Learning challenging L2 sounds via computer-assisted training : audiovisual training with an airflow model

Abstract

Background: Teaching Mandarin as a second language (L2) has become an important profession and an important research area. The acquisition of unaspirated and aspirated consonants in Mandarin has been reported to be rather challenging for L2 learners. Objectives: In the current study, a 3-D airflow model was integrated into the virtual talking head for audiovisual pronunciation training in these Mandarin consonants. Methods: Using the eye-tracking technique, Experiment 1 investigated L2 learners' general acceptance and gauged attention distribution online while learning with the system. To further compare training efficacy, learning outcomes were evaluated by randomly dividing Japanese learners of Mandarin into different training groups using the 3-D tutor with and without an airflow model, respectively, in Experiment 2. Results and Conclusions: Eye-tracking results showed that the talking head was well accepted, and was successful in directing L2 learners' attention to the visual modality of the airflow model marking the contrast in aspiration. Furthermore, the additional visual cue of the airflow model helped enhance their production of the aspirated Mandarin stops. Moreover, this computer-assisted training approach was shown to be robust as the advantage of training with an airflow model can be generalized to novel syllables with a change of tones or rimes.