Perceptual evaluation of Mandarin tone sandhi production by Cantonese speakers before and after perceptual training

Perceptual Evaluation of Mandarin Tone Sandhi Production by Canton- ese Speakers before and after Perceptual Training Bei Li Department of Chinese and Bilingual Studies, The Hong Kong Polytechnic University Hung Hom, Hong Kong beili@polyu.edu.hk Yike Yang Department of Chinese and Bilingual Studies, The Hong Kong Polytechnic University Hung Hom, Hong Kong yi-ke.yang@connect.polyu.hk Si Chen Department of Chinese and Bilingual Studies, The Hong Kong Polytechnic University Hung Hom, Hong Kong sarah.chen@polyu.edu.hk Abstract This study is the first to examine the effect of perceptual training on Mandarin tone sandhi production by Cantonese speakers. Auditory and visual inputs of tone sandhi contrasts were included in a short-term laboratory training, and the training was followed by an identifica- tion test. Twenty-four native speakers of Can- tonese participated in the study, which comprised the training session and a pre- and post-training recording session. There were 192 target stimuli of real words and wug words and 192 filler words in each recording session. Two native Mandarin-speaking linguists per- ceptually evaluated a total of 23040 syllables on a 101-point scale. The results show that Cantonese speakers may be able to improve their lexical word production in the context of T3+T1/T2/T4 by perceptual training or high fa- miliarity of stimuli, whereas the application of Mandarin sandhi Tone 3 slightly improves within a short time. Besides, the participants consistently apply Mandarin half-third tone sandhi rule with a lexical mechanism rather than a computational mechanism, whereas the preference to lexical mechanism is not found in the application of Mandarin third tone sandhi rule. 1 Introduction It has been shown that Cantonese speakers acousti- cally differed from Beijing Mandarin speakers in their application of two Mandarin tone sandhi rules (Chen et al., 2017). Specifically, Cantonese speak- ers preferred to neutralize the sandhi Tone 3 more with Tone 2 in isolation which indicated a more cat- egorical application of the third tone sandhi rule. Also, they tended to raise f0 value at the end of san- dhi Tone 3 which should be modified by half Tone 3 without the raising part. Mandarin and Cantonese are both tonal lan- guages, which discriminate lexical meanings by dif- ferent tones. Mandarin has four lexical tones (Chao, 1948): Tone1 (T1) high-level (55) “ma” (mother) Tone2 (T2) high-rising (35) “ma” (hemp) Tone3 (T3) low-dipping (213) “ma” (horse) Tone4 (T4) high-falling (51) “ma” (to scold) There are two rules of tone sandhi (Zhang & Lai, 2010) in Mandarin: • Third tone sandhi rule: T3 (213) → T2 (35)/__ T3 (213) lau 213 - pan 213 → lau 35 - pan 213 “boss” PACLIC 32 358 32nd Pacific Asia Conference on Language, Information and Computation Hong Kong, 1-3 December 2018 Copyright 2018 by the authors • Half-third sandhi rule: T3 (213) → 21/__ {T1(55), T2(35), T4(51)} lau 213 - ʂi55 → lau 21 - ʂi55 “teacher” lau 213 - liəŋ 35→ lau 21 - liəŋ 35 “old age” lau 213 - xua 51→ lau 21 - xua 51 “old saying” Six contrastive tones are involved in Cantonese (Bauer & …

Content truncated to 500 words