Yīn & Yáng
and the I Ching

In India the theory of the three elements in the Chândogya Upanishad led to the theory of the three forces, the guṇas, and to the later theory of five elements. In China, the theory of five elements coexisted early with the theory of two forces: and . These can also simply be called the "two forces," (where ch'i, , is the "breath" or vital energy of the body, but also simply air, steam, or weather). In the Spring and Autumn Period there was actually a Yin and Yang School. Later its theories were accepted by nearly everyone, but especially by Taoism. The implications of the theory are displayed in the great book of divination, the I Ching, , the "Book of Changes."

Yīn originally meant "shady, secret, dark, mysterious, cold." It thus could mean the shaded, north side of a mountain or the shaded, south bank of a river. Yáng in turn meant "clear, bright, the sun, heat," the opposite of yīn and so the lit, south side of a mountain or the lit, north bank of a river. From these basic opposites, a complete system of opposites was elaborated. Yīn represents everything about the world that is dark, hidden, passive, receptive, yielding, cool, soft, and feminine. Yáng represents everything about the world that is illuminated, evident, active, aggressive, controlling, hot, hard, and masculine. Everything in the world can be identified with either yīn or yáng. Earth is the ultimate yīn object. Heaven is the ultimate yáng object. Of the two basic Chinese "Ways," Confucianism is identified with the yáng aspect, Taoism with the yīn aspect.

Although it is correct to see yīn as feminine and yáng as masculine, everything in the world is really a mixture of the two, which means that female beings may actually be mostly yáng and male beings may actually be mostly yīn. Because of that, things that we might expect to be female or male because they clearly represent yīn or yáng, may turn out to be the opposite instead.

Taoism takes the doctrine of yīn and yáng, and includes it in its own theory of change. Like Anaximander and Heraclitus, Taoism sees all change as one opposite replacing the other. The familiar diagram of Yīn and Yáng, the , the "Great Ultimate" [Wade-Giles T'ai-chi] diagram, shows the opposites flowing into each other. The diagram also illustrates, with interior dots, the idea that each force contains the seed of the other, so that they do not merely replace each other but actually become each other. (The earliest attested example of the diagram, strangely enough, occurs on a Roman shield illustrated in the fifth century Notitia Dignitatum.)

Unlike Heraclitus, Taoism sees change as violent only if the Tao [Dào] is opposed: If Not Doing, , and No Mind, , are practiced, then the Tao guides change in a natural, easy way, making for beauty and life. Since trying to be in control is a yáng (or Confucian) attribute, Taoism sees Not Doing (and Taoism itself) on the yīn side of things; but since Not Doing does not literally mean doing nothing, Taoism can use the language of passivity and receptivity to mean something that is actually quite active.

That is especially obvious in the use of the term [Wade-Giles jou2], "soft, pliant, yielding, gentle." Róudào, the "yielding way," is read in Japanese as judô and is the name of a popular Martial Art. Judo doesn't look at all yielding or gentle, but it does employ Taoist doctrine in so far as it is not supposed to originate force or an attack but takes the attack of an opponent and uses its own force against it.

Thus the great economist F.A. Hayek invoked Taoism in the defense of capitalism, a system that does not seem particularly yielding or gentle, but is based on the principle that government should "leave alone" (laissez faire) private property and voluntary exchanges and contracts. The free market would thus be the Not Doing of government.

When it comes to the five elements, earth, water, and wood are clearly to be associated with yīn. Water, the softest and most yielding element, becomes the supreme symbol of yīn and the Tao in the Tao Te Ching. Fire (the hottest element) and metal (the hardest) both are associated with yáng.
Nevertheless, the Blue Dragon, , that symbolizes wood is a principal symbol of , while the White Tiger, , that symbolizes metal is a principal symbol of . This kind of reversal turns up frequently in the I Ching.

The I Ching, , is based on the principle of a broken line, , representing yīn, and an unbroken line, , representing yáng. During the Shang Dynasty (1523-1028 BC), questions that could be answered with a "yes" or a "no" were written on tortoise shells. The shells were heated, then doused in water, which caused them to crack. A broken crack, , was interpreted as a "no" answer, an unbroken crack, , as a "yes." The I Ching elaborates on this, by grouping the lines into sets of threes (the trigrams) and into sets of sixes (the hexagrams).

There are eight trigrams:
Among the trigrams it is noteworthy that in all the children, the sex is determined by the odd line, so that the trigrams are predominately the opposite quality from the sex of the child. Also, we expect water to be associated with yīn and fire with yáng, but water is the second son and fire the second daughter. The other children are associated with such things as we might expect, e.g. water turns up again in the third daughter as the Lake.

The arrangement of the trigrams around the compass reflects Chinese geomancy (), i.e. the determination of the auspicious or inauspicious situation and orientation of places (cities, temples, houses, or graves). Chinese cities are properly laid out as squares, with gates in the middle of the sides facing due north, east, south, and west. The diagonal directions are then regarded as special "spirit" gates: northwest is the Heaven Gate; southwest the Earth Gate; southeast the Man Gate; and northeast the Demon Gate. The northeast was thus the direction from which malevolent supernatural influences might particularly be expected. The situation of the old Japanese capital city of Kyôto is particularly fortunate. To the northeast is a conspicuous, twin-peaked mountain, Mt. Hiei (corresponding to the Mountain trigram), which is crowned with a vast establishment of Buddhist temples to guard the Demon Gate. Later, Tôkyô (originally called Edo) was laid out with temples to the northeast on rising ground in the Ueno district; but both the ground and the temples are now entirely surrounded and obscured by the sprawl of Tôkyô. [note]

The trigrams contrast the Moutain, , with the Lake, . A lake is essentially a valley filed with water (both with Yīn associations), and the mountain in general may be also contrasted with the valley, . We see this contrast in related characters, such as , "an immortal," and , "common, vulgar, worldly." Each of these contains the "mountain" and "valley" characters, respectively, with the radical for "person," . The idea seems to be that immortal beings live in the mountains, either because that is where the divine belong (as on Mt. Olympus) or because that it where Taoist adepts, who achieve immortality, practice their asceticism. Thus, Taoists themselves can be called , the "immortal-ists" or "school of the immortals." What is down in the valley is then common, mundane, and vulgar.

The I Ching uses the trigrams by combining pairs of them into 64 hexagrams. The hexagrams reuse the trigrams by combining pairs of them into 64 hexagrams. The hexagrams represent states of affairs, and the I Ching is consulted through the construction of a hexagram to answer one's question. The construction is carried out either through a complicated process of throwing and counting yarrow stalks, or by throwing three coins. The obverse (head) of each coin is worth 3 points (odd numbers are yáng), while the reverse (tail) is worth 2 (even numbers are yīn). Three coins will therefore add up to either 6, 7, 8, or 9. The numbers 7 and 8 represent "young" yáng and yīn, respectively. Starting from the bottom up, these add a plain yáng, , or a plain yīn, , line. The numbers 6 and 9, in turn, represent "old" yīn and yáng, respectively, and are called "changing lines." This illustrates an important aspect of the theory of yīn and yáng: Because the "Way of the Tao is Return," yīn and yáng, when they reach their extremes, actually become their opposites. The "old" lines therefore change into their opposites, giving us two hexagrams if any changing lines are involved: the first hexagram, representing the current state of affairs; and the second hexagram, after the changes have been made, representing the future state of affairs. Changing lines are usually denoted by writing for a 9 and for a 6. The text of the I Ching describes the significance of each hexagram and also the special meaning to be attached to the presence of any changing lines.

Fantasy Factorial Hexagrams

Chinese Elements and Associations

The Solar Terms and the Chinese Calendar

Psychological Types

Gender Stereotypes and Sexual Archetypes

History of Philosophy, Chinese Philosophy

History of Philosophy

Home Page

Copyright (c) 1997, 2002, 2003, 2006, 2007, 2008, 2009, 2011 Kelley L. Ross, Ph.D. All Rights Reserved

Yīn & Yáng and the I Ching, Note

As it happens, there is a conspicuous mountain north-east of Los Angeles Valley College. Indeed, there is a whole mountain range, the San Gabriel Mountains. Beyond the lower Verdugo Mountains in the foreground, which rise to 3126 feet, there is the conspicuous Mt. Lukens in the San Gabriels, which is 5074 feet high. Behind Mt. Lukens runs Big Tujunga Canyon. There are much higher peaks in the San Gabriels (up to Mt. San Antonio, "Old Baldy," at 10,064 ft., which is east and outside of the image provided here), as can be seen in the image, but these are hidden from the perspective of Valley College. Unfortunately, there are no Buddhist temples, as far as I know, upon Mt. Lukens. Los Angeles could use the protection.

Return to text

Categories of Chinese Characters

Chinese characters are the last ancient ideographic writing system that survives in modern usage. This was a close call. In Vietnamese, the Latin alphabet is used; in Korean, the Hangŭl phonetic system is now used. Japanese has its own syllabaries, the kana, which could easily replace characters altogether, as in the past they sometimes did. Both China and Japan were contemplating a transition to the Latin alphabet (the Pinyin system prepared the way for this in Chinese). Ironically, it is the most modern technology which has saved the most ancient writing. Computer assisted writing makes the use of characters relatively convenient, and the need for vast metal fonts for printing and even typewriting has now been eliminated.

Although Chinese characters are originally and basically ideographic, writing whole words, the language over time has become more polysyllabic and many characters now do not occur in isolation. The system thus can be said to have become morphographic, writing semantic elements of words, morphemes, rather than ideas or words as wholes. [note]

The characters and their definitions here are from Mathews' Chinese-English Dictionary [Harvard, 1972]. The pronunciation of each character, however, is rendered in Pinyin. There are, understandably, disputes over the classification system and over the assignment of individual characters. For instance, the very first example, , "big," is from the drawing of a man, and so can be considered "pictographic"; but since it doesn't mean "man," but "big," it might be considered "indicative" instead.

  1. Pictographic:  These are characters that originate with pictures of the objects in question. In the Shang Dyansty, these counted for 23% of all characters. By the Han they were down to only 4%, and during the Sung only 3%. The characters at right were all originally little pictures. "Great" was the picture of a man, while "mountain," "field," "woman," "horse," "shield," and "tree" were just that.

    John DeFrancis [The Chinese Language, Fact and Fantasy, University of Hawaii Press, 1984, 1986, & Visible Speech, University of Hawaii Press, 1989], one of the greatest scholars of Chinese, has the view that language (or meaning) is essentially spoken (i.e. sound) and that pictograms really stand for the words rather than for the things. However, it seems the most natural to say that a picture of a man, a woman, or a tree simply represents those things directly. While all writing systems, including Chinese, develop phonetic elements, the thesis that meaning is essentially sound is destroyed by the use of sign language among the profoundly deaf, for whom language and meaning have no aural component at all. At one time, it was not believed that the profoundly deaf had any true language, just because sign language was not taken seriously; but this view is now insupportable. Indeed, from Plato we already have the observation that the deaf sign and that this is a logical accommodation to that condition:

    SOCRATES:  Answer me this: If we had no voice [φωνή, phoné] or tongue [γλῶττα, glôtta], and wished to make things [πράγματα, prágmata] known to one another, should we not try, as mute [and deaf] people [ἐνεοί, eneoí; singular ἐνεός, eneós] actually do, to make signs [σημαίνειν, sêmaínein] with our hands and head and body generally? ["Cratylus," 433 E, Cratylus, Parmenides, Greater Hippias, Lesser Hippias, translated by F.N. Fowler, Loeb Classical Library, Harvard, 1926, 1963, p.133; translation modified]

    Sign languages are known to develop and exist with no connection to spoken language, and the form of signs has its own dynamic, unrelated to sounds. Thus, even as a Chinese character is classified by radical and phonetic, a sign can be specified by [1] the shape of the hand(s), [2] position(s), [3] orientation(s), and [4] motion(s) (if any).

  2. Simple Indicative or Ideographic:  Some abstract concepts can be suggested with certain diagrams, like simple lines for "one," "two," and "three." At right, we also have "under," "above," and "middle," all of which bear some relation, as diagrams, to the meaning. In the Shang Dynasty, only 2% of characters were like this. By the Han and Sung, it was down to only 1%. So these kinds of characters may be frequently used, but there aren't many of them.
  3. Compound Indicative or Logical Aggregates:  Multiple examples of the first two kinds of characters can be combined to suggest something semantically related to the original meanings. So at right, we see "sun" and "moon" combined to mean "bright," "light," or even "cleanse." Three "fields" can be combined to mean "fields divided by dikes." A "woman" under a "roof" means "quiet," "peace," "tranquility." Two "women" means "handsome" or "pretty," and also "cunning." This negative (misogynistic) suggestion emerges fully with three "women," which means "adultery," "fornication," "licentiousness," "debauch," "ravish." Two "trees" get us "forest," and three are "luxuriant," "overgrown," "dark." Three "stones" is "heap of stone, boulders." Note that there are altenative, radical and phonetic versions, given with the lei (boulders) and jiao (handsome) characters. In the Shang Dynasty, 41% of the characters were of this compound indicative type. In the Han it was 13%, and in the Sung only 3%. It is sometimes said that the Chinese character for "trouble" shows two women under one roof. Such a character is possible, and would look like this , but there actually is no such Chinese character, though I understand that the myth lives on the internet. Meanwhile, the character , which is a pig under a roof, means "a house, family, home, relatives," or a member of a class or school. We can imagine that this goes back to the conditions of rural life where people and farm animals might share the same dwelling, even as pork is still a conspicuous part of traditional Chinese cooking.

  4. The most common Chinese characters are of the Radical and Phonetic or Phonetic Complex form. These combine other characters either side by side or above and below. The constituent character called the "Radical" gives some clue about the meaning and, more importantly, is the basis for the listing of the character in Chinese dictionaries (where 214 traditional Radicals are used). The constituent character called the "Phonetic" gives some clue about the pronunciation, which is usually similar to that of the original character. In the Shang Dynasty, only 34% of characters (or 334 actual characters) were of this type. By the Han Dynasty, it was up to 82% (or 7697), the Sung up to 93% (21,810), and in the Ch'ing radical and phonetic characters were 97% (or 47,141) of the total. Clearly, this device becomes the most productive way of generating new characters in Chinese. It is also unique among Old World ideographic writing systems. Nothing similar is seen in Egyptian hieroglyphics, for instance, where the phonology of a word is indicated by writing extra, purely phonetic, glyphs. The exception, however, is in the New World, where Mayan glyphs, recently deciphered, include both ideographic and phonetic elements, just like Chinese characters. Mayan glyphs, however, fully specify the phonology (according to the current understanding), not just suggest it, as with the Chinese.

    In the diagram at right, the basic phonetic value of "horse" () turns up in the purely phonetic interrogative particle, and in a word for "mother." The character for "to tie, bind" occurred as a phonetic in the alternative character given above for "heap of stone/boulders" (lei). The "fields" compound character above (lei again) occurs as a phonetic with the character for "stone" to mean "roll stones down hill." "Shield" (gan) occurs with "sun" in "sunset," with "woman" in "crafty," villainous," "false," and with "tree" in "shaft of a spear," "pole." "Middle" occurs with the radical "heart," zhong, to mean "conscientious," "loyal," "honest," etc. It is these characters that provide some of the evidence for the reconstruction of the pronunciation of earlier forms of Chinese.

    Since radical and phonetic characters already exist in the Shang Dynasty, there clearly was a long period of development prior to this. But the evidence for this is scant, and the ultimate origin of Chinese characters is unclear.

Dialects of Chinese

History of Philosophy, Chinese Philosophy

History of Philosophy

Philosophy of History

Home Page

Copyright (c) 2000, 2005, 2006, 2010, 2013, 2015, 2016, 2017, 2019 Kelley L. Ross, Ph.D. All Rights Reserved

Categories of Chinese Characters, Note;
Ideograms vs. "Logograms"

Since Chinese characters originally wrote whole words, it is now fashionable to say that they are "logograms" (logos = "word") rather than "ideograms." On this view, Chinese characters (or the units of any such writing system) have no meaning apart from the words of Chinese. They are derivative of the words and are semantically, functionally, and even ontologically dependent on them. The notion that the characters could exist independently of the words, or of the Chinese language, is incomprehensible.

As noted, this is already rather behind the development of Chinese, where characters usually write morphemes. However, the principal reason for the change in terms is ideological rather than linguistic. Because of the influence of Ludwig Wittgenstein and Ferdinand de Saussure, the view has grown that language is a self-contained and self-referential system, without connection to the external world or to truth. Because of this, the notion that there are "ideas" or concepts that exist independently of language and embody meanings with a real relationship to the world has fallen into disfavor. So "ideogram" must go.

Unfortunately, those who are at pains to demonstrate their adherence to fashionable opinion have missed the point. The issue is not whether ideas or truth exist, but whether a writing system like Chinese characters directly matches up with spoken language. It doesn't. This is the most conspicuous in something like Ancient Egyptian hieroglyphics, where certain glyphs are "generic determinatives," which correspond to no words in the language but give a clue as to the general meaning of the word being written. As it happens, Chinese has something rather like generic determinatives, i.e. the "radical" which is that part of the character that gives a clue to the meaning and functions as the basis of classifying characters in a Chinese dictionary. These visual elements of the written language do a job where the written language may not fully represent the sounds of spoken language, which is what happens in Egyptian or Chinese. The written language does it in its own way, and so takes on a life of its own.

Since the fashionable view is that language is self-referential, we might wonder why opinion could not move over to the view that written language breaks away from the spoken language and takes on a self-contained life of its own. Clinging to the notion that written language refers to spoken language would seem to contradict part of the fashionable thesis, that there is no external reference. Indeed. But the move does not take place, perhaps because the connection of the written to the spoken language is too obvious (though one might think that their connection to the world would then be equally obvious, which it isn't to the bien pensants), but perhaps even more so because of an old prejudice that language can only exist as spoken language. This latter assertion is actually made by John DeFrancis in the work cited in the text above -- and reconfirmed to me in personal correspondence.

The notion that language can only truly consist of sounds is refuted by the existence of fully functioning sign languages among the profoundly deaf. Indeed, there are now cases where deaf children, with no previous contact with other deaf individuals, have been introduced together into new schools for the deaf and have spontaneously and quickly developed a completely new sign language between themselves. In the past, the possibility that sign languages could be the equivalent of spoken language was simply not believed, and even educators of the deaf thought that signs could properly only be used to spell the words of spoken languages. Word of the existence of true, semantically complete sign languages of the deaf has apparently still not reached everyone.

The truth is that visual (whether written or sign) and spoken languages match up to each other by way of meaning. There are ideas, concepts, and reference. That is why languages can be translated into each other -- though, indeed, there are philosophers, like W.V.O. Quine, in the self-referential tradition, who openly assert the "indeterminacy of translation," as though this were not contradicted by centuries of actual translating. The existence of meaning has been ably demonstrated by Jerrold Katz. Thus, Chinese characters, which write ideas, as spoken language speaks them (with, we might say, "ideophones" -- sounds that speak ideas), are ideograms. Since they historically correspond to Chinese words or morphemes, they can also be called logograms or morphograms. Since they often originally consisted of pictures of objects, they can also be called "pictograms," a term also in fashionable disfavor. If there are pictures of objects, after all, we might need to admit that there are objects, and that language has something to do with them. It is a shame when something so obvious becomes shocking to educated opinion.

Why there is now this ideological preference is a good question. Such theories, however, are conformable to the "deconstructionist" or "post-modern" view that everything is a matter of power relationships -- something about equally inspired by Marx and by Nietzsche -- and unrelated to any actual truth or reality, except a political reality. People writing about Chinese characters may not be aware of all the connections of the theories they promote, but it is usually the academic water within which they swim.

Return to text

The Dialects of Chinese

What are usually called the "dialects" of Chinese are really separate languages, all descended from the Chinese of the T'ang Dynasty. They are all about as far apart from each other now as English and Dutch. However, they are all written with the same characters (with some exceptions), which means that an educated person can understand (mostly) their written forms, and for cultural and political reasons, as well as their historical origin, are regarded by the Chinese as part of the same language. A new term has even been introduced for this unusual situation, calling the languages "topolects," i.e. speech of the "place," τόπος, topos. A Chinese equivalent term, , "speech of the place," not only is the official term for "dialect," but it is officially used for all the Chinese languages, whether they are actually languages or dialects.

The picture of the languages has changed somewhat over the years. Older sources (e.g. John DeFrancis, The Chinese Language, Fact and Fantasy, Hawaii, 1984; S. Robert Ramsey, The Languages of China, Princeton, 1987; and Nathan Sivin, editor, The Contemporary Atlas of China, Houghton Mifflin, 1988) say that there are seven different languages, or six, since sometimes Gan is linked with Hakka, or with Xiang. More recently, Lynn Pan, in The Encyclopedia of the Chinese Overseas [Harvard, 1999], lists ten languages, where Jin is separated from Mandarin, Hui from Wu, and Pinghua from Yue. Now, however, in The Sino-Tibetan Languages, edited by Graham Thurgood and Randy J. LaPolla [Routledge Language Family Series, Routledge, London, 2003], Jerry Norman ("The Chinese Dialects: Phonology") states, "If one takes mutual intelligibility as the criterion for defining the difference between dialect and language, then one would have to recognize not eight [or seven, etc.] but hundreds of 'languages' in China" [p.72]. This appears to resolve the issue. What previously were regarded as separate languages, like Cantonese, are in fact families of languages. It is therefore not surprising that the "splitters" (those who like to divide groups, as opposed to "lumpers," who like to combine groups -- a typological difference) should begin to divide the old languages into new ones. If there are really "hundreds" of languages involved, however, further splitting becomes pointless.

On the map at left, we see China of the late Empire divided by the ethnic principle of the "five peoples." While the Hui, , might be Turks or Uighurs, the term in general means "Muslims" and thus applies to ethnic Chinese Muslims. Those Hui speak Mandarin and tend to live in the area identified for the Han, , People on the map. Otherwise, the dialects of Chinese all refer to languages of the Han People. Manchurian has all but disappeared and been replaced by Mandarin.

Within each of the groups of Chinese languages, there are also true dialects, which means that they are mutually intelligible. In Pan's book and The Sino-Tibetan Languages many dialects are shown for the language groups. The confusion over all this -- couldn't everyone tell what forms of speech are mutually intelligible? -- was certainly due to the difficulties of doing research in China in the 20th century. From revolution, to war, to revolution, to totalitarianism, China until recently was not the best place for graduate students wandering around with tape recorders asking strange questions. Such behavior would often have evoked suspicion, arrest, or worse. Of course, there is also the problem of distinguishing dialects from languages in general, when dialects may be intelligible to those nearby, while those at the extreme ends of a range may be incomprehensible to each other.

The table gives a classification of languages and dialects based on a combination of The Sino-Tibetan Languages and other sources. The 10 languages identified on the map from Pan's The Encyclopedia of the Chinese Overseas are given in boldface; but the overall organization is in terms of the three groups and six "dialect familes" of The Sino-Tibetan Languages [p.6]. While Gan and Xiang and now definitely separated, Hakka has come to be included under Gan -- though this is not consistently seen in the book.

"Hakka" itself is an interesting term, in Mandarin, in Cantonese (客家, haak3gaa1), meaning "guest, visitor, traveller, stranger, merchant," or "customer." Althought there is a concentrated area of Hakka speakers, the language is otherwise spoken in widely scattered areas, where it has been taken by, indeed, Hakka traders.

"Mandarin" is a word from Sanskrit (, mantrin) by way of Malay (menteri) and Portuguese (mandarim). This meant "counselor." The word was applied because the Portuguese were originally dealing with traders along the southern coast of China, where, of course, many languages were spoken, but not Mandarin. When officials from the Capital came down to deal with the Portuguese, they spoke a different language, which the Portuguese had not otherwise encountered. Hence the name, the language of the "counselors". However, this may also have simply been a translation of what the counselors were calling their own language, which was the , "Official Language," or even "Language of the Officials," i.e. the Mandarins [note].

In Modern Chinese, official Mandarin is the , "Common Language," or the , "National Language." These same expressions are used in Cantonese, pronounced differently of course, where we also find , the "Beijing Language," 北京話, pronounced Bak1-ging1-waa6 -- or Beg7ging1wa6, or Bùk-gìng-wa̿ (see variety of Cantonese transcriptions below, including the a/e/u alternation we see here). As we have seen, is literally used for everything else in the country, whether language or dialect.

Some population figures are given for the older seven language classification. These are given as percentages of the total Chinese speaking population, as a number in millions (M), and, from another source, as a number in thousands (k). These count those for whom the languages are their first languages. The figure of 952,000,000 speakers for Mandarin given elsewhere is for people who speak Mandarin at all. This is considerably larger than the 715 million number below, not just because the population has grown in the last twenty years, but also because Mandarin in the national language of China, taught in schools around the country. Areas where the languages are spoken are given after the language name(s). Names of cities and provinces in Pinyin are given in italics. I have now added new population figures, after a dash, which are taken from The World Almanac and Book of Facts 2008 [World Almanac Books, 2008, p.728]. The Almanac gives the first figures I've seen for Puxian, which is now evidently often broken off of Min, as Jin is broken off of Mandarin.

Dialect FamilyInitialsFinalsTonesSyllables
Mandarin, 163942496
Gan, 195966726
Hakka, 176967038
Xiang, 233765106
Min, 155775985
Wu/Shanghai, 275079450
Yue/Cantonese, 205399540

It is noteworthy that the extension of Mandarin into the Southwest was in part the result of veterans being settled there after the Mongols were ejected from China and the Ming Dynasty founded.

The table is a comparison of dialect families from The Sino-Tibetan Languages [p.127]. The statistics, of course, are from representative languages in each group. I have rearranged the list to move the apparently more conservative languages towards the bottom of the table, though, of course, not all the indications are consistent. With the largest number of tones and of syllables, Cantonese wins as the most conservative, but then Xiang and Shanghai both have more initials than Cantonese -- and Hakka has an anomalously large number of finals and syllables. Mandarin has clearly undergone the greatest phonetic simplification.

For some idea of how the languages different, the character for "south" in Mandarin is , which, we see, is pronounced nán -- occurring in the table above and in the names of many dynasties. The same character and word becomes in Cantonese, in Hakka, in Southern Min, in Northern Min, in Eastern Min, and in Gan. In speech, one would be at a loss to identify most of these as the same word. With Gan, not a single letter is the same. With this going on, it is not hard to understand how the Chinese "dialects" are different languages. The character is borrowed into Japanese as nan (minami in Japanese) and into Vietnamese as nam (expressed as hướng nam, literally "south direction").

Back in the 1990's, I had a student from Singapore in one of my Introduction to Philosophy classes. She was a delightful person and enthusiastic student and decided that we should take Cantonese at UCLA. I had to decline the offer. I later learned that she had been "Miss Singapore" at some point. When the class got to Chinese philosophy, and I talked about the different "dialects" of Chinese, it turned out that the form of Chinese she learned growing up in Singapore was not Mandarin, but she did not know what it was. I don't think we were able to figure that out at the time.

Now I see that, although Singapore has officially been promoting Mandarin, Wikipedia says that among Chinese languages, "Hokkien (Min Nan) used to be an unofficial language of business until the 1980s. Hokkien is also used as a lingua franca among Chinese Singaporeans, and also among Malays and Indians to communicate with the Chinese majority." So my student almost certainly grew up speaking Hokkien.

The actress Michelle Yeoh (b.1962), well known for her role in Crouching Tiger, Hidden Dragon [2000], was herself born in Malaysia from a family of "Hokkien and Cantonese ancestry." She must have grown up speaking Hokkien and perhaps Cantonese also. She had to learn Mandarin for her part in the movie.

Hokkien is glossed as a version of Southern Min, . My older sources identified this as "Amoy-Swatow" and as based in the city of Xiamen, as in the list above. Wikipedia now says that Hokkien originated in "part of Fujian Province in Southeastern China and [is] spoken widely there. It is also spoken widely in Taiwan, where it is usually known as Taiwanese or Holo, and by the Chinese diaspora in Malaysia, Singapore, Indonesia, the Philippines and other parts of Southeast Asia and by other overseas Chinese all over the world." "Hokkien" itself is the Southern Min pronunciation of Mandarin Fujian, .

We then see that "The Amoy dialect is the main dialect spoken in the Chinese city of Xiamen (formerly romanized and natively pronounced as 'Amoy') and its surrounding regions of Tong'an and Xiang'an, both of which are now included in the greater Xiamen area."

In turn, we hear about Swatow thus: "The Shantou dialect, formerly known as the Swatow dialect, is a Chinese dialect mostly spoken in Shantou in Guangdong, China. It is a dialect of Chaoshan Min language." In turn, we then hear that Chaoshan "is a Southern Min language spoken by the Teochew people of the Chaoshan region of eastern Guangdong province, China, and by their diaspora around the world. It is closely related to Hokkien, with which it shares some cognates and phonology, though the two are largely mutually unintelligible."

Thus, we find that the umbrella of "Southern Min" encompases, not just different dialects, but a family of different languages. So my student perhaps would have been understood in Taiwan, but maybe not in some adjacent Mainland areas.

The language map for all of the Min languages is from the Chinese book, The Language Atlas of China [1987, 1989]. The language map above distingishes Puxian from the Min languages, but this map, and my other sources, do not. It is surrounded by Southern and Eastern Min.

Fukien Chinese Tones
Yīn Level, 55Yīn Leaving, 212
high levellow falling and rising
Yáng Level, 53Yáng Leaving, 242
high fallingmiddle rising and falling
Rising Tone, 33
middle level
Yīn Entering, 24Yáng Entering, 5
middle rising stoppedhigh level stopped
The treatment of the seven tones here is taken from a book, Fujianese Dictonary & Phrasebook [Hippocrene Books, 2014]. The book does not give an author, but it has a credit for "translated by" for Xiao Chu Chen. "Fujian," of course, is also called "Fukien" or "Hokkien," and is a branch of Southern Min, as we see on the map.

In the table, I have retained traditional characters; but Fukian uses simplied characters. Thus, 陰, yīn, occurs as 阴; and 陽, yáng, as 阳. The 3rd tone contains the character 聲, shēng; but this occurs in Fujian as 声. The other characters are traditional.

The Wikipedia page on the Fuzhou dialect uses the traditional characters. It does not number the tones and gives them in a slightly different sequence than Fujian. I have used the Wikpedia page for the numerical contours given for the tones. Wikipedia and Fujian use the same "description" for each tone, which frequently are not what one would expect from the traditional Chinese names, e.g the 3rd tone, called "rising" is actually level. If diacritics were to be used for the tones, the hachek, used in Pinyin, would be appropriate for the falling and rising 4th tone, while the circumflex would be appropriate for the rising and falling 5th tone.

Categories of Chinese Characters

Examples of Dialect Differences Between Peking, Shanghai and, Canton

Pronouncing Mandarin Initials

Mandarin Finals and Syllables

The Contrast between Classical and Modern Chinese

Philosophy of Science, Linguistics

History of Philosophy, Chinese Philosophy

History of Philosophy

Philosophy of History

Home Page

Copyright (c) 2000, 2006, 2007, 2008, 2017, 2021, 2022 Kelley L. Ross, Ph.D. All Rights Reserved

The Dialects of Chinese, Note

The word "Mandarin" has also been explained as derived from Chinese, as , "Manchu great man" [cf. Dah-an Ho, "The Characteristics of Mandarin Dialects," The Sino-Tibetan Languages, p.127]. However, this looks very much like a folk etymology, and an anachronistic one, since the Portuguese had been in China more than a century (since 1518) before the Manchus took over the country (1644). The language of the officials was going to be called something long before any officials were Manchurian.

There is also the problem of the pronunciation:   probably was not pronounced with an r in the era in question. The Wade-Giles writing of the syllable, jen, reflects an older pronunciation, which we see reflected as a y in Cantonese and an actual English-like j in Japanese. Indeed, this is probably why "Japan," , is pronounced in English as it is, with an older Chinese pronunciation -- in Japanese itself, the j/y/r can and does turn up here an n.

I have now found some good evidence of the anachronism, as I suggest, of this claim. "Mandarin" was used in reference to Chinese officials as early as 1552 by the Portuguese writer Fernão Lopez de Castanheda in his Historia do descobrimento e conquista da India. The text is cited by Yule & Burnell in their classic A Glossary of Colloquial Anglo-Indian Words and Phrases ["Hobson-Jobson," Curzon Press, 1886, 1985, "Mandarin," p.550-551].

Return to Text

Examples of Dialect Differences Between
Peking, Shanghai and, Canton

p-pu1 "wave"po1 [bō]
p'-p'u1 "slope"p'o1 [pō]
b-bu2 "old woman"p'o2 [pó]
t-tong1 "east"tong1 [dōng]
t'-t'ong1 "be open"t'ong1 [tōng]
d-dong2 "be alike"t'ong2 [tóng]
k-kuong1 "light"kuang1 [guāng]
k'-k'uong1 "frame"k'uang1 [kuāng]
g-guong2 "mad, wild"k'uang2 [kuáng]
-t/0kat7a "cough"k'e2 (sou4) [ké(sòu)]
-t/0pat7a "brush"pi3 [bǐ]
-t/0yüt7b/8 "moon"yüeh4 [yuè], 月
-t/0yat8/9 "sun, day"jih4 [rì], 日
-k/0paak7b/8 "hundred"pai3 [bǎi]
-k/0sik7a "color"(yen2)se4 [(yán)sè]
"national language"
kou23 [guóyǔ]
-p/0t'aap7b "pagoda"t'a3 [tǎ]
-p/0yap8/9 "enter"ju4 [rù], 入
-p/0sap8/9 "ten"shih2 [shí], 十
As we have seen above, Cantonese and Shanghai are really separate languages, distinct from the prestige and politically favored Mandarin of Peking. Cantonese, or Yuè, , is the language of Guǎngdōng, 廣東 (Kwang-tung), and Guǎngxī, 廣西 (Kwang-hsi), provinces. Shanghai, or Wú, , is the language of the city of Shànghǎi, 上海, and surrounding provinces.

Cantonese, also, is particularly the language associated wtih the capital of Guǎngdōng Province, namely Canton. That name, of course, derives from the name of the province. The city itself in Chinese is Guǎngzhōu, 廣州. The politically correct practice these days, of course, is to use the unpronounceable local name of a place, rather than the traditional name already familiar in another langugage. Viewers thus may encounter documentaries or news stories that avoid the name "Canton" altogether. However, those untrained in Chinese, or linguistics, will be unable to pronounce any of these Chinese names with the appropriate phonology.

In the table superscript numbers are the tones for Shanghai, Cantonese, and Peking Mandarin in the Wade-Giles system. Brackets contain Pinyin writings, which now on this page have been updated with the proper diacritics in Unicode. While Pinyin and the system from Teach Yourself Cantonese [1976] use diacritics to indicate the tones, other common transcription systems for Cantonese, except the Yale, use superscript numbers. Since many diacritics are now easily represented in Unicode, I don't know why superscript numbers still dominate.

The examples are from The Languages of China, by S. Robert Ramsey [Princeton University Press, 1987]. The Shanghai/Peking examples are from p.91; and the Cantonese/Peking examples from pages 102 & 105. If there are two numbers for the tones, the first is from this book, while the second is what we find, recounted below, with more recent sources.

The Wu () dialect of Shanghai is noteworthy because it retains the distinction between voiced and unvoiced, aspirated and unaspirated stops that existed in T'ang Chinese. In Mandarin and the other languages the voiced stops have disappeared. In these examples, the voiced stops have been assimilated to the aspirated ones.

Cantonese () is noteworthy because it retains from T'ang Chinese a greater variety of finals. In Mandarin, a syllable must end in a vowel or in n or ng. In Cantonese, syllables can also end in p, t, k, or m as well. Words borrowed from Chinese into Korean, Japanese, and Vietnamese often also preserve evidence of the older final consonants. Thus "China," 中國 (Mandarin Zhōngguó, "Middle Country"), in Korean is Chung-guk and in Japanese Chū-koku (Japanese syllables only end in a vowel or "n"). Both of them have an extra consonant in "country" where Mandarin doesn't -- but Cantonese (Jòong-gwōk, Zung1gwok3) does.

I had a lingustics professor once who said that you could get a kind of "instant Proto-Indo-European" by combining Greek vowels and Sanskrit consonants. Well, we can get a kind of "instant T'ang Chinese" by combining Shanghai initials and Cantonese finals. The evidence is poor for older versions of Chinese. Cantonese also preserves the larger number of tones that T'ang Chinese probably had. Mandarin only has four now, but Cantonese has six, or even nine if the tones of finals that end in stops are counted separately, which they usually are. The problem of the tones in T'ang Chinese is discussed below.
Mandarin Tones
High LevelRising

The most daring theory is that the Chinese of Confucius's day didn't even have tones. Evidence for this is that other members of the Sino-Tibetan language family do not have tones, while the nearby family of the Daic languages (like Thai) all have tones. In another adjacent language family, the Austroasiatic (Mon-Khmer) group, some languages have tones (like Vietnamese) and others do not. It is tempting to see the phenomenon as a South-East Asian Sprach Bund where the Daic tones have influenced some languages in the Sino-Tibetan and Austroasiatic families.
Cantonese Tones
1stHigh Rising,
司, 詩,
2ndMiddle Rising
3rdMiddle Levelsi̅
4thLow Falling
5thLow Risingsĭ
6thLow Levelsi̿事, 是
7thHigh Levelsìk色, 識
8thMiddle Levelsi̅k
9thLow Levelsi̿k

At left are examples of the Cantonese tones, using the notation in Teach Yourself Cantonese by R. Bruce [Teach Yourself Books, Hodder and Stoughton, 1970, 1976, pp.12-13]. Different tone symbols are not needed for the 7th, 8th, and 9th tones. Just as the diacritics for the 1st, 3rd, and 6th tones can be used for the 7th, 8th, and 9th, other systems actually use those numbers.

In other treatments, such as my source for the table above, The Languages of China, the 7th and 8th tones are styled 7a and 7b, while the 9th tone becomes the 8th. One case occurs with the character , "sun, day." This is in Mandarin. It was given in this source for the table as yat8.

The same word is transcribed ya̿t in Teach Yourself Cantonese, yed9 in A Concise Cantonese-English Dictionary, yaht in the English-Cantonese Dictonary, Cantonese in Yale Romanization [by Kwan Choi Wah, et al., The Chinese University Press, Hong Kong, 1991], and jat6 at Wiktionary -- where this matches the transcription in the ABC Cantonese-English Comprehensive Dictionary [Robert Bauer, University of Hawai'i Press, 2021]. Since the Yale notation (with h but no diacritic) indicates the 6th tone, these are all consistent, with either the 6th or 9th tone. We thus recognize that yat8 is actually the 9th tone, or elsewhere written, with a diacritic or otherwise, for the 6th.

The 7th, 8th, and 9th tones are all level. This is only a little confusing with the 7th, which writes the tone for the 1st, which itself is a little confusing, often called either rising or falling. But, as we will see below, the 1st tone rises slightly and then falls. In the 7th, the falling part is lost and the rising part flattened.

While this table originally used graphic images for the characters and the diacritics, I have now been able to substitute Unicode writings for all of them. The single and double overlines for two of the diacritics, and the breve for the 5th tone, are derived from the Uncode block of "Combining Diacrtical Marks" (#x0300). The breve with "i" can also be written "ĭ" from the "Latin Extended-A" block (#x0100).

Since the characters are now in Unicode, they can be copied and pasted into search engines, where each can be found with a Wiktionary page that gives information about meaning, pronunciation in different Chinese dialects, etc. The only really novel diacritic here is the double overline, but other transcription systems seem to be innocent even of traditional marks like the breve. Conspicuously missing is the venerable circumflex, already easily rendered in HTML, and which we might think particularly suitable for the 2nd tone. I don't know why it would not be used.
Cantonese Tones
Yīn LevelYīn HighYīn Leaving
Yáng LevelYáng HighYáng Leaving
High Yīn
Low Yīn

There seems to be a reluctance in many current systems to take full advantage of representing the tones with dedicated diacritics, even though Pinyin made full use of them, and Unicode easily makes many diacritics available.

The Chinese names of the Cantonese tones, at right, are themselves of interest. It is an exercise in the use of Yīn & Yáng. The first three tones are Yīn, 陰, and the second three Yáng, 陽. The 7th and 8th tones refer back to the 1st and 3rd, which are Yīn, while the 9th refers back to the 6th, which is Yáng. Because the 7th and 8th tones are both Yīn ("upper" and "lower"), this may be why some systems have called them "7a" and "7b." The "entering" tones, 入 (Mandarin , Cantonese ya̿p), have always gone with syllables closed with "p," "t," and "k," which is why we don't see them in Mandarin, where such syllables are gone.

Below left is a diagram off of Wikipedia showing the sound contours of the tones in pitch and duration, for the tones from the 1st to the 6th [Alexander L. Francis, 2008]. It is easy to match up the brief glosses from the tables with the actual tones. The characters used for the tones include the examples from Teach Yourself Cantonese, and I have added characters from the Cantonese page at Wikipedia. These are not the names of the tones but examples of characters with the given pronunciation.

The character , si̿, "affair, matter, undertaking, business," in particular catches my attention because of its importance in Chinese philosophy, correponding, among other things, to Greek πρᾶγμα. In Mandarin is shì, in Japanese ji (kun reading koto), in Vietnamese sự, and in Korean sa, [A Guide to Korean Characters, Reading and Writing Hangŭl and Hanja, by Bruce K. Grant, Hollym International, Elizabeth, NJ, Seoul, 1979, p.95]. Where has an extra meaning as a verb, "to serve," the Vietnamese-English Dictionary, by Nguyễn Ðình-Hoà, has separate entries for the two terms [Tuttle Language Library, 1966, 1991, pp.398-399]. Although sự is spelled in the same way, you would not know from that alone that they were from the same Chinese character.

These example words in the table will look different in A Concise Cantonese-English Dictionary by Yang Mingxin [Guangdong Higher Education Publishing House, 1999]. First of all, the latter uses an adapted Pinyin alphabet, where "x" is used for "s" and "g" for final "k." Second, although Pinyin introduced the use of Greek-like accents to show tones, the Dictionary reverts to the old Wade-Giles way of simply numbering the tones with superscripts. Also, the Dictionary uses simplified forms of some of the characters. I have used the unsimplified characters in Bruce where these are available. The Yale system of Romanization, with discussion of some alternatives (though not the Pinyin) is used in the English-Cantonese Dictionary.

Dictionaries or grammars of Shanghai Chinese in English seem to all be out of print. Grammar, however, can be found in many pages on line.

A nice example of a difference between Mandarin and Cantonese is a surname. This is in the former, Ňg in the latter (and actually O or Oh in Korean). The Cantonese name is one of many words that are simply a syllabic ng. There is also a syllabic in Cantonese, which is , "not," in Mandarin. That is the only word with that pronunciation in A Concise Cantonese-English Dictionary [pp.260-262]. Although it seems like there ought to be, there is no syllabic n in Cantonese.
There is more than one character used for the Cantonese surname. At right, we see the traditional character first, then a recent simplified one to the right of the pronunciation.

This was also the name of the Kingdom of Wu, one of the states of the Three Kingdoms Period in Chinese history, and of the modern language of Shanghai. At far right is an alternative character used, at least in Cantonese, for the surname. My only question is that the first character (with its simplification) and the second are pronounced differently. In Mandarin, the first has a 2nd tone, the second a 3rd. In Cantonese, the first has a 4th tone, , Ňg, the second a 5th, , N̆g (with the symbols used in Teach Yourself Cantonese). The contours are similar but the tone level a little different.

I originally learned of the two possible characters from a young woman whose name actually was Ng, but I didn't know then to ask about the different tones. Perhaps someone can help me out.

Note that the Cantonese spellings in the table above are from Teach Yourself Cantonese, while, as noted, A Concise Cantonese-English Dictionary uses a form of Pinyin adapted from Mandarin. Thus, words traditionally ending in t/k/p are written d/g/b in the latter.

Cantonese has a large variety of vowels and diphthongs. These are transcribed in confusingly different ways in different systems. Thus, the character , meaning the "Buddha, Buddhism," etc., is in Mandarin, butsu in Japanese, phật in Vietnamese, and bul, 불, in Korean [op.cit., p.78].

For Cantonese, is transcribed fu̿t in Teach Yourself Cantonese, fed9 in A Concise Cantonese-English Dictionary, faht in the English-Cantonese Dictonary, Cantonese in Yale Romanization [The Chinese University Press, Hong Kong, 1991], and fat6 at Wiktionary.

While we might not be surprised that the 9th tone in one source turned up as the 6th in another, it may be more puzzling why the very same vowel should be represented as "u," "e," and "a" in different sources. The actual vowel may be closer to the "schwa," /ə/, which is a reduced and indefinite vowel in relation to a full "u," "e," or "a." Unless a schwa is actually used, it is hard to say who has the best idea in this. Note that the Yale system tries to indicate the tones by diacritics with slightly different versions of the vowels -- "ah" without a diacritic is the 6th tone on "a."

The Yale system is also now used in the Pocket Cantonese Dictionary, Cantonese-English, English-Cantonese, by Martha Lam and Lee Hoi Ming [Periplus, Hong Kong, 2019], and the Complete Cantonese, by Hugh Baker and Ho Pui-Kei [Teach Yourself, Hackette, 2020]. Complete Cantonese has decided that there are seven tones, apparently by dividing the 1st into "high level tone" and "high falling tone," where the "high level" actually would be confined to what the other sources regard, indeed, as the actual 7th tone. But the tones are not numbered.

The Pocket Cantonese Dictionary confines itself to six tones, with the traditional numbering. The character there is the Yale faht, which seems to be missing from the vocabulary of Complete Cantonese. The connection of Complete Cantonese to the "Teach Yourself" books is all but buried, and the interesting system of the 1976 book is gone.

Restaurant Name in Cantonese

I have had some difficulty pinning down the number of tones in Shanghai Chinese. In the table above, we see Shanghai credited with seven tones [The Sino-Tibetan Languages, edited by Graham Thurgood and Randy J. LaPolla, Routledge Language Family Series, Routledge, London, 2003, p.127]; but at Wikipedia, I see five tones, as in the table at right. The most interesting feature of the Wikipedia treatment
Shanghai Chinese Tones
the T'ang Tones
Yīn, 陰1st, 53
2nd, 34
4th, 5ʔ
Yáng, 陽
3rd, 13
5th, 2ʔ
may be that the Yáng tones occur after voiced initials and the Yīn tones after unvoiced.

Over the years, I have had difficulty finding books in print about Shanghai Chinese. Now, however, I have in hand Shanghainese, Dictionary & Phrasebook, by Richard VanNess Simmons [Hippocrene Books, 2011], and A Chinese-English Dictionary of the Wu Dialect (Featuring the Dialect of the City of Shanghai), by Thomas Creamer [Dunwoody Press, 1991]. Shanghainese features a discussion of the tones, which, while giving five in number, nevertheless breaks them down into a total of eight overall. A Chinese-English Dictionary represents the tones with numerical contours, from 1 to 5. Thus, "53" is a tone that begins high, "5," and falls to mid-level, "3." This is a device commonly used, even where tones may otherwise be distinctively numbered.

While those numbers are the ones used in A Chinese-English Dictionary of the Wu Dialect, they are slightly different from what we find on the Shanghainese page at Wikipedia, where we have "52" for "53," "14" for "13," "44ʔ" for "5ʔ," and "24ʔ" for "2ʔ." These variations in sources are not unusual. What seems a little anomalous is that the "abrupt" Entering tones are represented as having the same duration as the other tones, while the dictionary has them cut off short.

Shanghai Chinese Tones
Yīn, 陰1a, High
2, Rising
& Falling
4a, Abrupt
5a, Level
Yáng, 陽1b, Low
4b, Abrupt
5b, Level
I have updated the table above in light of the analysis of the tones of Shanghai Chinese in Shanghainese. In three cases, we get a differentiation that occurs between voiced and unvoiced initials. With the 2nd tone, we don't seem to see this difference, but the 2nd tone is said to only occur in "words with three or more syllables" [p.14]. This is very intriguing since it begins to sound like the Greek case of only one tone occurring for a whole polysyllabic word.

This impression is reinforced in the discussion of the 3rd tone, where we see "all words with a falling accent (`) over the first syllable have a falling tone that spreads across the whole word (starting in the first syllable)" [ibid.]. With the "entering" tones, they occur, where we might expect, with "q" finals (glottal stops that indicate where the finals used to be "p," "t," or "k"), but they also occur in a polysyllabic word without "q," where the following syllable will retain its own tone.

I have not given numerical contours for the tones in this table because I have not found simple versions of them. The Shanghainese Wikipedia page has a large table for "tone sandhi" (Sanskrit sandhi for sound changes from one syllable to the next) across multiple syllables; but this has two numbers for each syllable, up to five syllables, which is not the simplicity I'm looking for. And there are anomalies, such as the 1st tone described as "rising," when the contour, "53," shows it falling. I don't have an explanation for that.

This raises all sorts of questions about the Wu/Shanghai language(s). There is even an issue concerning the voiced initials, which may also be aspirated. So it is a shame not to see more books in print about it. This may be because, unlike Cantonese, Shanghai is really just a spoken language. The written language in the place is Mandarin.

Pronouncing Mandarin Initials

Mandarin Finals and Syllables

Dialects of Chinese

Philosophy of Science, Linguistics

History of Philosophy, Chinese Philosophy

History of Philosophy

Philosophy of History

Home Page

Copyright (c) 2000, 2005, 2006, 2015, 2019, 2020, 2021, 2022 Kelley L. Ross, Ph.D. All Rights Reserved

Pronouncing Mandarin Initials

Chinese has the extraordinary structure that nearly every syllable has a semantic content, even if only a historical one. Each syllable is thus written with a Chinese character, which was originally a separate word.

Each syllable is analyzed into an "intitial" and a "final." The "final" contains the vowel, the tone, and the final consonant, if any. This structure is also applied to
Simple Initials
bpp, unaspirated (spot)
pp'ph, aspirated (pot)
dtt, unaspirated (stop)
tt'th, aspirated (top)
gkk, unaspirated (skit)
kk'kh, aspirated (kit)
Sibilant Initials
ztsts, unaspirated
ctsths, aspirated (hats)
Retroflex Initials
zhchṭṣ, unaspirated
chch'hṣ, aspirated
Palatal Initials
jchtš, unaspirated
qch'thš, aspirated (church)
Korean and Vietnamese, which borrowed Chinese writing and many Chinese words, even though neither language was even related to Chinese.

The "initials," apart from the tones, pose the greatest challenge for foreigners trying to pronounce Chinese. And now we have two common systems for writing Mandarin, the older Wade-Giles and the recent Pinyin. The greatest challenge is that Mandarin does not have voiced stops, like b, d, and g. These existed in T'ang Chinese (and have been preserved in the Shanghai or Wu language), but have been lost in Mandarin. Instead, Mandarin contrasts aspirated stops with unaspirated stops. "Aspirates" have breath coming out, "unaspirates" don't. In Wade-Giles, aspirates were indicated with an apostrophe, as in the name of the T'ang Dynasty. Sometimes it is said that an aspirated t is pronounced like the t in "hot house." This not quite right, since the t there is in a separate syllable, and a separate word, from the "h" aspiration. Instead, it should be noted that English contrasts, in certain environments, an aspirated from an unaspirated t. Thus the t in "top" is aspirated, and the t in "stop" is unaspirated. Holding a hand in front of the mouth can detect the breath expelled in one and not expelled in the other. The Chinese unaspirated t can be duplicated by pronouncing "stop" without the "s." Aspirations are indicated in the "pronunciation" column of the table with a superscript h.

Since there are no voiced stops in Mandarin, the Pinyin system conveniently uses the Latin letters for the voiced stops for unaspriated stops, and the Latin letters for the unvoiced stops for the aspirated stops. The English word "stop" thus could be written in Pinyin as "sdob," which looks very odd, and has a final consonant unallowed by Mandarin, but does use the proper values of the Pinyin consonants.

The "retroflex" initials have the tongue curling up, as in the similar series of sounds in Sanskrit and subsequent languages in India. But other Chinese dialects do not distinguish retroflex from palatal initials. In fact, even in Mandarin, retroflexes and palatals are really just different allophones (sounds) of the same phonemes, i.e. they do not occur in the same environment and so can actually be represented by the same signs (as in Wade-Giles). Retroflexes (and sibilants) occur only with a, o/e, and u finals. Palatals occur only with i and ü finals. The "i" written with sibilants and retroflexes, e.g. "si" and "zhi," does not represent a true i, but a "buzzing" for sibiliants and an r for retroflexes.

The Wade-Giles system represents Chinese more efficiently and familiarly. Pinyin, besides the phonemic redundancy, has the drawback that the sound of a number of letters (like q and x) has nothing to do with how they are pronounced in most Western languages. On the other hand, Pinyin makes a more elegant use of the Latin alphabet.

Mandarin Finals and Syllables

Dialects of Chinese

Philosophy of Science, Linguistics

History of Philosophy, Chinese Philosophy

History of Philosophy

Philosophy of History

Home Page

Copyright (c) 2000, 2006, 2019 Kelley L. Ross, Ph.D. All Rights Reserved

Mandarin Finals and Syllables

Simple Initials and Group-a Finals
Ø aanangaiao
f fafanfang 
Retroflex & Sibilant Initials and Group-a Finals
rri ranrang rao
Each syllable in Chinese is analyzed into an "intitial" and a "final." Initials of Mandarin are considered in the section
above. The "final" contains the vowel, the tone, and the final consonant, if any. The tables here show nearly all the possible syllables in the Standard form of Mandarin Chinese, i.e. the Mandarin of Peking (Beijing). This is not actually all the syllables because of the "Group-r" finals. Those are added either as "er" or "r" to the syllables shown. After a, o, e, u, and ng, "r" is added. After ai, an, and en, drop the i or n and add "r." After i and ü, add "er." and With "i," in, and un, drop the i or n and add "er."

The "Group-a" finals go with the simple, the retroflex, and the sibilant initials. The "i" final only occurs with the retrolex and sibilant initials, and represents a vowel with little kinship to an actual i. For the retroflexes, it is more of an r sound, while with the sibilants it is a vowel so reduced and indefinite that it is described as a "buzzing." Indeed, in the Yale system of transcription, the former is rendered with "r" and the latter with "z." Wade-Giles uses "ih" (or "tzu" for Pinyin zi, etc.), thus distinguishing it from the simple "i" used with the palatals. In this way, neither Pinyin nor Wade-Giles give much of a clue from English phonology how to make the sound. Since the "i" is the only letter i that is not used with the "Group-i" finals and the palatal initials, its presence rather confuses the symmetry of the system, although there is no ambiguity (I will not say confusion), since "i" does only occur with the retroflex and sibilant initials. It is a cleaner and more elegant solution than in Wade-Giles. Since Pinyin was willing to pick phonetic values of the Latin alphabet from different languages, the undotted Turkish I might have been considered for the "i" sound, though this is not available in HTML and is, as noted, unnecessary.

Otherwise, the vowels in the table are as they are in Wade-Giles. The syllabic m in included in the table just as a reminder that there is such a thing in Cantonese. In the index row, the tone is written over the vowel to show, where there might be ambiguity, which vowel is used.

Simple Initials and Group-o/e Finals
Ø eeneng ou 
bbo benbengbei 
mmo menmengmeimou
d de dengdeidoudong
tteteng toutong
n nennengneinounong
Retroflex & Sibilant Initials and Group-o/e Finals
zh zhezhenzhengzheizhouzhong
chchechencheng chouchong
rrerenreng rourong
ccecenceng coucong
With the "Group o/e" finals a major difference between Pinyin and Wade-Giles is that the latter writes the "ong" final as "ung." Since one may be used to seeing words like "Chung" in English, its absence from Pinyin is conspicuous.

Of priniciple interest here in the phonetic system is the lack of contrast between the o and e finals. Where the final o is used, e is not; and where e is used, o is not. That this was not always the case is shown with two anomalous syllables against a blue background. Pe and ho used to occur, but they do no longer. The only minimal pairs with o/e are those with contrasting eng and ong finals, though there are a good number of these.

The pe syllable is found in the name "Peking," which now, with the Pinyin Beijing being used, people might just think of as some kind of mistake. It is not a mistake, just a transcription of an older form of pronunication in Mandarin, where pe existed, and where the palatals in the "Group-i" finals had not yet developed from their original stops -- the word is still king in Cantonese and was borrowed as kyô into Japanese.

A difference between Pinyin and Wade-Giles that would also apply to the "Group-a" finals above is the initial r. In Wade-Giles, that is written j, which, pronounced r, must produce for Wade-Giles as much confusion as q and x in Pinyin. Again, this reflects some history. Since the r corresponds to a y in Cantonese (yat for ), and is often borrowed as (English) j into Japanese (e.g. jin for rén), writing j in Wade-Giles reflects the circumstance that this is pronounced y in German but j in English (the y pronunciation being the original value of j as a modification of Latin i). However, the letter is also borrowed as n into Japanese (e.g. nichi for ), and r itself does not look much like a natural derivative of either y or j. So there seems to have been something else going on in the original Chinese sound, which may have been more an ñ than a y.

Simple Initials and Group-u Finals
ddu duo duiduandun 
nnunuo nuan 
Retroflex & Sibilant Initials and Group-u Finals
chchu chuochuaichuichuanchunchuang
rru ruo ruiruanrun 
In the Group-u finals, uo often turns up as just o in Wade-Giles. Otherwise, we see a lot of possible syllables that are not used. A curiosity in both systems is that ui is actually pronounced more like ué (with the accent from French). Wei is written more like it is pronounced (with the anomaly that the tone goes on the e). That all this is the case may be because the Mandarin e in isolation has more of the reduced, schwa-like sound that is familar from many occurrences in English (the last a in "banana"), French (le), and German (Töne). We don't get a pure Italian e or French é in Mandarin.

The "Group-ü" finals feature the vowel ü, written and pronounced like the u-Umlaut in German (also used now in Turkish). This is the sound i with lip-rounding, and so, being a front vowel like i, is found with the palatal initials of the "Group-i" vowels, as given below.

Where Wade-Giles did not distinguish between retroflex and palatal initials with different letters, it did so by the circumstance that the palatals only occurred with "Group-i" and "Group-ü" finals. Thus the ü was always fully written.
Simple and Palatal Initials and Group-ü Finals

Since Pinyin does differentiate the initials with different letters, the need for the Umlaut, to separate "Group-u" from "Group-ü" finals, is mostly eliminated. However, some writers do not seem to realize that this is not universally the case. Where the initials are n or l, the Umlaut is still necessary. Thus, lü is sometimes improperly written as lu in Pinyin.

The retention of the Umlaut does create some graphic difficulties, since the tone must be written atop it in nü and lü, something that fonts may not often be called upon to do. Otherwise, its loss is a convenient simplification.
Mandarin Tones
High LevelRising
Unicode, however, allows us to use the tones even with nü and lü.

As we see here, at a convenient point, Mandarin syllables are distinguished by four tones, or five, if we count the "drop tone," where a syllable may be left without a tone. The tone is considered part of the Final, along with the other contents of this page.

Wide-Giles had indicated the tones with numerical superscripts, which was a cumbersome device and and easily lost for most non-linguistic purposes. Pinyin wisely returned to the origin of tone notation, in Classical Greek, and begins to write the tones with the equivalent of the accents that were invented to write Greek. The differences we see are the use of the macron for the high even tone, the 1st, while the 3rd tone, which is falling and rising, flips the Greek circumflex, which had been rising and falling, to show this. The result is the use of the Czech háček, which is applied to a vowel, rather than its customary use on Slavic consonants. Although this is now very convenient, we still rarely see a name like "Běijīng" written with the requisite tones.

Simple Initials and Group-i Finals
bbi biaobie bianbin bing 
ddidiaodiediudian ding
ttitiaotie tianting
Palatal Initials and Group-i Finals
With the "Group-i" finals, we see a number of systematic differences between Pinyin and Wade-Giles. Ian here turns up as ien in Wade-Giles, and iong as iung. Although written ian, the a is a reduced vowel pronounced still more like the e discussed above.

We also see the most unfamiliar use of letters in Pinyin, with q for Wade-giles ch' and x for hs -- which itself was simply an alternative to sh. X actually is used to write sh in some languages (e.g. Basque). I am not aware of q being used anywhere to write any variation of English ch. However, whether intentional or not, this evokes a bit of the history, since q usually is pronounced like k, and q in Pinyin is used with an initial that, although now a ch, was actually an original k. If that was the intention, in the use of q, it was cleverly done.

Pronouncing Mandarin Initials

Categories of Chinese Characters

Dialects of Chinese

Philosophy of Science, Linguistics

History of Philosophy, Chinese Philosophy

History of Philosophy

Philosophy of History

Home Page

Copyright (c) 2006, 2021 Kelley L. Ross, Ph.D. All Rights Reserved

The Contrast between Classical
and Modern Chinese

Although both ancient and modern Chinese are mostly written with the same characters, the modern daughter languages have become very different from the ancient one. One of the most conspicious differences is just that the terse, monosyllabic nature of Classical Chinese -- , "old writing," or , "literary language" -- has given way to many more particles, polysyllabic words, and periphrastic idioms.

The following story, given in both Classical Chinese and a translation into modern Mandarin, -- or the , "colloquial speech, vernacular" -- illustrates the difference. This is also a salutary example for one's view of government, as Confucius indeed makes clear to his students [I am unaware of the provenance of this text].

The modern Mandarin pronunciation is given for the Classical characters because the ancient pronuncation, indeed the pronunciation before the T'ang Dynasty, is unknown. Even that of the T'ang is reconstructed and uncertain. The extreme simplification of Mandarin phonology, which would render the Classical language ambiguous if used as a spoken language today (too many words now being pronounced the same), explains the polysyllablic character of the modern language and the reduction of many characters to morphemes.

Once when Confucius was passing near the foot of Mount Tai in a chariot, there was a married woman weeping at a grave mound, and dolorously too. Confucius politely rested his hands on the front rail of the chariot and listened to her weeping. He sent Zilu (Tzu-lu) to inquire of her, saying; "From the sound of your weeping, it seems that you indeed have many troubles."

Classical Chinese:

Mandarin Translation:

Then the woman said; "It is true. My father-in-law died in a tiger's jaw; my husband also died there. Now, my son has also died there." Confucius said, "Why do you not leave this place?" The woman said: "Here there is no harsh and oppressive government."

Classical Chinese:

Mandarin Translation:

Confucius said, "Young men, take note of this: a harsh and oppressive government is more ferocious and fearsome than even a tiger."
Classical Chinese:

Mandarin Translation:

Classical Chinese Read in Other Languages

The same Classical text that can today be read as Mandarin could as well be read with Korean, Vietnamese, or Japanese versions of the Chinese words, or the Korean, Vietnamese, or Japanese translations of the words. None of those languages is even related to Chinese, but since mediaeval, or even modern, Koreans, Vietnamese, and Japanese often wrote in Chinese, without, however, really speaking the language, their own renderings of the characters was customary. Since the ancient pronunciation of the Classical language is unknown (see below), Sino-Korean, Sino-Vietnamese, and Sino-Japanese readings are really just as "authentic" for Classical Chinese as a Modern Mandarin reading. Indeed, much of our evidence for the T'ang pronuncation of Chinese is from the Korean, Vietnamese, and Japanese readings, which were contemporary borrowings.

For example, the character for "mountain," now read shān in Mandarin, turns up as san in Korean, in Vietnamese as sơn or núi, and in Japanese as san, sen, zan, or yama -- the last versions in Vietnamese and Japanese being the native words.

Similarly, we find the name of Japan itself, "Sun Source," as Rìběn [Wade-Giles Jihpên] in Mandarin, Yatbóon in Cantonese, Ilbon in Korean, Nhật-Bản in Vietnamese, and Nippon or Nihon in Japanese. The Cantonese word is, of course, cognate to the Mandarin. The Korean, Vietnamese, and Japanese are all borrowings from Chinese, pronounced in the local manner.

Native words for "sun" are hae in Korean, mặt giời ("face of the sky") in Vietnamese, and hi in Japanese (e.g. hi-no-maru, "circle of the sun," "sundisk"). The Japanese borrowed word for "sun" in isolation is nichi, but this is just the pronunciation of niti, where the final i as been added because Japanese syllables cannot end in t. In compounds, the i can drop out, so nichi-hon (*hi-moto in the [largely] unused pure Japanese reading) becomes nit-hon. At that point different things can happen. The t can be lost in assimilation to the h, getting us Nihon, OR the h can revert to its original p, with the t getting assimilated and doubled with it, getting us Nippon.

Another example concerns the present capital of Japan. The Míng capitals of China were Nánjīng (Nanking) and then Běijīng (Peking), which simply mean, respectively, "Southern Capital" and "Northern Capital." The capital of Japan from 794 to 1868 was Kyôto, which meant "Capital District." Then the capital was moved to Edo, which was renamed the "Eastern Capital." In Chinese that would be Dōngjīng. In Japanese, however, that is pronounced Tôkyô. In Vietnamese it is Ðông-Kinh (or Tonkin). The Vietnamese version preserves more of the Chinese consonants, but both Japanese and Vietnamese versions reveal that "capital" originally started with a k, which has become palatalized (to a j) in Mandarin. The k is also preserved in early modern Western versions of Chinese names, like "Nanking" and "Peking" themselves -- whose use the politically correct now have rejected because of the idea that they are "wrong" and that the local pronunciation of place names must be used -- despite such people generally being unable to correctly pronounce Nanjing or Beijing and thoughtlessly continuing to say "Rome" instead of Roma, which has been the local pronunciation of the name of that city in Italian and Latin for more than two thousand years.

The Denial that Classical Chinese Was a Spoken Language

Chinese departments in colleges sometimes expect students to learn Mandarin even though they only want to read Classical Chinese or Sino-Korean, Sino-Vietnamese, or Sino-Japanese. This imposes a vast unnecessary burden on them, with the all the challenges of the tones and phonology of the modern language, but even some teachers and scholars of Chinese sometimes have trouble accepting that the ancient language is not the modern one and that the ancient language is part of the civilization of Korea, Vietnam, and Japan as much as of modern China. It is as though students of Latin were told they would have to learn Italian as well, even if they were Spanish or French.

The curious idea that something like Mandarin was already an ancient spoken language and that Classical Chinese is some sort of abbreviation or code derived from it can be found in various sources. For instance, Joanna C. Lee and Ken Smith, in their translations from the Analects, assert:

First of all, Confucius almost certainly didn't "say" these things, since written Chinese is scarcely a direct transcript of spoken language but a fundamentally different system with its own inner logic and grammatical structure. [The Pocket Confucius, Museworks, Hong Kong, 2010, p.9]

This is rather like a claim that the Romans already spoke Italian and that Latin was an artificial language only used in writing. Indeed, I have heard people say that Classical Latin could never have been a spoken language, because it is too difficult. This should give Russian, let alone Georgian, speakers a good laugh.

Now, Classical languages undergo their own development over time and diverge from their oral sources. But when this happens, we usually have texts attesting the original language and can follow the changes. Thus, Classical Sanskrit can be distinguished from Vedic Sanskrit, which has more in common with Old Persian and thus was certainly the original spoken language, although we cannot rule out some garbling in transmission, since documentary sources are late. Mediaeval Latin slowly evolved from Classical Latin, but the preservation of the older literature, like Cicero, made it possible to write a "purified" Latin prose during the Renaissance. Much the same thing happened in Greek.

But if we know that the texts of the Confucian corpus are in some sort of artificial language, a "fundamentally different system," it is hard to know what older literature or evidence is used to make this claim. There isn't any. Lee and Smith should reflect that if Classical Chinese has "its own inner logic and grammatical structure," that is because it is a different language, as different from Mandarin as Latin is from French or or Anglo-Saxon is from modern English. And it is not in the least surprising that the language Confucius spoke more than two thousand years ago should be quite different from any modern language. But if Beowulf had been written in ideograms that are still used to write modern English, the student could at least get the drift of the story, even if it would all look rather strange. That is what we are dealing with in Chinese.

While with Lee and Smith we get what is more or less a parenthetical comment by people who are not linguistic specialists, that is not the case in the treatment of Roger T. Ames and Henry Rosemont, Jr. in their The Analects of Confucius: A Philosophical Translation [Ballantine Books, Random House, 1998, 1999], which I have discussed as a translation elsewhere. Ames and Rosemont are academic professionals in philosophy, Chinese philosophy, and the Chinese language. However, they subscribe to this remarkable thesis that written Chinese was not in origin a spoken language -- "classical Chinese is a unique linguistic medium" [p.41] -- and, unlike Lee and Smith, they argue the issue at length with references to the literature.

Nevertheless, they cannot overcome the weakness of the case that separately attested examples of the spoken language do not exist for the Shang or even the Chou Dynasties, while the comparative evidence from Sumer and Egypt is that systems of writing develop as reflexes of speech. After noting that in Confucius, "virtually every passage is ambiguous," they say:

Many Western scholars have of course called attention -- often loudly -- to this ambiguity and lack of precision in classical Chinese, seeting it as a distinctive linguistic liability...

In those instances where detail or exactness of expression was necessary, we might assume Chinese thinkers availed themselves of their spoken language, wherein there is every reason to believe as much precision was as is possible as can be achieved in any other natural language. [p.42]

So the argument of Ames and Rosemont seems to be that because the written language of Confucius is ambiguous, by our standards, there must have existed a spoken language that was less ambiguous. This does not follow; and Ames and Rosemont cannot cite attested examples of "their spoken language" from the earlier periods.

And, for all their academic qualifications in philosophy, Ames and Rosemont don't acknowledge the ambiguity of moral and philosophical terms in other languages -- or any examples of other historical languages where an ambiguous written language corresponds to a less ambiguous spoken language.

Thus, the Egyptians had one famous word, , "Maat" or Muꜣꜥe, that can be translated "truth" or "justice," or the name of a goddess (of Truth and Justice). In turn, the Greek word δίκαιος, basically meaning "just" (Latin justus), is often better translated "right" (Latin rectus, although jus retains this sense also). We can clarify these ambiguities now because of a larger vocabulary, which Egyptian (or even Greek or Latin) clearly didn't have.

The Greek word λόγος has a range of meanings from "word" (Latin verbum) to "reason" (Latin ratio), such as to deceive David Bellos into thinking that Greek didn't have a word for "word." Indeed, I'm not sure that Greek actually has a word that just means "good will," Latin benevolentia. I have discussed this elsewhere. Modern French, at the same time, has a word, droit, that can be a "right" or a "duty," which morally and legally are actually opposites.

So Classical Chinese is not alone in posing the challenges of interpretation and translation that it does, especially as an ancient language; and the broad meaning or ambiguity of terms in Confucius, such as , "propriety, etiquette, good manners, ritual, the rites," is not unusal, even in some modern languages. At the same time, Chinese has always been able to resolve ambiguity, where desired, with expressions of two or more characters. Thus, my wife, who reads Classical Chinese, unlike me, likes to say, "There is no concept that cannot be expressed in four Chinese characters."

The "Philosophical Translation" of the Analects, by Ames and Rosemont

The Phonology of Classical Chinese

Chinese is certainly unique among modern languages with its continued use of ideographic symbols, so that each syllable of the language, even if it is now never used alone, has its own meaning. Japanese, as formerly was the case with Korean and Vietnamese, piggybacks on this system, combining Chinese characters with phonetic symbols, the kana syllabaries, to indicate inflections, particles, foreign words, etc. While Korean for a long time did something similar, Chinese has not resorted to this expedient

But the phonetic simplification of the Chinese language means that most symbols cannot be used alone in speech, because too many syllables sound alike -- they are homophones. Thus, Ames and Rosemont say:

This homophony is unusual among languages, but has existed in Chinese almost since its inception. Phonetically, most consonantal endings of syllables have dropped off over the centuries, but even when they were present, the number of homonyms was very high, with anywhere from two to seven different characters pronounced identically. [p.38]

Now, what is the evidence that this homophony "has existed in Chinese almost since its inception"? Ames and Rosemont cite no reference for this statement; and in fact separate examples of the spoken language are not attested for anything prior to the language of the Mediaeval era of the Sui and the T'ang. Yet the claim that "anywhere from two to seven different characters" were "pronounced identically" is something that would require a knowledge of the phonology of spoken Chinese that is totally absent for the Ancient language -- it is also senseless when we consider the implication that such confusing homophony existed in the spoken language, upon which Ames and Rosemont wish to rely to disambiguate the written.

Recent remarks about this situation can be found in A Student's Dictionary of Classical and Medieval Chinese, by Paul W. Kroll [Brill, Leiden, 2017]. The dictionary includes reconstructed readings for characters in Middle Chinese, i.e. the language from the period of the Northern and Southern Empire down through the T'ang Dynasty. Kroll explains why we do not see versions of the earlier language, i.e. "Old Chinese," as he defines it, from the Warring States Period to the Later Han:

Why, one may ask, are only the Middle Chinese readings supplied in the dictionary, and not the Old Chinese readings too? The simple answer is that there is a relatively stable consensus about Middle Chinese readings (though specialists argue about certain points of detail), whereas the linguistic reconstruction of Old Chinese continues to generate broad dispute and widely divergent representational models. Moreover, Old Chinese itself is in large part a back-projection from the established phonological values of Middle Chinese and is thus largely an abstract outgrowth (or perhaps we might say a pre-growth) of it. Given the relative certainty of Middle Chinese reconstruction and the relative uncertainty of Old Chinese, it has therefore seemed the safest course in a dictionary meant for general use to provide only the Middle Chinese readings for words. [pp.xii-xiii]

Noteworthy may be the cuttoff used by Kroll for Old Chinese, which does not even include the language used by Confucius in the age prior to the Warring States, i.e. the Spring and Autumn Period -- with the caution that even ideas about Old Chinese are not based on contemporary evidence, but only on the "back-projection" reconstruction from Middle Chinese. Translators of Confucius, whether Lee and Smith or Ames and Rosemont, thus do not have a clue, any clue, about the phonetics of his spoken language.

As we have seen above, Modern Mandarin allows 2496 possible syllables, but Shanghai (Wu) Chinese allows 9450 and Cantonese (Yue) 9540. And, while Ames and Rosemont refer to Mandarin losing its "consonantal endings," they have overlooked the circumstance that the Shanghai language preserves voiced initials, which have been lost elsewhere. Thus, since we can reconstruct something like the Chinese of the T'ang Dynasty from the current languages, we should expect at least half again as many syllables as are allowed by Cantonese, which has a large array of finals, but has lost initial voicing like Mandarin.

So perhaps T'ang Chinese had at least 14,310 possible syllables. That is already enough for a lot of unambiguous monosyllabic speech, and we are already more than a thousand years after Confucius (479 BC-618 AD), a lapse of time that puts the English language back to Beowulf -- and without a phonetic script holding a spoken language closer to a written language, we know that spoken languages can change quickly indeed.
T'ang Chinese Tones

A difficulty in that respect is the belief that T'ang Chinese only had four tones. This would substantially reduce the number of possible syllables, to perhaps fewer than we might want for semantic adequacy. Since these tones are based on contemporary descriptions, it might seem difficult to question it. However, it is a very strange business. The phonetic simplification of Mandarin in comparison to the other Chinese langauges is dramatic, in all respects, including with the tones. If we accept that the number of Mandarin tones have not been reduced since the T'ang Dynasty, the problem is that we must then explain why everyone else has so many more tones than Mandarin.

It seems like there would be at least four explanations:

  1. All the sister languages independently multiplied their own tones. This seems the least likely, when the overall tendency for Chinese languages is simplication, and it is hard to understand how this impulse would affect all the non-Mandarian languages equally and simultaneously.

  2. They are derived from an older version of Chinese that antedates the tone simplication that is attested for T'ang Chinese. This would require postulating that T'ang Chinese is not the ancestor of all the modern Chinese languages. Yet it is reconstructed using the evidence of all of them, and the only problem that seems to arise is about the tones.

  3. They derive the multiplication of their tones because of influence of the adjacent languages, in the family of the Thai-Lao group, for instance, that are all tonal, often with many more tones that the Chinese languages. This is a feature of the South-East Asian Sprachbund, where unrelated languages adopt features from each other. Thus, Vietnamese, now with tones, may not have had them originally. And,

  4. The T'ang analysis into four tones may have been incomplete, classifying what were different tones together. Both Mandarin the Cantonese classify tones with a Yīn/Yáng, 陰/陽, differentiation, doubling the number. With a "high/low," 上/下, division, as in Cantonese, the tones could be doubled again. This all depends on how acute we think that Mediaeval Chinese phonologists were. Indeed, if they were all that acute, we might expect the use of a phonetic system to represent all Chinese sounds. Just such a thing was being imported from India, where Sanskrit had been analyzed at a level no worse than in Modern linguistics. But that doesn't happen in China, leaving us to reconstruct the pronunciation of Chinese words from later and external evidence. This may give us some pause in accepting the completeness of the Chinese analysis of tones.

I cannot say which of these explanations is the more likely.
Shanghai Chinese Tones
the T'ang Tones
Yīn, 陰1st
Yáng, 陽
But a clue may be in the tones of Shanghai/Wu Chinese. Some sources say that Shanghai has seven or eight tones; but the Wikipedia page on Shanghai says five, which are arranged as in the chart at right. These are in relation to the tones of T'ang Chinese, with a Yīn/Yáng differentiation, where the Yáng tones occur after voiced initials and the Yīn tones after unvoiced. Shanghai, of course, preserves voiced initials (13 of them), most of which the other Chinese languages have lost. See further discussion of Shanghai tones above.

This is just the kind of thing we might wonder would have happened in T'ang Chinese, with the potential for no less than eight tones with a full doubling. The T'ang phonologists need not have recognized that doubling because the duplicated tones would be allophones, phonetically distinguished by their environment, the nature of the initials, not intrinsically. This all gives some weight to the 4th explanation for the puzzle of the number of tones from the T'ang to the present.

Indeed, the Wikipedia page on "Middle Chinese" says just this, that the four Classical tones were doubled to eight for just the reason considered here.

I might note that while Shanghai doesn't have the finals ending in p/t/k, it does have the "entering" tones because these sounds survive in reduced form, as glottal stops (IPA ʔ), which can be distinctively transcribed with a "q."

At the same time, this is all somewhat separate from the problem of the thesis that written Ancient Chinese was not a spoken language. There are many ways in which the language of Confucius can have been monosyllabic, even if it didn't have tones at all. Ames and Rosemont confidently claim to know things about Chinese "since its inception," and this confidence, let alone any such claim, is unwarranted.

Any thought that the spoken language of the Chou had not significantly changed by the time of the T'ang is ruled out by the comparative evidence of every language with an attested history of such length. The extraordinary claim of Modern Greeks that the pronunciation of their language has not changed since ancient times has been examined elsewhere. Ames and Rosemont, and whatever authorities they are using [Karlgen, Keightly, etc.], have no recourse but an argumentum ad ignorantiam, by which they make inferences from things that we do not know.

Classical Chinese, Ideographic vs. Phonetic

The irony of a thesis such as Ames and Rosemont support is that it destroys the principle behind the practice of Chinese departments that students studying Classical Chinese can do so only by also learning Mandarin. This is salutary, as we can affirm that Mandarin, except for the convenient pronunciation it provides for characters (Cantonese, etc. could do the same), is irrelevant to the Classical language.

Also, Ames and Rosemont defend the ideographic nature of Chinese characters and their independence, in the subsequent history of the written languages, from the spoken languages. This puts them at odds, as they realize [p.285], with some of the fads in linguistics, where spoken language is not only accorded a primacy, but the independence of the written language may be denied and Classical languages, which can exist entirely independent of contemporary spoken languages, positively disparaged, a matter I have addressed. Thus, Ames and Rosemont say:

A number of scholars have focused on the phonograms [i.e. the "phonetic" element in the "radical+phonetic" form of Chinese characters], however, and consequently have not given sufficient notice to the semantic properties of the graphs, dismissing the use of such terms as "ideographs" to denote them. John DeFrancis, for example, says that:

It should be apparent that there is much justification for considering the Chinese scipt to be basically -- that is, more than anything else -- a phonetic system of writing.

Similarly, William Hannas maintains that:

We can dismiss the fanciful notion that the units are icons of objects and concepts in the real and psychological worlds, i.e. that the symbols are pictographic.

We also reject the untenable assumption that Chinese characters are "ideographic," that is, relate to meaning directly without the intervention of language. [pp.295-296]

In this, Ames and Rosemont are more correct than DeFrancis and Hannas, although the existence of "radical+phonetic" characters, and their eventually total dominance in the vocabulary (97% by the time of the Ch'ing), does attest to the continued interaction of the written with the spoken language, which Ames and Rosemont may not sufficiently appreciate. But the idea that language is essentially spoken is itself an "untenable assumption" in terms of what should seem like the obvious fact that anything can be a symbol of meaning, from spoken words, to written words and other symbols (as in mathematics or street signs), to the forms of the hands used in sign languages (whose functional completeness as languages was long denied, to the grievous harm of the deaf). If a written language is meant to be read, and if written language is used to record speech, then there is going to be an intimate bond between them. But we have no difficulty discerning that the media create a different dynamic, and written languages also come to embody a conservativism that spoken languages easily escape.

Thus, Ames and Rosemont are correct that there has been "an independent life for the classical written language," probably even at "a very early state of Chinese history" [p.291], without, however, it's having somehow originated independently of the spoken language. This just quoted sentence continues:

...in that the excavated inscriptions are written on divinatory or other materials that are religious, not secular, in nature. That is to say, the archaeological materials at hand show clearly that whatever other uses it might have had, early writing was intimately bound up with ritual religious practices. [ibid.]

On the one hand, this is not surprising, since much of ancient life is going to be "intimately bound up with ritual religious practices"; but, on the other hand, Ames and Rosemont overinterpret the significance of this record. Thus, the survival of oracle bones and shells is no more surprising than the absence of "secular" materials on perishable media. In the flood plain of the Huang He River, early records on bamboo or wood will have disappeared, while early inscriptions on stone, which become common in Egypt, are about as rare as they are in Mesopotamia, where there were no nearby sources of stone, for construction or anything else.

In Mesopotamia, however, the adoption of clay tablets (i.e. mud pies) for ordinary writing enables us to see the bias in other archaeological records. Sumerian written documents are overwhelmingly commercial. Indeed, enough remains that it has been possible to reconstruct the origin of cuneiform writing from the earliest humble records of inventories, packing and shipping, debts, payments, and taxes. It is pretty dull stuff, but also revealing and persuasive. There is nothing equivalent from the early literate civilizations of Egypt, India, China, or Meso-America, where documents never used a material like clay, at once so humble and ubiquitous but also potentially (when fired) so durable. And it means that, looking at Chinese oracle bones and pronouncing that this is how Chinese writing began, perhaps because the Chinese "mind" was immune to commercial culture and its needs, is preposterous.

A revealing case may be in India. No commercial documents survive. Indeed, the only writing that does survive is mostly on cylinder seals, with usually no more than five characters in an inscription. But this does tell us something. We know from Sumerian records that there was a lot of trade with India. And the device of cylinder seals originated with the Sumerians. What did they use them for? Well, they sign or seal documents. Like commercial documents. Thus, it is pretty obvious what they were being used for in India, even if the documents don't survive. In Egypt, you signed with pen and ink, and the devices for signature thus disappeared with the documents. The nice thing about cylinder seals is that they could be used with equal facility on both clay tablets and, with ink, on other kinds of media.

It is a sensible hypothesis that all writing began in the same way; and the presence of numbers in all writing systems is itself evidence that, even were the use of writing to be primarily "bound up with ritual religious practices," those practices themselves rely on payments and commodities that fund and support the priests, temples, rituals, and sacrifices. Certainly Chinese seers did not perform their oracles for free. Even much later Egyptian temples, such as that of Ramesses III at Medinet Habu, feature long, long, long lists of offerings and requisitions brought to the temple, lovingly categorized, counted, and illustrated. This attests that much of the daily function of Egyptian religion was itself "ritual religious practices" built on a foundation of commercial culture, wealth, and accounting.

Thus, while I am happy to agree with Ames and Rosemont that written language, and certainly the Classical Chinese language, comes to lead an independent existence, the archaeological and literary record does not warrant their conclusion that the written language, and Chinese characters, led an independent, non-verbal existence from their inception. And the examples where they attempt to divine the meaning of characters from their graphic and visual constituents, or cite others attempting to do the same, as with , "benevolence" [p.48], amount to nothing better than folk etymologies, poorly motivating any interpretation, let alone the strange "authoritative conduct" reading that Ames and Rosemont offer for .
Ideograms vs. "Logograms"

Categories of Chinese Characters

Dialects of Chinese

Classical Languages

Philosophy of Science, Linguistics

History of Philosophy, Chinese Philosophy

History of Philosophy

Philosophy of History

Home Page

Copyright (c) 2000, 2008, 2016, 2020, 2021, 2022 Kelley L. Ross, Ph.D. All Rights Reserved