Beijing Dialect And Cantonese Speakers' Perception Of English Lexical Stress

English is a stress language, its lexical stress can be realized by a combination of pitch, duration, intensity and vowel quality, while Mandarin and Mandarin dialects are tonal languages. The prosodic differences between English and Mandarin Chinese dialects may bring extreme difficulties for Chinese learners. Mandarin dialects also differ in whether or not they have lexical stress. Beijing dialect has a stress pattern that contrasts full-full (e.g.,donglxil) and full-reduced (e.g.,donglxi) disyllabic word. The reduced vowel in the latter carries a neutral tone and it is shorter than the full vowel. These two stress patterns are thus signaled with duration and FO. By contrast, Cantonese does not have this stress distinction:Given the absence of the full-reduced stress pattern, FO may play a more important role than duration for lexical access in Cantonese. This thesis mainly explores the following research questions:1) What is the weight of acoustic cues (FO, duration, intensity and vowel reduction in lexical stress perception by native English speakers, Beijing dialect speakers and Cantonese speakers?2) In perceiving English lexical stress, are there any differences between Beijing dialect group and Cantonese group in terms of FO, duration, intensity and vowel reduction?3) Is there any initial stress bias or final stress bias for the three groups in terms of FO, duration, intensity and vowel reduction?The experiment involved three groups of participants:one group of native English speakers (E), one group of Beijing dialect speakers (B) and one group of Cantonese speakers(C).Each group is of 15 participants. They were asked to make stress judgments on 48 sequences of disyllabic re-synthesized nonsense words (four stimuli in a sequence). The 48 sequences were composed of words with stress suggested by "FO cues" only, "duration cues" only, "intensity cues" only or "vowel reduction cues" only. The stress judgment can be either stressed on the first syllable (Initial Stress) or on the second syllable (Final Stress).Mixed-effects models were conducted on the participants' sequence-encoding and word-encoding accuracies. The sequence-encoding accuracy was computed to examine the overall effect of prosodic cues on the processing of English stress; the word-encoding accuracy was computed to compare the processing of English stress for words with initial stress vs. words with final stress.Based on the data and analysis, I have obtained the following findings:1) The weight of acoustic cues in lexical stress perception by native English speakers is:Vowel reduction> Duration/F0>Intensity; for Beijing dialect speakers, the ranking is:Vowel reduction/F0> Duration>Intensity; as to Cantonese speakers, it is:F0/Vowel reduction>Duration/Intensity.2) Beijing dialect speakers performed significantly better than Cantonese speakers for words signaled by duration cues.3) Significant initial stress bias was found for English speakers under "F0 cues" condition, "duration cues" condition and "vowel reduction cues "condition, for Beijing dialect speakers under "FO cues" condition and "vowel reduction cues" condition, and for Cantonese speakers under "F0 cues" condition.The results from this study suggested that Chinese learners of English tended to apply their native phonological strategies and were more sensitive to acoustic cues that are activated in their native language in perceiving non-native English lexical stress.