In the vast territory of the Chinese Republic are found not only many dialects but also many languages of diverse structures, spoken by peoples of different cultures and customs. This constitutes a problem in which students of politics and education have to be interested. The National Language Movement with its hope of linguistic unification, the simplification of Chinese writing, the Romanization Movement, the giving of an orthography to language not having a writing of their own, the possibility of providing reading material as an aid to mass education, for speakers of languages and dialects very different from the national language—all these are problems which require a thorough knowledge of the linguistic situation in China.
But this condition in China is a very complicated one. Aside from the Chinese with its numerous dialects, there are many other languages, our knowledge of which is of an extremely uncertain quality. Some of them have been adequately studied, some of them are scarcely known to us, and many of them have not been sufficiently recorded. The material of these languages is there fore scanty, their history unknown, and their relation with other groups very vaguely understood For this reason we need not only specialists in Chinese dialects to clarify the dialect situation in China, but also specialists in other linguistic stocks to furnish us with material for a better classification of the languages now spoken in China and for a better knowledge of the distribution of these languages.
The following description of the languages in China groups them into branches and families, and names under each group some of those characteristics which distinguish it from the other groups. We need, therefore, a few words of introduction about the classification of these languages. Languages are classified into families, with the assumption that they are historically connected. Among a group of languages we find some similarities or correspondences in the word forms, in the grammatical elements such as the prefixes, the suffixes, the infixes, the vocalic alternations, etc., and in the general structure. These similarities or correspondences sometimes can be formulated into definite and exact statements. Particularly in the domain of sounds these statements are known as sound laws. With the relations among them thus formulated, it is apparent that these similarities cannot be due to chance or mere borrowing, but are due to the fact that these languages are descendants of a common parent speech. As time goes on, this parent speech splits into various dialects and through successive evolutions will develop into such different languages as English, Russian and Bengali. Sometimes the changes that these languages have undergone are far-reaching. The farther they have differentiated from each other the more difficult it is to trace their relations, unless ancient documents which reveal the older state are available. With languages which have no records and which are known only in modern forms, we encounter a great difficulty in establishing their relations. Many languages in China not only do not possess records of their own, but are known to us very fragmentarily. Their classification is therefore tentative.
INDO-CHINESE FAMILY
One of the largest families of speech in China is known as the Indo-Chinese (Tibeto-Chinese or Sinitic) family. Languages of this family are spoken throughout China Proper and Tibet, and extend into the Three Eastern Provinces (Manchuria) and Sinkiang, as well as to places outside of China such as French Indo-China, Burma, and Siam. One of the characteristics of this family is the tendency towards monosyllabicism. The Tibeto-Burman branch of this family still possesses some of the prefixes, sometimes asyllabic and sometimes syllabic, but the Chinese and the Tai Group have early lost all active use of the prefixes. It is probable that prefixes, suffixes, vocalic changes, and consonantal alternations have been in use in primitive Indo-Chinese to form causatives, denominatives, and to modify the meaning of the stem in general such as in classical Tibetan: hgns-pa, bkan (perfect), dgan (future), khon (imperative) "to fill," but such derivative processes have long been dead in most languages of this family, so that we have only inflexible monosyllabic stems such as are found in Chinese now.
The tendency to develop a system of tones is another characteristic of this family. We do not know whether tones existed in the primitive Indo-Chinese speech, and it is doubtful whether tones existed in classical Tibetan, but modern Chinese, modern Tibetan, Burmese and the Tai languages all have tones. These tones are further influenced by the nature of the initial consonant and are divided into two main categories: those with an original voiced initial consonant and those with an original voiceless initial consonant. Such has been found to be the case with Chinese, Tibetan, Burmese and the Tai languages, and is considered as the most powerful argument for the common origin of these languages.
Another phonetic tendency in common with this family of languages is the unvoicing of the original voiced initial consonants but this has not been carried out in all dialects. It happens in most Chinese dialects, except the Wu and the Hsiang in practically all the Tai languages and in many Tibeto-Burman languages, so that what was originally a voiced initial is only detected in the nature of the tone.
Aside from these there are of course many points in the vocabulary which seem to be common in this family. Exact correspondences of sounds, however, have not been worked out. Four main branches are known in this family: Chinese, Tai, Miao-Yao and Tibeto-Burman.
A. Chinese.—Chinese is the most important member of this family. The earliest records consist of numerous bone and tortoise shell inscriptions of around 1400 B. C. which are being systematically excavated in Honan. The reading of these inscriptions still presents many difficulties. Our knowledge of the archaic phonological system is largely derived from a study of the rimes in archaic texts, principally the Shih Ching, and of the phonetic compounds of the written characters. It has been shown that initial consonant clusters such as gl-, bl-, ml-, etc., and many final consonants such as -b, -d, -g, -p, -t, -k, etc., existed in archaic Chinese, but up to about 600 A.D., when we have the system of ancient Chinese well represented by the rime books such as the tsie-yun, the initial consonant clusters were already simplified and final -b, -d, -g, dropped. From the time on the Chinese language has gone through a series of evolutions such as the unvoicing of the initial sonants, the dropping of final -p, -t, -k, and the simplification of rimes. The modern Peiping dialect, which is taken as the national language (kuoyu) has now only some 400 possible syllables; and each syllable may have theoretically four tones. This phonetic simplification, which causes the existence of many homonyms, is counterbalanced by a great increase in the use of compounds, so that what was formerly expressed by one syllable must now be expressed in the colloquial by two or more syllables.
We may divide the Chinese dialects into the following groups:
1. The Northern Mandarin Group occupies a large area in North China, in the provinces of Hopei, Shansi, Shensi, Kansu, Honan, Shantung and extends into Sinkiang, Inner Mongolia and Manchuria in the north and into Hupeh, Anhwei and Kiangsu in the south. It is characterized by the unvoicing of the ancient voiced stops, affricates and spirants, and by the disappearance of the "entering tone." There are as a rule only four tones: ying-ping, yang-ping, shang and chu. The Peiping dialects (Pekinese) is the best-known of all the dialects. Further division into smaller groups is possible.
2. The Eastern Mandarin Group is spoken along the lower Yangtze Valley in the provinces of Anhwei and Kiangsu. It is differentiated from the Northern Group by the existence of the "entering tone" as a short tone, but the original final consonants -p, -t and -k, which accompanied the "entering tone," are lost. It presents therefore five tones.
3. The Southwestern Mandarin Group is a very uniform type of speech spoken in Szechwan, Yunnan, Kweichow and parts of Hupeh and Kwangsi. It has no "entering tone" just as the Northern Group.
4. The Wu Group of dialects is spoken south of the Yangtze in Kiangsu, Chekiang and in the east (in a few districts) of Kiangsi. It is characterized by the preservation of the ancient voiced stops, etc., as aspirated voiced consonants and by the preservation of the "entering tone," with the loss, however, of the final -p, -t, -k. It often presents six or seven tones.
5. The Kan-Hakka Group is spoken principally in the provinces of Kiangsi and Kwangtung. It is characterized by the change of the ancient voiced stops, etc., into aspirated surds in all four tones (aspirated only in ping-sheng in the three Mandarin groups). The "entering tone" is preserved and the final -p, -t, -k are more or less preserved according to dialects, and there are often six or seven tones. The Northern or Kan Group has the tendency to pronounce all aspirated surds as voiced in connected speech. The Hakka group preserves the final consonants, such as -m, -p, -t, -k, much better. Settlements of Hakka people can be found in various districts in K wangtung and Kwangsi.
6. The Min Group can be further divided into two subgroups. The Northern Group is spoken in the northern part of Fukien and the Southern Group is spoken in the southern part of Fukien, in the eastern part of Kwangtung, in Hainan Island and in parts of the Leichow Peninsula. It is characterized by the change of the original voiced stops, etc., as unaspirated surds, even in ping-sheng where the aspirated pronunciation is the prevalent one, by the preservation of the original palatal plosives t', t'c, d'c, as dental plosives, and by the preservation of final -p, -t, -k (sometimes in modified forms). It has as a rule seven tones. The Hainan dialects present many phonetic, peculiarities, possibly under the influences of an aboriginal speech, presumably a Tai language.
7. The Cantonese Group is spoken in the provinces of Kwangtung and Kwangsi. It is characterized by the preservation of the final consonants -m, -p, -t, -k. It presents a system of eight, nine or more tones. The distinction of long and short vowels as in Cantonese is also a special feature. Certain distinctions of tone depend on the length of the vowel.
8. The Hsiang Group is spoken principally in Hunan. The ancient voiced stops, etc., are as a rule kept as truly voiced consonants; the final -p, -t, -k are usually lost, but it often presents a system of six or seven tones.
9. Certain Isolated Groups, such as the dialects spoken in the southern part of Anhwei, certain dialects in Hunan and in the northeastern part of Kwangsi may be mentioned here.
Aside from the phonological features specific to the groups mentioned above, there are also elements of vocabulary more or less peculiar to each of these groups, but these are too minute to be gone into here. Among these various groups some are mutually intelligible, while some are quite unintelligible.
B. Tai.— The Tai group of languages is another important branch of the Indo-Chinese family, closely related to the Chinese. The Tai languages possess four tone categories analogous to the ping, shang, chu and ju of Chinese. These four tone categories are each further divided into two according to whether the initial consonant was originally voiced (low initial) or voiceless (high initial), so that the modern Tai languages often possess eight and sometimes nine or more tones as a further development according to vocalic lengths (the writer has counted those tones with a final -p, -t, or -k separately). It has two special plosives ('b-, 'd-, or m-, l- according to dialects) which are treated as high initials. Initial consonant clusters such as kl-, pl-, etc., are preserved by some dialects to this day, i.e., Siamese, and the Tai language of Wuming and Lungan in Kwangsi, but original voiced stops, etc., have practically all become voiceless in the modern dialects. Word order in Tai is also slightly different from the Chinese, the modifying word following the modified. For instance, "good man" in Chinese becomes "man good" in Tai The earliest Tai monument is an inscription of RamaKamheng of Siam in the 13th century. In China most of the Tai languages have no writing of their own, except some in Yunnan which employ either the Shan alphabet (derived from Burmese) or one closely related to the Southern Tai alphabet.
We may divide the Tai branch into two main divisions:
1. The Chuang Group consists of many dialects spoken in a great part of Kwangsi and in the southern part of Kweichow. The language of the Shu Li, spoken in the northern part (Linkao, Chengmai, and Kiungshan,) of Hainan Island also belongs here, but the Li dialects in the center and in the south of the island seem to be quite different from the ordinary Tai languages. Their relation to this branch is therefore doubtful. The languages of this group are characterized by the lack of aspirated surds such as p'-, t'-, k'-, the preservation of the distinction between original k'- and x-, g- and y-, and by the preservation of an original high initial (some sort of r-) as r- (in Wuming), l- (in Tienchow), s- (in Dioi of Kweichow), or y- (in Chienkiang), which appears as h- in Shan, Siamese, Lao, Nung, etc. The development of vowels also shows many peculiar features.
2. The Southwestern Group consists of some of the best known of the Tai languages and lies mostly outside of China. We may divide this group into four subdivisions: (l) Ahom, once spoken in Assam, (2) Kamti and Shan, etc., spoken in Burma and in Yunnan, (3) Siamese and Lao spoken in Siam and French Indo-China, (4) Tai Blanc, Nung, Tho, etc., spoken in French Indo-China and in the southwestern part of Kwangsi. This group is characterized by the preservation of all aspirated consonants such as p'-, t'-, k'-, by the charge of the original guttural a spirants into stops, by the appearance of the high liquid initial (some sort of r-) as h- (except Ahom where r- is preserved), and by a very uniform system of vocalic correspondences. In Yunnan, aside from the Shan dialects, some dialects of group (3) and of group (4) may be spoken in the southern part of the province, while some of the Chuang dialects may exist in the eastern part. The writer is also informed that a Tai dialect is spoken in the northwestern part of Hunan, but the details are not available.
C. Miao-Yao—The Miao-Yao branch of the Indo-Chinese family is monosyllabic like the Chinese and is known to possess tones, but the phonologic system and the relation between the Miao and Yao are not sufficiently known. This grouping is therefore tentative. Word order resembles the Tai. It is spoken by fairly primitive groups of mountaineers. Aside from the occasional use of Chinese characters, there is no writing of its own.
1. The Miao Group is spoken under various tribal names in the southwestern mountain regions of Hunan, in a large part of Kweichow, in the northeast of Kwangsi, and is found scattered here and there among the mountains of Yunnan, French Indo-China and Siam. It is characterized by the dropping of the final consonants, so that only -ng and rarely -n are allowed to stand in final positions. The number of tones varies from five (cf. M.F. Savina, Dictionnaire Miao-tseu-Francais) to eight (cf. Esquirol, Dictionnaire 'kanao-Francais et Francais-'ka-nao).
2. The Yao Group is also spoken under various tribal names in the northwestern mountain regions of Kwangtung and is scattered among the various mountain regions of Kwangsi, Yunnan, French Indo-China and Siam. It preserves the final consonant better than the Miao; final -m, -n, -ng, -p, -t, -k are all allowable. It is greatly influenced by the Tai and the Chinese.
D. Tibeto-Burman—This branch of the Indo-Chinese family is one which presents most clearly the use of prefixes, whether syllabic or asyllabic, alternations of voiced and voiceless consonants, and the use of suffixes such as presented by the classical Tibetan. Tones depend upon whether the initial is voiced or voiceless and are further influenced by the existence of prefixes, but the system of tones seems to be much simpler than that of Chinese or Tai (cf. Y. R. Chao's "Phonetic Notes to Love Songs of the Sixth Dalai Lama Tshangs dbyangs-rgya-mtsho," Monograph Series A. No.5 of the Institute of History and Philology, Academia Sinica). Word order is as a rule subject-object-verb, in contrast to the Chinese and Tai type of subject-verb-object. Four divisions are known:
1. The Tibetan Group is spoken principally in Tibet and Sikang, and extends into Chinghai and the western part of Szechwan. The earliest record of this group is dated from the ninth century, the alphabet being derived from the devanagari form of the Hindu alphabet. A great amount of literature, largely Buddhistic, exists. Three main groups of dialects may be distinguished. The Western Group, Balti, Ladak, etc., preserves more or less the prefixes, the initial consonant clusters, and the final stops -b, -d, -g, The Central dialects, including the dialect of Lhasa, are characterized by the loss of prefixes, by the simplification of consonant clusters, and the dropping of final consonants. The Eastern dialects, the Khams, preserve very faithfully the prefixes and the final consonants. Belonging to the Tibetan Group are some Tibeto-Himalayan dialects and some North Assam dialects spoken along the southern border of Tibet. Some form of the Tibetan dialect extends into Chinghai, Szechwan and Yunnan.
2. Katchin of the Bodo-Naga-Katchin Group, is spoken in the northwestern border of Yunnan.
3. Speakers of some dialects of the Burmese Group may be found in western Yunnan under various tribal names. The classification of these dialects remains to be done.
4. Among the Lolo Group, the Lolo, with its dialects, is spoken in a large portion of Yunnan, in the northwest of Kweichow, and also certain parts in the south of Sikang. It descends into French Indo-China and Siam. The Lolo has an in dependent syllabic writing of its own, for details of which see P. Vial, Dictionnaire Francais-Lolo. The Moso is spoken in the northwest of Yunnan, and extends to Sikang. It possesses two systems of writing, one hieroglyphic and the other syllabic like the Lolo. The Minkia and some minor dialects may also belong to this group. This group is characterized by great simplification of the phonetic system, such as the complete dropping of the final consonants and the rarity of diphthongs. Word order resembles the Tibetan.
AUSTRO-ASIATIC FAMILY
Of this large family proposed by P. W. Schmidt to which the Munda, the Mon-Khmer, and the Annamite belong, we may only mention the Mon-Khmer Group of which there are representatives in China. The earliest records of this group are some Khmer inscriptions of the seventh century and a Mon inscription of the 11th century, the alphabets being derived from the Hindu sources. This group of languages has no tones, and makes use of prefixes and infixes for the derivation of words. The stem is generally monosyllabic; the word order is subject-verb-object.
The dialects of this linguistic group spoken in China are the Palaung, the Wa, etc., along the Yunnan-Burmese border. The Palaung has been well described by Mrs. L. Milne in her "Elementary Palaung Grammar" and her A Dictionary of English-Palaung and Palaung-English. The language has no tones, and possesses a number of prefixes which are both syllabic and asyllabic, such as p-, pan-, ra-, kar-, for example: yam (to die), p-yam (to kill), pan-p-yam (the killing, one who is killed).
A special series of initials hl-, hr-, hm-, and hn- exist. This language shows close contacts with the Tai branch of the Indo-Chinese family.
ALTAI FAMILY
All along the northern territory of China from Sinkiang (Chinese Turkestan) through Mongolia to the Three Eastern Provinces (Manchuria) is spoken this family of languages. It extends further southwest to Asia Minor and northeast to Siberia up to the Arctic coast. This family consists of three main branches of languages, namely, the Turkish, the Mongolian and the Tungus. The relation of the three groups has not been established with certainty, although phonetic structure, syntax and vocabulary all show great resemblances. The exclusive use of suffixes either derivative or syntactical is one of the characteristics, so that the stem or root always remains at the beginning of a word. The suffixes are loosely joined one after another according to a defmite order—for example, Turkish baba (father), -lar (plural), -um (our), -dan (from), "from our fathers."
A specific phonetic feature is known as the vocalic harmony. The principle in short is that in a word all vowels must be either front vowels such as i, y, e, o, or back vowels such as i, u, a, o, for in stance, Turkish sev-il-dir-eme-mek, "Not to be cause to be loved" and jaz-il-dir-ama-mak, "not to be able to cause to be written." Different dialects of course present slight different rules for vocalic harmony. This harmony may further influence the consonants, i.e., palatals in the neighborhood of front vowels and velars in the neighborhood of back vowels.
Word order is subject-object-verb; the verb always staying at the end of the sentence. Modifying words are placed before the modified.
A. Turkish.— This branch of the Altai family is spoken in China in Sinkiang, in the northwest corner of Mongolia, and in certain parts of Kansu. The Turkish dialects are divided into several groups, but the differences among them are very slight. The Eastern dialects are characterized by the wide application of the rules of vocalic harmony and by the existence of only surds in initial and final positions and only voiced consonants in medial positions. Dialects spoken in the northwest corner of Mongolia (Tannu Urianghai) belong to this group. The central dialects possess voiced initials and have an indifferent i in regard to vocalic harmony. They are spoken chiefly in Chinese Turkestan, such as Tarantchi, dialects of Hami, of Aksu, of Kashgar, of Yarkand, etc. The dialects spoken in the northern part of Sinkiang, such as the Kirghiz, belong to the Western Group. The Southern Group of the Turkish dialects is not represented in China. Our oldest texts are some Siberian inscriptions from the eighth century, a Turkish-Arabic vocabulary from the 11th century, and a Latin-Persian-Turkish vocabulary from the 14th century. Several forms of writing have been known to be in use, including the Runiform, the Uigur, the Brahmi, the Tibetan, etc., but most dialects have adopted the Arabic alphabet.
B. Mongolian.—The Mongolian language is centralized in Mongolia and extends to Central Asia in the west, to Siberia in the north to Manchuria in the east and to the northern provinces in the south. We find here, as in Turkish, the use of suffixes; the vocalic harmony, although distorted in certain ways is still observable. There are several divisions of the Mongolian although the differences among them are very slight. The Khalkha Group occupies a vast area in Mongolia (Outer Mongolia); the Buriat Group is spoken chiefly in Siberia but also in certain parts in northern Mongolia and in the western part of Heilungkiang; the Kalmuck Group is spoken in western Mongolia and in the northern part of Sinkiang the Southern (or Eastern) Group is spoken in Chahar, Suiyuan, Jehol, Ninghsia and in some parts in Manchuria; some Mongolian dialects are spoken in Chinghai and Kansu; another group is spoken outside of China in Afghanistan.
The difference among these groups is slight. The most important one lies in the treatment of the palatal affricates dz- and ts- of the literary Mongolian. The Southern Group preserved the palatal position, the Khalkha changes them to dz- and ts- except before i, the Buriat changes ts-, ts into s- and s-, the Kalmuck changes dz- into z-.
The Mongols adopted the so-called Uigur alphabet in the 12th century. It is still in use now with slight modifications. It runs from the top to the bottom like the Chinese, but begins the line from the left side of the page. During the 13th and 14th centuries another alphabet known as the hphags-p. derived from the Tibetan alphabet was in use but was soon discarded.
C. Tungus.—The Tungus branch is spoken in eastern Siberia and northern Manchuria. An exact classification of the dialect is impossible on account of the lack of material. It is generally known to consist of two groups: the Northern Group and the Southern Group. Manchu, Goldi, Orochon, Daghor, and Solon, which form the Southern Group, are spoken in the provinces of Heilungkiang and Kirin. Manegir and Birar, of the Northern Group, are spoken in Heilungkiang also. Most dialects of this branch are spoken in Siberia. A small group of Manchu speakers is found also in Li in Sinkiang, being the descendants of earlier Manchu garrisons.
The best known language of this group is Manchu. The writing is derived in the 16th century from Mongolian with slight modifications. The vocalic harmony presents great fluctuations in Manchu; we may however summarize it briefly thus: Back vowels, a, o, o; front vowels, ä, u; indifferent vowel, i.
The writer has purposely avoided many of the tribal names, particularly of the southwest, which are very numerous and confusing. The writer is indebted to many authors for this brief account, but many points about the Chinese and the Tai dialects are from recent investigations and unpublished" sources. Aside from the living languages in China, there are several languages which have left records but are no more in existence. Among them we may mention Tokharian, an Indo-European language which was once spoken in Chinese Turkestan. It seems to form an independent group among the Indo-European languages. It was written with a kind of Hindu alphabet, and we possess a number of texts, mostly fragmentary. Another important language was Sihsia, once spoken in Kansu. It was written with a kind of character evidently modeled after the Chinese but very complicated. The deciphering of these texts, largely Buddhistic, is still in the early stages, but the language seems to belong to the Lolo Group. Others such as the language of the Kitan are still less well known, although some inscriptions in characters similar to the Sihsia have been found.