Does Orthography Affect L2 Tone Production and Perception?

Donghui Zuo, Qiuyue Chen, Peggy Mok Chinese University of Hong Kong Introduction

2 Chinese Orthography

• Chinese characters • Mandarin Pinyin • Major writing system • Official Romanisation for different Chinese system of Mandarin languages • Transparent (good • Opaque (no grapheme-to- consistent grapheme- phoneme to-phoneme correspondence) correspondence) • Tone is indicated by a • No information about number the tone • e.g., ‘horse’ ma3 • e.g., ‘horse’ 馬

3 Cantonese & Mandarin

• Shared orthography – Chinese characters

• Different phonologies, tones for example:

Tones Cantonese Mandarin T1 55 55 T2 25 35 T3 33 214 T4 21 51 T5 23 – T6 22 –

4 C-M tone correspondence

Tsang-Cheung (1988) Cantonese Tone Mandarin Tone %Correspondence T1[55] T1[55] 93% T2[25] T3[214] 89% T3[33] T4[51] 91% T4[21] T2[35] 93% T5[23] T3[214] 76% T6[22] T4[51] 94%

5 Production Model (Chu, 2011)

learners Advanced Advanced Research Question

• Whether activations of the L1 phonological system • Facilitates/hinders Mandarin tone production? • Facilitates/hinders Mandarin tone perception?

7 Method

8 Subjects

• 16 native speakers of Hong Kong Cantonese • 7 males, 9 females • 19-29 years old (mean = 23) • Students from the Chinese University of Hong Kong • Beginning learners of Mandarin (mean length of formal training = 2.6 years)

9 Materials - Production

• Pinyin task • mi [mi] and na [na] with all four tones • 2 syllables × 4 tones = 8 tokens • Written in Mandarin pinyin

• Chinese character task • ja [ja] and wu [wu] with all four tones (e.g., 壓 and 烏) • 2 syllables × 4 tones = 8 tokens • Written in Chinese characters

10 Materials - Perception

• Pinyin task • Disyllabic minimal pairs (e.g. can1 guan1 參觀 ‘visit’ vs. can1 guan3 餐館 ‘restaurant’) • 1 sets × 6 pairs = 6 minimal pairs • Written in Mandarin pinyin • Read by a Mandarin native speaker • Chinese character task • Disyllabic minimal pairs (e.g.,字跡 zi4 ji4 ‘handwriting’ vs. 自己zi4 ji3 ‘oneself’) • 2 sets × 6 pairs = 12 minimal pairs • Written in Chinese characters • Read by a Mandarin native speaker

11 Procedures - Production

• Recording • All stimuli produced in isolation • 256 tokens collected • 128 Pinyin tokens (2 syllables × 4 tones × 16 speakers) • 128 Chinese character tokens (2 syllables × 4 tones × 16 speakers)

12 Procedures - Production

• Transcription • Transcriber: 3 native speakers of Mandarin • Agreed transcription of at least two transcribers accepted as the actual tone produced No. of tokens Percentage 3 transcribers agree 173 67.6% 2 transcribers agree 80 31.3% No agreement 3 1%

13 Procedures - Perception

• Pinyin task • Write down the tones of the second syllables • Both words of each pair presented once • 192 responses (6 pairs × 2 words × 16 subjects)

• Chinese character task • Two-alternative forced choice task • Both words of each pair presented once • 384 responses (12 pairs × 2 words × 16

subjects) 14 Results

15 Production

• Pinyin task • %Err = 2.3% (only 3 errors out of 128 tokens)

• Character task • %Err = 43.4%

• t(15) = -7.678, p < 0.001

16 Production (cont.)

• Confusing T1 with T4, T2 with T3

45 40 35

25 pinyin %Err 20 character 15 10 5 0 T1-T2 T1-T3 T1-T4 T2-T3 T2-T4 T3-T4 Confusable pairs

17 Perception

• Pinyin task • %Err = 47.9%

• Chinese character task • %Err = 11.7%

• t(15) = 5.495, p < 0.001

18 Perception (cont.)

• Confusing T1 with T4, T2 with T3

45 40 35

25 pinyin %Err 20 character 15 10 5 0 T1-T2 T1-T3 T1-T4 T2-T3 T2-T4 T3-T4 Confusable pairs

19 Discussion

20 Error Patterns

• T2-T3 pair • Also found in learners of other language background • Possible accounts: • Acoustic similarity (Kiriloff, 1969) • Perceptual assimilation (Best, 1995) • T3 Sandhi

21 Error Patterns (cont.)

• T1-T4 pair • Not found in learners with some other backgrounds (cf. Kiriloff, 1969) • Possible account • Cantonese T1[55] has a falling allotone [53] (Mathews & Yip, 1994)

22 Pinyin vs. Character

• Production • %Err: Pinyin < character • Chinese character hinders L2 tone production

• Perception • %Err: Pinyin > character • Chinese character facilitates L2 tone perception

23 Production Model (Chu, 2011)

24 Production

• Pinyin • Directly access L2 phonological representations • May not reach the concept level

• Chinese character • Sub-lexical route through the L1 phonological system • Corresponding rules may not have been correctly established

25 Perception Model (Chu, 2011)

26 Perception

• Pinyin • Only need to reach L2 sub-lexical representation level

• Chinese character • Need to go through L1 phonological system

• However, pinyin %Err > character %Err • Possible account: top-down process in the Chinese character task

27 Conclusion

28 Conclusion

• Comparing with Pinyin, Chinese character • Hinders tone production • Facilitates tone perception

• Therefore, orthography plays an important role in L2 speech production and perception, especially when the learners’s L1 and L2 share the same orthography system.

29 References

1. Young-Scholten, M. and Archibald, J., “Second language syllable structure”, in J. Archibald (ed.), Second Language Acquisition and Linguistic Theory, 64-97, 2000. 2. Silveria, R., “Investigating the role of orthography in the acquisition of L2 pronunciation: a case study”, in M. A. Watkins, et al. (Eds). Recent Research in Second Language Phonetics/Phonology, 270-290, 2009. 3. Leung, A., Tonal assimilation patterns of Cantonese L2 speakers of Mandarin in the perception and production of Mandarin tones. Proceedings of the 2008 CLA Annual Conference, 2008. 4. Chu, P. C. K. Towards a Model of Second Language Word Production and Recognition in Mandarin. Young Scholar Award Competition for the International Conference on Chinese Language Learning and Teaching in the Digital Age, Hong Kong, China, 2011. 5. Tsang-Cheung, L.Y. The Phonological Correspondences between Cognate Morphemes in Cantonese and Mandarin. Unpublished M. Phil. Thesis, University of Hong Kong, 1988. 6. Kiriloff, C., “On the auditory discrimination of tones in Mandarin”, Phonetica 20:63-67, 1969. 7. Best, C. T., A direct realist view of cross-language speech perception, In W. Strange (Ed.), Speech perception and linguistic experience: Issues in cross- language research, 171-204, 1995. 8. Matthews, S., and Yip, V., Cantonese: a comprehensive grammar, 1994.

30 Limitations

• Different task settings: • Pinyin perception (transcription) vs. Chinese character perception (2AFC) • Production (monosyllabic) vs. perception (disyllabic) • Stimuli not well controlled • Word frequency • Segmental features • Character strokes • Cantonese-Mandarin correspondence 31