Multisensory pronunciation training in a video conference-based foreign language classroom

Speech production is multimodal and multisensory. We present a novel approach to segmental pronunciation instruction that directs foreign language (L2) learners to simultaneously monitor the visual, tactile-articulatory, and auditory information of their articulators' movements in a video conference-based classroom. We reasoned that this would allow L2 learners to heighten their proprioceptive awareness for speech motor actions and thus facilitate pronunciation gains. We present preliminary results from a study with German learners of English who participated in a pronunciation training session with their video self-view activated. The focus was on the L2 /v-w/ contrast in singletons and /Cw/-clusters. Each participant completed one small-group lesson and a pre-, post-, and delayed post-test on Zoom. Tasks were designed to elicit data from controlled to spontaneous speech. Here, we report preliminary results from the controlled word-list reading task. L1 English raters judged words both a) categorically, i.e., whether speakers produced /v/ or /w/, and b) on a six-point goodness scale. Most participants started near ceiling; individual speakers with the most room for improvement showed slight gains from pre- to post-tests; /kw/-clusters had the lowest ratings at the delayed post-test. Although no significant differences between time points were found in the preliminary data, the results suggest that L2 learners might benefit from pronunciation training that facilitates the integration of multisensory feedback, in particular visual and somatosensory feedback.