Back to list

Detail of contribution

Auteur: Yizhou LAN

Co-Auteur(s): OH Sunyoung

A probe into the nature of L2 interlanguage through articulatory gestures: examples from alveolar clusters by Cantonese speakers of English

Abstract/Résumé:   Cantonese speakers of English are reported to substitute initial tr/dr clusters to [tʃ] category. According to a pilot study, tr/dr are produced with significant affrication by both native Cantonese and English speakers, and the F3 values of [r] transitions are also significantly different from other two places of articulation. Theoretical models of L2 sound acquisition may have different approach to learning of alveolar clusters. Speech Learning Model (SLM) suggests that L1 and L2 categories are merged as L2 learning proceed, which underlies that L2 productions are actually impoverished versions of L1 and/or L2. However, Major proposed Ontogeny Phylogeny Model (OPM) which stipulates that "interlanguage (IL) is composed of L1, L2 and linguistic universals (U)". Present study investigates whether the unlikeness of IL to L1 and L2 can be contributed to merger of L1 and L2, or it allows factors in U.   To find out the nature of IL's dissimilarity to L1 and L2, two experiments were done. Firstly, five native American English speakers and five native Cantonese college students participated a perception test. Synthesized minimal or near-minimal sets of [tr]/[tʃ] initial cluster and single plosive-vowel consonants (train, chain tain; trip, chip, tip) were repeated as stimuli. Identification and differentiation rates imply that Cantonese speakers can correctly identify and discriminate the [tr]/[tʃ] contrast almost as English speakers do. This suggests that another cue other than F3 was utilized by Cantonese speakers to discriminate the contrast.   Secondly, a production test with same participants and stimuli was done. Results show that affrication of Cantonese speakers' [tr] is realized as an affricate with lip protrusion ([tʃw]) with almost no F3 dip in [r] transition. It is different from native English speakers' realization (L2), which involves TT constriction followed by a clear TT curling gesture which resulted in a dip of F3; it is also different from Cantonese plain alveolar affricate (L1) where no lip gesture is found. So IL is different from L1, L2, and their middle point. Results further suggest that although English [r] have both [lip] and [TT] gestures, only the [lip] gesture, not [TT], was captured and utilized by Cantonese speakers, hence a higher F3. The findings partly support Major's hypothesis that IL phonology is not a simple merging of L1 and L2 imperfections, but a consistent grammar with features from U. The revealing of gestural cues reliable for speakers' production may contribute to the concrete content of U.