Does high variability training improve the learning of non-native phoneme contrasts over low variability training? A replication

Acquiring non-native speech contrasts can be difficult. A seminal study by Logan, Lively and Pisoni (1991) established the effectiveness of phonetic training for improving perception: Japanese learners English were trained to perceive /r/-/l/ using minimal pairs over 15 sessions. pre/post-test design learning generalisation. In a follow-up study, Lively, Logan (1993) presented further evidence which suggested that talker variability in stimuli was crucial leading greater These findings have been very influential “high training” is now standard methodology field. However, while general benefit well replicated, an advantage high lower remains mixed. large-scale replication original studies updated statistical analyses we test whether generalise more after multiple talkers single talker. We find listeners learn both conditions. training, no difference how vs low training. When comparing generalisation novel relation pre-training accuracy, ambiguous high-variability low-variability training: This means if exists, effect much smaller than originally thought, such it cannot detected our sample 166 listeners.