Hardison, D. M. (2003). Acquisition of second-language speech: Effects of visual cues, context, and talker variability. Applied Psycholinguistics, 24, 495–522.
Presentation: Belinda
Summary: Sally
In this study, the effects of modality (audio-visual (AV) versus audio-only (A-only)), vocalic context (English cardinal vowels), and word position (a 2-by-2 combination of initial-versus-final and singleton-versus-cluster) were investigated on Japanese (Experiment 1) and Korean (Experiment 2) learners’ identification of American English /r/ and /l/. Possible transfer from perceptual training to production performance was also examined. The same pretest-posttest design (with a 3-week training session with feedback in between) was adopted for both experiments. The only difference was that in Experiment 2, Koreans learners were further divided into single- and multiple-talker groups to examine for the effect of talker variability. Results showed that both Japanese and Korean learner groups who received AV training outperformed those who received A-only training. In terms of vocalic contexts, stimuli followed by rounded vowels were consistently harder for both learner groups. Differences were found mainly in word position: /r/ and /l/ in word-initial positions seemed to be more difficult for Japanese learners, whereas those in word-final positions were harder for Korean learners. Specifically, visual information contributed more significantly to the most challenging phonetic contexts: /u, o/ in initial clusters for Japanese learners and /i, ɪ/ in final singletons for Korean learners. Differences were found among individual talkers regarding the intelligibility of the stimuli they produced. However, the effectiveness of High Variability Phonetic Training (Lively et al., 1993) could also be reached by the combination of visual and audio modalities in this study. In the generalization tests of Experiment 2, Korean learners assigned to the single-talker condition performed equally well as those assigned to the multiple-talker condition in AV training. Finally, transfer from perceptual training to production performance was observed. For both Japanese and Korean learner groups, posttest recordings were consistently judged as more intelligible than pretest ones by native English speakers.
