Talker-variability in visual feedback articulatory training: Transfer to words, long-term maintenance and generalization

An important challenge in second language acquisition is learning how to produce novel sounds. Visual-acoustic feedback training has been shown to improve second language production, but it is still unclear whether training on isolated sounds transfers to words and is maintained over time. Forty-six Spanish speakers were trained to produce unfamiliar French vowels /e/-/ɛ/ through exposure to either single- or multiple-talker stimuli. On each trial, they heard a vowel, produced it, and received feedback showing the position of their vowel and the target in the acoustic space. Both groups showed improvements in vowel production, which transferred to words and were maintained twelve days later. However, better generalization – more accurate production when repeating vowels after an unfamiliar speaker - was maintained only in the multiple-talker group. This study shows that visual-acoustic feedback training is an effective method for improving second language production as the gains of such training demonstrate three key features for pedagogical practice: transfer, generalization and maintenance.