| Summary: | 碩士 === 國立清華大學 === 語言學研究所 === 96 === The aim of this thesis is to investigate the influence of the native phonological system on the production and perception of non-native speech. In order to address this issue, three experiments were conducted to examine how Taiwan Mandarin (TM) speakers produce and perceive English word-initial consonant clusters. The first experiment is an acoustic study of the production by TM speakers of English consonant clusters. Results show that the majority of the production corpus has been correctly produced. However, errors have been found: TM speakers employ multiple strategies and the most frequent strategies are vowel epenthesis between two consonants (CCVCCV) and liquid replacement (/bl//br/). Moreover, speakers are more accurate on some clusters than others. One interesting finding is that the liquid duration ratio of TM speakers is larger than that of the native speaker and the F3 value of replacement (/l//r/) is similar to that of /r/ (i.e. the characteristic feature of /r/ is low F3). The larger ratio of liquids implies that the TM speakers needed more time to coarticulate the articulatory gestures between segments. The lower F3 value may be due to the coarticulation of the previous stop consonants (coarticulation effects, i.e. labialization in /b/- and /p/-clusters and velarization in /g/- and /k/-clusters). The second and third experiments explore the perception of TM speakers on English CCV clusters and their counterpart disyllabic C«CV stimuli. Results of perception experiments demonstrate that participants could perceptually distinguish CCV and C«CV stimuli. Occasionally, participants hear illusory vowels and misperceive /l/ for /r/. The results are not consistent with the prediction that native phonotactics plays an important role in non-native perception. Thus, it is suggested that, in addition to phonotactics, perceptual cues play substantial roles in the perception of non-native syllable structures. Results of the present study support the notion that there exists a close relationship between production and perception despite the individual variations. Speakers with more production errors tend to misperceive the stimuli more frequently and they tend to employ the same strategies in both production and perception. Moreover, the fact that participants can produce and perceive English consonant clusters indicates that native phonotactics does not play a very important role in the non-native speech. The results of these studies not only contribute to the understanding of the interactions of native and non-native languages on the levels of phonemes and phonotactics, explore the relationship between speech production and perception, but also provide implications for the data analyses and formulation of hypotheses for future studies.
|
|---|
