A LONGITUDINAL EXAMINATION OF ENGLISH VOWEL LEARNING BY MANDARIN SPEAKERS

摘要

Current models of second language (L2) speech learning were developed lo account for various types of difficulties that L2 learners face in the acquisition of specific vowels and consonants. Flege's (1995) speech learning model (vSLM), for example, proposes that accurate production of L2 phones ultimately depends on how well learners are able to distinguish them perceptually from similar LI sounds. According to Major's (1996) Similarity Rate Differential Hypothesis (SDRH), phones dissimilar to LI sounds are acquired more rapidly than those that are similar. However, u detailed evaluation of the relative difficulties presented by L2 sounds and of their rate of acquisition requires a wide range of data, including data collected at different times duirng L2 learning. Despite the large number of studies of L2 perception and production conducted in recent years, relatively few have examined the acquisition of categories by low-proficiency speakers from a longitudinal perspective. Most studies with a longitudinal component have considered only a limited range of speech sounds or have considered learners over a short period of time. This study addresses the gap in current research by evaluating a broad range of English vowel productions of recently-arrived Mandarin speakers during their first year in Canada.