New Book Preview: The Pronunciation of Chinese

This English version was translated from the Chinese original with AI assistance. The Chinese text is authoritative.

(This text is excerpted from Design Principles of Chinese Characters and Classical Chinese by Li Xutong, forthcoming)

2.3 The pronunciation of Chinese

2.3.1 The historical development and characteristics of the pronunciation of Chinese characters

The earliest recorded pronunciation of Chinese is Yayan (the “elegant speech”). Yayan was very probably the pronunciation of Classical Chinese and of Chinese characters; together with Classical Chinese and Chinese characters, it came from a highly developed culture or from the divine realm and was introduced by the ancestors of the Huaxia. It is a beautiful and complete system of pronunciation.

This Yayan was inherited by the Xia, and became the Xia speech. There are many examples of “雅” (yǎ) and “夏” (xià) being interchangeable in ancient times: the “Gongzi Ya” of the Zuo Commentary is written “Gongzi Xia” in the Hanfeizi; in the Guodian Chu bamboo slips’ Confucius’s Discussion of the Odes, “Daya” and “Xiaoya” are written “Daxia” and “Xiaoxia”; the “Daya” quoted in the Mozi is also written “Daxia.” Especially, the Xunzi, “On Honour and Disgrace,” says “The people of Yue are at ease in Yue, the people of Chu are at ease in Chu, the gentleman is at ease in Ya (雅),” while “The Teachings of the Ru” writes it as “living in Chu he becomes Chu, living in Yue he becomes Yue, living in Xia (夏) he becomes Xia.”

The Yayan sound system is the Old Chinese sound system: one character, one sound, with a beautiful, soft and smooth pronunciation. The Xia dynasty set its capital at Luoyang, and the Yin (Shang) later also set its capital around Luoyang, so the standard pronunciation of Yayan was handed down through the ages in the Luoyang area. Old Luoyang speech, influenced by this Yayan, took in more and more Yayan elements, so that even today people can faintly sense the flavour of Yayan in the Luoyang dialect.

This Yayan spread from the centre of the dynasty to the surrounding regions and gradually became the common speech of ancient times, or the “elegant sound.” The Zhou also inherited the Yayan of the Shang-Yin. When Confucius taught in the state of Lu, his three thousand disciples came from all directions, and Confucius also taught in Yayan. The Analects, “Shu Er,” Book 7, says: “The Master used the elegant speech (yayan) for the Odes, the Documents and the performance of rites: all in the elegant speech.”

After the Zhou, the orthodox Han Chinese dynasties all spared no effort in promoting Yayan. Yayan reached its peak in the Tang and Song periods, when Tang poems and Song ci lyrics emerged in great numbers and the surrounding countries all vied to learn the Yayan of the Central Plains. But as each dynasty moved its capital, Yayan was also modified under the influence of local dialects.

With the change of dynasties, many loyalists of the previous dynasty left the Central Plains and migrated to what are now Zhejiang, Fujian, Guangdong and Guangxi, and even to Yunnan and Guizhou. When such loyalists of a former dynasty settled together in remote, closed areas, they often kept the pronunciation of their time even after thousands of years. So it can be surmised that in some remote mountain areas of Yunnan and Guizhou there may still be people using the language and pronunciation of the Xia and Shang, or of even more remote antiquity.

In the Tang and Song, Chinese phonology reached its peak and developed and perfected the system of eight tones. The eight tones take their model from the Eight Trigrams of the Book of Changes. The eight tones are divided into yin and yang, and yin and yang are each further divided into the four tones “level, rising, departing and entering.” That is: “yang level,” “yang rising,” “yang departing,” “yang entering,” and “yin level,” “yin rising,” “yin departing,” “yin entering.”

The eight tones sound roughly as follows: “The level tone is even, neither low nor high,” that is, the sound is resonant and long and relaxed; “the rising tone is called out high, fierce and strong,” referring to a sound that rises and is heavy and turbid; “the departing tone is clear and mournful, going far,” referring to a sound that is light and sharp with a long tail; “the entering tone is short and abrupt, quickly stored away,” referring to a sound that is extremely light and quickly cut off.

After the Tang and Song, the Mongols, who had long lived in bitterly cold lands, entered and ruled the Central Plains bringing a more primitive sound system; the gentry fled south, the Five Barbarians tinged the Chinese, and the pronunciation of Chinese in the Central Plains changed greatly, falling back from eight tones to four. In the late Yuan, the Chinese pronunciation of the Yuan capital Dadu (today’s Beijing) became the standard, called the “common speech of the realm.” This pronunciation evolved into the Mandarin (guanhua) of the Ming and Qing dynasties, and eventually into the National Language (guoyu) of the Republic and the Putonghua of mainland China after 1949.

The eight tones, and today’s four tones, are a main feature of Chinese, and this tonal feature of Chinese is very special among all the languages of the world. Although Korean and Japanese both contain Chinese vocabulary, neither has the tones of Chinese. In fact, apart from Chinese, the pronunciation of the main languages popular in the world today, such as English, French and Japanese, distinguishes only stressed and unstressed. In terms of Chinese yin–yang theory, they have only distinguished yin and yang and have not yet developed to the Four Images and the Eight Trigrams.

Although the eight tones have not survived in today’s Putonghua, they are fully preserved in Cantonese and other dialects of south-eastern China. Because Cantonese is rich in tones, it has few of the homophones found in Putonghua, and in many cases expresses the same meaning in shorter words. When classical poems are recited in Cantonese, their original sounds and rhymes come through better.

Until modern times, scholars in ancient China always chanted classical texts. This chanting was more like singing, and was certainly not the word-by-word staccato reading of articles today. Chanting classical poetry and prose in Cantonese or other dialects is full of ancient flavour, and the meaning lingers endlessly in the rise and fall of cadence.

In Chinese medicine there is a method of treatment called “Sanskrit-sound vibration.” For certain illnesses, one can pinch particular acupoints while voicing particular sounds, and this can have a curative effect. Here is an example: for people with excessive liver fire or a weak liver, press repeatedly with the thumb at the base of the ring finger until you find a tendon; hold it, coordinate it with your breathing, and repeatedly voice “ge—, he—,” drawing the sounds out appropriately; this can nourish the liver and drain liver fire.

There are many records of treating illness with music in ancient Chinese books. The Record of Music says: “When music is practised, human relations become clear, the ears are keen and the eyes bright, and the blood and qi are balanced.” The Yellow Emperor’s Inner Classic explicitly puts forward the theory of “treating illness with the five notes,” and classifies human constitutions according to the notes to which they belong as the “twenty-five types of yin and yang people.” The Zuo Commentary goes further, saying that music, like medicine, has flavour and can keep people free of all illnesses, healthy and long-lived. The famous Yuan physician Zhu Zhenheng stated plainly: “Music is also medicine.”

Many other classical texts also mention that when people voice certain character sounds, these can vibrate and stimulate the internal organs, and that particular sounds can regulate the body and drive out pathogens. When Cangjie created the characters, he had already left people this hint. Among Chinese characters, if the grass component 艹 of herbal medicine is added on top of “樂” (yuè, “music,” traditional character), it becomes “藥” (yào, “medicine,” traditional character), which shows precisely that “樂” and “藥” share the same origin.

Certain sounds can not only vibrate the internal organs, but in ancient times could also communicate with heaven and earth. According to ancient historical records, Cangjie was born in an age when humans and gods coexisted; the gods of the heavenly realm often walked in the world, and humans could also communicate with the heavenly realm. Legend has it that Cangjie had “double pupils and four eyes,” one pair looking at heaven and the other at the earth; drawing on the writing and pronunciation of the heavenly realm, he created writing for the human world. Therefore composing and reciting texts in this writing could communicate with heaven and earth. Hence the Huainanzi, “The Fundamental Warp,” records: “In ancient times, when Cangjie created writing, heaven rained millet and ghosts wailed at night.” The writing Cangjie created had this power to communicate with the divine and drive away ghosts, so for a long time after it was created, high demands were placed on those who mastered it: they were all required to be people of noble morality, selfless benevolence and reverence for the gods.

Legend has it that every Chinese character Cangjie created corresponds to a god in heaven. To write or read the character is to call upon that god; to use the character in an improper place, speaking or using it carelessly, is to blaspheme that god. Among scholars of the past, the carved bamboo slips and books were treasures passed down through the generations, and scraps of paper with writing on them could not be thrown away carelessly, but had to be burned with thanks to the gods; that is, after using the character of this god, one returned it to the god and expressed one’s gratitude.

By the time of Zhuanxu, more people knew these characters, their pronunciation began to change, and at the same time people’s hearts in the world were no longer as pure as in ancient times. So Emperor Zhuanxu sent the gods Chong and Li to cut off the passage between heaven and earth, thus “severing the communication between heaven and earth,” and reciting these characters could no longer easily communicate with the heavenly realm. But in every dynasty there remained a very small number of people who mastered the original pronunciation of this writing and had the ability to communicate with the gods; these people became what later generations called priests. The pronunciation of the writing they mastered, which could communicate with the gods, was called incantations by later generations.

By the Spring and Autumn and Warring States periods, the great Dao had been abandoned and rites and music had collapsed. So Confucius taught everyone without discrimination and spread this writing widely. This writing then basically became a tool for recording information and communicating in the human world.

Even so, scholars of the past would first bathe and change their clothes, calm their minds and regulate their breathing, and sit upright to read respectfully. By reading the books of the sages for a long time, their minds became clear and penetrating, so that they could cultivate themselves, regulate their families, govern the state and bring peace to the world.

Classical Chinese poetry and prose were chanted, sung aloud. When people chanted the classics with the ancient sounds and rhymes, they were not only pondering the meaning of the text, but this also directly drove out pathogens and strengthened the body.

2.3.2 Marking the pronunciation of Chinese characters

It is generally held that in language, speech sounds come first and writing afterwards. For all phonetic scripts this should be beyond doubt, but for Chinese characters it may not be so. If Classical Chinese and Chinese characters came from a highly developed civilization or from the divine realm, then the sound, form and meaning of Chinese characters may have been transmitted to people as a whole.

In ancient China, the more common methods of indicating the pronunciation of characters were the analogy method, the “read like” method, the direct-sound method and the fanqie (spelling) method. The analogy method describes the pronunciation of a character by the manner of articulation, for example by analogy with the shape of the mouth, the position of the tongue, the speed of aspiration and the length of the tone. The Huainanzi, “Cultivating Effort,” says: “Among the Hu people there are those who know profit, yet people call them 駤 (stubborn).” Gao You’s commentary: “駤 is read like 质, spoken with slow breath; it is obtained at the tip of the tongue.”

The “read like” method indicates pronunciation with a character of the same or similar sound. For example, the Shuowen Jiezi: “唉 means to answer. From 口 with 矣 as phonetic; read like 埃.” “鼾 is breathing while lying asleep; 干 as phonetic; read like 汗.” “Read like” is also written as “read as,” “read for,” “read called,” “sound same as,” “sound close to” and so on. The direct-sound method indicates pronunciation with a character of exactly the same sound. For example, in the Book of Han, “Annals of Emperor Gaozu,” “Lord Lü of Shanfu was on good terms with the magistrate of Pei,” the commentary says: “单 is pronounced 善 (shàn); 父 is pronounced 甫 (fǔ).”

The fanqie method spells out a new pronunciation from the initial of the first character and the final of the second character. The fanqie method was at first called “反” (fǎn); for example, the Eastern Han scholar Fu Qian wrote in his commentary on the Book of Han, “惴 is pronounced 章瑞反.” In the reign of Emperor Daizong of the Tang, because the character “反” (also meaning “rebellion”) was taboo, it was changed to “翻” (fān), and later “翻” was changed to “切” (qiè). Volume 25 of the Compendium of Materia Medica says “糗, 去九切.”

In the surviving classical texts, the analogy method, the “read like” method and the direct-sound method appear earlier. Yang Shen of the Ming says in Danqian Miscellany, “The Ancients Often Used Analogy”: “Before the Qin and Han, the language of books was mostly analogical, and must be sought beyond the words.” By the Sui dynasty, the phonologist Lu Fayan discovered that the pronunciation of Chinese characters can be divided into two parts and that the four tones are distinguished mainly in the final, and so invented the “fanqie method.”

In modern times, Zhang Taiyan, imitating the Japanese kana script, used the method of “simplified side components,” taking the structure of small-seal characters and using the most basic characters as reference sounds, such as “丂” and “日,” as letters, and created a set of phonetic letters called the “initial script” and the “final script.” On this basis, in 1912 the Ministry of Education of the Republic of China convened a provisional education conference and passed the “Resolution on Adopting the Phonetic Alphabet (Zhuyin Zimu).”

In 1958 mainland China promulgated the Hanyu Pinyin scheme. This Hanyu Pinyin scheme is not only used to mark the pronunciation of Chinese characters, but is now also the international standard for the Latin transcription of Mandarin Chinese and the unified norm for the romanization of Chinese personal names, place names and Chinese documents.

2.3.3 The influence of Hanyu Pinyin and Zhuyin on the pronunciation of Chinese

There are many formulations of the relationship between the sound and meaning of Chinese characters, such as: same sound, same meaning; same sound, connected meaning; the law of same sound and connected meaning; homophones share meaning; cognates share sound, and so on. Whatever the differences in wording, these formulations all share one idea: there is a connection between the sound and meaning of Chinese characters.

In Mr. Wang Li’s On Cognate Characters, cognate characters are defined as follows: characters whose sound and meaning are both close, whose sounds are close and meanings the same, or whose meanings are close and sounds the same are called cognate characters. The method of explaining the meaning of a word with a character of the same or similar sound is called sound glossing. Sound glossing is used a great deal in Han-dynasty works such as the Erya, the Fangyan and the Shuowen Jiezi, and by the end of the Han, Liu Xi’s Shiming uses sound glossing throughout.

Cognate, synonymous or semantically related characters use the same or similar pronunciations, and the pronunciations of these characters are marked using the “read like” and direct-sound methods. So when learning these characters, one naturally comes to understand how they are related. This should be regarded as a shortcut left for later generations, to make it easier to learn Chinese characters, when the characters were created.

This approach established a relative system of pronunciation. People in different regions often have different features of pronunciation; for example, as latitude increases, people’s articulation gradually shifts toward the back of the mouth. And people’s pronunciation also tends to change over time. If cognate characters keep the same or similar pronunciations, then even if these pronunciations change with time and place, the relationship between sound and meaning does not change, so it does not affect people’s understanding and learning. This also ensured that throughout the long history of the development of Chinese civilization, the relationship between the sound and meaning of Chinese characters could always be preserved.

Marking the pronunciation of Chinese characters with the fanqie method, while paying more attention to the laws of pronunciation itself, to some extent severed the relationship between the sound and meaning of Chinese characters. And the modern Zhuyin and Pinyin, in which Western missionaries who were expert in phonetics but completely ignorant of Chinese marked the pronunciation of Chinese characters using the pronunciation rules of alphabetic scripts, completely cut off the relationship between the sound and meaning of Chinese characters.

Compared with the damage to the rules of Chinese characters done by the well-known simplified characters, this way of marking the pronunciation of characters with phonetic symbols has damaged the internal structure of the Chinese language at least as seriously. It is an absolute system of pronunciation that takes no account of the differences in the pronunciation features of different regions and peoples, and it places certain constraints on people’s pronunciation and emotional expression; but under the conditions of modern audio and video media, it has indeed made communication and exchange easier.

2.3.5 The phonology of ancient Chinese

Phonology (yinyunxue), also called the study of initials and finals, has a history of more than a thousand years in China. It is a traditional discipline that studies the systems of initials, finals and tones of ancient Chinese in each historical period and the laws of their development, and is an important part of ancient Chinese studies. In its methods, traditional phonology mainly used yin and yang, the Five Phases, the sun and moon, heaven and earth and the like to express the inner laws of initials and finals; this now often seems incomprehensible, but in fact it really does express the inner laws of Chinese phonetics.

The different understandings of the pronunciation of writing in ancient Chinese phonology and in the phonetics of alphabetic scripts arise from the fact that their most basic understandings of speech sound took different paths.

Wikipedia explains the phoneme and the syllable as follows:

A phoneme (English: Phoneme), also translated yinsu, is the smallest unit of sound in human language that can distinguish meaning, and is the basic concept of phonemic analysis. A character or word may consist of one or several syllables, and a syllable may consist of one or several “segments” (vowels, consonants, etc.). Phonemes are very similar to segments, but the basic definition of a phoneme is that it must be able to distinguish meaning: if two sounds represent the same word with the same meaning, the different sounds can be regarded as the same phoneme; conversely, if any phoneme of a word is replaced by another, it is no longer the original word, and its meaning changes accordingly. Meaningful words can all be composed of phonemes, but substituting any of their phonemes does not guarantee a meaningful word; it may also become a meaningless string of sounds. Each language has its own set of phonemes, which is the phonological system of that language; phonemes can be used to study how sounds are combined into words in a particular language. Phoneme is sometimes translated as “yinsu,” but the use of the word yinsu in Chinese is rather confused and does not always refer to the phoneme.

A syllable (English: syllable) is a unit that makes up a sequence of speech sounds, and is also the most natural structural unit of speech. For example, the English word “water” (/ˈwôtər/) consists of the two syllables “wa” (/wô/) and “ter” (/tər/). A syllable usually contains a syllable nucleus (usually a vowel), and may also have margins at the beginning and end of the syllable (usually consonants). The syllable is generally regarded as the “building material” of words at the phonological level, and it affects the rhythm, prosody, metre and stress patterns of the language to which it belongs.

This method of dividing the pronunciation of writing into phonemes and syllables is consistent with the whole Western system of thought; in principle it belongs to reductionism. These phonemes exist independently, and phonetics must study how they combine into syllables, how syllables combine into the pronunciation of words, how the pronunciations of words combine into the pronunciation of sentences, and so on.

The key point is that these phonemes and syllables are a static description. For example, in languages such as English and French there are the concepts of vowels and consonants. A vowel can be understood as a sound produced while the shape of the mouth, the position of the tongue and the exhalation remain unchanged, and a consonant as a sound produced when these factors change. That is, the static is called the “primary” sound (vowel, yuanyin) and the dynamic the “auxiliary” sound (consonant, fuyin).

In Chinese, the syllable of a character is divided into two parts, the initial and the final, which correspond to the initials and finals of today’s Hanyu Pinyin. The initial is the sound produced at the start of a syllable, and the final is the way this sound is prolonged, giving the sound its flavour. So this understanding of sound is based on the dynamic: its basis is to distinguish the ways different articulations are connected, and to take these modes of connection as the basic units of speech.

Because the final of a Chinese syllable is regarded as the prolongation of the initial’s sound, there are different ways of prolonging it, rising and falling, and these form the tones, such as the four tones of today’s Putonghua and the eight tones of Cantonese. When syllables are divided into vowels and consonants, the vowel is primary and unchanging, so there are no tonal changes.

In accord with the Book of Changes, in which the Taiji divides into two to give the Two Modes, the Two Modes divide into four to give the Four Images, the Four Images divide into eight to give the Eight Trigrams, the Eight Trigrams are doubled to give the sixty-four hexagrams, and each hexagram has six lines, sixty-four times six giving three hundred and eighty-four lines, Chinese characters are basically one syllable per character. For example, all the pronunciations of today’s Putonghua divide into about four hundred syllables, each with four tones, making more than a thousand pronunciations, which together with homophones represent the pronunciations of all Chinese characters.

Following the ancient Chinese way of thinking that divides from the whole, at first the pronunciation was broken down only as far as the syllable, giving rise to the direct-sound and “read like” methods. Later the pronunciation was divided further into the two parts, initial and final, giving rise to the fanqie method of marking the pronunciation of characters. On the surface, the next step in this process would seem to be the understanding of phonemes and the formation of pinyin; but if one thinks this way, one has in fact already lost one ultimate characteristic of the pronunciation of Chinese characters.

The romanization schemes for Chinese characters produced in modern times, whether Roman pinyin or the Zhuyin phonetic symbols, are all based on this phonetic-script understanding of speech. This approach makes it easy for Westerners to remember the pronunciation of Chinese characters, but it has damaged the characters’ own system of pronunciation.

When the pronunciation of Chinese is broken down according to the rule of dividing one into two, there is also a relationship of mutual generation between the different parts; that is, the syllables of Chinese character pronunciation do not exist independently, but have specific associations. In that case, one need not consider certain invalid connections between syllables; or rather, for connections between certain phonemes or sounds, there are specific relations of mutual generation and specific connections. These associations and relations of mutual generation are the basic objects of study of Chinese phonetics.

By contrast, in the connection of phonemes and syllables in phonetic scripts, because phonemes exist independently and have no inner connection, connecting them requires all kinds of rules and the study of these rules; this is the content of phonetics (Phonetics) and phonology (English: Phonology or Phonemics, also called the study of sound systems, phonemics, or yinyunxue).

The achievements of ancient Chinese research on the pronunciation of Chinese are mainly embodied in the thirty-six initials (zimu), the system of letters used in the rhyme tables of the Song dynasty to represent the initial classes of Middle Chinese. As to who invented the thirty-six initials, Wang Yinglin in Jade Sea held that they were made by Shouwen. At the beginning of the twentieth century, two documents recording Tang-dynasty initials were discovered at Dunhuang: one is Examples of the Thirty Initials, and the other is a fragmentary scroll of Shouwen’s phonology. Both record “thirty initials” with largely the same content, showing that in the Tang there were only thirty initials, and that the Song thirty-six initials were an expanded version based on them, adjusted to the phonetic features of the time.

The thirty initials of Shouwen, as corrected by later generations, are shown in the table below. Labials: 帮 滂 並 明. Linguals: 端 透 定 泥 来 are tongue-tip sounds; 知 彻 澄 are tongue-top sounds. Velars (“molar sounds”): 见 溪 群 疑 and the like. Dentals: 精 清 从 心 邪 are tooth-tip sounds; 审 穿 禅 照 日 are proper dental sounds. Gutturals: 晓 影 are guttural sounds, clear; 匣 喻 are also guttural sounds, muddy.

According to the Rhyme Mirror (Yunjing), the system of thirty-six initials is as follows (reconstructions follow Wang Li’s History of Chinese Phonetics):

Phonological termsClear and muddy (voicing)
  Fully clear    Secondary clear    Fully muddy    Secondary muddy  Fully clearFully muddy
Five notes (places of articulation)MannerStopAffricateStopAffricateStopAffricateNasalApproximantLateralFricative
PlaceSibilantNon-sibilantSibilantNon-sibilant
LabialsHeavy labialsBilabial帮[p]滂[pʰ]並[b]明[m]
Light labialsLabiodental非[p̪]敷[p̪ʰ]奉[b̪]微[ɱ]
LingualsTongue-tip soundsAlveolar端[t]透[tʰ]定[d]泥[n]
Tongue-top soundsAlveolo-palatal知[ȶ]彻[ȶʰ]澄[ȡ]娘[ȵ]
DentalsTooth-tip soundsAlveolar精[ts]清[tsʰ]从[dz] 心[s]邪[z]
Proper dental soundsAlveolo-palatal照[ʨ]穿[ʨʰ]床[ʥ] 审[ɕ]禅[ʑ]
Velars (“molar sounds”)Velar见[k]溪[kʰ]群[g]疑[ŋ]
GutturalsGlottal影[ʔ]   
Palatal 喻[j]
Velar   晓[x]匣[ɣ]
Semi-lingualsAlveolar   来[l]
Semi-dentalsPalatal   日[ɲ]

Comparing the two, it can be seen that the thirty-six initials have six more than the thirty: “非, 敷, 奉, 微, 床, 娘.”

These thirty-six initials can be assigned to the two tone classes yin level and yang level according to whether the initial is clear or muddy; this is the “division of the level tone into yin and yang” in Chinese phonology and Chinese dialectology. Of the Middle Chinese level-tone characters, the initials that go to yin level are: 帮, 滂, 非, 敷, 端, 透, 知, 彻, 精, 清, 心, 照, 穿, 审, 见, 溪, 影, 晓. The initials that go to yang level are: 并, 明, 奉, 微, 定, 泥, 澄, 娘, 从, 邪, 床, 禅, 群, 疑, 喻, 匣, 来, 日.

Clear and muddy are further divided into the Four Images of fully clear, secondary clear, fully muddy and secondary muddy. Matching these Four Images of clear and muddy with the five notes of the Five Phases, labial, lingual, dental, velar and guttural, gives the thirty-six initials. Matching the Four Images, which represent cyclic nesting, with the Five Phases, which represent cyclic enumeration, can represent the laws of all things, and so can correspondingly represent the syllables that express these laws.