Consonants

The consonant inventory of Changsha Chinese comprises 19 phonemes. The list above presents corresponding key words/bound morphemes of the consonants. The plosives and affricates show a general two-way distinction between voiceless unaspirated and voiceless aspirated.

As shown in Figure 2, the voice onset time (VOT) of unaspirated plosives, that is, /p t k/, is shorter with the release and voicing onset occurring within approximately 20 ms. On the other hand, the VOT of aspirated plosives, that is, /p^h t^h k^h/, is about 90 ms. The VOT of aspirated or unaspirated affricates is significantly longer than the VOT of plosives due to their frication phase. For instance, the mean VOT of /tɕ/ and /tɕ^h/ (both based on eight tokens) is approximately 66 ms and 160 ms, respectively; while the VOT of /t/ and /t^h/ (both based on eight tokens) is approximately 14 ms and 92 ms, respectively. Moreover, the VOT of affricates have a larger standard deviation of values than the VOT of plosives, whether aspirated or not.

Figure 2

VOT of plosives and affricates in Changsha Chinese, the unaspirated and aspirated are indicated by distinct colors. The mean and standard deviation were calculated using eight tokens of each plosive and affricate from a single speaker.

There are four fricatives in Changsha Chinese, namely, /f/ (as in /fa³⁴/ ‘flower’), /s/ (as in /sa³⁴/ ‘sand’), /ɕ/ (as in /ɕø³⁴/ ‘book’), and /x/ (as in /xo⁴³/ ‘fire’). Figure 3 shows the spectra of the frication (made with a 23 ms window centred at the peak of noise intensity) in four fricative samples. There is a correlation between the frequency region with the greatest energy and the place of articulation. The farther back the fricative’s articulation, the lower the frequency region with the greatest energy tends to become. However, there is also variation between velar frication and glottal aspiration, both for the fricative /x/ and for aspirated stops. For instance, in the passage ‘The North Wind and the Sun’, /x/ in /pə²²³ xən³⁴/ ‘north wind’ (北风) was pronounced as [pə²²³ hən³⁴] and /t^ho²⁴ uai³¹ i³⁴/ ‘undress’ (脱外衣) was realized as [txo³⁴¹ uai³¹ i³⁴], as shown by Figure 4.

Figure 3

FFT (Fast Fourier Transform, blue) and LPC (Linear Predictive Coding, red) spectra (made with a 23 ms window centred on the peak of noise intensity) of the frication in /fa³⁴/ ‘flower’, /sa³⁴/ ‘sand’, /ɕø³⁴/ ‘book’ and /xo⁴³/ ‘fire’.

Figure 4

Waveforms and spectrograms of [txo³⁴¹ uai³¹ i³⁴] ‘undress’ (脱外衣) and [pə²²³ hən³⁴] ‘north wind’ (北风). FFT (Fast Fourier Transform, blue) and LPC (Linear Predictive Coding, red) spectra were made with a 23 ms window centred on the midpoint of the fricative duration in [txo³⁴¹ uai³¹ i³⁴] ‘undress’ (脱外衣) and [pə²²³ hən³⁴] ‘north wind’ (北风), respectively.

Some conventions regarding Changsha Chinese consonants should be clarified. /tɕ tɕ^h ɕ/ are in complementary distribution with /k kʰ x/ and /ts ts^h s/. While /tɕ tɕ^h ɕ/ are followed by only /i/, /jV/ and /ø/, as in, for example, /tɕja⁴³/ ‘fake’, /ɕjo²²³/ ‘to peel’ and /ɕø³⁴/ ‘book’, /k kʰ x/ and /ts ts^h s/ are followed by other vowels besides /i/, /jV/ and /ø/, as in /ko⁴³/ ‘fruit’, /k^ho⁴³/ ‘approve’, /xo⁴³/ ‘fire’ and /tsa²²³/ ‘tea’, /ts^ha²⁴/ ‘thrust’, /sa³⁴/ ‘sand’. To a certain extent, then, it is reasonable to treat /tɕ tɕ^h ɕ/ as allophones of /k kʰ x/ or /ts ts^h s/. However, /tɕ tɕ^h ɕ/ are treated as distinct phonemes in keeping with the Chinese tradition in this Illustration (Bao et al. Reference Bao, Cui, Shen and Wu1998, Lee & Zee Reference Lee and Zee2003, Zeng Reference Ting2019), same as /tɕ tɕ^h ɕ/ in Mandarin (Lee & Zee Reference Lee and Zee2003).

Changsha Chinese has three nasals: /m n ŋ/, /m/ (as in /ma³⁴/ ‘mom’) and /ŋ/ (as in /ŋo⁴³/ ‘me’) occur only in syllable-initial position, while /n/ (as in /nø⁴³/ ‘women’ and /pin³⁴/ ‘soldier’) can occur both in syllable-initial and syllable-final position. It should be noted that, when occurring in syllable-initial position, /n/ is followed by only /i/ and /ø/, while /ŋ/ is followed by other vowels besides /i/ and /ø/. Thus, the contrast between /n/ and /ŋ/ is neutralized in syllable-initial position. Moreover, it is important to mention that when followed by /i/ or /ø/, /n/ is realized as an alveolo-palatal, which could be transcribed as [ȵ] (a symbol which presently does not occur in the IPA).

/ɹ/ occurs in isolation in both syllable-initial and syllable-final positions. However, there are some distributional restrictions. /ɹ/ can only occur in syllable-initial position if preceding /a ə/ and in syllable-final position following /ts tsʰ s/. /ɹ/ in isolation (as in /ɹ²⁴/ ‘sun’) or in syllable-initial position (as in /ɹə⁴³/ ‘provoke’) is an alveolar approximant. When occurring in syllable-final position after /ts tsʰ s/ (as in /tsɹ⁴³/ ‘purple’, /ts^hɹ³⁴/ ‘defect’ and /sɹ³⁴/ ‘silk’), /ɹ/ is realized as a syllabic apical (or laminal) denti-alveolar approximant, which is the same case with Putonghua (or Standard Chinese, Lee & Zee Reference Lee and Zee2003), and could be transcribed as [ɿ] – another symbol widely applied in the Chinese phonetic tradition which normally appears after /ts tsʰ s/ and does not have an IPA counterpart.

Figure 5

Waveforms and spectrograms for /ɹ/ in isolation, syllable-initial, and syllable-final positions. The red dot on the spectrogram marks the formants.

Figure 6

F1/F2 and F2/F3 of /ɹ/ occurring in syllable-initial and syllable-final positions. The mean and standard deviation of each formant were calculated from seven tokens. /ɹ/ in isolation was not included, because there is only a single token in Changsha Chinese, i.e. /ɹ²⁴/ ‘sun’.

Figure 5 shows the waveforms and spectrograms of /ɹ/ occurring in isolation, syllable-initial, and syllable-final positions. The formants of each syllable are marked by red dots on the spectrograms. Figure 6 reveals acoustic plots of /ɹ/ occurring in syllable-initial and syllable-final positions, and each category was calculated from seven tokens. F2 of /ɹ/ occurring in syllable-final position (as in /ts^hɹ³⁴/ ‘defect’) is lower than /ɹ/ occurring syllable-initial position (as in /ɹə⁴³/ ‘provoke’), suggesting a more retracted tongue position. F3 of /ɹ/ occurring in syllable-final position is higher than for /ɹ/ occurring in syllable-initial position, suggesting less lip rounding syllable-finally. /ɹ/ in isolation was not included in Figure 6, because there is only a single token in Changsha Chinese, i.e. /ɹ²⁴/ ‘sun’.

Vowels

Monophthongs

There are six monophthongs in the basic inventory of Changsha Chinese. /i ø ə a/ occur in open syllables, including syllables closed by a nasal like /pi²²³/ ‘skin’ and /pin³⁴/ ‘soldier’, while /u o/ occurs only in open syllables as in /ko⁴³/ ‘fruit’ and /k^hu⁴³/ ‘bitter’. There are two nasal vowels in Changsha Chinese: /õ ə̃/, which can only occur in open syllables, as in /sə̃²⁴/ ‘kind’ and /k^hõ⁴³/ ‘a sum of money’.

Figure 7

Acoustic plots for the vowels of Changsha Chinese, F1, F2 and F3 of each vowel were based on mean formant values of eight open syllables.

Figure 7 shows the formant plots for the vowels of Changsha Chinese. F1, F2 and F3 of each vowel were based on the mean formant values of eight open syllables. There might exist some controversy about the symbol used for /ø/, since /ʏ/ is also a rational choice based on auditory quality and the formants shown in Figure 7. In fact, /ø/ functions phonologically as a close-mid front vowel, in terms of the behavior of the nasal onsets and the triggers for onset velars becoming alveolo-palatals (as in /tɕø³⁴/ ‘pig’ and /tɕ^høn³⁴/ ‘spring’). /a/ is positioned to the lower right because F2 of /a/ is lower, so it is also reasonable to transcribe /a/ as /ɑ/. However, /a/ is adopted here typographically for convenience and as a symbol for two allophones: when occurring in open syllables, /a/ is realized as [ɑ]; when /a/ is not preceded by /i/ and occurs in syllables closed by a nasal, /a/ is realized as [ɐ] (as in /pan³⁴/ ‘squad or class’ or /kuan³⁴/ ‘close’).

Additionally, /u/ is adopted here typographically for convenience since it is a symbol for two allophones. When occurring in non-open syllables, /u/ is realized as [u]; when occurring in open syllables, /u/ is realized as [ʊ]. Therefore, the position of /u/ in Figure 7(A) is much more fronted than expected as the eight tokens for /u/ are all open syllables. This further explains why /u/ and /o/ are positioned relatively lower than /i/ and /ø/ in Figure 7(A); when /u/ is realized as [ʊ] with a lower F1 compared to the close front counterpart /i/, /o/ is also affected compared to its close-mid front counterpart /ø/.

Diphthongs

As shown in the above vowel diagrams (rising and falling diphthongs are presented separately to provide a better view), Changsha Chinese has 11 diphthongs, illustrated in the left-hand-side list above. Note that /ja jo je jen/ are only preceded by /tɕ tɕ^h ɕ/. The [i] quality arises from coarticulation with the alveolopalatal onset. Among the 11 diphthongs, /ja jo ai au ei əu/ occur only in open syllables, and /je ua uə øa øe/ occur in open syllables with syllables closed by /n/ (as in /tɕ^hjen³⁴/ ‘gun’, /kuan³⁴/ ‘close’, /kuən⁴⁵/ ‘stick’, /tɕ^høan²²³/ ‘bed’, /øen³³/ ‘injustice’).

It is important to note that /ə/ occurring as an allophonically nasalized vowel before a nasal consonant (as /kuən⁴⁵/ ‘stick’) is different from /ə/ occurring as a single vowel (as /pə²⁴/ ‘north’) and as a nasal vowel (as /sə̃²⁴/ ‘kind’). In fact, phonemic nasal vowels do not occur before nasal consonants, but oral vowels are contextually nasalized before nasal consonants. As shown by Figure 8, F1 of /ə/ in /ən/ is significantly higher than F1 of /ə/ and /ə̃/, while F2 of /ə/ in /ən/ is significantly lower than F2 of /ə/ and /ə̃/. Significant difference between the formants of /ə/ and /ə̃/ is not present.

Figure 8

F1 and F2 of /ə/ (as single vowel, nasal vowel, and allophonically nasalized vowel before an underlying nasal consonant). The mean and standard deviation of each formant were calculated from seven tokens.

The vowel /e/ appears only in specific diphthongs (i.e. /je øe/) and triphthongs (/uei/, as in /kuei⁴³/ ‘ghost’). Its origin is unclear but may be due to the influence of Putonghua since most native speakers speak both Putonghua and Changsha Chinese in their daily life. In fact, /e/ is in a complementary distribution with /ə/. /e/ occurs only in /je øe uei/, while /ə/ occurs in isolation or diphthongs and triphthongs except /je øe uei/.

Triphthongs

As shown in the above triphthong diagram, Changsha Chinese has four triphthongs: /iau/ (as in /piau⁴³/ ‘surface’), /iəu/ (as in /niəu²²³/ ‘cattle’), /uai/ (as in /kuai⁴³/ ‘crutch’), and /uei/ (as in /kuei⁴³/ ‘ghost’). All triphthongs occur only in the open syllables.

Tones

This study confirms previous studies (Zhong Reference Zhong2003, Li & Liu Reference Bing and Yanni2006, Xiang & Shi 2007) that there are six citation tones in the inventory of Changsha Chinese. Though the specific Chao numerals for each one might be unique, essential differences do not exist between this transcription and the previous studies. Tone1, conventionally named Yingping, is a mid rising tone that has a stable pitch in the middle of the pitch range and is transcribed as 34. Tone2, conventionally named Yangping, is a low-rising tone with a longer duration and is transcribed as 223. Tone3, conventionally named Shangsheng, is a mid-falling tone transcribed as 43. Tone4, conventionally named Yingqu, is a high-rising tone transcribed as 45. Tone5, conventionally named Yangqu, is a medium-falling tone transcribed as 31, which reaches the lowest point of the pitch range. Tone6 is a rising tone and is transcribed as 24. Note that Tone6 is conventionally named Rusheng (also known as ‘entering tone’ or ‘checked tone’), which is commonly represented by one digit in Chinese literature, due to its short duration which is normally about 200 ms; while in the case of Changsha Chinese, Rusheng is represented by two digits since its average duration is about 430 ms and not significantly different from other tones, as shown in Figure 9.

Figure 9

The pitch trajectory of tones in Changsha Chinese in terms of absolute duration (based on mean pitch values of eight randomly selected words).

Figure 10

Creaky voice of rising tones in Changsha Chinese (/ma²²³/ ‘hemp’ and /ma²⁴/ ‘to wipe’ are taken as examples here). Creaky voice causes a dip in amplitude (as shown by yellow line) and f0 could be discontinuous (as shown by blue line).

The pitch trajectory of each tone in terms of absolute duration is shown in Figure 9. The f0 values of each tone were extracted using the Praat software. It is noteworthy that the pitch trajectory of tones in Figure 9 might contradict the transcription in terms of Chao numerals. For instance, in Figure 9, the pitch trajectory of Tone4 is rising-falling, so in addition, Tone3 could be transcribed as 454 using Chao numerals. However, since the falling part of Tone3 is ignored by listeners, it is preferable to transcribe Tone3 as 45.

Creaky voice tends to be associated with the two rising tones, that is, Tone2 and Tone6, in Changsha Chinese. A period of creaky voice appears to take place in the middle of the vowel, in the form of voicing irregularities and a dip in amplitude. As shown in Figure 10, the f0 of /ma²²³/ ‘hemp’ is discontinuous when the amplitude contour has a dip. However creaky voice does not always occur with rising tones in Changsha Chinese, since some words have not shown any kind of dip in amplitude or voicing irregularity, more work is needed to determine the factors that condition its occurrence.

There tends to be creaky voice associated with the low-pitch targets of tones 31 and 43, As shown in Figure 11, the end of f0 contour of /ma³¹/ ‘curse or condemn’ and /ko⁴³/ ‘fruit’ are missing because f0 algorithm has failed as the falling f0 contour reaches a certain threshold. The waveforms and spectrograms showed certain voicing irregularities, and the pitch periods in the pink region were significantly larger than that in other parts of the syllable.

Figure 11

Creaky voice of falling tones in Changsha Chinese (/ma³¹/ ‘curse or condemn’ and /ko⁴³/ ‘fruit’ are taken as examples here).

Sandhi patterns for bisyllabic words of Changsha Chinese are complex. For more information, please refer to Zhong (Reference Zhong2003) and Li & Liu (Reference Bing and Yanni2006).

Syllable structure

The syllable structure of Changsha Chinese is (C)(C)V(C), so consonant clusters are allowed only in syllable-initial but not syllable-final position. The nucleus (vowel, or V) must have at least one vowel and cannot exceed three vowels since Changsha Chinese has only monophthongs, diphthongs, and triphthongs. The other elements in the syllable are optional. All Cs (consonants) occur in syllable-initial position but the only permissible syllable-final consonant is /n/ (as in /pin34/ ‘soldier’). Moreover, it is important to mention that all words in Changsha Chinese are monosyllabic. Here are examples of all possible syllable structures in Changsha Chinese:

Further phonotactic restrictions may be noted in Changsha Chinese. First, while /tɕ tɕ^h ɕ/ are followed by only the high front vowels/glides (/i ø j/), /k kʰ x/ and /ts ts^h s/ are followed by other vowels besides the high front vowels/glides (/i ø j/). Second, while the alveolar sibilants /ts ts^h s/ are not allowed before the high front vowels/glides (/i ø j/), /ɹ/ occurs precisely after these consonants or occurs in syllable-initial position by preceding /a ə/. Third, among all three nasals, /m ŋ/ occur only in syllable-initial position, while /n/ can occur both in syllable-initial and syllable-final position. When occurring in syllable-initial position, /n/ is followed by only the high front vowels/glides (/i ø j/), while /ŋ/ is followed by other vowels besides the high front vowels/glides (/i ø j/). Fourth, among all six monophthongs, /i ø ə a/ occur in open syllables, including syllables closed by a nasal, while /u o/ occur only in open syllables. The two nasal vowels /õ ə̃/ can only occur in open syllables. Fifth, among the 11 diphthongs, /ja jo ai au ei əu/ occur only in open syllables, while /je ua uə øa øe/ occur in open syllables and syllables closed by /n/. Sixth, the four triphthongs (/iau iəu uai uei/) occur only in open syllables. Finally, the new vowel /e/, which might come from borrowed words, appears only in /je øe uei/ and /e/ is in a complementary distribution with /ə/. /e/ occurs only in /je øe uei/, while /ə/ occurs in isolation or diphthongs and triphthongs except /je øe uei/. However, /e/ is not taken as an allophone of the phoneme /ə/ because its auditory quality sounds quite different, even to native speakers.

Transcription of the recorded passage

The passage is the Changsha Chinese version of ‘The North Wind and the Sun’ story. It is transcribed phonemically. The tones are marked for each syllable and transcribed in terms of Chao numerals. The symbol ‘ǀ’ marks the end of an intonational phrase within an utterance, whereas ‘ǁ’ marks the end of an utterance.

pə²²³ xən³⁴ | y³¹ t^hai⁴⁵ ien⁵³ ts^hau⁴³ ǁ k^han⁴⁵ na⁴³ i⁴³ kə³¹ ti³¹ pən³⁴ sɹ³¹ tai³¹ ǁ tɕ^hja²⁴ xau⁴³ | iəu⁴³ ləu³¹ ɹən²²³ | tɕ^hyn³⁴ kə⁴³ uai³¹ t^hau⁴³ | tsəu⁴³ ko²⁴ li⁴³ ko⁴³ sən²²³ǁ t^ha³⁴ mən⁴³ lien³⁴ kə⁴³ san³⁴ lən⁴³ ǁ k^han⁴⁵ na⁴³ i⁴³ kə³¹| na⁴³ i⁴³ kə³¹ɹən⁴³ ɹan³⁴ ləu³¹ ɹən²²³ t^ho²⁴ uai³¹ i³⁴ ǁ t^ha³⁴ tɕjəu⁴³ sɹ³¹ in²⁴ tɕja³⁴ ǁ y²⁴ sɹ⁴³ lə³⁴ | pə²²³ xən³⁴ tso²⁴ sɹ³¹ ti⁴³ kua²⁴ tɕ^hj⁴⁵ lai⁴³ ǁ uie²⁴ kua²⁴ uie²⁴ li⁴⁵ xai⁴³ ǁ ko³⁴ ləu⁴³ ɹən²²³ | pa³⁴ i²⁴ fu⁴³ pau⁴³ tə³⁴ pən⁴⁵ tɕjn⁴³ ti⁴³ ǁ tsei⁴⁵ xau⁴³ ǁ pə²²³ xən³⁴ tsɹ²²³ xau⁴³ fan³⁴ tɕ^hi43 ǁ ko²⁴ xa⁴³ tsɹ³¹ | t^hai⁴⁵ ien⁵³ fan³⁴ sɹ⁴³ ti⁴³ tsau⁴⁵ tɕ^hi³⁴ lai³¹ ǁ ko²⁴ xa⁴³ tsɹ³¹| ko³⁴ ləu⁴³ ɹən²²³ ɹə²⁴ ti⁴³ pa⁴³ i²⁴ fu⁴³ | tsɹ²²³ xau⁴³ t^ho²⁴ liau⁴³ ǁ tsei⁴⁵ xau⁴³ | pə²²³ xən³⁴ | pei³¹ p^hə²⁴ tsən²⁴ ɹən⁴³ | t^hai⁴⁵ ien⁵³ | pi³⁴ t^ha³⁴ ti⁴³ pən³⁴ sɹ³¹ tai³¹ǁ

Orthographic version