High variability identification and discrimination training for Japanese speakers learning English /r/–/l/

Abstract

Second-language (L2) learners can benefit from exposure to phonetically variable speech during computer-based training. Moreover, this training can be effective even for L2 learners who have extensive exposure to their L2 in daily life, suggesting that there is something specific about the training task that aids learning. The present study compared traditional identification training with discrimination training to evaluate whether discrimination training could be effective, and whether different types of focused attention (i.e., on categorization vs. perceptual differences) could combine to provide a greater increase in learning. Adult Japanese speakers were given 10 sessions of identification and discrimination training, with pre/mid/post tests of identification, auditory discrimination, category discrimination, and /r/–/l/ production. The results demonstrated that both identification and discrimination training increased accuracy of Japanese speakers’ perception and production of English /r/–/l/ in similar ways, but that there was little added benefit to using the two training methods in combination. It thus appears that identification and discrimination training have similar effects in second-language learners, as long as both training methods incorporate high variability.