Phonomotor Training for Accent Modification

Abstract

There is a lack of evidence regarding efficacious training programs for modifying accent. The program developed for this study is based on a parallel distributed processing model of phonology and is expected to improve production of sounds in non-native speakers of English, and generalize to measures of intelligibility. The data gathered from this treatment study will inform models of language processing as well as inform the development of subsequent trials of the same program. The following research questions asked if intelligibility and listener effort improved following training and whether there was a difference in production of trained sounds and untrained contexts following the training program. Post-training results revealed a large effect size for phonemes in isolation thus supporting the hypothesis that non-native speakers can improve phoneme level production with targeted phonomotor training. However, generalization results beyond phonemes in isolation were variable suggesting that future participants would benefit from an increased training period and further program development to promote transfer of skills.