CANTONESE1 Transcription is IPA. Like all Chinese languages, Cantonese is a tone language. Tones are shown by “tone letters” following the words. The vertical stroke represents the pitch range of the speaker’s normal speaking voice. The horizontal line represents the tone as a relative pitch or pitch change within that range. For example, [˥] = a tone with level pitch at the top of the speaking range, [] = a tone rising from a mid-level pitch to the highest pitch, etc. 1. kʰok ˧ accurate 37. tin ˥ mad 2. t!ʃøn allow 38. si market 3. pʰɪŋ ˩ apple 39. si ˨ matter 4. t!sʰa ˥ bad 40. t!sʰan ˥ meal 5. t!sʊk ˥ bamboo 41. kun ˥ observe 6. t!ʃʰúk ˧ bird 42. t!sʰaŋ orange 7. ʃy ˥ book 43. kɪŋ ˥ pass 8. t!sʰe ˥ car 44. tut ˥ pay for train w. debit card 9. kʰat ˥ card 45. hʌŋ ˩ permanent 10. pin ˧ change, become different 46. lɪk ˥ pick up 11. løt ˩ chestnut 47. t!ʃy ˥ pig 12. t!ʃúŋ ˥ Cheung (proper name) 48. t!ʃʰy ˧ or place 13. t!ʃi ˨ Chinese characters 49. si ˥ poem 14. t!sʰɪŋ ˥ clear 50. kʊŋ ˩ poor 15. so˥ comb 51. sa ˥ sand 16. kʰʌt ˥ cough 52. tyn short 17. jim dye 53. tʰin ˥ sky 18. t!saŋ ˥ elbow 54. pʰin ˧ slice 19. t!ʃøn ˧ enter 55. ʃyt ˥ snow 20. kok ˧ feel 56. ʃyn ˥ sour 21. pɪŋ ˩ flat 57. t!ʃit ˧ stanza of poem 22. t!sʊŋ ˥ flush 58. t!ʃʰøn stupid 23. sʊŋ ˧ food for cooking a meal 59. sam ˥ three 24. t!ʃʰøt ˥ go out 60. tap ˧ take (bus, etc.) 25. t!sa ˥ grab 61. si ˩ time, period of time 26. lʊk ˨ green 62. t!ʃʰi ˧ time, occasion 27. pun ˧ half 63. tʰap ˧ tower 28. kin ˨ healthy 64. fu ˧ trousers 29. sʌm ˥ heart 65. si ˧ try 30. si history 66. sʊk ˥ uncle 1 The starting point of the data was a 120A paper by Marissa Tse. Additional data come from Stephen Matthews and Virginia Yip, Cantonese; A Comprehensive Grammar, Routledge, 1994, Virginia Yip and Stephen Matthews, Basic Cantonese: A Grammar and Workbook, Routledge, 2000, Keith S.T. Tong and Gregory James, Colloquial Cantonese, Routledge, 1994, and Kwan Choi Wah et al., English-Cantonese Dictionary, The Chinese University Press, 199131. húŋ ˥ Hong (proper name) 67. haŋ ˩ walk 32. t!sʰʊk ˥ quick 68. t!ʃúk ˧ wear 33. t!san ˥ real 69. t!sɪk ˥ weave 34. wut ˧ live 70. seŋ ˩ finished 35. t!ʃʰúŋ ˩ long 71. t!sʰo ˧ wrong 36. mɪk ˧ look for 72. júŋ ˨ yield Analyzing the data: 1. Aspiration: Is aspiration vs. non-aspiration of stops and affricates distinctive? One approach to answering this would be to set up a complete distribution table for each of the aspirated and unaspirated consonants. However, a short cut would be to simply scan the data for MINIMAL PAIRS, which would immediately demonstrate CONTRAST. The following minimal pairs are found in this data, showing that (non-)aspiration is contrastive at all points of articulation for voiceless stops and affricates: 10. pin ˧ change 54. pʰin ˧ a slice 37. tin ˥ mad 53. tʰin ˦ sky 25. t!sa ˥ grab 4. t!sʰa ˥ bad 5. t!sʊk ˥ bamboo 32. t!sʰʊk ˥ hurried 12. t!ʃœŋ ˥ Cheung (name) 35. t!ʃʰœŋ long 20. kok ˧ feel 1. kʰok ˧ accurate 2. Vowels: What are the vowel phonemes? Step 1: Scan through the data and make a list of all the vowels that you find. Then arrange them in a standard vowel chart this will help identify sets of vowels which look suspicious in terms of whether or not they contrast. Vowels found in the data: [ø, ɪ, a, ʊ, œ, y, e, o, ʌ, u] Front unround Front round Central unround Back round High (upper) i y u High (lower) ɪ ʊ Mid (upper) e ø ʌ o Mid (lower) œ Low a2 Step 2: Make a distribution table to see whether distributional patterns show COMPLEMENTARY DISTRIBUTION, i.e whether certain vowels can be grouped as allophones of a single phoneme. SUGGESTION: There are three kinds of environments: preceding, following, and tone. Just to get an idea before painstakingly filling out all these 2 Descriptive practice in Cantonese refers to the vowel represented as [a] here as “long”, and it may well be long compared to other vowels. However, Cantonese does not have a vowel length distinction that cannot be predicted either by environment (all vowels are “long” word final) or quality, e.g. “a” and “ʌ” are distinct by height, and perhaps backness--“a” may be more accurately represented [ɑ], in which case it would be back.environments for all the vowels, take a pair of “suspicious” vowels, e.g. [i, ɪ] and quickly scan through the data for preceding environment, following environment, and tone. If one of these looks more promising than the others, concentrate on that one. Preceding [i], we find [p, t!ʃ, j, k, s, pʰ, t!ʃʰ] Preceding [ɪ], we find [pʰ, t!sʰ, m, k, l] This doesn’t look very promising, e.g. both can be preceded by [ph, k]. Tones with [i], we find [. ˧. ˥, , , ˨] Tones with [ɪ], we find [˥, ˧] This also doesn’t look very promising—both occur with [˥, ˧] Distribution table for following environment: [i] [ɪ] [e] 10, 28, 37, 53, 54 13, 30. 38, 39, 49, 61, 62, 65 17 57 ___n ___# ___m ___t 3, 14, 21. 43 36. 46, 69 ___ŋ ___k 8 70 ___# ___ŋ [u] [ʊ] [o] 27, 41 34, 44 64 ___n ___t ___# 5. 26, 66 22, 23, 32, 50 ___k ___ŋ 1, 20 15, 71 ___k ___# [y] [ø] [œ] 7. 47, 48 52, 56 55 ___# ___n ___t 2, 19, 58 11, 24 ___n ___t 6, 68 12, 31, 35, 72 ___k ___ŋ [ʌ] [a] 16 29 45 ___t ___m ___ŋ 4, 25, 51 9 18, 42, 67 33, 40 59 60, 63 ___# ___t ___ŋ ___n ___m ___p WHAT IS THE PHONOLOGICAL ANALYSIS?! (See the last page.) 3. Coronal obstruent phonemes: What are the coronal obstruent phonemes? (coronal sounds = dentals, alveolars, alveopalatals; obstruents = stops, affricates, fricatives) Step 1: List all the coronal obstruents in the data. Coronal obstruents in the data: Word initial [t!ʃ, t"sʰ, t"s, t"ʃʰ, ʃ, s, t, th] Word final [t]Step 2: Make a distributional table for each of the coronals. Since only [t] appears word final, and since its presence or absence is not predictable, we assume that we need to set up a word-final phoneme /t/. Since all the others appear only word-initial, we need only consider following environment. t!s t!sʰ t!ʃ t!ʃʰ 5, 22 18, 25, 33 69 ___ʊ ___a ___ɪ 4,