaker accents, regardless of how accurate they were in producing native-like consonants and vowels.