The Relationship Between Speech Perception And Production In Adult Learners If English As A Foreign Language: Implications For Phonetic Training

In China, phonetics is an area that has not been adequately studied, however, crucial to the fundamental function of language—communication. This study examines perception and production skills in adult English-as-a-Foreign-Language (EFL) learners by testing their behavioral plasticity in intensive perceptual training. The participants were 39 Chinese college students who had received at least eight years of EFL education. Comprehensive assessments of perception and production of all English phonemes were administered before and after training. The digitally recorded production data were rated on a scale of 1-5 by two native speakers of English with formal graduate-level training in phonetics. Based on the pre-test results, a training software program was developed to incorporate enriched linguistic exposure, including systematically controlled temporal and spectral exaggeration, multi-talker variability, audio-visual presentation, and adaptive listening (Zhang et al., 2009). Four pairs of English contrasts that are known to be difficult to acquire for Chinese learners, namely, /?/ vs. /i/, /?/ vs. /?/, /n/ vs. /?/, and /r/ vs. /l/ were used in training. The participants were divided into three groups respectively going through multi-talker training, single-talker training, and no training,. Identical perception and production procedures were used in pre-post tests to assess learning. The findings suggested a significant relationship between speech perception and production at a p<0.01 level. Participants had significantly better perception and production on phonemes in initial position than in final position of the word. Significant perceptual improvements and transfer of learning to novel words were consistently found in the experimental groups as compared to the control group. The multi-talker group had significantly better performance than the single-talker group. However, the training effects on production were not retained in the second post-test five months after training. Furthermore, no consistent correlations between perception and production on the contrasting segments were found. A large correlation was only found for the /?/-/i/ pair in the single-talker group in post-test 2 (r=0.75), and the /r/-/l/ pair in the multi-talker group in the first post test (r=0.74). The results indicate the complex nature of behavioral plasticity in adult learners with non-uniform perceptual-motor links and large inter-subject variability. These findings have important implications for ESL/EFL phonetic training and learning in terms of optimizing training effects and reducing accent in oral English.