"Clear" and "turbid" in Chinese phonology

« previous post | next post »

A key concept in traditional Chinese phonology is the distinction between "clear" (qīng 清) and "muddy / turbid / murky" (zhuó 濁).  Although it is mainly applied to the sounds of language, the qīng 清-zhuó 濁 distinction also has applications / implications for music.

Roughly speaking, the linguistic and musical correlations are qīng 清 ("clear; high pitch") and zhuó 濁 ("muddy; low pitch").  Also applicable to music are the wǔshēng 五聲 ("five musical tones [of the pentatonic scale])": gōng 宮, shāng 商, jué 角, zhǐ 徵, and 羽 — equivalent to do, re, mi, sol, and la in western solfège. (source)

I've often wondered how and when these terms arose, how they function in historical phonology, and how they correlate with usages in modern linguistics.  I asked several specialists in Chinese historical linguistics their opinion on these matters.

David Prager Branner:

How the terms function in historical phonology. Qīng 清 "clear" and zhuó 濁 "murky" describe absence or presence of voicing in syllable initials. Qīng means the absence of voicing and zhuó means its presence. Most syllable-initials in the Yùnjìng 韻鏡 (1161; 1203) system are classified as either qīng or zhuó, although there are also some classified as "qīngzhuó," about whose intended meaning there is argument.

Correlation in modern linguistics. We normally explain the analogy by way of "murmur" or "breathy voice," something still widely heard in, famously, many dialects of the Wú group. This kind of initial-voicing sometimes appears to be a phonetic feature distinct from the segments of the syllable.

When and how they arose. The contrast and the two specific terms qīng and zhuó seem to have originated in the late Nánběi Cháo period (386-589) — they appear together in Yán Zhītuī's (531-591) Jiāxùn 家訓, for instance. Just how they arose in phonetics is uncertain, but they are also known in musical terminology, from a much earlier period — they occur together in the Lǐjì, for instance. I have heard it said that they refer to normal vs. flatted musical pitch, which would offer a situation parallel to the effect of voicing on tone contours — syllables with a murmured-voice initial tend to begin with lower pitch than syllables in the same tone category but with a voiceless initial. But I am not certain this explanation is agreed on universally.

Tsu-Lin Mei, quoting Jerry Norman, Chinese (1988). 

On p. 33, there is a photo reprint of a page of the Yunjing rhyme table (a Song work).  Reading from right to left, there are 4 categories of initials, qīng 清, qīng 次清, zhuó 浊, qīngzhuó 清浊。

[VHM:  respectively voiceless unaspirated, voiceless aspirated; voiced; nasal stops.  (source)]

On p. 30 there is this explanation:

     The primary division of the initials is into qīng 'clear' and zhuó muddy'. Qīng designates the voiceless initials; the qīng initials are further divided into quánqīng 'completely clear', the voiceless unaspirated initials, and ciqing 'secondarily clear', the voiceless aspirated initials.  Quanzhuo 'completely muddy' refers to the voiced obstruents (stops, affricates and fricatives); there is evidence that in the late Tang standard language the quanzhup initials were pronounced as voiced aspirates or murmured stops, much like their descendants in certain contemporary Wu dialects.  The cizhuo or 'secondarily muddy' initials are the voiced sonorants.

How it correlates with concepts in modern linguistics?

Qing 'clear' correlates with voiceless initials.  Zhuo "muddy' correlates with voiced (obstruent) initials.  Cizhuo 'secondarily muddy' is no longer used.  Instead there is biyin 鼻音 nasals, bianyin 边音 laterals, etc.

Since the "clear" (qīng 清)-"muddy / turbid / murky" (zhuó 濁) distinction in phonological analysis emerged around the 4th-5th century when so many other important phonological concepts arose in China under the impact of Indian language science (brought primarily with Buddhism), I suspect that the qīngzhuó 濁 distinction also has an Indian basis, but I've never been able to determine exactly what it was.

I asked Indian language specialists, "Does this "clear"-"muddy / turbid" distinction ring a bell with you in Indian language science?"

Patrick Olivelle replied:

Could this have something to do with the division of consonants into “sonant” ghoṣavant, and surd aghoṣa?

That seems like a reasonable supposition to me.  Now we need to undertake how, when, and in what form this distinction was transferred to China.  It would also be desirable if there was anything in the etymological basis of ghoṣavant, and aghoṣa that would have suggested "clear" and "turbid / muddy / murky" to those who wished to translate them into Sinitic.  Since the qīng 清-zhuó 濁 distinction already occurred in musical analysis in China during the classical period, it must have been taken over ready-made for phonological analysis.

This post has raised a lot of questions, answered some of them, and points to further questions and answers from interested readers.  Above all, its main purpose is to introduce a key concept in historical Sinitic linguistics to modern linguists who previously may have been completely unaware of it.  Hopefully this will lead to fruitful interaction between the two fields.


  1. Victor Mair said,

    November 29, 2020 @ 8:31 am

    From a specialist on Mesopotamian music:

    The most important comment in the post for me is, of course, David Prager Branner's comment:

    "Just how they arose in phonetics is uncertain, but they are also known in musical terminology, from a much earlier period — they occur together in the Lǐjì, for instance. I have heard it said that they refer to normal vs. flatted musical pitch, which would offer a situation parallel to the effect of voicing on tone contours — syllables with a murmured-voice initial tend to begin with lower pitch than syllables in the same tone category but with a voiceless initial."

    That is, as you know, exactly the same meaning of the two terms in Mesopotamian music theory. Can we ask David where he "heard it said" and try to trace this use of the two terms as far back as possible?

  2. Bathrobe said,

    November 29, 2020 @ 8:14 pm

    Just as an aside, Japanese has borrowed the Chinese terminology: 清音 sei-on 'clear sound' for voiceless and 濁音 daku-on 'muddy sound' for voiced. The two dots that are placed next to kana syllables with unvoiced onset to show voicing (e.g., く ku –> ぐ gu) are called 濁り nigori 'muddying'.

  3. Wolfgang Behr said,

    November 30, 2020 @ 4:03 am

    As Victor and David Prager Branner said, the qingzhuo terminology is attested in Chinese musical terminology early on, where it is widely used (104 "pre-Qin & Han" ctext hits) to refer to the balancing of the tones on a pentatonic scale (wuyin 五音), the tuning of pitchpipes, the equilibrium sough in singing etc. and as a metaphor for political and cosmological "harmony" or even the selection range in divination. Clear and muddy are, so to speak, "negative polarity items" opening a space from which the ear has to make a selection to produce well-tuned music — and the ruler to produce well-adjusted political measures.
    The locus classicus for the term occurs in Zuozhuan (《昭公》 20), but the earliest passage is probably that in the Yizhou shu "Guan ren jie" 《官人解》 chapter. There, qingzhuo is used to refer to a distinction naturally arising in all objects through the workings of primeval qi ("pneuma", "odem"):

    "When pneuma first produced objects, the objects, once produced, had sound; and when there was hard and soft, clear and muddy among the sounds, good and bad concomitantly emerged from these sounds."

    Later on, from the Eastern Han onwards, there are also medical usages referring to the quality of blood or the directionality of pneuma passing "upward through the lungs" (qing) or "downwards through the stomach" (zhuo), but there doesn't seem to be a direct connection to breathiness or aspiration in articulation, as far a I can see.

    It is also hard to tell whether there is any direct Buddhist influence on the harnessing of these musico-cosmological terms for their applications to phonological distinctions. In any case it is clear that a lot of ancient musical terminology was recycled in the technical language of medieval Chinese phonologists, a point made very eloquently long ago by Zhang Qingchang 张清常(1915-1998) in the following article:

    Zhang Qingchang 张清常 (1944), "Zhongguo yinyunxue suo jieyong de yinyue shuyu" 中國聲韻學所借用的音樂術语 [The musical terminology borrowed by traditional Chinese phonology], in: Renwen kexue xuebao 人文科學學報, repr. in his: Yuyanxue lunwenji 語言學論文集, Beijing: Shangwu 商務 1993: 209-228. (I'll be happy to send out a [scribbled] copy to those of you interested).

  4. Victor Mair said,

    November 30, 2020 @ 11:38 am

    From Hiroshi Kumamoto:

    These terms seem to originate in musicology. Kǒng Yǐngdá 孔穎達’s (574-648) subcommentary (shū 疏) to Lǐ jì zhèngyì 礼記正義 (Correct Meaning of the Book of Rites; on Yuè ji 楽記 [Record of Music] = Lǐ jì 礼記 [Record of Ritual], Ch. 25) says: Zhèngyì yuē. Yán shēng zhě. Shì gōng shāng jiǎo zhēng yǔ yě. Qū zhuó zhě wéi gōng. Jí qīng zhě wéi yǔ 正義曰.言声者.是宮商角徴羽也.曲濁者為宮.極清者為羽.* That is: in the octave, the lowest is called 濁, the highest 清.

    That's because phonetically voiceless onsets are relatively high-pitched and voiced ones low-pitched across languages. In Modern Lhasa Tibetan this contrast has become phonological; voiceless onsets in Classical Tibetan are now high-toned, voiced ones low-toned, and voice distinctions are now lost (or irrelevant).

    [*VHM: "The Correct Meaning says: When speaking of sounds, they are gōng, shāng, jiǎo, zhēng, and . The {most} bent / curved turbid one is gong, the extremely pure one is ." Gōng, shāng, jiǎo, zhēng, and are the notes of the Chinese pentatonic scale. The basics of Chinese music theory are spelled out here. Be sure to follow the links, especially the one to the twelve pitches (shíèrlǜ 十二律). The most difficult part of the quotation from the Correct Meaning for me is how to interpret the significance of the word 曲 ("bent; curved"). I think that it may refer to the dipping quality of the lowest possible tone in the human vocal speech register. This is how excellent Chinese language teachers instruct students to pronounce the third tone of Mandarin, which has a "dipping" contour, and is so graphically represented in the third tone mark: ǎ ě ǐ ǒ ǔ ǚ Ǎ Ě Ǐ Ǒ Ǔ Ǚ.]

  5. Victor Mair said,

    November 30, 2020 @ 5:16 pm

    In 2008, Tsu-Lin Mei showed that the source of “clear “ and “muddy” alternation in intransitive/ transitive verbs is the causative *s-prefix, e.g. 败 *b-/ 败 *p < *s-b and s- has the capacity to devoice the obstruents as well as the sonorants. His 2008 article has become the standard view in China, Taiwan, America, and France. 梅祖麟,2008,〈上古漢語動詞濁清別義的來源——再論原始漢藏語*s-前綴的使動化構詞功用〉,《民族語文》2008.3:3-20。 See also: Mei, Tsu-Lin. 1989. “The Causative and Denominative Functions of the *s-prefix in Old Chinese.” Proceedings of 2nd International Conference on Sinology (Section on Linguistics and Paleography), 33-52, Taipei: Institute of History and Philology, Academia Sinica.

    And, most recently:

    Géshì lùnxué sānchóngzòu ──yǔ Dǒng Tónghé, Lǐ Fāngguì, Mǎ Xuéliáng sān wèi xiānshēng lùnxué de huíyì


    Shēngyùn lùncóng‧dì èrshí'èr jí Zhōnghuá mínguó shēngyùn xué xuéhuì yè 1~19 Táiwān xuéshēng shūjú 2019 nián 6 yuè


    pdf (not sure that will work)

    "The Source of Voiceless Nasals in Old Chinese──and the Route of Discovery from Li Fang-kuei to Ma Xue-liang"

    Mei, Tsu-lin


    This paper discusses the views of Li Fang-kuei, Tung T’ung-ho and Ma Xue-liang concerning the source of voiceless nasals in Old Chinese, and shows that the *s-prefix can devoice nasals (s-m > hm-, s-n > hn, s-ng > hng) as well as obstruents (s-b > p, s-d > t, s-g > k). Finally the paper attempts to show that the causative *s-prefix is the source of voicing alternation (敗*b-/敗*p-) in Old Chinese.

    Keywords: OC voiceless nasals *hm-, *hn-, *hng-, *s-prefix in Old Chinese, Voicing alternation in Old Chinese (敗*b/敗*p), Li Fang-kuei, S. E.Yakhontov, Ma Xue-liang

    The author thanks F. K. Li, Ma Xue-liang, Sergei Yakhontov, Gong Hwang-cherng, Chang Kun and Betty Shefts Chang, August Conrady, and Dai Qing-xia.

  6. Chris Button said,

    November 30, 2020 @ 8:26 pm

    and shows that the *s-prefix can devoice nasals (s-m > hm-, s-n > hn, s-ng > hng) …

    Here, I agree

    … as well as obstruents (s-b > p, s-d > t, s-g > k).

    Here I disagree:


  7. Lane Greene said,

    December 2, 2020 @ 7:31 am

    Since I'd never heard of the Chinese pentatonic scale before this post, I went looking for it – I couldn't find an English Wikipedia article on it but the Italian Wikipedia says its tones are do, re, *fa*, sol and la, not do re *mi*. But that seems to be wrong. German Wikipedia agrees with Victor, saying CDEGA, so maybe Italian Wikipedia was wrong.

    But (sorry if this is somewhat off topic) what makes it the Chinese pentatonic then? CDEGA is the western major pentatonic. It sounds like westerners and the Chinese just discovered the same pentatonic, no?

  8. Chris Button said,

    December 2, 2020 @ 8:04 pm

    Here I disagree:


    Incidentally, I believe Baxter & Sagart would disagree with both sides of Mei's argument.

    Personally, I have more sympathy with what Mei is trying to achieve with an *s- prefix than Baxter & Sagart, but I still don't buy Mei's argument for *s- devoicing voiced stops.

  9. Nick Williams said,

    December 6, 2020 @ 7:43 pm

    Wolfgang Behr's comment is extremely clarifying: "a lot of ancient musical terminology was recycled in the technical language of medieval Chinese phonologists."  Professional linguists are often less attentive to this kind of cultural "recycling" than they should be.

RSS feed for comments on this post