| Perform an experiment investigating perception and production of Cantonese by native speakers of English and Mandarin using a) percetpual discrimination, categorization, and goodness rating tasks; and b) dynamic articulatory production data including simultaneous ultrasound images of the tongue, video of the lips, and acoustic speech recordings. |
| Compare patterns of perceptual assimilation with dynamic articulatory production data, in order to test the hypothesis (H1) that perceptual assimilation of nonnative speech sounds is grounded in articulatory and gestural similarity, rather than acoustic-phonetic similarity, such that nonnative speakers attempt to produce articulatory rather than acoustic vowel targets. |
| Carry out dynamic articulatory analysis of tongue and lip gestures using generalized additive mixed models (GAMMs) in order to test the hypothesis (H2) that L2 speakers learn to successfully perceive nonnative articulatory gestures, but fail to produce them in a native-like manner due to difficulty with gestural timing. |
| Compare native speakers of English and Mandarin who have no prior experience with Cantonese, in order to test the hypothesis (H3) that naïve L2 listeners perceptually assimilate vowels with L1 categories on the basis of articulatory similarity rather than acoustic similarity, and therefore produce nonnative vowels by re-using articulatory gestures used for their L1 vowel categories. |
| Compare native speakers of English and Mandarin who are advanced L2 learners of Cantonese, in order to test the hypothesis (H4) that the presence of the high front round vowel [y] in the phonological inventory of Mandarin confers an advantage to Mandarin speakers in perceiving and producing the Cantonese mid front round vowel [œ], due to its similarity in articulatory configuration and timing. |
| Analyze the results of the perception and production experiments in terms of three theoretical models of L2 phonological acquisition: the Speech Learning Model (SLM), the Perceptual Assimilation Model-L2 (PAM-L2), and Direct Mapping from Acoustics to Phonology (DMAP). Disseminate the findings through conference presentations and journal articles. |
| Obtain and disseminate articulatory data on L1 and L2 vowel production for a large sample of Cantonese, Mandarin, and English speakers (20-40 speakers per language); establish baseline L1 articulatory and acoustic measurements for comparison with L2 learners and clinical speaker populations. |
