ASIA Stljata, Neva Ijnrh CHARLES WILLIAM WASON COLLECTION CHINA AND THE CHINESE THE GIFT OF CHARLES WILLIAM WASON CLASS OF 1876 1918 Cornell University Library The original of this book is in the Cornell University Library. There are no known copyright restrictions in the United States on the use of the text. http://www.archive.org/details/cu31924023552510 THE LANGUAGES OF CHINA BEFORE THE CHINESE. LINGUISTIC AND ORIENTAL WORKS/ By Prof. Dk. TEKRIEN DE LACOUPEPIE. Du Langage : Essai sur la Nature et l'Etude des Mota et des Langues. — Paris- Leipzig, 1867, 1 vol., 8vo. Lea Noms proprea. — Le Havre, 1868. Early Hiatory of the Chinese Civilization (with plate of early Chinese and Babylonian characters). — London, 1880, 8vo. - The Silver Coinage of Tibet (with plate).— London, 1882, 8vo. (Repr. from JVumismatic Chronicle). A Lolo Manuscript written on Satin.— London, 1882, 8vo. (Repr. from J.R.A.S.) Paper Money of the Ninth Century, and supposed Leather Coinage of China. — London, 1882, 8vo. (Sepr. Num. Chron. ) On the History of the Archaic Chinese Writings and Texts. — London, 1882, 8vo. (Repr. from J.S. AS.) Orientalia Antiqua (edited by T. de Lacouperie). — London, 1882, 8vo. The Oldest Book of the Chinese and its Authors (Jour. Soy. Asiat. Soc, 1882, vol. xiv., pp. 781 -815 ; 1883, vol. xv. , pp. 237-283. 484). — Second edition in the press. The Old Numerals, the Counting Bods, and the Swan-pan in China. — London, 1882, 8vo. (Repr. N. C.) ■ The Affinity of the Ten Stems of the Chinese Cycle of Ten with the Akkadian Numerals. — London, 1883 (Repr. Academy). The Chinese Mythical Kingand the Babylonian Canon. — London, 1883 (Repr. Acad). Traditions of Babylonia in Early Chinese Documents. — London, 1833 (Repr. Acad.) ■ Chinese and Japanese Coins (Historical Sketch of), pp. 198-235, Coins and Medals by the authors of the British Museum Official Catalogues.— London, 1885, 8vo. The Cradle of the Shan Bace. — London, 1885 (Introd. to A. B. Colquhoun's Amongst the Shans). The Sinim of Iaaiah, not the Chinese. — London* 1887 (Repr. Bab. and Or. Record). Did Cyrus introduce Writing into India? — London, 1887 ('Repr. B. O.R.) Babylonia and China. Western Origin of the Early Chinese Civilization. — London, 1887 (Repr. B.O.R.) Beginnings of Writing (with six plates).— London, 1887, 1 vol., 8vo. The Languages of China before the Chinese ( Repr. from the Address of the President to the Philological Society). — London, 1887i 1 vol., 8vo. Formosa Notes on MSS., Languages and Eaces. — London, 1887 (Repr. J.R.A.S.) Ideology of Languages, and its Belation to History. — London, 1887, 1 vol., 8vo. .In the press. The Coins of China, from the British Museum and other Sources. — Vol. i., 4to. In the press. The Science of Language, chiefly with regard to S. E. Asia. — 1 vol., 8vo. Soon in the press. The Chinese and their Future Prospects (J. Soc. of Arts, July, 1880).— The Lolo and Mosso Writings (Proc. R. Q. S. , 1882, supp. pap. I. )— On a Bagspa Legend on Coins of Ghazan (Catalogue of Oriental Coins, British Museum, vol. vi. 1881). — Chinese Inscriptions (Academy, 24th Sept., 1881). — The Chinese name of the Koman' Empire (ibid., 1st October, 1881). — The Sumerian and Accadian Dialects (ibid., 24th Jan., 1882). — Lolo not connected with Vei Characters (Athenasum. 23rd Sept., 1882).— The Th-king (ibid., 21st Jan 9th and 30th Sept., 1882).— Chinese and Accadian Affinities (Academy 20th Jan., 1883).— The Shifting of the Cardinal Points, as an illustration of the Chaldeao-Babylonian Culture borrowed by tbe early Chinese (abstr ibid 12th May, 1883).-Early Chinese Literature (ibid., 28th July, 1883) —Chinese and Siamese (ibid., 11th August, 1883).— Tin-yfit not India (ibid 2nd Mav 1885).— India from China (ibid., 5th Sept., 1885).— Babylonian and Old Chinese Measures (ibid., 10th Oct., 1885).— Indo-Chinese Philolosv fibid 24th Oct., 1885).-The Nestorian Tablet (Times. 4th Feb 1885) —The Tn Indigenes of Tungking (Proc. R.G.S., April, 1886).— Babylonia and China (Academy, 7th Aug., 1880).-The Nestorian Tablet at Si-ngar Tfu (Time, Z Sept., 1886).— Comparative Ideology (Academy, 4th Sept., 1886) —Akkadian and Sumerian in Comparative Philology (Bab. and Or Record Nov 18861 —The Kushites, who were they P (ibid., Dec, 1886).— A New Writing frn™ S. W. China (Academy, 9th Feb., 1887)—A Native Writing in Formos^ (ibid., Aug., 1887).— Les Langues de la Chine avant les Chinois (Le Mu*l„ n Janvier, Avril, Jnillet, 1887). K fusion, THE LANGUAGES OF CHINA BEFORE THE CHINESE. EE8EAECHES ON THE LANGUAGES SPOKEN BY THE PEE-CHINESE EACES OF CHINA PEOPEE PEEYIOUSLT TO THE CHINESE OCCUPATION. TERRIEN DE LACOUPERIE, DOCTOR IN PHILOSOPHY AND IN LETTERS ; PROFESSOR OP INDO-CHINESE PHILOLOGY (UNIVERSITY COLLEGE, LONDON) : . M. OP COUNCIL ROYAL ASIATIC SOCIETY AND PHILOLOGICAL SOCIETY J CM. ACADEMIE EE STANISLAS OF NANCY, PEKING ORIENTAL SOCIETY, ETC. LONDON : DAVID NUTT, 270, STEAND. 1887. PL330I {Included in the Address of the President to the Philological Society, 1886.) THE LANGUAGES OF CHINA BEFORE THE CHINESE. TEERIEN DE LACOUPERIE, Ph. & Litt.D., PROFESSOR OP INDO-CHINESE PHILOLOGY (UNIVERSITY COLLEGE, LONDON) ; M. OF COUNCIL ROYAL ASIATIC SOCIETY AND PHILOLOGICAL SOCIETY ; M. SOCIETY OF BIBLICAL ARCHEOLOGY, LONDON, SOCIETE ASIATIO.UE DE PARIS, ETC. ; CM. ACADEMIE DE STANISLAS OF NANCY, PEKING ORIENTAL SOCIETY, ETC. (Reprinted from the Address of the President to the Philological Society, 1886.) LONDON: DAVID NTTTT, 270, STKAND. 1887. THE LANGUAGES OF CHINA BEFOEE THE CHINESE. Part I. The Data and their Treatment, §§ 1 — 12. I. Data. 1. The languages mentioned in these pages are not all of them those, or the representatives of those, which were spoken in the Flowery Land when the Chinese made their appearance in that fertile country some four thousand years ago. The Chinese have only occupied it, slowly and gradually, and their progressive occupation was only achieved nominally during the last century. Some portions of the S. and S.W. provinces of Kueitchon, Szetchuen, Yunnan, Kuangsi and Kuangtung 1 are still inhahited by broken and non-broken tribes, representatives, generally cross-bred, mixed and degenerated, of some former races who were once in possession of the country. Therefore the expression pre-Chinese languages of China implies an enormous length of time, which still continues, and which would require an immense study should the materials be available. 2. Unhappily the data are of the most scanty description. They consist of occasional references given reluctantly and contemptuously during their history by the Chinese them- selves, who were little disposed to acknowledge the existence of independent and non-Chinese populations in the very 1 The only peculiarities of transcription are the following: a, e, i, o as in Italian; «=the French ou; «=the French u ; sh=sch All., oh Fr. ; tch = tsoh All., eh Engl. midst of their dominion. Though they cannot conceal the fact that they are themselves intruders in China proper, they have always tried the use of big words and large geographical denominations, which blind the unwary readers, to shield their comparatively small beginnings. Such indications can be obtained only by a close examination of their ancient documents, such as their histories, annals, and the local topographies, where, in the case of the annals, they have to be sought for in the sections concerning foreign countries; an arrangement somewhat startling,. though not unnatural when we consider the real state of the case from a standpoint other than the views entertained by the ancient sinologists on the permanence and the ever-great importance of the Chinese nation. But the Chinese, though careful to inscribe in one or another part of their records all that occurred between themselves and the aboriginal tribes, and all that they could learn about them, were not enabled to know anything as to the events, linguistical and ethnological, which took place beyond their reach. So that displacements of the old races, as well as the arrival of new ones, have taken place in the regions non-Chinese, now part of China proper. Foreign linguistic influences have also been at work, and of these we have no other knowledge than that deduced from the traces they have left behind them which enable us to disentangle their peculiar characteristics. 3. Little attention has hitherto been paid to the ethno- logical and historical importance of the pre-Chinese popu- lations. Series of short notices from Chinese light works or illustrated albums, compiled for the sake of curiosity, about the modern tribes, remnants of these populations, have been translated into English by Bridgeman, Lockhart, Clark, "Wells "Williams, etc., and into German by Neumann. More elaborate notices concerning the ancient tribes have been published by Dr. Legge and the late Dr. Plath, with- out, however, any reference to their parentage with the modern tribes. But the first who recognized the great part played by some of them during the Middle Ages was Prof. Marquis d'Hervey de St.-Denys in a short paper read by him at the first Congres des Orientalistes held in Paris in 1873. Nothing has been done in the way of tracing out the limits of the territories occupied by the different races and tribes in former times; so that we shall have to draw the information when required from the materials compiled for a work still in MS. on China before the Chinese. 4. The linguistic materials are very meagre, and any grammar is out of the question. They consist only of 38 mere lists of words of various lengths. The longest embraces 242 words, the shortest one word only. Their direct value is unequal, inasmuch as their sources are most curiously mixed, perhaps more so than in any other linguistic document hitherto studied. Some of them are made up of the words occasionally quoted in the Chinese records, from where we have collected them ; others are lists made purposely by the Chinese, and extracted from their local works on topography. Others again were collected by European travellers, such as Mr. E. Colborne Baber, Father Suchier, M. Hosie, Father Desgodins, the late Francis Gamier and others. 1 The pre- ciseness of the vocabulary of the Lolos of Szetchuen compiled by Mr. Baber is the best specimen of all. 5. As to the Vocabularies compiled by the Chinese, their value cannot be otherwise than indifferent from a scientific standpoint, and their use for the sake of comparison, lacking in accuracy, cannot in many separate cases be accepted other- wise than as provisional data. As a matter of fact, they are the worst materials that could possibly be placed in the hands of a philologist: Written with the ordinary ideo- graphical symbols of the Chinese, they are now read with the current pronunciation of the Mandarin language. So were transcribed the 14 intended vocabularies of Chinese origin which were published some eighteen years ago at Fuhtchou by the Rev. J. Edkins. We have not here access to the Chinese originals, and are therefore compelled to trust to the transcriptions of this zealous but careless missionary and 1 As the references are given further on in every ease, it is not necessary to quote them here. scholar ; consequently the original mistakes and misunder- standings of the compilers, the slips of the pen of the transcribers, coupled with the Chinese and European mis- prints, form, when taken together, a not unimportant amount of possible errors. 6. But this is not all. There is another unknown quan- tity which must not be neglected in our estimates. These compilations were made in different times and different regions by different people not speaking the same Chinese dialect, and we have no information as to the details of these peculiar circumstances. The bearing of the dialectal characteristics for the region or the time being, in the Chinese transcriptions of foreign words and names, has never been understood until the present day, and I am afraid, excepting in one case, so far as I know, it has never been applied. The students of Chinese Buddhism have not advanced beyond the pre-scientific methods of Stanislas Julien, though much information could be derived from its study, coupled with that of the Prakritic peculiarities of pronunciation of the early Buddhist missionaries in China. Applied to the ancient geographical notions contained in the Chinese records I find it invaluable. But it is difficult to get at the proper information. In the present state of the Chinese vocabularies of non-Chinese words, we cannot in the case of those already published go to the source. The difficulty, however, is only temporary, and personal to us as far as concerns the present paper. 7. For comparative purposes the range of affinities for every word may run within the variants of pronunciation offered by the principal Chinese dialects which may have been used in their case, namely, the Pekinese, the Old Mandarin, and the Cantonese. They may also run beyond the phonetic limitations of these dialects, and present such letters as r, v, and the sonants which generally are missing in their phonetic systems. 'Moreover, these dialects are affected by wear and tear, and as the age of the vocabu- laries is not ascertained, though they are not generally older than the twelfth century, there is still present an un- known quantity of small amount, which, however, has to be taken into account. Therefore the probabilities are in these last respects, that the proportion of affinities detected in comparing these languages, as shown by the documents of Chinese origin, is larger than smaller so far as glossarial similarities are concerned. II. Methods op Classification. 8. The means I shall make use of for determining the respective places of the native dialects in the general classi- fication of languages are their affinities of vocabulary and of ideology. The latter is notated with a few figures which must be here explained summarily, as well as what is Ideology or, better, Comparative Ideology, and its purpose. 1 9. It is concerned with the order of words in the sentence. The only question with which it deals is the order of succession, in which the ideas in different languages must be expressed in order to convey the same meaning ; for the truth is, that languages are unmistakably framed on several plans of thought, some of which seem altogether inexplicable and unintelligible to our minds. Several of these may be explained by a difference of standpoint : one language, for instance, considers the word of action as a passive qualitative of the object ; while another makes it a noun expressing the activity of the subject on the object ; and in the third it is a qualitative of the subject. But all these subtleties do not alter the fact that all languages, to express a similar statement, make use of different schemes of thought, some of which are unintelligible. But where is the justification for any of these explanations P Are they not de facto vitiated for this reason — that we ourselves intro- duce the difficulty by our own scheme of thought, which is but one out of six in existence ? Therefore, we must, for the present, confine our aspirations to empirical methods of comparison. 1 Cf. my article in the Academy, 28th August, 1886, and my hook, Ideology of Languages, and its Relation to History (London, 8vo.,.D-. Nutt)- 10. Comparative ideology does away with the inveterate and unjustifiable prejudice of the Aryan school of philology, of permanence of grammar, which most of us have been brought up to regard as one of the fundamental axioms of the science of language. The fact (still unpalatable to many) is that grammar does mix, though with greater diffi- culty than any other elements of speech. 11. In lecturing last winter at University College on ' The Science of Language with reference to South-Eastern Asia/ I ventured to show that comparative ideology might be made a useful instrument for ethnological research for the genealogical classification of languages, and the history of the human mind. With this object in view, I tried to reduce the difficulty to the most simple facts, considering only the normal arrangement in different languages of the proposition, and the respective positions of the noun, genitive, adjective, and of the object, subject and verb in the sentence. Though inadequate to satisfy precise requirei- ments, and not answering the reality of facts in languages where the categories of speech are of different development to ours, the grammatical terms may be used for their equivalents with the restrictions here indicated. 12. In order to render practical the notation of these simple facts of ideology, and to permit their comparison on a large scale, I have designed the following formulae : of Arabic numerals, 1 to 8 for the minor points of word-order, and of Roman numerals, I. to VI. for the syntactical arrange- ments. The possibilities are the following : a) For the word-order or separate points of ideology : 1. Genitive + noun ; 2. Noun + genitive; 3. Adjective-fnoun; 4. Noun + adjective; 5. Object+verb; 5. Verb + object; 7. Yerb+ subject; 8. Subject+verb. By this distribution all the prepositional cases are marked by the uneven, 1, 3, 5, 7, and the postposing by the even numbers, 2, 4, 6, 8. b) For the syntactical order of the subject, verb, and object, six arrangements are met with : I. Object+subject+verb; II. . Object+verb + subject ; III. Subj ect +obj ect + verb; IV. Verb +subject+ object; V. Verb + ob j ect + subj ect ; VI. Subj ect + verb + object. In the arrangements I., II., III. the object precedes, and in IV., V., VI. it follows the verb ; should the relative posi- tion of the object and subject be taken as the standard, the above arrangements would also form two series, namely : I., II., V. where the object precedes, and III., IV., VI., where it follows the subject. As to the relation between the separate minor points of ideology and the syntactical indices, it may not be useless for practical purposes to remember that 5, 7, imply II. only ; 5, 8, imply I. or III. ; 6, 7, imply IV. or V., and that 6, 8, imply VI. only. So that the ideological indices of any language may be expressed with five figures only, four Arabic and one Roman. Description is carried further with the help of diacritical marks and small additional letters, which it would take too long to explain the use of here. Part II. Aborigines and Chinese, §§ 13-19. III. Aemval op the Chinese. 13. The fertility of China, which has earned for the country the appellation 'Flowery Land,' and for which it is indebted to the Loess geological formation, covering a large part of its area, was always for that reason highly attractive to the populations wandering temporarily or other- wise in the cold and barren lands of Central Asia. When the original Chinese nucleus, consisting of about a dozen 10 Bak tribes from the west of Asia, 1 reached the country, some twenty-three centuries before the Christian era, the region was already inhabited by several races. Altaic tribes from the North had come South to the basin of the Yellow River, and had fallen in with populations of southern origin. The arrival of the Chinese was no more than a repetition of previous events, followed by many of the same kind. They came, acording to all probability, slowly along the north- west route through the modern province of Kansuh ; 2 but they could not pass the southern bend of the Yellow River, as they were prevented from so doing by the stronghold of former invaders from the north, the Jungs. They were compelled to turn northwards, and they then crossed the river about the latitude of Tai-yuen, from whence they established themselves in Shansi and ~W. Tchihli, with the eastern course of the same river as southern boundary, for several centuries. 14. When Shun, the semi-mythical emperor (2043-1990 b.c.), 3 whose deeds form the second chapter of the Shu-King, - made his famous tour of inspection in the South, he did not go further south than was permitted by the bend of the Yellow River. The region within this extreme corner (S.W. Shansi), whence the natives had been dislodged by his predecessor Yao (2146-2043 B.C.), became the favourite seat of successive leaders. The sea-shore was not actually reached before the beginning of the eighteenth century, and the power of the new-comers began only to be felt south of the Yellow River under the reign of the great Yii and in a limited area, though the river had been crossed before his time under the fourth leader, Kao-sin. But we have not to relate here the history of the growth, so remarkable, though so slow, of the Chinese nation, and we are concerned with it 1 Some archaic inscriptions on rocks in Southern Siberia, near Abalansk on the banks of the upper course of the Yenisse'i, may be traces of their passage east- wards. These inscriptions, still undeciphered, are written in Chinese of the most archaic kind. They were published by J. Spassky, He Antiquis quibusdam sculpturis el inscriptionibus in Siberia repartis, Petropoli, 1 822. 2 The burial-place of their first leader in China was near the modern Ning on the common south border of Kansuh and Shensi. 3 According to the chronology built up from the Annals of Bamboo books. 11 only so far as we can find some information concerning the languages of the former occupiers of the soil. We are also concerned with the Chinese languages only so far as in ancient and modern times they show traces of influence of the aborigines. 15. The position of the early Chinese emigrants (the Bak tribes) towards the native populations was peculiar, and explains away many of the illusions long entertained by their descendants regarding the supposed greatness of their begin- nings. Unlike the other invaders from the North, they were civilized. It is now well shown that in their former homes i in S.W. Asia, west of the Hindukush, the Bak tribes had been under the neighbouring influence of the civilization of Susiana, an offshoot of that of Babylon. Through an inter- course of some length, they, or at least their leaders, had learned the elements of the arts, sciences and government, among which the writing, which we are now enabled to identify as a derivate of the cursive and not of the monu- mental cuneiform style, was conspicuous x 16. Their comparatively high culture when they settled in the Flowery Land, and the better organization which ensued, soon secured for them a dominant standing and position over the native tribes, occupying as they were a lower standard in the scale of civilization. Some tribes acknowledged readily their supremacy, and were befriended from the beginning, while others strongly objected to any interference on the part of the new comers. Their names 1 Cf. T. de L. : Early Sistory of Chinese Civilization (London, 1880, 8vo.). The Th-King, in The Athenceuin, 21 Jan., 9, 30 Sept., 1882. Chinese and Akkadian Affinities, in The Academy, 20 Jan. 1883. Marly Chinese Literature, ibid. 28 July, 1883. The Affinity of the Ten Stems of the Chinese Cycle with the Akkadian Numerals, ibid. 1 Sept., 1883. The Chinese Mythical Kings and the Babylonian Canon, ibid. 6 Oct., 1883. Traditions of Babylonia in Burly Chinese Documents, ibid. 17 Nov. 1883. The Oldest Book of the Chinese and its Authors, in J.lt.A.S. vol. xiv. part iv. ; vol. xv. parts ii. and iv. Babylonian and Old Chinese Measures, in The Academy, 10 Oct., 188-5. Babylonia and. China, ibid. 7 Aug., 1886. Beginnings of Writing,!. § 50; II. § 114, etc. This discovery, important for the philosophy of history, of the non-indigeneousness of the Chinese civilization, and its derivation from the old Chaldaso-Babylonian focus of culture by the medium of Susiana, is scientifically established in the above publications. However, in order to make it more accessible than it may be in these scattered papers, I will soon put forward all the proofs together in a special book, with many more facts than those hitherto published. 12 appear successively in history in proportion as the Chinese advanced either by their political domination or by intrusion as colonists. We cannot here enter into the details of the inquiry, upon which we have been able to ascertain, in many cases with probability, their place in a classification. It requires a study of their modern representatives, coupled with that of the fragmentary traditions, small historical facts, and scraps of information gathered about their racial and linguistic characteristics. An exposition of all these makes a volume of itself, so that we are compelled to curtail our remarks more than the comprehensiveness of the case would require. IV. Chinese aud Aborigines. 17. The policy of the Chinese towards the previous occupiers of the soil, which was imposed upon them as a necessity by the surrounding circumstances, and which has so much contributed to the formation of their national character, has always been, with few exceptions, strictly followed. They have, as a rule, always attempted to befriend them, and they had recourse to coercion and conquest only when compelled to do so by the aggressiveness of the tribes. It must be admitted in favour of the latter that the exertions of Chinese officials in later times, where and when they had accepted the Imperial protectorate, have often caused them to rebel. As soon as they arrived in the Flowery Land, the Chinese began to spread individually or in groups according to their well-known practice of gradual occupation by slow infiltration. It is by this slow and informal advance of colonists among the non-Chinese populations of the country, and their reporting to their government, that some glowing accounts were got up of the Chinese dominion on large tracts of country over which they had no hold whatever. 18. Should we be satisfied, considering them as repre- senting the primitive population of the Flowery Land, to take notice of the tribes as they came successively under 13 the Chinese ken in proportion to their advance east and south, the chief difficulty would consist only in the scantiness of information ; but the obscurities and difficulties are com- plicated by the continuous arrival of northern tribes. They could slip through the scattered settlements and strongholds of the Chinese, and those of them who objected to accept the Chinese yoke were compelled to go southwards, where they could either swell the number of those banished or of others who were discontented with Chinese authority, or join the independent native tribes. Those among these tribes, recently arrived in the country or not, who were settled among the Chinese scattered posts and strongholds, or who were in proximity to their dominion, used to satisfy the proud authority of the Celestial government by an apparent sub- mission and acknowledgment more or less sincere of its suzerainty. 1 19. They were divided into small principalities, whose chiefs generally enjoyed Chinese titles of office or nobility, and which occasionally, or better frequently, could form an offensive coalition when their independence was imperilled by the pressure of the Chinese growth and power. The pressure, however, became too strong for them and they had to yield before the Chinese advance, though always attempting by compromise or open resistance to hold their own ground on some point or other, more south or south-westwards. Those who objected to absorption were partly destroyed, partly expelled, and progressively driven southwards. 2 Some were removed by the conquerors, and many tribes, now broken and scattered away far apart from each other, were formerly members of an ethnical unity. Such, for instance, were the Gyalungs, now on the Chinese borders of Tibet, whose language isolated there presents such curious affinities with those of Formosa, of the Philippine Islands and also of the 1 The relative isolation of the Chinese during a long period resulted from the fact that they were encircled by semi-Chinese or non-Chinese states which, receiving the outside communications or making them, produced the effect of buffers, through which the external influence had to pass before reaching the Middle Kingdom. We are kept in the dark about many of these communica- tions by the disparition or the non-existence of records of the border states. 2 Cf . T. de. L. , The Cradle of the Shan Race, passim. 14 Toungthus of Burma, and whose location would seem inexplicable, should we not be able to connect it with an historical event, as we shall see hereafter. The majority of the population of Indo-China is made up of ethnical elements which were formerly settled in China proper. The ethnology of the peninsula cannot be under- stood separately from. the Chinese formation, and the in- tricacies of one help pretty often to make intelligible the complication of the other. Part III. The Aboriginal Dialects in the Chinese Language and Ancient Works, §§ 20-61. V. The Chinese Language affected bt the Aborigines. 20. The succession of races and the transmission of lan- guages, two facts which are not correlative, render it difficult to follow the linguistical history of any country, and often leave open the question of identity of a race always speaking the same language. In the present case the earlier data are ethnological ; the linguistic information does not exist beyond that which we can derive from the influence of the native languages on the speech of the Chinese intruders. 21. The language of the early Chinese or invading Bak tribes was entirely distinct from that of the Aborigines of China, excepting, of course, the speech of the Northern tribes, which had preceded them in the Flowery Land and apparently belonged to the Altaic or Turko-Tartar races. It was not with the Altaic division that the early Chinese language was more closely connected, but with the western or TJgric division of the Turanian class-family, and in that division it was allied with the Ostiak dialects. Its ideo- logical indices were probably those which are common to all the TJgro- Altaic when undisturbed, namely, 1 3. 5. 8. III., but we have no texts still in existence continuously written with that ideology. 1 1 Instances of the Illrd syntactical order occur in ancient texts, like survivals and as such almost always limited to the position of the objective pronoun, placed before the verb. 15 22. The modem formula is 1. 3. 6. 8. VI. in all the Chinese dialects, but traces of an occasional older one, 1. 3. 5. 8. I., are found in the more archaic of these dialects, such as those of Fuhtchou, Canton, and Tungking ; and in the Confucian, as well as in the Taoist classics, there are not a few survivals of the primitive ideology 1. 3. 5. 8. III. The ring of the Chinese linguistic evolution and formation is not, however, complete with these three formulae. 23. In some older texts there are occasional instances of 2. 3. 6. 7. which are very remarkable. The indices 6. 7., which show the postposition of the subject, and imply a syntax IV. or V., appear in early texts of the Hia dynasty, about 2000 B.C., namely in some parts of the Yh-Eing, 1 and in the ' Calendar of the Hia dynasty.' Now the latter was compiled at a time when the founder of the said dynasty advanced like a wedge into the S.E. towards the mouths of the Yang-tze Kiang, which most likely he reached, but from whence he was never able to return. The result of this advance was for a time an intermingling of the language of the conquerors with that of the previous inhabitants. As the above calendar, containing useful information, was written and diffused for the sake of the intermingled popu- lation, it was necessarily written in the most intelligible way for their wants ; and so it happened that the discordances it presents with the pure Chinese of the time being, must have corresponded to the linguistic features of the region. These are peculiar to the Tagalo-Malayan languages, and , cannot be mistaken ; since the most prominent feature, namely, the postposition of the subject to the verb, does not appear 1 I have established, I think, beyond doubt that the Yh-kmg, the most sacred book of the Chinese, is nothing less than a collection of old fragments of various kinds, lexicographical, ethnographical, etc., whose original meaning had been lost sight of, and which for that reason were looked upon as mysterious, supposed to be imbued with a deep learning and knowledge of the future, and therefore of great importance for divination. Through the transformations of writing and the ideographical evolution which took place after the renovations of 820 b. c. and of 227 B.C., both resulting in the addition of silent ideographical signs to the phonetic word-characters of antiquity, some continuous meanings were sought for in the rows of symbols of the sacred book, but unsuccessfully, as shown by the 2200 attempts made in China to unravel the mystery. Ten European translations, all at variance one with the other, have told the same improbability. Cf . my special work, The Oldest Booh of the Chinese and its Authors, London, 1882-83. 16 in the other formations which have influenced the evolution of the Chinese. 24. The postposition of the genitive to its noun, which occurs not unfrequently in the popular songs of the Book of Poetry, where it cannot possibly be looked upon as a poetic licence, belongs to an influence of different origin, and is ) common to the Mon and Taic languages. The same must be said of the preplacement given to the object, an archaism still preserved occasionally in the S.E. dialects mentioned above. And for the postposition of the object to the verb, and the syntactical order of the VI. standard, in contradistinction to the unadulterated indices of the Ural-Altaic, which it formerly possessed, there is no doubt that the Chinese language was indebted to the native languages of the Mon, and subsequently of the Taic-Shan formations. So that the Ideological indices 13 5 8 III., 13 6 7 IV., 13 5 8 1., 2 3 6 8 VI., and 13 6 8 VI., permit us to follow the rough lines of the evolution and formation of the Chinese ideology. 1 25. The phonesis, morphology, and sematology of the language bear, also, their testimony to the great influence of the native tongues. The phonetic impoverishment and the introduction and growth of the tones as an equilibrium to make up deficiencies from wear and tear, are results of the same influence. In the process of word-making, the usual system of postplacing particles for specifying the conditions in space and time common to the Ugro-Altaic linguistic alliance has been disturbed in Chinese, and most frequently a system of preplacing has been substituted for the older one. And, finally, in the department of sematology, we have to indicate, also, as a native influence on the language of the Chinese, the habit of using numeral auxiliaries, or segregative particles, otherwise classifiers, which, if it had not been altogether foreign to the older state of the lan- guage, would not have taken the important place it occupies in the modern dialects. 1 We must also mention here the postposition of the adverb to the verb, which, contrary to the Chinese habit, is frequently resorted to in the Taoist books. I take it as a Taic-Shan influence, to which, as we know, Taoism was much indebted during its beginnings. 17 26. The vocabularies which, contrary to the usual habit, have not been first considered, here come at one pace with the preceding alterations. The loan of words has been ex- tensive on both sides, native and Chinese, and reached to a considerable amount. VI. The Aboriginal Languages in Chinese History. 27. The written documents of the Chinese concerning their early settlements in the Flowery Land are so short that it would be surprising to find in them any information concerning the languages of the aborigines. The most im- portant struggles that occurred between them are noticed in a few words, but nothing more. It is only in later times, when the records are more copious, that we are enabled to draw from them a few linguistic data. 28. In the most valuable chronicle of Ts'o. Kieu Ming, a young disciple of Confucius, which accompanies the dry ephemerides or Tchun tsiu of his master, there is a most positive statement that some of the non-Chinese tribes inter- spersed with the Chinese in the small area then occupied by them, were speaking different languages. The statement concerns only the Jung, a race whose tribes had advanced into China from the north-west, before and after the im- migration of the Chinese Bak tribes. One of their leaders, Kin-tchi of the Kiang Tung tribes, took part in a covenant between the Chinese princes of the Eastern principalities to whom the ruler of the non-Chinese state of Ngu 1 had applied for help against the encroachments of the State, also non- Chinese, of Ts'u. 2 It was in the 14th year of the Duke Siang of Lu, otherwise 558 B.C. The Jung Yiscount Kin-tchi, i previous to his admission to the covenant, said : " Our food, our drink, our clothes, are all different from those of the Flowery States; we do not exchange silks or other articles of introduction with their courts ; their language and ourS do 1 Now Wu in Mandarin pronunciation. Corresponding roughly to the Maritime ( provinces of Kiang-su and Tchehkiang. 2 Eoughly S. Honan, Hupeh, Anhui, and N. Honan Prq-vinces. 18 not admit of intercourse between us and them." 1 The Jung, as a race, apparently belonged to that which is represented nowadays by the majority of the Naga tribes. 29. Though there is no other allusion to the foreign languages of the non-Chinese tribes so precise as the pre- ceding, there is no doubt that the other races did not speak Chinese. Some of them, like the Jung, were interspersed with the thin Chinese population, not as intruders, but as occupiers of the soil. They were more or less completely under subjection to the Chinese yoke, which they could have escaped by migrating southwards as so many of their brethren had done. It is to the influence of the intermingling with these well-disposed tribes that we must attribute the early native influence of foreign languages on that of the Chinese. And we have seen that this influence proved to be that of idioms proper to the Mon linguistic formation ; an inference which other sources of information confirm plainly. 2 30. Those on the borders organized into states, large and small, under the Chinese rivalry and influence, were more important for the people of the Flowery States. Their independence and occasional aggressiveness compelled the Chinese to take notice of their languages. While the natives settled within the Chinese dominion were in the necessity, by consideration for their power, to learn to speak Chinese, besides their own language, as was the case with the Jung Viscount mentioned above, those of the outside were in a different position. "We know, for instance, by the Li-Ki, 3 that during the Tchou dynasty, 1050-255 B.C., or at least during the second half of that period, there' were in the machinery of Chinese government some special interpreters, whose title of office varied according to the region with which they were concerned. ' In the five regions of the Middle States (or Chinese principalities) of the Y (or Eastern 1 Tso tchuen, Siang Kung, 14th year, § I. Chinese Classics, edit. Legge, vol. v. p. 464. 2 Cf. T. de L., The Cradle of the Shan Saee, pass. 3 Dr. J. Legge has just published a complete translation of this important ■work, -which, finally compiled about the Christian era, is made up for the most part of older documents. 19 barbarians), of the Man (or Southern barbarians), of the Jung (or Western barbarians) and of the Tek (or Northern barbarians), 1 the languages of the people were not mutually intelligible, and their likings and desires were different. To make what was in their mind apprehended, and to communicate their likings and desires (there were officers), — in the east called transmitters ; in the south, representationists ; in the west, Ti-tis; and in the north, interpreters.' 2 31. During the reign of the above dynasty, on the immediate south of the Chinese principalities, was the great state of Tsu, which had grown into civilization through the civilizing influence of its northern neighbours. Yet it remained non-Chinese, in spite of its entrance into the sort of confederation formed by the Flowery States under the nominal suzerainty of the said dynasty. It covered the south of Homan, Hupeh, and a waving and ill-defined territory all round. Towards the end of the fourth century b.c. the chattering philosopher Mencius, speaking of a man from that State, calls him ' a shrike-ton gued barbarian of the south,', and on another occasion he alludes to the languages of Ts'i (W. Shantung) and Ts'u as quite different from one another. 3 32. It was not, however, the first allusion that was made to the language of Ts'u. In the chronicle of Tso, already mentioned, in 663 B.C., two words are quoted in support of an interesting legend similar to others well known elsewhere. The scene is in Ts'u (i.e. Hupeh). A male child was thrown away by his mother's orders in the marsh of Mung : there a tigress suckled him. This was witnessed by the Yiscount of Yun, whilst hunting, and when he returned home in terror, his wife (whose son the child was) told him the whole, affair, on which he sent for the child and had it cared for. The people of Tsu called 1 I have added the information in brackets, in order to make the matter clearer. Roughly speaking, the T- corresponded to the Tagalo-Malays, the Man to the Mons, the Jung to the Nagas, and the Tek to the Turko-Tatars. 2 Li-Ki, tr. Legge, I. pp. 229-230. 3 III. 1,4, 14; III. 2, 6,1. 20 ' suckling ' tou or nou, and ' a tiger ' they called toiirtu, hence the child was called ' Tou-w.utu,' and he became subsequently Tze-wen, the chief minister of Tsu. 1 33. The nearest approximation to these words are found in the Ta'ic-Shaii vocabularies, where ' suckle or suckling ' is called dut (Siamese), and 'a tiger' is htso, tso, su, 2 etc. The connection here suggested by these vocables is further promoted by this fact that a large proportion of the proper names of that same State of Ts'u are preceded by tou, which seems to be a sort of prefixed particle. This is also a ►peculiarity of the Tchungkia dialect of some tribes still in existence in the south-west of China and formerly in Kiangsi, where they represented the ancient ethnic stock of the State of Ts'u. And this Tchungkia dialect is Taic-Shan to such an extent that Siamese- speaking travellers could without much difficulty understand it. "We shall have again in the sequel occasion to speak of the language of Ts'u. 34. On the east of Ts'u were the states of Wu and Yueh, covering the modern provinces of Kiang-su and Tcheh-kiang. • The former, which appears in Chinese history about 584 B.C., was conquered by its southern neighbour of Yueh about 473 b.c. 3 As could be expected, the Chinese language was not spoken there, and although we have no record dealing with the fact, we are made aware of it by the non-Chinese appearance of their kings' names. This fact has not escaped the atten- tion of commentators, and one of them, Ein Li-ts'iang, has remarked that such names as that of Tan-tchih of Yueh must be read as one single word, in accordance with the syllabic method of the west. 4 . 35. On the other hand, it has also been remarked that the names of the kings of Wu have decidedly a non-Chinese 1 Cf. Tso-tchuen, Tchwang Kung, year XXX. ; and Siun Kung, year IV. Chinese Classics, edit. Legge, vol. v. pp. 117-118, 295 and 297. 2 Apparently decayed forms. 3 The State of Ts'u warred against the two states for centuries, and finally extinguished that of Yueh circa 334 B.C. * Chinese Classics, ed. Legge, vol. iii. intr. p. 167, n. 2. The Chinese scholar means that no signification has to he sought for in each of the Chinese symbols employed to transcribe these foreign names. 21 appearance. 1 The . finals ngu, ngao, etc., are singularly sug- gestive of a known adjective meaning 'great/ and postplaced according to the genius of the language which would have belonged to the Indo-Pacific linguistic formation. It is; again, by the use of characteristic prefixes, that we find a confirmation of this surmise. Of course it is only in the case of proper names, as common words do not appear in these documents. We. find Kon prefixed to personal names, Kon Ngu, 2 Kon T&ien, of that region mentioned in Chinese records. 36. But the majority of the names are generally pre- ceded by wu, written as in wutu ' tiger,' in Ts'u. These prefixed words are the well-known auxiliaries which are employed for all living beings in the Ta'ic-shan and other cognate languages ; they are occasionally used in some lan- guages as some kind of articles, but their use is generally limited to the case of auxiliaries to the lower numerals. Of the languages spoken in the border states of the south- west and west, nothing is known during the period of which we are just speaking. VII. Ancient Chinese Works on the Old Dialects. 37. The gradual absorption by the Chinese of the aboriginal tribes interspersed among them, and their pro- gressive extension on a larger area, made itself felt by the introduction of foreign words here and there into the general^ language, as well as the appearance of provincialisms and local pronunciations of some words of their old stock. This fact could not fail to attract the attention of a careful ruler, jealous of his own power all over the Chinese apglomeration . In 820 B.C., during a phasis of revival of power of the Tchou dynasty, a wilful ruler, Siuen "Wang, with the help of a great minister, tried to ensure for ever the intelligibility of his written communications and orders to the various parts of his dominion, whatsoever might be the local variations of speech in words or' in sounds. His important reform, which 1 Dr. Legge, ibid. vol. v. introd. pp. 107, 135. 2 Cf. Mayers, Chinese Readers' Manual, N. 277. 22 has left for ever its mark on the writing of China, being repeated on two occasions later on in accordance with his teachings, has exercised, undoubtedly, a great influence on the future enlargement of his country by the facility it afforded to the propagation of the Imperial orders. It con- sisted in a partial re-cast and simplification of the characters of writing, in order to give a predominant and extensive position to the silent ideographs, suggestive of meaning, which hitherto were not much used coupled in one and the same groups of signs indicative of sound by syllabic spelling or otherwise. 38. The effect produced, which could not be enforced everywhere at that time, by impotence of the central authori- ty, did not keep up, as was expected, the general language on the same level, nor prevent the introduction of foreign words : some other means had to be found in order to make the central government aware of the new words gradually introduced. The records of the time are silent on the subject. We only know that the sacred books were explained in the various states by special men sent for the purpose ; and we also hear of the complaints made against the independence shown by these states individually, in their customs as in their words. 39. In the Fung su t'ung, compiled by Yng Sha'o (second century a.d.), it is said that it was the custom for the sovereigns of the Tchou (1050-255 B.C.) and Ts'in (255-206 >b.c.) dynasties to send ' commissioners or envoys travelling in light chariots ' 1 yeo Men-tchi she, on an annual circuit of the empire during the_ eighth moon of the year to inquire for the customs and forms of speech (or words) used in various regions, 2 On returning, these messengers presented to the Emperor reports, which, at first preserved in the house of 1 The Rev. Dr. W. "W. Skeat reminds me here of the words of the poet : " where Chineses drive With sails and wind their cany waggons light." — Milton, Paradise lost, iii. 438. which, however, was not an allusion to these 'light chariots,' still unknown in Europe, but referred to the ordinary ' wind barrow ' often used in China, 2 Cf. Mayers, Chinese E.M., N. 918. 23 Archives, were afterwards scattered and lost. 1 When the practice began exactly, and what use, if any, was then made of these reports, containing as they did so many interesting ' data for the history of the language, are not stated. But I . have a strong suspicion that one or more parts of the old dictionary Erh-ya was made by means of help derived from them. 40. The Erh-ya is a work of the Tchou dynasty ; it is divided, according to its subjects, into nineteen sections, out of which the first three stand apart, because of their lin- guistic importance. The first section, Ski Ku, the author- ship of which is attributed, perhaps rightly, in part to the celebrated Duke of Tchou, who, by his genius and adminis- trative capacities, was the real founder of the dynasty. It consists of small lists' of words arranged according to their related meanings. The second section, Shi-yen, is also made up of lists of words, the last of which gives the meaning of the others : its composition is generally attributed to Tze-hia, a disciple 2 of Confucius. The following section is made up of couplets arranged in pairs, with their explanation. This class of double-words, which are a characteristic feature of the Taic-Shan languages, are frequently met with in the popular songs of the Shi- King, or Classic of Poetry ; and there is no doubt in my mind that they have crept in there through the influence of the native dialects of this family on the speech of the Chinese. 41. The purpose of the Erh-ya is said to be a dictionary of the Classics, but it goes beyond that, and notwithstanding q the loss of some parts of these classics, it contains many words which do not seem to have ever been used in any Chinese text properly so called. They are regional words borrowed from other stocks of vocables, and they could be expressed in Chinese writing only by the use of homonyms as phonetic exponents. When the Erh-ya was annotated by Kwoh-p'oh (276-324 a.d.), this great scholar, well acquainted 1 Though apparently made use of by some scholars, but not preserved in their integrity and original shape. 2 B. b.c. 507. 24 with the regional words, was enabled to add not a few remarks on some correspondence referring to such vocables, with many examples, in the said dictionary. There are no less than 928 words, or about one-fifth of the general stock, which do not appear anywhere else than in the Erh-ya. 1 Therefore, it seems to me that, if it is not an ascertained fact that the compilers of this work have made use of prepared lists of local words like those collected by the yeo hien-tchi she, it looks like it and seems very probable. 42. But the most important work of its kind, and, I think, that which is unique in antiquity,, is the vocabulary of regional words compiled by Yang-hiung (B.C. 53-18 a.d.). The whole title is Yeo Men she tche tsiiieh tai yu shi pieh kicoh fang yen, generally simplified into Fang yen, and may be translated : ' The language of former ages from the envoys in light- chariots, with regional words from various states explained.' This title would show that the author has used the lists, or at least some of the lists, made by the envoys mentioned above. Much attention was paid to local words about the time of this author. A countryman of his, Yen Kiiin p'ing, of Shuh (Szetchuen), had collected more than a thousand words used in dialects. Liu-hiang, the scholar who was commissioned to draw up the catalogue of the books preserved in the 1 According to the Wu King wen tze, the Five Kings, or Canonical books, con- tain only 3335 different word-characters. They are the Yh-King, Shu Xing, Shi King, Liki and Tehun tsiu. Adding to these the Sue Shu, or Four books, namely, the Ta hioh, Tchung yung, Lun-yu and Meng tie, the total of words reaches only 4754. The great collection of the Thirteen Kings, Shih sail King, which, besides the preceding, includes the I-li, Tehou-li, Hiao-King, Ko-liang, Kung-yang, and the Erh-ya, the great total is 6544 different words, including those which appear exclusively in the latter. Cf. G. Pauthier, Dictionnaire Chinois-Annamile Latin- Frangais, p. xv n. (Paris, 1867, 8vo. 1st part only published}. The non-existence in the Erh-ya of modern characters found in some Taoist books, such as the Tao-teh-Mng, does not imply that these books, or the passage where these characters occur in them, have a later origin than the Erh-ya. A not unimportant cause may be that this vocabulary belongs to the Confucianists, and therefore that a recension of the books of other schools may not have been made when its various parts were successively compiled. A more important cause is that the Erh-ya, and the said books, were independently tran- scribed from their original style of writing, Ta tchuen into Siao-tchuen and modern characters. Whence some differences. Much of the obscurity of the Tao-teh-king may be explained in that way; for instance, the symbols g and @ of the original were both rendered by jjj^ . 25 Imperial collection, and father of Liu-hin, who achieved the task (b.c. 7), laboured on the same subject. Lin-lu and Wang ju-ts'ai, engaged in. similar studies, made use of what they called Keng Kai tchi fah, or ' General lists.' Yang- hiung greatly appreciated these documents, and worked upon them for twenty-seven years. During the same time he diligently consulted persons of repute all over the country, and compiled his work, which contained 9000 words arranged by order of subjects in fifteen sections. 43. It is nothing less than a comparative vocabulary, and we must recognize in him a predecessor in the science of language. Unfortunately his book has not been trans- mitted to us as he left it. As we now possess it, there are only thirteen chapters and over 12,000 words. It has been augmented by one- third, and consequently these, or at least many of these, 3000 words, being, additions of instances of later times, when many changes had occurred in the respective position of several of the non-Chinese tribes, present many inconsistencies. A critical edition made by European scholars might lead to some better readings and emendations. The Chinese themselves have begun the task. In the Imperial edition of the present dynasty, the editors have followed the text preserved in the great collection of the fifteenth century called Yung-loh ta tien, 1 restoring to order and correctness the common editions of the work. The most valuable . commentary was that made by Kwoh-p'oh, the same scholar who annotated the Erh-ya and other works. Yang hiung was enabled by his efforts to include, in his vocabulary, words from over forty-four regions, 2 many of which were Chinese only in name, and others not Chinese at all, though within the modern area of China proper. 1 It is a collection in 22,877 books with sixty books of Index, preserved in the Han-lin College, and compiled in a.d. 1407. It contains long extracts from works which have now disappeared, and it has never been printed. Cf. W. F. Mayers, Bibliography of the Chinese Imperial Collections of Literature, in China Review, vol. vi. Jan.-Feb. 1878. 2 Dr.' Edkins, who has written a short notice on this work of Tang hiung, in his Introduction to the Study of Chinese Characters, append, pp. 40-44, to which I am indebted for several facts mentioned above, quotes only 24 of these regions, out of which he identifies only seven. 26 44. The dialectic regions which occur the most fre- quently in Tang-hkmg's comparative vocahulary hear the following names, to which I add a short indication of their approximate correspondence on the modern map of China : 1. N."W. Ts'in, in Shensi. 2. N.E. Yen, in N. Tchihli. 3. C.N. Tsin, in Shansi. 4. C.N. Fen, in W. Shansi. 5. N.E. Im, in S."W. Shantung. . 6. N.E. Ten, in S.W. Shantung. 7. C.N. 2M«o, in E. Shansi. 8. C.N. Wei (anc. Ngu), in N.E. Honan and S. Tchihli. 9. C.N. Ki, in W. Tchihli. 10. C. San, in S.E. Shansi and N. Honan. 11. C. JETo-nei, in Honan. 12. C. Tchen, in C. Honan. 13. C. Tchou, in N. Honan. 14. C. Wei, in S. Shansi and N."W. Honan. 15. C. Nan Wei, south of preceding. 16. C.E. Sung, in E. Honan and "W. Kiangsu. 17. C. Tehing, in C. Honan. 18. C. Juh, in 0. Honan. 19. C. Tng, in C. Honan. 20. W. Mien, in S. Shensi. All the above regions were names of States of the Chinese confederation, and were Chinese. 45. The following were on the borders and Chinese only in parts, or non- Chinese at all : 21. S. King, or Hupeh. 22. S. Tsu, in Hupeh and neighhouring region. 23. E. Sail and Tai (between), in Shantung. 24. E. Tung Tsi, in N. Shantung. 25. E. Siu, in N.W. Kiangsu. 26. E. Tunghai, in N. Kiangsu. 27. E. Kiang and Ewai (between the), in S."W. Kiangsu. 28. E. Tsing, in N. Anhui. 29. S.E. Wu, in Kiangsu. 30. S.E. Wuku or Five Lakes, in S. Kiangsu. 31. S.E. Bui-M, in N. Tchehkiang. 27 32. S. Tan-yang, in S. Anhui. 33. S. Yang, in 8. Kiangsu and Kiangsi. 34. S.E. Yueh, or Tchehkiang. 35. S. Nan Tsu, Hunan. 36. S. Siang, or S. of Tung-ting lake, C. Hunan. 37. S. Ling, or C.E. Hunan. 38. S. Nan-yueh, or Kuang-tung. 39. S. Kwei-lin, or W. Kuang-tung. 40. S."W. Shuh, or Szetchuen. 41. S.W. Yh, or C. Szetchuen. 42. "W. Liang, or N. Szetchuen. 43. "W". Lung-si, or S. Kansuh. 44. N.E. Leh river, N. of Tchihli and others. From trie last region of the list, the words which are quoted are Korean and may often be still assimilated to modern Korean words. 46. Some of these regions are specified in history as those where removal of populations took place before the time of Yang-hiung,' and we do not know how many of the new data, which he was able to gather, and join to his former documents, were affected by these events. We have good reason to suppose that" they were so affected, otherwise no such regions as those of Mien (20) or of Kiang Hwai (27), which are virtually included in other names, would have been quoted as dialectic centres. 47. And it is rather curious that the region of Mien, an old name of the Han river, in Hupeh, should appear distinct from the region of Ts'u in which it was included, as the event which made it conspicuous in that respect occurred long after the overthrow of the said state of Ts'u by its powerful and successful rival kingdom of T'sin in their struggle for the empire (222 B.C.). In a.d. 47, some thirty years after the death of Yang-hiung, the Luy-tsien Man 1 and other southern barbarians of the Tu moun- tains (East Szetchuen ?) rebelled against the Chinese yoke ; the rebellion was crushed by a Chinese army, and seven thousand prisoners were removed to the Kiang-hia 1 Sou Ban sMtj Nan Man tchuen, Bk. 116. 28 region, on the left banks of the Yang-tze, otherwise in the region of the Mien river, where they developed and asso- ciated themselves with cognate tribes. Unless a critical analysis of the words which appear under that Mien label, in the Fang yen, should prove them to be utterly distinct from those of Ts'u (which case is not apparent, and not likely so to be for the reason that all these, tribes were kindred), it will be difficult to know whether they are interpolations of later date, or data obtained by Yang hiung about his time, when the word Ts'u as a geographical denomination was no longer in use or was too expansive in meaning. 48. The other name of region which we have singled out is less open to criticism. It is that of the country between the Yang-tze and the Hwai rivers, which corresponds to the south of Kiangsu. In 138 B.C. the state of Tung (or Eastern) Ngou (in Tchehkiang), being repeatedly assailed by that of \Min-yueh (in Fuhkien), removed a portion of the population of the latter, and expelled them 'to the northern side of the Yang-tze in the said Kiang-hwai region. Later on (110 B.C.) the same country received, with the help of decked-boats from the south, another population or portion of population from the Min yueh, at' the request of the same Tung Ngou state, then an ally'of the Chinese, who came to its rescue and saved it from the attacks of its obnoxious neighbour. 1 These events, which had certainly attracted the attention of Yang hiung and others, similarly engaged in seeking for curious forms of speech, prove, in all probability, that the entries of words under this title were not extracted from earlier documents when the region was hardly accessible and little known. 49. The probability already put forward that Yang-hiung has used, for his compilation, documents of various dates, including some lists of the * Envoys in light-chariots,' is shown at large by the fact that his geographical nomencla- 1 Ban shu, Si-nan Man tchuen, bk. 95. Li lai Ti Wang nien piao tsien ban, fol. 6 and 9i\ 29 ture, always excepting the names of new regions, does not belong to the Han dynasty, but to the Tchou dynasty, and more especially to the contending states period, namely, 481-255 b.c. It is shown also by several other facts. He makes use of the name of the state of Tsin in Shansi, which was destroyed in 436 b.c, and partitioned by the states of Wei, Han and Tchao, which also appear in his list. The name of the great state of Ts'in in Shensi, which extended his power over all the other states in the third century, appears as a name for the region of Shensi only, and we meet also with the name of Kwan ' the gate,' which was that of the capital of the state and the neighbouring region in the third century. And also the names of Nan-yUeh, Kuei- lin, Siang, which did not exist before the latter part of the third century B.C. 50. All this plainly shows that the words given in this remarkable work cannot be considered as belonging to the same period, and that their collection represents several centuries. Such being the case, the phonetic rendering of the Chinese symbols employed in the work is a matter of serious consideration. Chinese symbols were attached to the rendering of foreign sounds by the successive transcribers who noticed these sounds, and, consequently, according to their own pronunciation ; and, as this pronunciation varied in time, as in space, there is no uniformity of rendering. This is made apparent by this fact, that differences of pronun- ciation are often indicated by symbols whose sounds have for long been homonymous. However, the best means to start with, and subject to the least proportion of ulterior modifications, . are the sounds preserved in the Sinico-Annamite, the most archaic of the Chinese dialects. The only reservation to be made, is that the hardening and strengthening which this dialectal pronunciation indicates goes perhaps beyond the mark, and that half of its strength might be due to local peculiarity of the dialect. 51. A few examples will be better than any further explanation, and show the average of information which is to be obtained from this remarkable work : 30 ' Hu 1 (a tiger) is in Tcheng, "Wei, Sung and Tsu called Li-fu ; 2 between the Kiang and Hwai and in Southern Tsu it is called Li-ni, s sometimes Udu ; * on the east and west of Kwan it is sometimes called Bak tu.' 5 — Bk. viii. 'North of Tsin, of Wei,' and of the Ho- within, to say lam 6 (to beat, to kill) they utter tan ; 7 in Tsu, tarn ; 8 in Southern Tsu and between the Kiang and Siang, K'e? Kwoh P'oh's commentary : Now west of Kwan, the people say lam 6 for ta 10 (to beat).' — Bk. iv. ' D&u u (to confer), lai, 12 thu. n Outside of Southern Tsu they say lai; Ts'in and Tsin say t/w.' — Bk. ii. ' Tieu u (to covet) lam, 15 tan. 16 Tcheng and Tsu say lam. 15 — Bk. ii. 'Vien 11 (a pole), between Tsu and Wei (anc. Ngu), is called chu.' 18 — Bk. ix. 52. These examples will suffice to show how the state- ments of the Fang yen are arranged, and how far the regional forms and the non-Chinese words are intermingled. The insufficiency of the glossarial data of the native dialects does "tit- 2 5^S ■$£. Of. Cantonese : Lofu. Manyak : Lephe. 3 2ig Jjl. Cf. Burmese nari ; Kiranti dial. nyor. i ^ fnMi- ^ s ™ *^ e legend quoted above, and some Taic-Shan words. 5 f6 SIS- Cf - S S al1 Karen: Bautho. 6 ■Jtt '»*> Sin. An. lam. 7 3§it ts , an, Sin. An. tan. Cf. Outtihn in Shan and Siamese. 8 'fi Van, Sin. An. tham. Cf. Shem, in Annamite. 9 fit KH, Sin. An. Khe. Cf. Kha, in Siamese and Laocian, Kai in Tchung Miao. 10 fT in, but was probably tin as suggested by the phonetic. 11 ^p yu, Sin. An. dzu. 12 j|l lai, Sin. An. lai. Cf. Ann. ««y. 13 |H tch'ou, Sin. An. thu. 14 P|] <'«o, Sin. Ann. Wew. 15 I'fcfc '<»•> Sin. Ann. lam. Cf. Shan, to ; Annamite them lam. 16 :§§ t^'an, Sin. Ann. ten. 17 Hi 2W"> Sin. Ann. wm. 18 f| TcAom, Sin. Ann. chu. Cf Annamite *'««. 31 not often admit of our finding their corresponding words to those of Yang-hiung, though the reverse happens not un- frequently ; but we are more often enabled to trace out the corresponding words or forms of words in the languages cognate to the native dialects. ' On the other hand, the tendencies exhibited by the phonetic equivalents found in the examples of the Fang yen, show ' themselves corre- sponding to some extent with those existing between the reciprocal loan-words in Chinese and the said southern lan- guages. Therefore, the probabilities are, that within China proper during the slow Chinese conquest, these same equiva- lents of sound were caused by the reciprocal influence of the ancient Chinese and the native dialects representative of or antecedent to these languages. 53. An analysis of a large number of the statements in Yang-hiung's work has shown me some equivalents of frequent repetition, the most important are the following in the range of initials : N.W. and N. D-, Dz- - L- of S.E. and S. N.E. N. and N.W. Tch-Sh- = L- of S.E. N..W. and E. L- = K, R, of S.E. and S. N.W. and C. Ng. = M- of E. W. N- = D-ofE. N.E. Toh- = J-ofS. C. and E. Si-, Dzi- = Ei- of S.E. and S. W. K-, H- = T- of C. and E. K., E., W. F- = SA-,S-,tt-ofS.E.andS. N.C. M- = Sh-, 8- of S.E. and S. W.C. T-, Teh-, Ts-= P-, J?-ofS.E. 54. Let us compare with the last two of these equiva- lents, the following which are frequent between Mandarin, Chinese, Sinico-Annamite, and Annamite. It will be remembered that the latter is a language of the Mon family. Chinese. Sinico-Annamite. Annamite. M- = Dz- = M-, Dz- P- = T- - T-, Ch-, P'- = T- = M- 32 Teh- = Tr- =. Tr-, Bl L- — 8k- = Sh- K- •=. Ch; Sh- = Sh- Hw- — V- = V- B-, Y- = Hw, Ho- = V- P- — B- = V- 55. And also the following equivalents of most frequent occurrence between the Ta'ic languages and the Mandarin or Standard Chinese : Chinese E- Kw- Hw- =z V- Taic L-, H-, = K; H- H Sh- ,J- = L- >t Tel ,-, 8- = Th- S T- 3» P'- ,F- W- = P- )t Sh ;Ts- = Pr- » T-, Tch- ,s- = B- Jf J-, N- = N- JJ T-K-, K-, H- = Ng- >1 which appear in the reciprocal loan-words between the two ; the proportion of their respective loan-words reach a total of 325 out of one thousand words which I have compared. And these equivalents are also in existence in the broken dialects of the natives of China according to their respective relationship, as we shall see hereafter. 56. The following list of a few frequent equivalents in Cantonese and Mandarin Chinese will prove interesting, in parallel to the preceding : F-, W-, in Cantonese. Y- K-, H-, T-, 8-, 8-, 8h, Teh-, Ts-, "„ P'; F; W-, f- '. -£"-> » Chinese Hw-, Kw ») J- j> Lw- >; Lai- >> Li- j) Lo- ») L- >> T- and others. 33 57. Also that of a few frequent equivalences with the dialect spoken at Tcheng-tu (Szetchuen) : Chinese K- (-e, -a, -u) = Er- at Tcheng-tu. „ Kw- = K- „ „ IT- = L- » &• = If- (-», -«) „ T- = Ng-, N- „ Teh- = Sh- And a few various ones : 1 Chinese 2V- = L- at Nanking. „ -in = -% „ „ -a«^ = -«ra „ „ J"- = Z- at Tsi-nan fu (Shantung), » F- = L- „ „ Y- = J"- in Kweitchou. „ Tc/j- = Ts- at Tientsin (Tchihli). „ Sh- = S- „ „ )SA- = i^- in Kansuh. „ J- = M- in Szetchuen, etc. 2 1 The Cantonese equivalences (§ 56) have been noticed in perusing Dr. J. E. Eitel's Chinese Dictionary in the Cantonese Dialect, parts i.-iv., Hongkong, 1877-1883. The equivalences in § 57 are extracted from Dr. J. Edkins' best work, A Grammar of the Chinese Colloquial Language commonly called the Mandarin Dialect, 2nd edit., Shanghai, 1864, pp. 69-71, 35-37. 1 When the time comes for making a scientific study of the Chinese characters, and especially of the class of those commonly known as ideo-phonetic, which is by far the most numerous (circEl 1250 a.d. it included, according to Tai-tung, 21810 characters out of a total of 24235), these local equivalences of sound, and such others which are not given here, will be found of great help in discriminating the variation of sound, especially initial, in phonetics. Many of the so-called ideo- phonetics do not deserve this appellation, and ought to be classified differently, as they belong to distinct systems of making up the characters. There are the com- pounds where the two or more characters employed have each a part in expressing the sound, I.) by a rough system of aerology and syllabism as in the oldest Ku-wen compound signs, II.) by a juxtaposition of two or more phonetic signs, either a) of different value in order to express a bisyllabic or trisyllabic word afterwards contracted and crippled into a monosyllable, or b) of homonymous signs explaining one another phonetically, with or without any ideographical meaning, these types a, i, extending to the intermediary period, viz. of the late Ku-wen and of the Ta-tchuen style ; and III.) the phonetic compounds made before and after the sixth century of our era, and composed of two symbols which both con- tribute to the sound according to the fan-tsieh method, i.e. by, the, initial sound of one and the final of the other. Once all these supposed ideo-phonetic characters discarded, and not before, it is possible to study the ideo-phonetic compounds, properly so called, and made of an ideographic symbol suggestive of idea addi- tional to a character suggestive of sound. But it would be a great mistake to 34 (58y Now let us remember that the court language for the time being has always exercised a powerful influence in China. And as this court language is and always was that of the capital, it changed as often as the capital itself, which does not mean unfrequently. For instance, with reference to the present time, the dialect of Peking became the court dialect since 1411 a.d., under the Ming dynasty, when during the reign of Yung-loh the court was transferred there and has there remained since. The Mandshu conquest, and the estab- lishment of their sway all over China, did not change this state of affairs otherwise than in this way, that it has helped greatly to corrupt the former language,- and that it is this rapidly-decaying form of speech which now has the lead over the other Chinese dialects; Ki, tsi, tchi, are now all tchi ; si, hi, are now a medial sound usually transcribed hsi, and so forth. 1 When the N.W. state of Ts'in, the most powerful principality of the Chinese confederation, conquered the Chinese parts and some others of the Flowery Land in the admit bluntly the doctrine put forward by some sinologists, that the ascertained old sound of a phonetic gives the ancient sound of the words expressed by its ideo- phonetic compounds. The application of this doctrine thus formulated is simply disastrous and antiscientific. It leads to the grouping of supposed forms of words which have never existed, and brings forth this chaotic result, exemplified in some papers of a well-known sinologist, of typical sounds having each of them all sorts of meanings, and of any certain thing or idea expressed by many of these apocry- phal typical sounds. The aforesaid principle is only true when worded and restricted as follows : The ideo-phonetic characters may have their old sound indi- cated by that possessed by their phonetic at the time and in the region of their formation, and in these conditions only, for instance, numerous phonetics with a dental initial which they have preserved in some ancient, and which exist also in some of their later compounds, have produced some compounds expressing sounds beginning with a labial during the period of labialization (of. §§ 58, 59). Others, originally dental, have passed for ever to the labial series. Phonetics give us in their compounds such equivalences of sounds as the following ; T=K, T=P, K = L, K=P, K=M, L = D, L=N, N = D, L = Sh, S=H, etc., but their relative position here is not suggestive of their historical succession. We find, how- ever, transitions such as T = S=H=K, K=Tch = S=T, T=Tcli = Dj=H=E, T = Tz=F=P, K = Kw=V=M, etc. Some of these equivalences are easily explained by the everlasting degeneration and wear and tear, and some by the action of easing, which means a facilitation proper only to its authors as a facilita- tion for some may be an increase of difficulty for others. The regional phonetic preferences will contribute to the latter explanations. All these do not preclude the existence which I have been able to disclose of polyphonic characters among those which the ancestors of the Chinese civilization have brought with them from S.W. Asia, and also the substitution of characters only homophonous temporarily, which often took place in the course of history. 1 This is the reason why the use of the Pekinese pronunciation by the European scholars and officials in China who write about historical and ancient geographical' matters cannot be too strongly deprecated. 35 third century b.c, it brought along with its sway a strong current of labialization all over the country. The pronuncia- tion was carried from the teeth to the lips. The capital was then in Shensi, and remained there during the first Han dynasty. The same phonetic influence, with perhaps less energy than in the beginning, was continued until the transfer of the capital in Honan, with the establishment of the Eastern Han dynasty (25 a.d.). 59. The above tables show that the Taic equivalents and also the Annamese are older than those of the Cantonese. The Mon-Annam loan-words from the Chinese have kept the dental sounds which preceded the labialization, brought in by the T'sin and Western Han dynasties. And the Taic, in their migration southwards, have .preserved the phonetic peculiarities which used to characterize the regions of E. and S.E. China, where we know them to have been settled. Some more information is given by the same tables, concerning the multiplication of the written language of China. They show that the partial polyphony, or better, the variation of initials, which are frequently met with in ideo-phonetic word- characters containing the same phonetic element, have arisen from the various circumstances in time and region of their formation and entrance in the Chinese vocabularies. 60. The Fang-yen of Tang-hiung is not the only work 1 in which some information is to be obtained on dialectal forms and regional words. It is the sole work in existence speci- ally written on the subject, but occasional indications are met with in another important dictionary of the same period, the Shwoh wen. Its author, Hii shen, 8 who lived in the first century a.d., was like Yang-hiung a great scholar, and, in addition to the said work, wrote a most valuable treatise on the 'Different Meanings of the Five Canonical Books.' His dictionary was only presented to the Emperor after his death, namely, in 121 a.d., and the just reward of his labours, a shrine among those sages admitted into the Temples of 1 In the I-hai chu teh'in collection are two works, entitled Suh fang yen, and Suhfang yenpu toh'ing, which I have not seen. 2 A hiography of Hii shen has been compiled by Mr. T. Vatters, in his excellent Guide to the Tablets, in a Temple of Confucius (Shanghai, 1879, 8vo.), pp. 98-100. 36 Confucius, was granted to him in 1875 only, i.e. eighteen centuries after his lifetime ! 1 61. The Shwoh-wen, which contained 9353 words, has re- mained the standard work of Chinese lexicographers, and was in fact the first work deserving the name of dictionary, as the Erh-ya, of which we have spoken above, was not more than a glossary, classified ideologically, without definitions. Hii shen collected in his work all the signs of the Siao tchuen style (the Small Seal character), which he considered the best framed ; and he gave also about 441 symbols from the oldest style (Ku-wen) of the writing, which, it cannot be repeated too often, has nothing whatever to do with the grotesque pictorial signs, long supposed and always quoted wrongly as the originals of the Chinese characters, instead of what they are in reality, corrupted and fanciful forms. 2 Yang hiung's Fang yen is not quoted eo nomine in the Shwoh wen, which, as we know, was compiled some forty years jafter- wards, perhaps because copies of the work were not yet in circulation. Hii shen speaks, however, of Yang-hiung in his introduction as the author of a sort of vocabulary of all the Chinese word-characters known in his time, some 5340 altogether, entitled the Instructor. 3 Many dialectal forms and regional words are quoted in the Shwoh wen, many of which are met with in the Fang-yen, while many others are ■ not. It looks as if the author was enabled to make use of some of the same materials as Yang-hiung, supplemented by later documents. 1 Chinese scholars value the Shwoh wen highly, and many of them have ex- pended great learning and industry in confirming and illustrating its explanations and derivations, which are far-fetched and often worthless, so far as they bear on late and secondary forms, intentionally altered, and not on the genuine old forms of the word-characters. If anything can be learned from the ancient writing of the Chinese on their beginnings, it is only from an analysis of the oldest symbols. Cf. Terrien de Lacouperie, On the Archaic Chinese Writing and Texts (London, 1882, 8vo.), and Beginnings of Writing, §§ 46-SS (London, 1887, 8vo.). Dr. John Chalmers, author of an epitome of the K'ang hi tie tien, phonetically arranged, has published an able translation of a late edition of Hii shen's work : An Account of the Structure of Chinese Characters, under 300 primary forms ; after the Shwoh wan, 100 a d., and the phonetic Shwoh worn, 1833 (London, 1882, 8vo.), where the etymological processes remind us singularly of our own etymologists of the pre-scientific period. 2 Cf. my Beginnings of Writing, I. § 48. ' ffl 3K Hi Sitmtawanpim, lit- 'Teaching collection book.' 37 We have already mentioned the interest displayed towards these sorts of words by Kwoh P'oh, the great com- mentator of the third century. 62. By putting together all the. data contained in the Erh* pa, the Fang-pen, the Shwoh wen, and the commentaries of Kwoh P'oh on the first two of these works, and a critical arrangement of them by region and by date, as far as it may be possible from the succession of the geographical nomen- clature, much light would be thrown on the linguistic history of China between 500 b.c. and 250 a.d. But such a work would require a great deal of time, and somewhat long pre- paratory study, to be successful. Part IV. The Extinct and Surviving Aboriginal Languages and VIII. Families op Languages. 63. A complete survey of all these languages is out of the question within the limits of the present work, for two opposite reasons. Some are known by mere inferences which require long and complicated expositions, as we had occasion to show above (§ 23) ; and the data concerning many others, deficient and unequal as they may be, would, however, form together a mass of a certain length much beyond our possi- bility of dealing with in these pages. Therefore we are only permitted to examine briefly a few of them, in order to show what sort of documents we have available for study,' and to give a short statement "of the facts about the others, together with the necessary references. 64. We shall enumerate them according to their relation- ship with the two great linguistic stocks to which we find they belong ; namely to the 1) Indo-Chinese division and its two branches Mon-Khmer, and Tdi-Shan ; and #lso to the Interoceanic division, Indonesian branch, of the Indo-Pa- cific stock ; and 2) to the Tibeto-Burmese and other divisions of the Kuenlunic stock of languages. A prime distinction being made between the Pre- Chinese aborigines and the Pre- Chinese ; the latter being distinguished from the former, 38 for the reason that they have entered into Pre-Chinese lands in historical times. (Cf. above § 1.) 65. The fragmentation of tribes from the various original races, and the subsequent reunion of some of these broken tribes into new units hybrid in character, have been of frequent occurrence amongst those remnants of the former population of the Flowery Land, under the continuous pressure of the Chinese growth and extension. Therefore several of the following entries are probably provisional, as the greater number of the surviving dialects are either mixed or hybridized when they are not altogether hybrid. The distinction) carried by these qualifications is this, that mixed implies only a mingled composition of the vocabulary, while the two other terms indicate the state of the grammar, which is hybridized when a part of it has been altered by inter- mingling with a foreign grammar, and hybrid when the language is the result of a new unit made up of various sources. 1 Therefore the languages are classified in the follow- ing pages according to the greater number of their affinities. IX. The Pbe-Chinese Aboriginal Mon-Tai DiAiects. a) Unmixed and Mixed. 66. Tbe Pong j|J or Pan hh H 2 U race was pre- dominant in Central China, i.e. south of the Yellow River, when the original Chinese or Bak tribes migrated into the country. Their leader named Pong, about whom various legends cropped up afterwards, was established in the N.E. of Szetchuen and in "W. Honan, and was friendly with the Chinese from the outset. In fact, he helped them against the Jungs of JSTaga race coming continuously from" the N.W. Many tribes claim to be descended from him, and not a few worship and venerate his memory. Their generic name was Ngao ' powerful,' now degraded into Yao? 1 For the sake of brevity M. =mixed, hybrid is H., and hybridized is Hd. * Or g* 3 As already remarked, the present work being exclusively devoted to languages, all the historical and ethnological researches and demonstration are forcibly left aside. 39 The Pan-hu race was a branch of the Mon race from the south-west, which had occupied a large part of China before the arrival of the Chinese, consequently before the twenty- third century B.C. It is from this branch, and as a result of their intermingling with Northern, otherwise Kuenlunic tribes, that the Ta'ic or Shan-Siamese populations have evolved, < some of which, migrating southwards in the course of time under the' Chinese pressure, spread into Indo-China, 1 and developed into several states. 67. The Pan-hu language is only known through the inference to be derived from the dialects of the tribes which have sprung from it. Its main characteristic was its ideology (2468 VI.), nearly opposite to that of the Kuenlunic languages (13 5 8 III.). The oldest relics of their speech are those which were preserved by the Chinese writers of the Han dynasty, notably in the Annals of the Eastern Han. 2 Some older traces exist in former works, and we have been enabled to point out more than one in a previous part of this paper, but they are quoted only with a geographical indication, and we have to draw our own conclusions as to the race from whose speech they were quoted ; whilst in the present instance the words are quoted with precision as those employed by the Yao of the Pan-hu race, 3 and this makes all the difference. These are only a few of them : Puk-kien, i.e. ' to tie the hair in a knot.' , Tuk-lih, i.e. ' sort of cloth.' Tinh-fu (t&ing-fu), i.e. ' chieftain.' Eng-tu, used in addressing each other. Pien-kia '.a cross-bow.' 4 Tiao-tsiang ' a long spear.' Tcho kou ' a dog.' Tu pei ' great chief they worshipped.' Puk-kien is undoubtedly the same as the Siamese p'uk ' to 1 Cf. below, §§ 116, 117, and The Cradle of the Shan Mace. 2 Mm Ban Shu, bk. 116; 3 Then in Hunan. 4 The two latter words are not derived from the same source as the others ; they are given by Fan ch'eng ta in his Kuei hai yu he'ng tchi (twelfth century). 40 tie,' and k'on ' hair.' The tuk-Uh cloth was a hair-cloth, as shown by the Siamese sakalat l ' woollen.' Eng-tu is the Siamese eng 'self and tu 'I.' Tinh-fu is tsing-fu 2 in the same language. Pien-kia is given in Chinese notices of the Kiu ku Miao of "W. Kueitchou of the same race as their own term for a cross-bow ; 3 but in Malay panah is ' a bow ' according to Crawfurd. Tcho kou finds its correspondent in the Kambodian toh . ke.* And as to Tu pel, I suppose that tu is the class article for proper names and living beings, which we meet in many of these languages. 5 68. The Yao-jen \% A, also called Fan-k'oh f# *g, 6 were an important people of the Pan-hu race, whose name has been preserved with some alteration in their own appella- tion. They are now broken up into many tribes, several of which oome under our notice, because something is known of their language. They have preserved some specimens of an ancient writing of their own, which was derived from the old Chinese characters, and of which a specimen has lately reached the British Museum. 7 69. The Pan-yao $££ $§, also oalled Ting-Pan-yao and Yao-jen, now removed southwards, are found in Kuangsi and Kueitchou. ' "We have only of their language a short list of 2\ words and the numerals, collected by a French missionary, 8 as follows : father, Ma ; 9 mother, ma ; son, tonh ; 1 Cf. Burmese thek-ka-lat. 2 Unless it be the Burmese htoung bo, as in ta-htaung lo ' colonel,' hrelc-htoung bo ' general,' in D. A. Chase, Anglo-Burmese Sand-book, part iii. pp. 51-62 (Maulmain, 1852)." 8 Miao Man hoh tchi, bk. iv. f. 6-7. These famous cross-bows of six or seven feet long, which require three men to string them, appear in a picture of "men of this tribe, reproduced from a Chinese album on native tribes, by Col. H. Yule, in his Marco Polo, 2nd edit. vol. ii. p. 68. 4 Cf . Lu-tze, degui ; Burmese, tan hTmay ; Mon, ta km ; Toungthu, httee ; Sgo Karen, 'twi, htwi; Pgo KaTen, twi. » Cf. §§ 65, 70, 105, 108, 109. • Iiuh Tze-yun, Tung k'i sien tchi, i. 8. Ts'ao Shu k'iao, Miao Man hoh tchi, i. 1 ; iv. 14. ' Cf. my Beginnings of Writing, ii. 176; and my article on A New Writing from South- Western Chinaia. The Academy, I 'eb. 19, 1887. Also below, § 70 n. 3. 8 M. Souchieres, in Be quelques tribus saitvages de la Chine et de V Indo-Chine : Zes Missions catholiques, Lyon, 1877, vol. is. p. 126. 9 The spelling is French. 41 daughter, min-sye ; man, tou mien ; woman, tou mien ao ; male, tou mien ngou ; female, tong niey ; house, nam plao ; earth, da o ; water, nam ; fire, teou : wind, dgiao ; sky, nam long ; dog, teou klou ; cat, tou mi lorn ; tree, ty dh'eang ; rice, blao ; bamboo, tylao ; hand, pou ; foot, &e£ sao ; 1, yatf ; 2, y ; 3, jwm ; 4, j»fe" ; 5, jj&j ; 6, klou ; 7, sy-a ; 8, yet ; 9, cfow ; 10, tchep; 100, ya£ pe; 1000, y«£ (ftbw. The construction is stated to be similar to that of the French (2468 VI.). The vocabulary is Mon-Ta'ic, the numerals belong to the Mon type. Tou is a visible class-prefix. 70. The Pan-y shan-tze 38E ^C Ul -p or ' Pan-y hill- men,' also called Siao Pan, 1 and Mo-yao H {|j, are known in history under the latter name since the sixth century, when they were settled in Tchang-sha kiun, Hunan, 2 He. in Central China, which was still independent. They are now refugees in the mountains of Kuangsi on the Tungkinese frontier, and they have been lately described 3 by a mis- sionary, . M. Souchieres, who had collected the following small vocabulary of their language: Father, tao fa i ; mother, dji ; son, ton ; daughter, mon cha ; man, tou moun ; woman, tou moun ao ; male, man pha ; female, tong niey ; house, sen piao ; earth, ngi ; water, nom ; fire, teou ; wind, djiao ; sky, tou ngong ; dog, tou klou ; cat, tou meou ; tree, ty ngiang ; rice, biao ; bamboo, tchey lao ; hand, pou ; foot, chey sao ; 1, a ; 2, y ; 3, po ; 4, piei ; 5, pia ; 6, Mo ; 7, ngi ^ 8, yet; 9, dou ; 10, chop ; 100, a pe ; 1000, n diou. The language is Mon, but the ideology is not made visible in any of these few instances. A determinative prefix or article tou is the 1 Or ' Lesser Pan,' as a distinction from the Pan-yao. 3 In the Sui shu or ' Annals of the Sui dynasty.' Miao Man, hoh tehi, i. 8t. 3 De quelques tribus sauvages de la Chine et de Vlndo- Chine, in Les Missions Catholiqttes, 1877, vol. ix. p. 114. Speaking of their costume the author says: " Us portent assez volontiers auteur du cou un fichu brode, auquel ils enfixent souvent un autre qui pend par devant. Ces fichus sont brodes, partie en caracteres chinois, partie en caracteres bizarres, qui se sont transmis de generation en genera- tion, et aont personne ici ne connait le sens. On voit des caracteres identiques fort bien brodes sur le bonnet des enfants," etc. Would not these unknown- characters belong to the writing of the Yao jen ? Cf. my notice on A new Writing from South-Western China, in The Academy, 19 Feb. 1887. The tutu of the Eeh Miao are perhaps similar to those bits of cloth they wear in front. Cf. § 68. * The spelling is French. 42 only characteristic of importance. The construction is stated to be like that of the French, whence the Indices 2 4 6 8 VI., and the language is a sister-dialect of that of the Pan-yao. 71. The Ling Kia Miao $$ ^ "$, also called Ling Jen, of S. Kueitchou, speak a cognate dialect to those of the Pan-yao and Mo-yao, 1 who understand it. X. The Pre-Chinese Aboriginal Mon Tai Dialects. i) Hybridized and Hybrids. 72. The T'ung Jen $|? J\„ or Tchuangjen, belong also to the Pan-hu race. One of their chief tribes, the Huang, 2 appears at a very early date in Chinese history, as they came in contact with the emigrants when the latter advanced to the W. borders of Shantung, where their settlements stood at first. Under the Chinese pressure they moved southwards and remained settled and independent in the S.W. of Hupeh until their subjugation by the state of Tsu in 648 B.C. The collapse of their conquerors under the successful attacks of the Ts'in in 222 b.c. made them pass under the nominal rule of the Ts'in and following dynasties. In 450 a.d. we find them in open rebellion with the other aborigines of Hupeh, Hunan, and W. Szetchuen provinces. The Chinese armies sent to subdue them were repeatedly beaten, and the result of a protracted struggle was the acknowledgment by the Central Government of a state of things equivalent to their independence. The T'ang dynasty repelled them within the basin of the Wu and Yuen rivers in Kueitchou, from whence they advanced still further south. We know very little of their language, i.e. not more than a small number of words. 73. Fan Tch'eng-ta, Chinese Kesident at Tsing-Kiang, the modern Kuei-lin in N.E. Kuangsi, in 1172 a.d., has given in one of his works, 3 a description of these tribes, and has occasionally quoted the following words from their language : 1 Fung Wi sien tchi, i. 10. 2 Two others were the Wei or Nguei and the Nungs. 5 Kuei hai yii heng Uhi. — Miao Man hoh tchi, bk. i. S. 3, i. 43 i P Tchii-hu, 1 'chief (elected).' M P£ Ti-to, 2 • people.' ^ ;§JJ Kia-nu, 3 'slave.' ^ "J* Kia-tingf ' servant.' EB "J* ffl ~T Tien-ting tien-ting, 5 'servant (of a higher grade).' >?§ "iHf $M Ma-tsienpai, same as preceding. }|p| "J* Tung-ting, ' a common man.' H I! "Ma-lan? ' house.' |H #)| Mei-niang, 1 ' wife.' Out of these nine words, three at least present Anna- mese affinities. Nineteen words from the language of the same tribe are provided in the Chinese 'Statistical Account of the province of Kwangsi' : 8 sky, men; sun, ta wu (ngu); moon, tch'en, loan; wind, hi; father, ha; mother, mi; elder brother, pi; younger brother, nung ; elder sister, a da; younger sister, a mi; son's wife, p'a; mother's father, ch'ia hung; mother's mother, ch'ia pu ; water, tch'o; wine, ley; drink wine, heng lau; rice, 'hen; flesh, no; I, ku; thou, meng. From this list, the words for mother, elder brother, wine, drink wine, sun, etc., belong to the Mon-Annam formation ; the pronouns are Siamese-Shan, otherwise Taic. 74. The proportion of Mon-Annam in the two lists com- prising together 28 words is ten, or more than one-third. The Chinese symbols employed in the rendering of the foreign words give them a different complexion from their inseparable ideographical meanings, which in such cases play the part of popular etymologies. Therefore it is more prudent 1 Cf. tehao ' king * or ' chief ' in the Shan language of Nantchao (§ 103). 2 Cf. Annamite ddy t6 'menial.' ' Cf. Annamite gia no ' servant.' 4 Cf. thtmg, the Annamite appellation for servants. 6 In Ma Tuanlin's Wen Men t'mg k'ao, this expression occurs as Tien tze Ma ; cf . d'Hervey de St. Denys, EthnograpMe despewples Strangers de Matouanlin, vol. ii. p. 259. 8 They are on piles. Cf. Luh Tze-yun, T'tmg k'i sien tchi, f. 14«i. 7 Cf. Annamite volon ' wife.' The Blue Miao say Mai niang for ' father's younger brother's wife.' 8 Kttang si tfmg tchi ; extracted by Dr. J. Edkins, in The Miau-tsi, o.c. 44 to consider these symbols as meaningless signs' and simple exponents of sounds only. Kia- and Ma- or Mei-, which, occur two and three times respectively, look like definitive prefixes. The ideological indices are not all exemplified; only the first three are shown to be 1 4 6, which proves the hybridity of the language and displays strong Chinese influence, which has led to the altering of the position of the genitive. 75. The Miao-tze ^g ■? of • Ta shui tcheng, in S.E Szetchuen, speak a dialect cognate to that of the following Peh-Miao of Kueitchou and Yunnan. 1 A list of 112 of their words was collected by Mr. Hosie in 1882. Numerals and pronouns are missing, but the similarities in words are con- clusive. Class-prefixes are employed, such as lu-, fan-, tu-, and ng-. ' Tea ' is hou cha ; ' hot water ' is houtliku ; ' cold water ' is houlitsa ; ' to light a fire ' is chou tou ; ' to shut the door ' is ko chung. These instances and some others display the ideological indices 2 4 6. 76. The Peh Miao g | or 'White Miao/ a few centuries ago in the centre and west of Kueitchou, 2 have now partially migrated in the S.E. of Yunnan.? A vocabulary of 148 words was formerly taken by the Chinese in W. Kueitchou. 4 While the numerals and many words belong to the Mon- Khmer family, with which they prove a deep aflinity, not a few vocables are Lolo-Nagas and Chinese, and an equal proportion, including the pronouns, is Taic-shan. Determinative prefixes are in use, such as kai, variously rendered in the Chinese transcriptions by kah, ke, kai, Mai, etc., and lu or le. The latter is common with the Seng Miao, Blue Miao, and Hua Miao dialects. The only ideological indices which can be perceived are 2 3 6 0, where the Chinese influence is felt by the pre-position of the 1 Notes of a Journey through the Provinces of Kueichow and Yunnan, p. 31. 2 Miao Man hoh tchi, iv. 4. 3 A. E. Colquhoun, Across Chryse, i. 333, 335, 347, 356, 389, 392, 393 ; ii. 302. 1 Extracted from the Ring-yfu tchi, or " Topography of the Prefeotural City Hing-y," by J. Edkins, A Vocabulary of the Mian Dialects. 45 adjective more completely than in the Blue Miao and Hua Miao dialects. 77. The Hcja Miao ^g ]jg according to the Chinese descriptions are interspersed all over the province of Kuei- tchou and the N.B. of Yunnan. 1 European travellers have met with some further south in the latter province on the borders of Kuangsi. 2 A vocabulary of 112 words without numerals has been collected in W. Kueitchou at Ta shui tching by Mr. Alex. Hosie. 3 The words to the extent of 25 per cent, are similar to those of the White Miao. Prefixed classifiers lit and tu, also ng-, are in use. The pronouns and other words are Ta'ic- shan and many words belong to the Mon stock. The visible ideological indices are 2^60. 78. The Miao of Yaop'tt tchang, S.W. of Ngan shun in Kueitchou, is known through Mr. Alex.' Hosie,* who collected a list of 110 words, besides the numerals, in 1882. Their affinities prove the language to be closely related to those of the White or Peh Miao and Hua Miao, though the vocables do not offer any apparent classifiers, which are so conspicuous in the other lists of words. But their absence may be simply an affair of interpretation. The ideological indices made visible are 2^60 only. 79. The Leng-ky Miao, or Miao-tze of Leng-ky, in the north of Yunnan, were seen by the late Fr. Gamier, 5 who collected the following thirty-three words from their vocabulary : Sun, tchan to ; moon, ka ly ; earth, 14; mountain, hem tao; forest, ma Id. Man, tsy nd; woman, po; child, to; father, tsy; mother, na ; daughter, ku ; brother, ty. Rice, kia ; cook the rice, a kia. House, tchud ; wood, kd. To eat, loo (Ida) ; to drink, heou ; to sleep, tcheou jou ; to run, mou M ; to come, 1 Miao Man hoh tehi, vr. 3 ; iii. 11». 2 A. E. Colquhoun, Across Chrysl, i. 334, 347. 3 Notes of a Journey through the Provinces of Kueichow and Yiinan, p. 31. 1 Notes of a Journey by Mr. Hosie through the Provinces of Kueichow and Yiinan, pp. 11 and 31 (Parliamentary Papers). Mr. A. Hosie was then H.B.M. Consul at Tchung-king in Szetchuen. 6 Voyage d' exploration in Indo-Vhine, vol. ii. pp. S09-617. 46 ya mou te ; to call, tchao tchang ; to sit, ta ta ; to enter, niao. 1, «"; 2, aou; 3, pii; 4, plaow, 5, tchoui; 6, teow; 7, Man chet ; 8, «7o ; 9, kia ; 10, keou. These words evince a strong affinity with the Peh Miao and Hua Miao dialects, and the only point of ideology which can be perceived is the position of the object after the verb, or 6. Class-articles are not apparent. The numerals are interesting ; 1, 2, 3, 4 are cognate with those of the Peh, Hua, Seng Miao, Pan-y, Pan-yao, in their affinity to the Mon formation ; 6, 8, and 10 also belong to the older strata of the group ; 5, with its palatal initial, sides with the Hin, Huei, Souc, Ka, Nanhang and other Cochin- Chinese dialects of the same formation ; 7 is peculiar, as made up of two words, chet the second which is Chinese, and Man similar to the Peh and Hua Miao. , 80. The Min-kia tze 1 Jj; jf; -p, or Peh-jin ^ A> now intermingled with the other population of the neighbouring region of Tali-fu in C. W. Yunnan and the S.E. of the Province, claim to have come from S. Kiangsu near Nanking. They are much mixed in race, and their language bears the same testimony; we have a vocabulary of 110 words, includ- ing numerals, published by Father Desgodins, 2 and another series of numerals by the late Prancis Gamier. 3 Chinese, Mosso, Lolo and Tibetan words have been adopted instead of the original vocables, but the Mon character of the language is still recognizable in many words, and the positions of the genitive and of the adjective (2 4) are in accordance with this indication. Categorical particles are apparently used not as prefixes but as suffixes only, somewhat as in Chinese. K'u, K'ou seem to be attached to all names of things high or large, de appears at the end of words for animals, and qualities. All the 1 The Min Ha, ' a race with features more European than Chinese,' Alex. Hosie, Report of a Journey through the Provinces of Ssu-eWuan, Yunnan, and Kueichow, p. 37. Parliamentary Papers, China, No. 2, 1884. 3 Mots principaux des langues de certaines tribus qui habitent Its bords clu Zan- tsang Kiang, du Lou-Ize Kiang et Irrawaddy (Terkalo, 26 Mai, 1872), in Bullet. Societede Geographie, Paris, 6th ser. vol. iv. July, 1875. ' Voyage or 'Black Miao' tribes, so called from the usual colour of their garments, are scattered all over the province of Kueitchou ; the greater number of them were subdued in 1735, and those who, still independent not many years ago, were called Seng Miao £f£ "jij, or Raw, i.e. untamed or independent Miao, used to be found in the western part of the same province. Their language is known only through Chinese sources, which give us a few 1 Cf. Miao Man hoh tchi, bk. iv. fl. 9-10. s At P'ing yuen. 3 And published from the Miao fang pei Ian, by Dr. Edkins, o.e. 4 The pronouns, wei ' I,' mu ' thou,' ngo, hai ' he,' are Mon. 6 As the ideological indices of the Lolo class are 1 4 5 8 III., the alteration of the second indice is most probably a result of Chinese influence (13 6 8 VI.). 50 isolated words collected at first by officials, 1 and afterwards a small vocabulary of 120 words. The isolated words are the following : Ah-mei, 'woman.' Ma-lang-fang, 'a bachelor's house,' 2 by the Pah-tchai tribe. Tu-tu, 'an embroidered square on the stomach.' 3 Lo-han, ' unmarried man,' by the Tsing kiang tribe. Lao-pei, 'woman,' by the same. Mei-niang, ' wife,' by the f|ijj tribe. 4 Tung-nien ' those of same name.' Ma-lang, 'youngsters.' 5 These words require a few remarks, as they present some inconsistency resulting from the broken and intermingled genealogy of all these tribes. Ah-mei ' woman,' and mei- in mei-niang, ' wife,' are similar to the Siamese me 'mother, wife,' and to the Laocian ime * mother,' while mei-niang has already appeared as proper to the Pan-hu race (§ 73). Ma-lang fang is said to be literally ' young men's house,' so that we have here a Chinese word, fang, and a pre- position of the genitive. 86. The larger vocabulary which has been published by Dr. Edkins is instructive. There we find some of the same words as those above quoted : for instance, ami, ' mother ' ; tung nien, 'friend.' There are two class-articles, or deter- minative prefixes : ta- for animals, and Icuo-, ho-, ha-, a-, for all that is human, -pel or -pa in the above lao-pei, ' woman,' 1 Miao Man hoh tchi, iv. 6. 2 As among the Tsing Tchung Kia, the Huang and Nung trihes, all belonging to the Mon-Taic stock. Among the Heh Miao the custom is peculiar to the Pah- tchai tribe only. The latter's name is written in the Chinese documents with two symbols meaning the ' eight stockades,' which is the name of a place probably derived from the name of the tribe, and in which transcription we may see nothing more than a foreign graphical folk etymology.- The name Pah-tchai is apparently cognate with the Siamese p'u tehai (Pallegoix, Bid. Zing. Thai, pp. 180, 687), meaning 'the ancients, grandees.' As to the practice of the Bachelor's house in every village to stay at night, it is well known in India ; there we find the dekachang of the Garos, the dhangar bassa of the Bhuiyas, the dhunkiiria of the Oraons, and also among the Paharias-Malers, the Gonds, the Kandhs, etc. Cf. Col. E. D. Dalton, Ethnology of Bengal, pass. 3 Cf . $ 70 n. 3. * Luh Tze-yun, T'ung k'i sien tchi, f. 20. 5 Ibid. f. 20». 51 and in te-p'a, 'daughter,' seems to be a feminine word of gender. The numerals are Mon. Tchim nung, ' eat rice,' tarn net, ' carry water,' lieu pu ' ascend a hill,' pe teu 'light A fire,' indicate the position of the object after the verb. Ha-mei-la, 'first day of the month,' where ha is 'first' and la 'month,' shows a pre-position of the adjective and a post- position of the genitive. The adverse position of the genitive exemplified in ma-lang fang is also evinced by other instances, so that the ideology of the language is hybridized. The indices are \ \ 6 0. The position of the subject is not evidenced. 87. TheYAOMiN J|g Jg tribes inhabiting the mountainous region of the N~.E. Kuang-si and N.W. Kuangtung provinces, in the conterminous prefectures of Kuang-yuen and Lien- tchou, speak a mixed and hybrid language. We have as sole data a vocabulary of 65 words from Chinese sources, and extracted from the Kwang si t'ung tchi l as follows : I. Objects in Nature. Sky, ngang ; moon, fa ; star, hang ; wind, Wang ; clouds, kia ling ; earth, lie ; road, kwo ; fire, tan. II. Man and Family. Man, kuei ; father, pa ; mother, ma, man ; father's father, pan ; father's mother, pan man ; father's elder brother, pi ; elder brother, Ian pa ; younger brother, Ian ti ; husband, Mnan ; wife, a ; elder sister, ko ; younger sister, liau kuei ; son, tang ; daughter, pi ; grandson, tang sheng ; wife's father, ta ; wife's mother, tu ; wife's elder brother, liau shu ; wife's younger brother, tang ahu. III. Metals. Silver, yen. IV. Animals. Fowl, kiai; pig, mien; dog, Hang. Y. Parts of the Body. Hair, pienpi; eyes, tsi kang mien ; ears, tsi kia pa. VI. Food, Eating. "Wine, tieu ; rice, hai ; flesh, yen yen ; vegetables, ts'ai, wei. VII. Implements, Clothing, etc. Table, t'ai tdu ; bed- stead ; t'ai tsung ; stool, t'ai Mai ; clothes, au ; petticoat, teng li. 1 A Statistical Account of the Province of Kuang-si, in Dr. Edkins, The Miautsi. 52 VIII. Agriculture. Grass, wu. IX. Pronouns. I, ye ; thou, meu. X. Numbers. 1, ki ; 2, i ; 3, fom ; 4 k s«; 5, wu ; 6, 7, A wo ; 8, jmw<7 ; 9, hung ; 10, sAe. XI. Verbs. Drink, hau ; eat, nang ; sleep, j»et ; die, t'ai ; bury, y. XII. Sentences. Eat rice, yen nun ; take a wife, shau ■ 7m^; marry out a daughter, liau pi; have a son, tfwrag' to^; to face the fire, fo fera. 88. The glossarial affinities are composite ; out of 55 words, 14 or one-fourth are Taic, and their nearest cognates are in the Seng Miao, Tchung Miao, Kih lao, etc., dialects. The next elements of importance in the vocabulary are Chinese and Tibeto-Burmese. The numerals 1, 2, 3, are similar to those of the Nagas of N.E. India, Khari, Nam tang, and Tablung tribes ; 4, 5, 10 are simply Chinese ; 6, 8, and 9 are alterations from the same stock nasalized. The pronouns are Mon. Only slight traces of class- articles. The ideological indices which can be detected are 1, 4, 6. The genitive precedes, and adjectives follow their nouns, and the object follows the verb. XI. The Pke-Chinese Aboriginal Mon-Khmeb Dialects. 89. From internal evidence, which agrees with the fore- going facts, the ancestors of the language and civilization of the Annamites, and partially also of their race, must be sought for in Central and Eastern China. We hear from history that the former population of the south, between the Kuangtung province and Tungking, both inclusive, was generally displaced by, or intermingled with, half a million of colonists drawn chiefly from the region of modern Tchehkiang and its west, by Jen Hiao in 218 B.C. 90. The traditions set forth at the beginning of the Annamite history, however completed they may have been subsequently, conceal under a native dress, several proper names which read in Mandarin Chinese turn out to be 53 familiar to us as belonging to Chinese heroes and to the aforesaid region of Central China and the South. We can only allude here to the matter which we have treated at length in China before the Chinese, where we have shown as a great probability that they date from the beginning of the Chinese Empire, end of the 3rd century B.C. Kinh-vuong vuong, in Mandarin King-yang wang, or 'King of King-yang,' the name of their first legendary king, is borrowed from King-yang, the name of a locality in proximity to the capital of the Ts'in Empire, Kuan, now Si-ngan in Shensi. He is reported to be the son of a Chinese Prince by a girl of the race of immortals (the race of Peng or Pan-hu), near the ngu lanh, Mandarin Wu-ling, otherwise the 'five mountain ranges,' a name given to the mountainous southern boundary of the new Empire. The same prince married a wife from Dong dinh quan, Mandarin Tung ting kiun, otherwise province of the Tung ting lake (in Hunan, ~N.), and belonging to the dragon, otherwise the Lung race, well known in the non- Chinese ethnology of the country. The king Lak-long, issue of this union, was the first of a series of eighteen rulers, the last of whom ended in 257 B.C. At the rate of twenty- five years a reign, the highest average possible, these specu- lative data lead to circa 800 B.C. as the probable date of these beginnings, which therefore would have taken place when the state of Ts'u in Hupeh and Hunan S. was in full prosperity. 91. The boundaries of the kingdom of these early Anna- mese rulers were, according to tradition, on the east the sea, on the north the Tung ting lake, on the west Pa and Shuh, ( both names for Szetchuen. The second dynasty goes by the name of Thuc, in Mandarin Shuh, the name of Szetchuen, with one ruler whose reign of fifty years ended in 202 B.C., when the third dynasty begins. The latter is no less than that founded by a successor of Jen Hiao, Tchao T'o, a rebel Chinese general who established his sway all over the maritime provinces of the south, extending from Fuhkien to Tungking; it lasted with five rulers until 112 B.C., when it submitted to the Chinese dominion, which, however, was merely nominal in some parts, and not at all established 54 on the east. It was recognized in Tungking from that date, with the exception of three years (39-42 a.d), until 186 a.d., when a native king, Si-nhip, ruled for forty years. It was this king who introduced the Chinese literature, and pro- hibited the use of the phonetic writing hitherto employed by the Annamites. 1 92. Two languages are used in Annam. One employed by the literati only is pure literary Chinese, 8 with the old sounds of the Ts'in period attached to the written characters. 3 It is the Sinico-Annamite, this very dialect, which, with a necessary allowance for decay and • self-divergence, rightly deserves the qualification of the most archaic of the Chinese dialects. 4 93. It is a curious fact, that its existence was not, in the minds of many scholars, separated from that of the other language, the vernacular Annamese or Cochin-Chinese, which belongs, as recognized by John Logan, and though full of Chinese idioms, to the same family as the Mon or Peguan. s The Annamite has been largely studied, and numerous are the grammars and extensive vocabularies of this lan- guage. 6 We need not enter here into details, and it will be sufficient to state that the ideological indices of the Annamite are 2 4 6 8 VI. 1 On this writing, cf. Beginnings of Writing, i. 44. 2 A short grammar of this language is .given in Notions pour servir * ne Miao who inhabit the prefecture of An-shun in the centre west of the province of Kuei- tchou. The affinities are Mon-Tai, with a decided leaning towards the Tai, as shown by the numerals and the pronouns, which evidently belong to this family. As in all the other Miao vocabularies, there has been a not-unimportant absorption of Chinese and Lolo words. The use of determinative prefixes is revealed by the known tu- before the names of animals. Ideological indices, 2 4 6 0. 107. The Tchiing-kia tze <$ j^ -jF, also Tchung Miao, 1 Miao Man hoh tchi, it. 4. s Extracted by Dr. J. Edkins, in his Vocabulary of the Miau Dialects, from the Eing-y fu tchi. 3 Kev. Brounton, in China's Million, 1883, p. 62. 1 In the Hing-y fu tchi, and translated by J. Edkina in his Vocaiulary of the Miau Dialects. 62 or Y-jen, whose own name is Pu-y, 1 speak a Tai language so closely connected with the Siamese that Mr. Abrand, a missionary who had resided in Siam, was soon enabled with- out great difficulty to understand them in Kueitchou. The Chinese notices about them state that they migrated north- wards from the region of Yung, S. Hunan, in the eleventh century. 2 On the other hand, their traditions say that their ancestors were originally from the Kiangsi province (E. Chiua), and they have kept in great respect the memory of Ma Wang. 3 The latter apparently refers to the same migration as that reported by the Chinese, but taken from a more eastern point of departure, where the two provinces are conterminous. They are now in scattered settlements over four prefectures of Kueitchou province, and also in the north of Kuangsi. 108. "We are indebted to the missionaries who have furnished the notes with which Mr. E. Lasserteur has written the article of the Missions Catholiques, quoted in the foot- notes, for some grammatical information on the language of those of Kueitchou, and to Deka in Notes and Queries on China and Japan i for eight words of those settled in Kuangsi, whom he calls Tchung tze. 1 E. Lasserteur, Be quelques triius samages de la Chine et de VIndo-CMne, in Zes Missions Catholiques (Lyon, 4to.), 1878, t. x. p. 308. 2 At the time of the five dynasties (a.d. 907-959), when Ma yn was king of Tsu, they migrated from the government of Yung (Yung kuan). Cf. Miao Man hoh tchi, bk. iv. f . 1 . The localization of Tung knan is not identified, hut the connection of Ma yn as the cause of the migration may aid in the solution of this little geographical problem. Ma yn ruled over Hu-nan as king of Tsu from a.d. 908 till 951 ; and in A D. 928 he successfully attacked King Nan or south of King (S. Hupeh), and in a.d. 941 the Man of Ki tchou (N.W. Hunan) were pacified. His rule had nothing to do south of the Mei-ling and Nan ling, where the southern Han dynasty was established (a.d. 917-971). Therefore, as the migration of the Tchung Ma tze happened during and under Ma yn's rule, they must have crossed westwards along N. of the mountain ranges, S. of Hunan ; and Tung kuan, some- times written also Yung yng, must not be mistaken for Yung tchou, now Nan ning fu, in S. Kuangsi. On the wars made by Ma yn, and the dates, cf . Li Tai Ti Wang nienpiao, Wu tai. Luh tze-yun, writing circa 1650, says in his T'ung k'i sien tchi, f. 2 (Shwoh ling collect, bk. 29), that Ma yn drove away these people until Tching-tu in Sze-tchuen. 3 "La famille du martyr Jerome Lou de Mao Keou pretend posseder la table genealogique de ses aieux depuis l'epoque de leur emigration du Kiangsi au Kongtcheou." E. Lasserteur, JDe quelques tribus sauvages de la Chine et de I'lndo- Chine, in o.c. 1877, t. ix. p. 149. 4 Spoken Language of the Miau tsu and other Aborigines, in N. and Q., Hong- kong, 1867, vol. i. p. 131. 63 The words are the following : 1) no mung 'pork.' 2) to ma ' dog.' 3) to wai ' ox.' 4) to pit 'duck.' 5) to mo ' pig.' 6) to ki ' chicken.' 7, 8) lean ngai ' eat rice.' The words 2, 5, 6, and 7 are of Siamese parentage ; the categorical prefix to needs no comment, and as kan of kan ngai means ' eat,' the object follows the verb. 109. The missionary notice says that in Tchung Kia the adjective and the genitive follow the noun, unlike the Chinese, and like the Siamese and Annamite. It has no declensions nor conjugations, and this usual statement of persons unfamiliar with comparative philology goes on with the also usual mis-statement that the same word may be noun, adjective, verb, etc. ; the position of the words in the sentence and the use of particles determinate the con- ditions of space and time of the action. There are tones and accents which diversify words otherwise apparently the same. In Tchung kia frequent use is made of categorical prefixes, such as : Tu- for living beings : tu wen 'man,' tu-kai ' fowl.' x JDant- or Da- for objects : dant-tcho ' table.' Leg- or lee- for all that is born from, or produced : leg mi ' a boy,' leg beng ' a girl.' Pu- for reasonable beings: Pu-ha 'the Chinese,' Pu- yah ' the brigands,' Pu-y, ' themselves.' a- for the proper names. 2 1 1 0. Besides these data, there is a vocabulary of Chinese origin, 3 of 234 words, compiled in the S."W. and S.E. ' Which Deka writes to hi, as we have seen ; in Siamese tua hdi, so says the notice ; in the vocabulary of Chinese compilation, tu kai. s Cf. E. Lasserteur, o.c. p. 186. 3 It is one of the vocabularies given by J. Edkins, Vocabulary of the Miau Dialects, from the Sing-y fu tehi. 64 of the province of Kueitchou, which agrees with all the above statements, without however furnishing any other basis for extending our knowledge of the ideological indices of this important language beyond 2 4 6 0. 111. The T'u-jen £ \, whose settlements extend from the east and centre of Kueitchou to the west of Kuang-si provinces, also speak a Taic language. 1 We have a short notice and a vocabulary of thirty-three words from the same source as the data about the Tchung-kia, whose original speech may have been strongly influenced by them. The Chinese notices say nothing of former residences, and their name means properly 'aborigines' in its Chinese garb; so far it is not an ethnic, and may have no other signification than the relative antiquity of residence of these tribes with reference to that of the others more recently arrived. Here is the vocabulary with its French spelling, and the com- parisons of Siamese made by the same author : 2 father tou-peu Siamese pho. mother tou-mei }> me. son tou-lak }> luk. daughter lak-sao i> luk-sao. man oug-hon 9t ong- or khon-manut. woman lak-mei >> mia. male tou-tak 97 toua-phu. female tou-mei )> toua-mia. house an-loun earth thomh water nam >> nam. fire . foi » fai. wind lorn >l lam. heaven au-boen dog tou-ma >> tou-ma. cat tou-meou )) tou-meou. tree keu-may )) tou-mai. rice hao 9> khao. i Miao Man hoh tehi, bk. iii. f 3r, iv. f . 9». 2 E. C. Lasserteur, o.e. 65 bamboo keu-may Siamese mai-phai. hand au-moy » mu. foot an- ten >t tin. one yt two ngioi three . mm >j sam. four soi » ti. five ha » ha. six lok » ok. seven t&it j» chet. eight pet 9t pet. nine koou }) kao. ten chip )! sip. 112. The numerals are Chinese like in Siamese, where, however, the first two were preserved from the older state of the language. 1 Adjective and genitive come after their noun, as in Siamese. Nothing is said of the position of the subject, nor of the object in relation to the verb, except that the construction of the sentence is analytical, as in French. Only a few names can be used separately without prefixed class- articles, such as : tou-, as in Tchung-kia, Pan-y, Yao-jen, and cognate languages. 2 ong- for 'men,' 3 the significance is precise enough to save the use of the word hon ' man ; ' for instance, ' how many men ' moi ka lay ong, where hon, which ought to come at the end of the sentence, is dropped because the article is sufficient and does not permit of any misunderstanding. mak- for ' fruits.' . an- for 'objects.' 4 ty- for ' woodwork.' 5 1 Cf. infra, § 117. 2 In Siamese tua, in Shan to. 3 In Siamese onle. * In Shan an. 6 In Shan hsilc. 66 The Tu-jen language, says the missionary, gradually mingles with those of the Tchung-kia and of the Miao-tze. The construction is similar to that of the French, whence the indices 2 4 6 8 YI. 113. The Pai-y 1 so called are now chiefly met in the south and west of Yunnan, where their name has become the generic appellative of the Shan tribes still living there. They are undoubtedly, with such transformations in race and language as have resulted from subsequent inter- minglings, the descendants of the old Pa £, people of Eastern Szetchuen and Western Hupeh, known to the Chinese since 1970 B.C., when ' a Chinese envoy was sent to them to preside over litigations.' 2 The link can be traced through ages, and. has never been obliterated. 114. We know nothing of their original language, as no specimens have been preserved. We only know it from a recent document. It is a vocabulary compiled by the Chinese. The teaching of this language along with that of the Pah peh sih fu, s another Shan dialect, was added, after 1644, to those previously taught 4 at the Translatorial 1 Their name is variously -written in Chinese now-a-days J|| ffi , *jj ^ , •lb S > jJ J^a > f3 31 ' e * c ' ■*■* 011 S n * to be written [^ ^ . 2 Tchuh shu k'i nien, Ti k'i, 8th year. 3 71 "|j jjft iff , Ht. ' eight hundred wives,' so called, say the Chinese, from the fact that their Tu-yu or chief had this number of wives, each of them having a separate encampment. They use to tattoo flowers and birds between the eyebrows (cf. Luh Tze-yun, T'ung k'i sien tchi, i. 7v). The legend may have resulted from the name, in its Chinese form, and this form may have been a play made by the Chinese in transcribing a foreign name. Mr. Ney Elias, in his Introductory Sketch of the History of the Shans in Upper Burma and Western Yunnan, Calcutta, 1876, p. 3, supposes them to be Karens, but the specimen of their language which has come to us, along with that of the Pai-y, goes against this view, as it is that of a Shan dialect. The following list shows it plainly : fa, heaven ; mo, cloud ; lie, sun ; leng, moon ; huan, smoke ; fating, cloud's colour ; falanq, thunder ; fen, rain ; naotchanglung, polar star ; nao, star ; luklie. hail; Itm, wind; famie, lightning; mohmm, cloud;' meinung, snow; meikan. ice; nankang, dew ; me'i, hoar frost ; lang, rainbow ; nam, water; menam, river : menamfa, river of heaven or milky way ; se, spirit ; pi, devil ; lai, mountain ; ITn. stone ; na, field ; pahmai, forest ; nung, sea ; tin, earth ; menamkung, lake pulatchao, king ; pi, elder brother ; mong, younger brother'; luk, son ; lu, head na, face ; .st- lichen Asiens, pp. 281-295 of Techmer's Zeitschrift fur Sprachwissenschaft, vol. i. Leipzig, 1884. And also Prof. G. von der Gabelentz, Sur la possibility de prouver une ajiuite ginialogique entre les langues diles Indo-Ohinoises (IV. Congr. Oriental, Firenze, 1878, p. 283) ; Dr. Em. Forchhammer, Indo-Chinese Languages (The Indian Antiquary, July, 1882). Also Brown, Comparison of Indo-Chinese Languages, Calcutta, 1837. 1 Spoken Language of the Miau-lsst and othir Aborigines, in Notes and Queries for China and Japan, vol. i. pp. 131-132 (Hongkong, 1867, 8vo.). 71 119. The more numerous affinities of this vocabulary are common to the Seng Miao, Kih lao, Tsing Miao, and Peh Miao. Class-articles, if any, are not apparent, excepting for the parts of the body, whose names in several instances are preceded by a-. The numerals 1 to 7 are Mon; 8 is a variant of the type common to Sheng, Peh, Blue Miao, Pan-y, Pan-yao and Long-ky Miao. The numerals 9 and 10 are mere variants of the Kuenlunic. As ideology, the postposition of the object to the verb is the only certain point ; the postposition of the adjective is also pretty sure, and the preposition of the genitive is only made probable. The indices will apparently be 1 4 6. 120. The Hotha Shan tribes, on the S.W. borders of Yunnan, ' a not very tall people,' speak a Shan dialect, lately hybridized by Chinese influence. A vocabulary of 179 words including the numerals and four short sentences have been compiled by Dr. J. Anderson of Calcutta. 1 It is mixed up of words from neighbouring tribes, Kakhyen and Li- so, besides the Chinese. The ideological indices 14 6 only are illustrated in the aforesaid documents. 121. This dialect of the Shan is not the only one in this region which displays proofs of hybridism. The Khamti, 2 who have extended into Assam, have been strongly impressed in race and language by the Kakhyens or Singphos, 3 amongst 1 Report on the Expedition to Western Yunnan vid. BhamS, pp. 99-101, 40 1-409. " For the Khamti, cf. William Bohinson, Khamti Grammar, pp. 311-318, of Journ. Asiat. Soe. Seng. 1849, vol. xviii. ; Eev. N. Brown, Khamti Vocabulary, pp. 342-349, ibid. ; Pere Desgodins, Vocabulaire Kham di Mou oua, in Mots principaux des langues de certaines iribus qui habitent les bords du Lan-tsang Kiang, du Lou-tze Kiang, et Irrawady, Bullet. Soo. Geographie de Paris, 1872. Mr. Edouard Lorgeon, who seems to have known nothing of the Khamtis, wrote about the latter some Remarques relatives au vocabulaire du Mouhoa, p. 28 of Bulletin de la Soeiete Acadimique Indo-Chinoise, vol. i. Paris, 1881. Major C. B. Macgregor has published a vocabulary of 500 Khamti words, along with the Singpho Vocabulary which accompanies his Singpho Grammar. 3 For the Kakhyen or Singpho. cf. William Babinson, Singpho Grammar, and Vocabulary, Journ. Asiat. Soc. Bengal, 1849, vol. xviii. ; J. N. Cushing, Gram- matical Sketch of the Kakhyen Language, pp. 395-416 of Journ. Roy. Asint. Soc. 1880, vol. xii. ; C. B. Macgregor, Outline Singpho Grammar, and Vocabulary, Shillong, 1887 ; P. A. Bigandet, Comparative Vocabulary of Shan, Ka-kying and Palaong, and J. E. Logan, The Kakying, in pp. 221-232 of Journal of the Indian Archipelago, Singapore, 1858, N.s. vol. ii. ; Dr. J. Anderson, a Kakhyen vocabulary of 250 words, in his l/eport, above quoted, pp. 400-408. Also some grammatical remarks by Dr. Forchhammer, in Notes on the Languages and JJialeUs Spoken in British Burma, Rangoon, 1884. 72 other tribes of cognate stock. Though their vocabulary is still strongly Tai'-shan to such an extent that it cannot be denied, their ideology, following the admixture of blood, has deviated from the ideological standard 2 4 6 8 VI. of the Ta'i-shan family, and displays the indices 2 4 5 8 III., which show a large adoption of the Kakhyen ideology 14 5 8 III., which is also that of the Tibeto-Burman group. 122. The Li Jen ft A, also ¥u Jen |J| A, 1 of the great island of Hainan, are apparently a mixed population made up chiefly of refugees from the non-Chinese tribes of the Chinese continent. The following vocabulary is due to the exertions of Mr. Robert Swinhoe : 2 Heaven, lai fa; 3 sun, tsa van; 4 " moon, leu nan; star, ta plao; 5 earth, fan; 6 water, nam. 1 Father, pah ; 8 . mother, may, pai pai ya ; 9 elder brother, / yong ; younger brother, ko ong ; elder sister, k'au ; younger sister, hu ong ; son, ta bo man ; daughter, ta bo pai ko. Bird, tat ; 10 sheep, ch'i ; cat, ping nai. u Head,/« tcu, dau, wa la ku gan; eyes, ucha; ears, m, t-sun sha, seng sha ; mouth, mom, pom ; hand, tarn ; foot, k'ok. Eat rice, k'an ka, lu.t'a; to smoke, luju; tobacco pipe, t'auja. Knife, kliu ka; one bow, van vat; two arrows, ten pun tiek ; man's house, hau po plungao. I, pun, 12 hau or ho ; thou, meu ; he, pun ; this, pai heu ; that, pai nei. Many, tai; few, to; bad, teh tuy. Sixty years, tvmfo tai; morning, leu ; evening, ko fan. Have, du ; shoot, cheu. Yes, man ; no, wei. 1, van, ku, ch'it ; 2, tow, do, tau ; 3, tsu, su, fu ; 4, Wo, san, shao ; 5, pah, ma, pa ; 6, turn, nom, turn ; 7, to, situ, ten ; 8, ho, du, geu : 9, fan, fen, fai ; 10, lapoom, pu tit, fu tit ; 1 On the Li-jen, cf. Luh Tze-yun, Fung k'i sien tchi, S. 10-12. 2 Mr. B,. S. handed his notes to Dr. J. Edkins, who inserted the words in his Voeabulary of the Miau Dialects. 3 Cf. Siamese fa. 4 Cf. Siamese sawan, heaven. 5 Cf . Siamese daw. 6 Cf. Siamese fun, dust. 7 Cf. Siamese nam. 8 Cf. Siamese p'e. 9 Cf. Siamese p 1 u ytng, woman ; m4, mother. 10 Cf . bird in Old Chinese and Kareng. 11 Cf . Kareng menyaw, maing-yaw, cat. 12 Cf . Chinese /^ . The other pronouns are Ta'i-S.han. 73 11, la pun wu ; 12, la puk lau ; 13, la pun pih ; 14, la pu k'o ; 15, la pu ch'i; 16, lu pu ch'it; 17, l.a pu tu; 18, la pu tau; 100, lau van ; 1000, longeen. How old are you ? meu pu tala hoe pone ? 123. The glossarial affinities exhibit a Tai-Shan parent- age, which cannot be denied, though the language is certainly not pure, ' and shows traces of other influences. In the numerals, for instance, which are given in two or three series, similarities exist with those of some tribes of Formosa. But they are remote, and do not come from a direct relation- ship ; they are • apparently survivals of a former state of things, previous to their respective migrations, when their various ancestors had relations between themselves on the continent. From the above list of words, three of the ideological indices may be inferred. They are 14 6 0. Unfortunately the fourth indice, that which refers to the relative position of the subject and verb, is not exemplified. 124. The Li are reputed to have known the art of writing, which they seem to have forgotten. Capt. J. Calder 1 has found near Yu-lin kan some characters scrawled on the walls of a temple, which I think may have belonged to the writing of Tsiampa. 2 We know that several migrations from the latter country to the island of Hainan took place in the tenth century. 3 In some parts of the island, ' the Li women carry a piece of lacquered wood, on which are written several lines of a ballad ; the writing however is like the wriggling of worms, and cannot be deciphered.' 4 125. Another list of words of a possibly different dialect, that of the Loi, has been written down in Annam by M. J. Moura, 5 from the mouth of a Chinaman, who had 1 Notes on Eu'inan and its Aborigines, where a facsimile is given. China Review, 1882, vol. xi. pp. 41-50. 2 Cf. my Beginnings of Writing, i. 45, ii. 235, 236. 3 D'Hervey St.-Denys, Ethnographie des peoples etrangers de Matouanlin, vol. ii. p. 547. ' B. A. Henry, Zingnam, or Interior Views of Southern China, including Travels in the hitherto Unlraversed Island of Hainan, London, 1885, 8vo. 6 J, Moura, Le Royaume du Cambodge, i. 513. 74 spoken the language some forty years before, and whose memory was not quite safe : Man, nam ; woman, sdbo ; wife, moa ; body, nga ; hand, chhean ; finger, sean ; mouth, mok ; leg, hong. Bull, ngealc ; buffalo, ngalc ; horse, hi ; duck, hek. Sky, thoang ; water, tut ; fire, fat ; cold, koa ; hot, nguon. Formerly, kou ; to-morrow, maso ; now, hau ; far, Aow^ ; much, tot. Tree, chheong ; bam- boo, tfAoa; flower, ho; house, su; -was., mac. Beautiful, phia ; white, mac ; blue, suum ; great, tang. Buy, peang ; love, &w<7 ; go, pdan ; be afraid, kea ; drink, c^ifo ; break, thoac ; understand, khean; ask, tho; give, «'o; sleep, ma; go in, seang; eat, cAea. There are many Ta'i-shan affinities in this list, but there are also numerous words which differ from the Li dialect. There are no instances illustrating the ideology. 126. But the distinction between the Li and the Loi is by no means established, and may be simply a local, difference of pronunciation for the same name ; in the same way, the differences of vocabulary may be simply regional peculiarities of one and the same language, unsettled and not fixed, as becomes to languages without literature or writing. XV. The Pre-Chinese Aboriginal Negritos. 127. The languages spoken by the tribes of this dwarfish race, which formerly were settled in China proper, have not left any modern representative that we know of. Such tribes fell in proximity to the Chinese Bak tribes, about 2116 B.C., when the latter already immigrated into the Flowery Land, and advanced eastwards of the great southern bend of the Yellow River. Some tribes of the same race are spoken of in the fabu- lous geography of the Shan ha'i king, a few centuries previous to the Christian era, and in later writers, circa 285 a.d., the Chinese advanced in the region that is now the S.E. of their An-hui province, and met there again some of these dwarfish tribes. 1 "We hear no more of them in the Chinese annals ; 1 I have collected all the available information on the matter in Chinese sourceB in a special paper : The Pygmies of the Chinese ; A Contribution to the Study of the Neyrito Race. 75 but Friar de Odoric de Pordenone, about 1330, mentions them in the relation of his journey. 128. Nothing is said of their language, and no landmarks exist as yet, which may permit us to discover any survival of its former influence, if any. Their race has shown frequently a great propensity to give up their own language and learn that of their neighbours, as they did apparently in the Philippine Islands; 1 also in the Shan and Cambodian regions. Elsewhere, in the Malayan peninsula, the Mergui Archipelago, and the Andaman Islands, their dialects are, perhaps, more like one another thau could be expected from their respective surroundings ; but this is not sufficient to permit any inference as to the original characteristics of the pristine language of their race. As a matter of fact, the dialects spoken by the Negrito tribes still in existence form two like groups widely distinct, 2 and nothing beyond what we recall below shows exactly which of the two groups was formerly represented in the Flowery Land. The historical probabilities are in favour of the group to which the Kamuks, Canchos, and Tjraos 3 belong; the latter name, Tjrao, is but slightly different from the old appellative Tiao-yao, or ' Dark Pygmies,' of the Negritos of China. But in that case they must have had their language strongly modified by their M5n-Khmer and Tcham surrounding neighbours. 1 " The Negrito languages of the Philippines appear, judging from the scanty specimens in my possession (says the eminent Prof. Georg von der Gabelentz, of Leipzig), to enjoy grammatical systems very similar in fullness, richness, and in the phonetic means employed, to those of their light-coloured neighbours." — The Languages of Melanesia, in Journ. Roy. Asiat. Soe. 1886, vol. xviii. pp. 489-490. Dr. H. Kern, of Leide, in the Bijdragen tot de taal- land- en vol- kenkunde van Nederlandsch-Indii, 4th ser. vol. vi. pp. 243-264, claims for their language a character thoroughly Malayo-Polynesian, and largely connected with the Philippine dialects, with certain admixture from more remote members of the family. 2 This is shown by their ideological indices. «) Andamanese, 14 5 8 1. III. ; Silungs, Samangs, 14 5 8 III., and, so far like the Papuas of New Guinea. b) The Kamuks, Kameits, and Tjrao have the indices 2 4 6 8 VI. of the Mon- Tai'. Should the Negritos of the Philippines have the same indices as the Tagala, these would be 2 4 6 7 IV. , therefore belonging to the second group. a A vocabulary collected from the Kamuck has tieen collected by Mr. Holt Hallet and will soon be published. One from the Concho has appeared in J. Moura, Le Camhodge, vol. i. pp. 439-447 ; and one of the Tjrou by U. Aymonier in Excursions et Reconnaissances, No. 24 (Saigon, 1885), pp. 316-316. 76 XVI. The Pre-Chinese Aboriginal Indonesians. 129. In the second part of the present work, 1 we have called attention to the striking influence exercised by a special stock of languages on certain ancient texts in Chinese character, written at a time and in a region where the Chinese, in their advance towards the west of China proper, had fallen into contact with some native tribes, Indonesian in their customs, and therefore, judging from this influence, Indonesian also in language. The former population of the country has been so completely swept away by the Chinese, that no distinct remains have been hitherto pointed out. However, we do not despair of hearing of some surviving tribe of Indonesians still taking refuge in an out-of-the-way corner of the mountains, such as in the Peiling range, west of the province of Fuhkien, and having preserved something of their older language. 2 Anyhow, in the present state of 1 § 23, above. 2 On the surviving rude tribes of this part, cf . Geo. Minchin, A Race of Men and Women Living at Pei-ling, in Notts and Queries for China and Japan, 1870, vol. iv. pp. 121-122. From their reported tradition these seem to belong to the Pong race. Eev. E. N. Lion, Notes of a Tour in South Chekiang, in The Chinese Recorder and Missionary Journal, 1875, p. 261. Mariini in the seventeenth and Marco Polo in the thirteenth century had mentioned such tribes. But nothing has hitherto been said of their language. The Rev. F. Ohlinger has published A Visit to the ' Dog-eared Barbarians ' or Sill-People l]j 5* near Foo chow, in 1 886 {ibid. vol. xvii. pp. 265-268), who call themselves Sia-bo, and are apparently the same as those described by Mr. Geo. Minchin. They claim to have come from the Canton province. The following triple list of words, to which I subjoin the so-called Chinese, and the Hakka numerals, come from the above-quoted article : English. SlA-EO. FuHTCHOTT. Hakka. Chinese. tea teh'6 da ts'a tcha water ssii jui shut shui house lau teh'io wuk wu{k tree shii tehtu shu shu earth t'i de t'i ti heaven t'ang t'ieng t'en Hen man nging nbng ngin jen (njen) sit teh'6 801 ts'6 tso sedan Ic'iun gieu yi kino pen bik bek yut pitt paper /» jai tshi tche word tch'i J* s tze wheat ma mah mah mai{k cow, 01 ngaou ngu ngu niu hat mo mo man mao hand ssiu tch'iu shu shou foot giok k'a kyok kio(k 77 our knowledge, we must be content with the indirect linguistic proof here mentioned, besides a knowledge of several disjecta membra, now representatives scattered west, south, and east of the former Indonesian dialects on the soil of the present China. 130. The geographical position of the Gyarung, common on the north borders of Tibet and China, so far apart from their cognates in language, and also their smaller stature compared with that of their neighbours, shows them to be a displaced population. Let us remember that at the time of the leadership of Shun, when the Chinese made their first advance. into Honan, some tribes hostile to them had to be removed to the west, 1 and we may accept the Chinese statement that they are the descendants of these removed tribes, now, much altered by the surrounding influences. The removed people were a northern portion of the San Miao, whose central quarters lay around the Po-yang lake, and also to the north of the Yaitg-tze kiang. The descendants of these Miao now in the S.W. prove in their languages to have common features with that of the Gyarung, a name which, English. SlA-BO. FUHTCHOU. Hakka. Chinese. mouth joi tcftoi tsoi tsui field Vang tch'eng t'en tien rice mei mi mi mi bridge k'iii gio k'yau kiao stone sshiah sioh shak she(k vegetable tch'oi tch'ai tsoi tsai boat sshiong sung t'yang tchuen Girl S. jit niong giang Boy S. diang buo giang F. bung ngiikjdi F. toh'iong bujoi H. a-tsyau-moi H. a tsyau (tsai) C. ku-mung C. turip i i.ze one ek sioh yit yi{t two yong long nye erh, (m) hang three sang sang sahm san (sam) four si se see sze five ng ngo ng wu (ngu) six liik Uk luk lu{k seven tch'ik tch'ek ts'ut tsi(t eight bah biak paht pa{t nine giu gau kiu kiu ten hsilc sek ship shi(p Nothing is said of the ideology. The words show that this people had given up their own language and adopted dialectal Chinese. 1 Shu-king, II. i. 11. 78 by the way, is said to be nothing else than & Tibetan sur- name. 1 131. The chief, and so far as I know the only, data on their language are those given by Mr. Brian Hodgson in 1853. 2 These consist of a vocabulary of 176 words and a few remarks, given at the same time as six other vocabularies of Northern Tibet, and also in another paper. The venerable scholar, to whom we are indebted for so much material about the little-known languages of that region, was struck by the complicated system of prefixes and infixes affixed to the verb in Gyarung similar to that of the Tagala, and in order to show their relationship, in speaking of the Gyarung verb, he made use purposely of the statements of Leyden 3 about the said language spoken in the Philippines. 132. Owing to its long continuance under Tataric influences, the Gyarung has adopted the ideological indices of the Altaic class, namely 13 5 8 III., as well as something of the phenomenon known as vowel-harmony by the harmoni- zation of the prefixed vowel with that, of the significative vocable. And through the paramount influence of Tibetan, it has adopted the Kuenlunic numerals ; though preserving its older ones, as is often the case, for special uses of one and two. But it has preserved eastern features other than the complicated prefixed system. The numeral auxiliaries, or segregatives, and the class-prefixes inherent to the former state of the language, have not disappeared. 133. They have been strongly assailed by decay from wear and tear in unfavourable surroundings, but something of them remains, and cannot be mistaken. There is no list 1 Though perhaps a compound of Gya ' foreign,' in Tibetan, and a native name Lung or Rung. 2 Journal of the Asiatic Society of Bengal, vol. xxii. p. 121, reprinted in Selections from the Records of the Government of Bengal, No. xxvii. p. 173 (Cal- cutta, 1857); and in his book on The Languages and Literature of Nepal and Tibet (London, 1874, 8vo.), part ii. pp. 65-82. And also instances given in his essay on the Mongolian Affinities of the Caucasians, in the Journal of the Asiatic Society of Bengal, vol. xxii. pp. 26-76. They display the ideological indices 13 5 8 III. 3 In his paper still valuable, On the Languages and Literature of the Indo- Chinese Nations, in Asiatic Researches, 1808, vol. x. p. 209. Reprinted with bibliographical notes by Dr. E. Eost in Miscellaneous Papers relating to Indo- China (London, 1886, 2 vols.), I., pp. 84-171. 79 given of the segregatives, and only one instance is set forth by Mr. Brian Hodgson with the statement that they exist in the language. But the class-prefixes are largely ex- emplified in the vocabulary, yet the compiler has not con- sidered them with the attention they deserve. He has not seen that the present state is one of disorder arising from a decay not balanced by the usual regeneration resulting from linguistic surroundings of an identical stock. He has also only seen some of them. As a matter of fact we recognize four prefixes to nouns k-, t-, b-, and r-, corresponding in a rough way, yet not in all the individual cases, to the k-, t-, p-f and I- of the Miao languages of China. In some cases the vowel of the prefix has been changed, and even nasalized, to match with that of the vocable of signification. In cases of an s initial to the latter, the t prefix has fused into it. Another alteration is the preponderance taken by the dental prefix over the others ; not only superseding the less frequent r-+ and b-+ by superaddition and the k--\- altogether, but also showing a remarkable tendency to become mere signs of parts of speech; the t--\- absorbing all the nouns and the k--\- remaining with a more extensive field the prefix for adjectives and verbs. 1 134. In glossary the affinities of the Gyarung are with the Miao (Blue) of Kueitchou, the Tayal of Formosa, and especially the non-Kareng substratum of the Toungthus of Burma, all apparently branching off a long time ago from a former common stock. The resemblances are extensive with the Toungthus, amounting to 25 per cent, of the vocabulary, except this, important difference, that the latter under the Sgau-Kareng influence has apparently given up the use of class-prefixes. 135. Another remark may be added in winding-up that concerns the Gyarung; the monosyllables seem to be strongly objectionable to its linguistic sense, and almost in every case where a vocable is diminished to an open syllable by contraction or wear and tear, analogy or otherwise, a prefix 1 Nearly all the adjectives exhibit the prefix h-, while 75 per cent, of the verbs nave the prefix ta-. 80 is sure to be added. A single syllable in this language, as in so many others, does not present a sufficient basis for the, mind to stand upon and admit from its single sound, the concrete meaning generally required by a low standard of mental force. These remarks apply to the so-called monosyllabic languages of S.E. Asia, at large and in particular. 1 136. The Toungthus of Burma are generally looked upon as Karengs for all purposes; but this handy explanation of a difficult problem of ethnology has become insufficient. 2 Proofs may be adduced to show that their Kareng character is not more than a varnish which disappears under the pressure of criticism. The Sgau-Kareng affinities are ap- parently loan-words of recent date, and the other Kareng affinities are words borrowed in common by the Karengs and the Toungthus from the Mon stock. 3 137. They call themselves Pan-yau (also Pa-o), a name which recalls to mind the Pap-yao 4 tribes, and also Pan-yu, the old name of Canton territory. And they claim Thatun 5 as their ancient capital, wrested from them by the Mon- Peguans ; they declare that its name is derived from Tha-too, the word for 'laterite' in their language, and of which mineral the hills in the vicinity of this old city chiefly consist. They must have come down at an early period, about the Christian era, into the peninsula. 6 1 On the illusions about monosyllabism, cf. below, § 204. 2 On the Toungthus, cf. Dr. Ad. Bastian, Journ. Roy. Asiat. Soc. 1868. British Burma Gazetteer, i. 186-188. 3 Dr. Forchhammer, in Notes on the Languages and Dialects Spoken in British Burma, p. 11. The Eev. C. Bennett, to Rangoon, p. 15, and Mr. P. H Martyr, of Myaungmya, p. 16, testify in the same place to the mixed nature of the Toungthu dialect in comparison with the Kareng dialects. * Cf. above, § 69. 6 Situated on a tidal creek opening into the Gulf of Martaban, and formerly an important sea-port. Cf. Sir Arthur Phayre, History of Burma, pp. 27-28. Dr. Em. Forchhammer, in his Notes on the Early History and Geography of British Burma, Rangoon, 1883, vol. i. p. 3, claims the Pali name Saddhamanagara as the original appellative of ThatSn, but this is merely, as in many other instances, the grafting of a Pali etymology upon a foreign and previous name. We hear of this port from Chinese sources about the Christian era as Ta-tan, afterwards corrupted in Chinese phonology into Ta-tsin, and one of the several appearances of this geographical denomination. About l'a-tan = Ta-tsin, cf. Terrien de Lacouperie, The Sinim of Isaiah not the Chinese, p. 46 of The Babylonian and Oriental Record, January, 1887. 6 Albert Fytche, Burma Past and Present, vol. i. p. 341. 81 138. Their glossarial affinities with the non-Chinese tribes of China are similar to those of Qyarung, with whom the large proportion of similar words indicates a close relation- ship, though remote in time, as we have seen in a previous page (§ 134). 139. The Tatal of Formosa, 1 and probahly some other dialects of the same great island opposite the Chinese coast of Fuhkien, are entitled to a place in this survey, not because these dialects, ipsis verbis, were formerly spoken on the mainland of Pre- China, but simply for the reason that they are made up greatly of linguistic elements which have heen carried thence to the island. The uninterrupted introduction of linguistic and ethnical elements from the Philippines, the Celebes, the Liu-kiu islands, etc., have blended, often beyond recognition, all the original characteristics. Add to these the important Chinese influence always proceeding, also the • temporary and casual Dutch, Spanish, Malay influences and others which, though left undescribed, must never be neglected when dealing with uncultured tribes; and we may easily imagine that in some cases the relationship of the broken dialects of broken tribes must be given up in despair. The problems are further complicated like those in question here, when the superimposition and intermingling of languages take place, for the most part, between dialects and tongues, variously derived and diverged from branches of one common stock. 2 140. In addition to the aforesaid difficulties, we have still to mention the defective character of the documents which we have to work upon. A dictionary of the Favorlang dialect, compiled in the seventeenth century ; grammatical notices of the same; a vocabulary with sentences in the 1 On the Tayal and its group of dialects, of. M. Guerin, Vocabulaire du dialecte Tayal ou Aborigine de Vile de Formose, in Bullet. Soe. Geogr., Paris, 1868, xvi. pp. 466-495 ; L'Abbe Favre, Note sur la langue des Aborigines de Vile Formose et remarques sur la precedent vocabulaire, ibid. pp. 495-507. Mr. E. C. Taintor, in his valuable paper on The Aborigines of Northern Formosa (Journ. North China Branch Boy. Asiat. Soc. 1 875, vol. ii. pp. 63-88) has given a vocabulary of the Kabaran Pepohwan, and also one of the Tukan-Tayal. 2 All these remarks apply as well to the greater part of the languages spoken, of in this work. 82 Tayal dialect ; texts in Favorlang, Sideic and Old Pepohwan, with short lists of words from the twenty and odd other dialects, form the whole of the material at the disposition of scholars for study. 1 141. In the search after the relationship of languages, historical information and geographical convenience must not be neglected. The great island of Formosa, in proximity to the mainland, could not have received its populations exclu- sively from the outside islands. Whatever facilities the currents and winds may have given to the numerous migra- tions which, willingly or unwillingly, have reached its shores south, east, and north, the nearness of its coasts, in full view of the mainland, was a sufficient temptation for the inhabit- ants of the Chinese coast to venture on the journey. The pre-Chinese tribes of the maritime provinces of China were addicted to seafaring, 2 and their roving habits were for long an obstacle to the Chinese advance. When compelled to interfere because of internecine wars between the local princes of native states in 110 B.C., the Chinese troops withdrew as soon as they could, 3 leaving the country ' vacant/ as they say, or ' to itself,' as we understand it. The soldiers of the Son of Heaven remained several centuries before penetrating there again, as it was not before the fifth and sixth centuries that the country of Fuhkien was~seriously colonized and incorporated into the empire. It was during this intermediary period that some native tribes, driven gradually by the Chinese private colonists, were induced to pass over the channel and establish themselves on the great island opposite their coasts. 142. Shortly after their definitive settlement in the region of Fuhkien, at the beginning of the Sui dynasty (circa 593 a.d.), the Chinese were struck by the sight of the great island and the reports which reached them concerning it. They undertook two successive expeditions ; the first of 1 For the detailed bibliography, cf. my Formosa Notes. 2 "They are practised in aquatic Warfare, and skilful in the management of boats," says a report to the Chinese Emperor in 135 B.C. 3 Cf. the foregoing § 48. 83 these, an unsuccessful one, enabled them to learn that the language of the natives could be understood by the Kiien-lun men, and consequently led them to take such men as inter- preters for their second expedition. But there were already several languages spoken in the island, and the interpreters could not make themselves everywhere understood. This is the first intimation that a connection existed between some of the dialects of Formosa and of the continent, as 1 have shown elsewhere 1 that the aforesaid Kiien-lun men were none other than the non- Chinese tribes of the mountain range north of Kuang-si and Kuang-tung provinces. 143. Glossarial affinities and similarities in grammar and morphology have long permitted the recognition of a re- lationship between the Favorlang and the Malayan languages. The Tayal was subsequently admitted into the same group, and the Philippine languages were those with which the connection was admittedly the nearest. I shall not inscribe myself against this opinion, which is well supported, and with which I agree, though with some restrictions. The ideological indices of the Philippine dialects, such as Tagala 2 = 67 IV., Bisaya 2 4 6 7 IV, Pampanga 2 4 j> 7 II. IV., are peculiar, and the latter finds its nearest approximate in the Formosan Side'ic 2 4 | 7 II. IV. But the Formosan Tayal and Favorlang, each belonging glossarially to a separate group, exhibit symptoms of another kind. Their common indices are 1 I 6 7 V. which, as may be seen, are somewhat different from those of the Philippines, and their similarities in words are not restricted to the Tagala-Malayan languages. They go further than that, and they prove rather extensive, along with the Pre-Chinese dialects of the great Mon-Ta'i family, as well as with several Pre-Chinese dialects of the Kiienlunic stock of languages. Taking them in the order of their greater number of affinities, we find the T'u-man, the Kih-lao, the Ngan-Shun Miao, the Blue Miao, the Miao- tze, the Tchung-kia Miao, the Li of Hainan, and the Gyarung dialects. The use of prefixes of various sorts, so 1 Cf. my Formosa Notes, Linguistic Division. 84 conspicuous in the Philippine languages, and visible also in Tayal and Favorlang of Formosa to a lesser extent, are met with in the aforesaid non-Chinese dialects of China, though, perhaps to a smaller extent than they are in reality, because of the deficiency of documents. They exist on a large scale in Gyarung, as we have seen in the section devoted to this interesting language. 144. The remarkable ideological indices 6 7, common to the Formosan Tayal, Tsoo, and Favorlang, and to the Tagala, Bisaya, and Pampanga of the Philippines, which refer to the postposition of the noun-subject and noun-object to the verb, are illustrated in the Pre-Chinese soil by the influence it has left in ancient texts, as reported in § 23 of the present work. They are not evidenced in the aforesaid native dialects of China, because the linguistic data at our disposal stop short at the third indice, and show only the postposition of the noun-object, without any indication as to the position of the subject. The ideological indices of these dialects, 1 1 4 6, or 1 J 6, agree thus far with those of the Tayal and Favorlang of Formosa. Part V. The Pre-Chinese Intrusive Languages. Extinct and, Surviving Kiienlunie Dialects, §§ 145-179. XVII. The Kabengs or Btjrma and theik Pke-Chinese Cognate Dialects. 145. The Karengs of Burma are entitled to a special . entry in these pages. They have preserved some curious traditions, presenting several characteristics of a peculiar nature, in which they claim a northern origin and a connec- tion with China, and a relationship with the Chinese as well. They call them their younger brothers. However worthless such traditions may be in Indo-China, amongst nations deeply impressed with the greatness of the Chinese power, and rather open to the adoption of traditions from one another, as I have shown elsewhere, it is nevertheless a fact that a real connection has existed at a certain time between 1 For the Gyarung, cf. \\ 130-135. 85 the Karengs and several of the non-Chinese populations of pre-China. The evidence, interesting for us here, is that of the languages which point to a relationship now distant with some Kiienlunic tongues of the non-Chinese. 146. The glossarial affinities and the ideological indices indicate a formation by itself, on the same principles and with ethnic and linguistic elements partly similar to those of the Chinese, though distinct from its infancy and developed separately. Their individuality and separateness do not, however, exclude frequent intercourse and intermingling with the Chinese during the necessarily long period of their infancy, as they were living in proximity to the Middle Kingdom. Their formation, according to all probabilities, took place within the dominion of the non-Chinese state of Ts'u 1 (1050-223 b.c., in Hupeh, Hunan, etc.), and they were driven into the south-west at the time of the kingdom of Nan-yueh 2 (218-206 b.c). 147. Though it is difficult to 'know precisely how their ori- ginal dialects were different from their present state, 3 we may assume with great probability that they were vTibeto-Burman, and that their distinct and modern characteristics, such as their five tones and their ideology (indices 14 6 8 VI.) have been acquired in the course of their evolution, or, better, of their formation. Their nucleus belonged to the same group of inchoate dialects which, descended from the north, have evolved into the Nagas and the Burmese groups of languages. Kareng ancestral tribes, though arriving in China after 1 Cf. above, §§ 31, 96, and also The Cradle of the Shan Race, p. xxxviii. 2 Cf. below, § 194. On their subsequent history of., with caution, Mr. Holt S. Hallett, Historical Sketch, I.e. And on the .history and languages, cf. Major Spearman, British Burma Gazetteer, i. 162-173. 3 On the Kareng languages, cf. J. Wade, Karen Vernacular Grammar, with English interspersed, for the benefit of foreign students, in four parts, embracing terminology, etymology, syntax, and style. Maulmain, 1861 ; J. "Wade, Karen dictionary, Tavoy, 1842 (unfinished) ; F. Mason,' Synopsis of a Grammar of the Karen Language, embracing both Dialects, Sgau and Pgho, or Sho, Tavoy, 1846, 4to. ; F. Mason, Journal of the Bombay Asiatic Society, 1858, 1868 ; Brown, On the Sgau and Pwo Karens, in Journal of the American Oriental Society, vol. iv. , etc. Cf. also E. L. Brandreth, On the Nan- Aryan Languages of Jndia, in Joum. Boy. Asiat. Soe. 1878. And J. E. Lo^an, On the Ethnographic Position of the Karens, pp. 364-390 of Journal of the Indian Archipelago, Singapore, n.s. 1858, vol. ii. 86 the Chinese, have slipped through the border states on the west of their dominion, then exclusively restricted to the north, and established themselves to the south on pre-Chinese soil. They came into slight contact with Indonesian tribes, and intermingled to a large extent with Mon tribes, whence the ancient absorption into their vocabulary of many words belonging to the Indonesian and Mon formations. They received also many words from the Chinese, and also from the same northern sources as the latter, a circumstance which has contributed to multiply the Chinese-Kareng glossarial affinities. Their primitive ideology, of which the indices were most probably 13 5 8 III., was much modified by the entrance into their formation of numerous elements of uncouth Mon-Ta'i tribes, which, according to the usual case formulated into a law of comparative ideology, have imposed upon them their ideology of the verb, and their indices were finally 14 6 8 VI. The development of tones arose from a similar necessity, as in the case of the Chinese and other languages. As we have several times explained their formation, we need not again insist here on the point. 148. The linguistic affinities of the Kareng dialects with the pre-Chinese languages are various. Their glossarial similarities are numerous with the Yao, Kih-lao, Ngan-shun Miao, Blue Miao, and Miao-tze, and the relationship is also established by an identity of ideological indices 14 6. But these affinities, which display a social intercourse and inter- mingling for a certain length of time, are not however sufficiently extensive to justify any other connection than that which is described in the previous section. The only remnant of the Kareng linguistic formation, distinctly recog- nizable among all the other broken dialects, is that of the following T'u Man. Some others may exist, but they are not represented in any of the scanty documents at our disposal. 149. The language of the T'u Man ± §| is only known through a list of 102 words collected by the Chinese at Tan-kiang in Tu-yun fu, south-east Kueit- 87 chou. 1 The affinities of 1 words are chiefly Kareng, with a strong mixture of words belonging to the Tibetan, Burmese, Chyin, and Kachari-Koch groups, and a few proper to the Lolo group. The ideological indices which the aforesaid glossary discloses are 14 6 only, and these agree also with those of the Kareng group (14 6 8 VI.). XVIII. The Junos, Nagas, and Lolos. 150. The non-Chinese nature of the language spoken by the Jungs, nomads and invaders, has been mentioned in a former part of this work (III. § 28), to which we may refer. No actual specimen of their speech, ipsis verbis, has been preserved, to our knowledge, in ancient Chinese documents. But the names of several of their tribes have been kept, and some important surviving populations are their descendants, more or less mixed and unmixed. The Jungs penetrated into the Flowery Land from the north-east and east of Tibet, before and after the arrival of the Chinese civilized Bak tribes. They were therefore in- truders like the latter, but having gone over to some regions of the country west and south previous to the Chinese advance, they are entitled to a place among the pre-Chinese. 8 151. Their tribal denominations are singularly sugges- tive of the same relationship, Burmo-Naga, 3 possessed by the languages of their descendants. They agree rather interestingly with the names of the tribes forming the Western Naga division, as arranged by the late Gr. H. Damant a few years ago, in a valuable paper published after 1 Extracted from the Miao fang pei Ian, by Dr. J. Edkins, A Vocabulary of the Miau Dialects. 2 The late Dr. J. H. Pleyte, of Munich, had collected all the historical state- ments concerning the Jungs in his memoir, Die fremden barbarischen stamme in Allen China (Munchen, 1874, 460-522), pp. 477-495. And Dr. James Legge had done the same, for the Tchun tsiu period only, in his introduction to his Chinese Classics, vol. v. pp. 122-136; cf. pp. 123-126. 3 On the Burmo-Naga connection, cf. Capt. C. J. Forbes, On Ttbeto-Surman Languages,™ Journ. Roy. Asiat. Soc. 1878, vol. x. pp. 210-227; and also his posthumous Comparative Grammar of the Languages of Further India, a frag- ment, London, 1881, pp. 32-76. 88 his death. 1 For instance, these Naga names are Mao, Jemi or Yemi, Yang, Li-yang, Meng, Quoi-reng, and others, while the tribal names of the Jungs were Mao, Yam, Yun, Jung (for Rung?), Li, Lo-huei, etc. And as the complete lists would include only a few more names on each side, it cannot be denied that there are some strong similarities between them. Combined with the linguistic affinities, these show a real relationship in former times, whatever may be the divergences occurring in the course of time. 152. The Laka, or Lolo, 2 language of south Szetchuen, spoken over a large area by a race numbering some three millions of souls, is that of which we have the most precise vocabulary. It was compiled with a visible great care for accuracy in 1877, by my learned friend, E. Colborne Baber, of H.B.M.'s Consular Service in China. It numbers 200 words, besides the numerals and a few short sentences ; and it refers to the region on the left bank of the T'ung river, an affluent of the Min river in C. Szetchuen. 3 Mr. Alex. Hosie in March, 1883, at Hai-t'ang, also in C. Szetchuen, but more on the west, compiled a short vocabulary of 75 Lolo words, besides the numerals 4 of a regional variation of the same language. A list of 80 words, including numerals, was compiled by the Chinese in the district of "Wei-ning in W. Kueitchou 5 in proximity to Szetchuen, and these again prove to be a regional variation of the same language. At Yuen-kiang, in the south centre of Yunnan, a vocabulary of 140 words, including numerals, was collected by the late Doudart de Lagrde, commanding the French ex- ploration in Indo-China, 6 which is another instance of the relative unity of this important language, which extends in S.W. China between 30° and 23° of latitude. 1 'Notes on the Locality and Population of the Tribes dwelling between the Brahmaputra and Ningthi Risers, by the late G. H. Damant, Political Officer, Naga Hills, in Journ. Roy. Asiat. 8oc. n.s. vol. xii. 1880, pp. 228-268. 3 Also called Lo-huei as one of the Jung tribes. 3 Travels and Researches in Western China, pp. 73-78, in Supplementary Papers, Royal Geographical Society, vol. i. parti. 1882. 4 Report of a Journey through the Provinces of Ssu-cKuan, Yunnan, and Kuei chou, pp. 62, 73 {Parliamentary Papers, China, 1884, No. 2). 6 Hingyfutchi; transcribed in J. Edkins, Vocabulary of the Miau Dialects. 6 Voyage a" Exploration en Indo-Chine, Paris, 1873, vol. ii. pp. 509-817. 89 153. In Chinese descriptions of this people from former travellers and officials, a few words have been casually given, and as they do not appear in any of the above vocabularies, we give them here together : Nai-teh, ' chief woman ' ; toiu- ko, ' bachelor ' ; tchai-tchu, ' man ' ; sabohwa, ' chief.' Teh, Peh-ma, or Pai-ma, 'sorcerers'; also Teu-muh, Keng-tsui, Moh-kuei, Tchoh-kuei, and Heh-tcha, all titles of officials. The vocabularies, which show a large relationship with the Burmese and the Mo-so, exhibit the ideological indices 14 5 8 III. of the Tibeto-Burman group of languages. There are class-particles and' tones ; these have been noticed by Mr. E. C. Baber, who has identified them with the 1 3 4 of modern Pekinese, besides the abrupt tone. 154. The Laka-Lolos occupy an important position in the ethnology and history of S."W. China, but space forbids here more than a few words on the matter, 1 Their name, formerly Lo-kuei in Chinese, altered into Lu-luh, and now Lo-lo and Ko-lo, has become a by-name for many of the mixed tribes which in the S.W. provinces owe their origin to the intermingling with tribes of the Ta'ic and Mon and other stocks. The variants in their name have come from the influence of the Ta'ic-Shan phonology, which makes h or k equivalent of I in its adaptation of foreign words begin- ning with the latter consonant. 2 The Laka-Lolos were a south-eastern extension of the populations of north-eastern Tibet, which used to recognize the sovereignty of the woman and, accordingly, were ruled by queens. 8 The Laka-Lolos, like their brethren the Mosos, have preserved some survivals of this old custom. They were known to the Chinese S.W. of the present Shensi in the twelfth century B.C., but we have, no information on their movements towards the south. Some of their tribes were still in the N.W. of Szetchuen until the eighth century of our era. But some of them had reached the present N.E. of Yunnan and the west of 1 Their numerals are Kiienlunic. E. C. Baber (I.e., p. 71) learned that the first three numerals were formerly tit, fan, yi, but have been changed. * Cf. above, §§ 55-S6. 3 Their gyneoocratic habits gave rise to the numerous stories of Amazons in Central Asia. Cf. Terrien de Lacouperie, The Cradle of the Shan Race, p. 20. 90 Kuangsi, and in the third century they formed part of the Tsuan state (divided in two, circd 575 a.d.), which was conquered by Kolofung, King of Nantchao, in 778 a.d., who drove a great many of them back to S. Szetchuen. 1 155. They have swarmed from time to time and mixed with the neighbouring tribes, and they cover now a large area as indicated above. They have preserved the knowledge of the Tsuan writing, which, on the examination of several specimens and texts, bilingual and others, I have been able to recognize as alphabetic, and related to the oldest writing of India. 2 ] 56. The Y-Kia, a mixed race of Lolos and Chinese, on the borders of Szetchuen and Yunnan, have a language which belongs to the same group, so far as we can judge from the twelve words, including the numerals, which Fr. Gamier has collected and published from their speech at Ma-shang, 3 as follows : cato ' take some fire ' ; tcho tcho ' to eat ' ; 1, amo ; 2, mi mo ; 3, so le ; 4, Mien ; 5, ngou mo ; 6, tchou mo ; 7, seu mo; 8, ha mo; 9, kou mo; 10, tseu mo. With their postfixed class-articles, these numerals belong to the Lolo-Kato-Ho-nhi forms. In 2, 3, and 9 the class-article leu is the same as that added to the Lolo numerals, while -ma of the others is similar to the -mo of the Ho-nhi, and of the Man-tse (Lolos) in Garnier's lists. Tcho-tcho is the Lolo tzei fed or zozo le {le is final, frequently used for verbs), though ultimately of Chinese derivation. And the fact that such a word should be found there, far away from any European influence and Pigeon- English practice, shows that the speculations put forward concerning its origin, as caused by European alteration of a Chinese word, are not adequate to the fact. 157. The Liso, whose name is variously written, 4 and 1 Cf. Tang shu; Tu she, T'umg tien; Tdi-ping yii-lan, bk. 701, f. 12. Yuen kirn hi han, bk. 232, fl. 34-35. Miao Man hoh tchi, bk. ii. ff. 1-4. 2 Cf. Terrien de Lacouperie, On a Lolo MS. written on Satin, Journ. Boy. Asiat. Soc. vol. xiv. 1882 ; Beginnings of Writing, §§ 38, 226-232 ; also 156-158. 3 Voyage d' Exploration en Indo-Chine, vol. ii. pp. 509, 513, 517. * Lih-so -ft &, Li-su ;§|£ Jg or $% $l> Li-teheh $g ^ in the v ' Chinese sources ; Leesaw in Dr. John Anderson's report ; Lei-su in T. T. Cooper's Travels of a Pioneer of Commerce, p. 337. 91 who call a man Letcho, whence probably their denomination, speak a language cognate to the Burmese, and sister-dialect to the Laka-Lolo in the north-west of Yunnan. 1 They are not homogeneous as a race, but specimens of their speech taken from two tribes of various aspect are similar. A list of 107 words has been collected by Pere Desgodins, ? and another one of 166, with a few sentences, by Dr. John Anderson, 3 both lists comprising numerals. The ideological indices illustrated are 14 5 8 III., the standard ones of the Tibeto-Burman group. The proportion of similar words in Moso, Laka-Lolo, Liso, etc., and Burmese is very large. Many adjectives in Liso have -aw as a final. Class-prefixes seem to be known ; for instance, parts of the body begin with pah-, haio-, or bay. In the words for ' woman, wife, young, hand, man,' etc., la- is the common prefix. Latchoe ' man,' or better Letcho, would therefore leave -tcho as the proper name for ' man,' cognate to Lolo tou, which has the same meaning. 158. The Mo-so Jg ^Sf, who call themselves JVa-shi, and to whom the Tibetans give the name of Djia, belong to the current of migrating tribes from the north, which on the western borders of China proper have successively, since the prehistoric period, made their way towards the sunny region of the south. They are known in the Chinese records since the eighth century. 4 We have now a few data on three branches of them. 159. The northernmost and the older branch, in the N.W. of Yunnan and S.W. Szetchuen, on the borders of Tibet, has been made known by Father Desgodins. Their medicine-men have a hieroglyphic writing, of which I have 1 Miao Man hoh tchi, iii. 3. 2 Mots principaux de certaines tribus qui habitent les bords du Lan-tsang Tciartg, flu Zou-tze-kiang et Irrawaddy (Yerkalo, 26 Mai, 1872) ; Bulletin de la Soeieii de Giographie de Paris, vi. ser. t. iv. s Report on the Expedition to Western Yunnan vi& BhamS (Calcutta, 187) , 8vo.), pp. 136, 401 sq. 4 I have collected all that the Chinese sources- say ahout them, and the infor- mation of modern travellers, in Beginnings of Writing, part i. §§ 56-82, where the history, description, writing, linguistic and ethnology, of their northern division are successively studied. 92 published two MSS. 1 The above missionary has printed a vocabulary of some 200 words of their language, from notes taken by his colleagues, Fathers Gr. Biet, F. Biot, and J. Dubernard. 2 The late Francis Gamier, in the records of the French Exploration in Indo-China, was not enabled to give more than one word and one sentence from the same people. 3 160. Of the second branch, or Mu-tse, originally from Wei- Yuen in Southern Yunnan, and now settled in the Muong Lim territory in N. Indo-China, 4 we have a vocabu- lary of 151 words, collected by the late Doudart de Lagr^e, and published also by Fr. Gamier. 5 161. The Musurs, who call themselves Lahu, likewise originally from China (N.W. Yunnan), and form the third division, were not known before the last expedition to the Shan states by Mr. Holt Hallett. 6 This traveller has collected 148 of their words, and a score of their sentences, which I am preparing for publication. They are settled between Kiang-hai and Kiang-hoen, consequently eastwards of the Mu-tse. 162. The three vocabularies prove to be dialects of one language, and from the one sentence of the Na-shi and those of the Lahu, their ideological indices are 14 5 8 III., namely, Tibeto-Burman. Their glossarial affinities, which concur to the same position, show moreover that they belong to the Laka division of that family, with the Lolos, Liso, etc. 163. The Ho-ni fu fit, which the Chinese write variously Ngo-ni, Ho-ni, O-nhi, and more often Wo-ni ^ ^g, 7 is the name of a group of tribes and also of a language spoken 1 Ibid, plates i. ii. iii. 8 Mots principaux de eertaines tribus qui habitent les bords du Lan-tsang Iriang, du Zan-tze kiang et Irrawaddy, by l'Abbe Desgodins, Missionnaire au Thibet (Yerkalo, 26 Mai, 1872), in Bulletin de la Societi de Qiographie de Paris, vi. ser. t. iv. 3 Voyage d' Exploration en Indo-Chine, vol. i.'p. 520n. 4 MeLeod's and Richardson's Journeys, pp. 58, 60 {Parliamentary Papers, 420 Keturn, East India, 1869). 6 Voyage d' Exploration, vol. ii. pp. 508-516. 6 Exploration Survey for » Railway Connection between India, Siam, and China, p. 8, in Proc. Soy. Geogr. Soc. Jan. 1886. 7 Miao Man hob, tchi, iii. 2. 93 in S. Yuunan, which is a dialect of the same family as that of the Laka-Lolos, Mosso, Khos of Paleo, etc. They have come from the north, 1 at an unknown date, and now extend in the Shan states. A list of 125 words, including numerals, of their language has been compiled by the late Doudart de Lagr^e, in the Yunnan district of Yuen-kiang. 2 164. The K'ato -fr |Jj| of the Yuen-kiang department, in S. Yunnan s (and probably the same as the Kado of Burma), speak a language of the Lolo family, closely connected with that of the Ho-ni, whose name is casually given to them. A vocabulary of 139 words, including numerals, has also been collected at Yuen-kiang by the late Doudart de Lagr^e. 4 The similarity of words shows the parentage, but there are no instances of their grammar ,and ideology. 165. The Kho tribes, generally called Khas Kho, now in Indo-China, were formerly in China and claim to be a colony migrated from the Tien tsang mountain, W. of the lake of Tali fu in W. Yunnan. Their language, which belongs to the Lolo group, is only known through a small vocabulary of 138 words, numerals included, from which no indication may be easily drawn of the ideology. It is due to the devotion of the same chief of the Exploration of 1867, who collected it at Paleo, near the Mekong (lat. 21°) . 5 166. All these languages and dialects constitute a sub- group by themselves, as they resemble each other more than any of their cognate languages and groups. And they take place as a connecting link between the Burmese and Naga sub-groups. 6 167. The Ltj-tze }£ J-, on the two banks of the Lu- tze kiang, western frontier of China conterminous to Tibet, ! Some Ho-ni lean tze tribes are still met with north of Ta-tsien lu on the Tibeto-Chinese frontier. Cf . the great Chinese map 1'a tsing i tung yii t'u, nan iv. si 5. 2 Voyage d' Exploration en Indo-Chine, ii. 509 sq. 3 Miao Man hoh IcM, iii. 2. The No-pi and Seh Po, in the same department, belong to the same race and speak the same language. 4 Voyage d" Exploration en Indo-Chine, ii. 509 sq. • Voyage d' Exploration, ibid. Cf. also i. 373, 392. 6 For the scheme of classification, cf. below, §§ 229, 231. 94 call themselves Anungs, or Kanungs, 1 The Tibetan name them Quia. The name of Lu, similarly written, is known of old in Chinese history. A tribe of that name was still settled in Shan-si in 593 B.C., at Lu-ngan, when the state of Tsin destroyed them as an independent power, and com- pelled them either to be absorbed or to migrate southwards, two courses which they partially followed as usual in the case of other tribes. The Lu-she were not indigenous in China; they belonged to the Red Tek, 2 who, with the- "White Tek, formed the two branches of a people who had appeared on the N."W., near the seat of the Tchou in S."W. Shen-si, circd 1300 B.C., and spread afterwards through the Chinese- states, among which some of their tribes kept their independence until the period of the civil wars, while some others penetrated to the still pre- Chinese regions of Central and Western China. 168. The Lu-tze may or may not be the altered de- scendants of the Lu-she, 3 but the general relationship which their language discloses seems to indicate that they are. "We have a list of 111 of their words, published by Pere Desgodins, 4 which shows a strong connection with, or in- fluence of, Tibetan, as thirty-nine words are similar. The remaining words are Chinese, Lolo, Moso, Khanti, and also Kakhyen, with some Khyeng and Burmese connection. 1 They figure as Kunungs on the upper courses of the Nam Tisan and Nam Dumai or Phungmai, northern affluents of the Irawadi, and the Norkan of Nognum range, on the map of the country between the Brahmaputra and Tipper Irawadi, in the Journey of an Expedition under Colonel Woodthorpe, from Upper Assam to the Irawadi, and return over the Pakkoi Sange, by Major C. Reginald Macgregor, Proe. Soyal Geograph. Soc. January, 1887, pp. 19-41. a The Tek J^, modern Ti, Sin. -Ann. dicli. In Ku-wen, or oldest system of Chinese writing, this foreign name was transcribed ^j ifl^ , to be read from right to left. Dek THuen is De-k. This peculiar spelling from right to left with a repetition of final, shows that this peculiar word belongs to the later period of the Ku-wen style, when the use of ideographic determinatives was growing and had not yet reached its mute stage. For the Ku-wen forms of this name, and the graphical recast (which substituted tfc for ;jjt in order to obtain a contemptuous meaning ' fiery dogs,' and an allusion to this 'Bed' division), vid. Min tsi k'ih, Luh shu t'ung, bk. x. f. 1 8«. 3 Dr. J. H. Plath, Die fremden barbarischen Stamme in Alien China, in Sitzungsber. d. philos. philol. CI. der Akad. d. Wiss. 1874, pp. 457-471. Mots prineipaux des langues de certaines tribus qui habilent les bords du Lan- tsanq kiang, du Lou-tze kiang et Irrawaddy, loc. cit. 95 The numerals 3 to 9 present the peculiarities of the Kakhyens. No phrase nor text is given, and we are left with the short vocabulary for the supply of the ideological indices. The genitive follows the noun ; for instance, in the word for ' door ' (which, by the way, shows that T. T. Cooper is right in stating that they do not build houses ! ) nam Mm, the first word is 'sun' or 'light,' and hum is 'house,' lit. ' the light of the house,' which, apparently, has no windows, and is simply a hut. The adjective follows the noun : Re-me ' River ' is ' water,' as in Burmese, and me ' great ' as in Ta'ic ; grams lit. ' sword great ' ; chiam Men is ' iron hard ' for 'knife.' 169. The nearest language to that of the Lu tze is that of the Melam, who formerly belonged to their tribe, but now are a part of the Tibetan district of Tsa-rong. A small vocabulary of 58 words, and a set of nine short sentences, have been published by the aforesaid missionary, who states that the languages of the Lu-tze, of the Pa-gny or GMen, 2 of the Telus, and of the Bemepans, are nearly the same as that of the Melams, and form a linguistic family by themselves. "We are indebted also to the same zealous missionary for a few remarks on these tongues, which we here record. 3 In sentences, the subject comes first, then the direct object, the indirect object, and at last the verb, which is always at the end. Many words are borrowed from the Tibetan, but they pronounce all the letters of the written Tibetan which have disappeared from the usual vernacular of Tibet. For instance, the Tibetan slop- 'to learn,' is so pronounced by them, while it is only lob- in the common vernacular of Tibet. These languages make use of suffixed particles and various finals for the cases of nouns and modes of verbs, but the missionary has not classified them. The verbal tenses are not well marked ; however the past is 1 Travels of a Pioneer of Commerce, p. 310. They used knives for money, as did the ancient Chinese. Cf. Terrien de Lacouperie, The Old Numerals, the Counting -rods, and the Swan-pan in China, p. 14. 2 Inhabiting Pa-yul or Kiang-yul on the Tibetan, borders of Assam. 3 C. H. Desgodins, Le Tibet d'aprk la correspondance des Missionnaires (Paris, 1885), pp. 372-377. 96 characterized by tdne (Tibet, thun) or by bi ; the future by pon-ona ; the imperative by pon prefixed. When the verb indicates a movement towards an object, ngal is postposed to the imperative ; if a movement towards the subject, jd is used. The Melam language is not monosyllabic ; its pronunciation is not smooth and uniform as in Tibetan ; though not rough, it comes by jerks ; each syllable has a stress upon it, so that when they talk with animation and quickly, it seems that they are stammering. 1 170. The ideological indices are therefore complete, 2 4 5 8 III. They show an interesting displacement of the genitive from the usual Tibeto-Burman standard 14 5 8 III., which was most likely that of the language at a former period, as suggested by its glossarial affinities. The post- position of the genitive is not probably due to a Khamti influence, because of the late arrival of this branch of the Shan race in contact with them. Such an effect of inter- ference with the ideology of a language requires a protracted and powerful impression. Therefore it was probably effected within China proper when Lu-tze and Mon-Taic tribes re- mained in contact for many centuries. 171. The Lu-tze, as a part of the Teks, have hardly any claim to be classified among the pre-Chinese. Like the Jungs they were invaders, and not aborigines of the Flowery Land, whose title to that appellative consists in the fact that the date of settlement goes back to prehistoric times. Unlike the Jungs, they had not begun to enter into China previously to the Chinese civilized Bak tribes. But as some of them after their entrance into China spread over parts of the country before the Chinese themselves, they may be, for the sake of convenience, placed among the pre-Chinese. XIX. The Si-Fan and Tibetans. 172. The Si-fan and Tibetan, dialects mentioned in the following §§ 173-179, belong to the Kiang or Tibetan tribes, and to those of the Jung tribes, who, during the whole 1 La Mission du Tibet, p. 374. Cf. the remarks of Capt. W. Gill, below. 97 of the Chinese history,. were like thorns on the western side of the Chinese dominion. The history of their early and incessant raids against the Chinese is somewhat mixed with that of. the Jungs; 1 a confusion which the relationship of all these tribes easily explains. 173. The Meniak, or Menia tribes, south and west of Darchiendo, on the Tibeto- Chinese frontier, speak a language which is known to us through two vocabularies ; one of 185 words collected by Mr. Brian Hodgson 2 in 1853, and another of 232 words and a few short sentences collected by Mr. E. Colborne Baber 3 in 1878, both lists of words comprising the numerals. There are class-articles, and three tones, namely, the first and second of the modern Pekinese and the abrupt tone. 4 The ideological indices as disclosed by the instances are 14 5 8 III., in other words the standard type of the Tibeto- Burmese group, 5 in which it occupies a special position, as it has been strongly influenced by the Chinese in its vocabulary. 174. The Sung pan Si fan % $f "g #, or ' Western aliens of Sung-pan ting,' in the N.W. of Szetchuen, on the Tibetan borders, speak a language known to us only through a short list of words 6 collected on the spot by the late Capt. W. Gill, who has left me his MS. notes on the subject. man, laru, yam. 1 woman, maru. water, che. Cf. Tibet, tchu, tchab. mountain, heureux (French). Cf. Tibet, hbrog, 'mountain pasture.' cold, chdque (French que). Cf. Tibet, k'yags-. hot, drogue. Cf. Tibet, dropo. 1 Cf. Si Hang tehuen, in JECou Ran shu, bk. cxvii. 2 On the Tribes of Northern Tibet and of Sifan, in Journal of the Bengal Asiatic Society, 1853, vol xxii. p. 121. Probably tbe Mi-nok ij|| p§ of the Nan y tehi, in the Tdi ping yii Ian, bk. 789, f . 5. 3 Travels and Researches in Western China, pp . 73-78. 4 If there are others, they are not notided in the vocabularies. Mr. E. C. Baber has noticed the 1 and 2, and Mr. B. Hodgson the abrupt tone. 6 As in Tibetan, the negation is placed in the middle of compound verbs, or before simple verbs. 6 The numerals 1-12 and 20 only have been published by Col. H. Yule in his Essay Introductory to Capt. Ojll's Jowmy, o.c. ' Cf . Mongol era, ere. 7 98 eat, zdmdzo. Cf. Tibet, bza-ba. human beings, ngue. name of a Lama, nawa. yes, ddri. no, ddmdri. one, hi. two, nye. three, song (very nasal, o like o in soft). four, hgherh? five, te. six, drk. seven, tenit (French exactly). eight, gye (e very short), nine, kur. ten, chithambd. eleven, kitze. twelve, chunye. thirteen, chusong. fourteen, chuugurh. sixteen, chudru(k. seventeen, chutenit. eighteen, chukye. nineteen, chuque (French que). twenty, nyiketdmbd. thirty, songitdmbd. forty, hghtyitdmba. fifty, knachitdmbd. sixty, drukhit&mbd. hundred, chid or jiatdmbd. 175. There are no instances which permit any inference as to the ideology of the language, but the numerals and 1 " The letter r is rolled in a very pronounced manner, a striking contrast to the way in which this letter is slurred over hy the Chinese, who in many cases cannot pronounce it, as, for instance, at the beginning of a word before a or i, when the r is changed into I. Yet in other cases they are capable of producing the sound, as, for instance, in the word i-ran." The lamented traveller has remarked also about this list of words : ' ' This orthography can convey but a feeble idea of the astounding noises the people make in their throats to produce these words."— Capt. "William Gill, The River of Golden Sand (London, 1880, 2 vols.), vol. i. p. 378. — Similar remarks were made by Mr. E. 0. Baber about the Lolos (§§ 152-154).— "The speech of the independent Lolos is harsh, abounding in gutturals and strange vibrating consonants. The Welsh aspirated I frequently occurs, as in hlopo ' moon,' but it is not so easy to aspirate an m, as in hnabi ' nose.' There is a labial sound which might be written bwrbwru, pronounced as if the speaker were shivering with cold, and which is not difficult to imitate ; but when the same process of shuddering has to be applied to a lingual, as in the word ' iron,' which I have despairingly written shu-thdhru, an English tongue is dumb-foundered. Happily for strangers, these old vocables are freely modified into much simpler sounds without danger of misapprehension." Travels and Researches in. Western China, p. 72 Compare with the Lolo hlobo ' moon ' ; written Tibetan zlava, Limbu lava, Lepcha lavo, Chepang lame, Pahri nhiba, Kiranti dial, ladipa, ladiba, ladima, etc. ; and with the Lolo shu-thdhru ' iron,' Bodo eh'&rr, shirr, Dhimal chirr, Garo shurr, Kachari sorr, Kiranti syal, syel, sel, Thochu sor-mo, Mandshu tele, etc. 99 the majority of the words are Tibetan, 1 with differences. The Lama named Kawa, who gave these data, wrote the numerals and a few words on the note-book of the traveller, in the TJmin or cursive Tibetan characters. 176. The language of the Outside Man-tze, in other words the Man-tze of the outside west of Lifan fu, in W. Szetchuen, is only known through a few words still unpub- lished and the numerals, some of which have been published, 2 which have been collected by the late Capt. W. Gill when travelling there. I reproduce them from the stray leaves of his note-book, which he has left to me : yes, ngus (like the final ng in ' thing,' joined to English us). no, miak. man, latzye (the ye very short). woman, temek (k scarcely perceptible). water, teche. mountain, hang re (roll the r ; the ng scarcely perceptible). cold, ko-ad-ri. hot, ko-as-ti. eat, kdz-ye 3 (the e like e in ' yet '). one, argu* (all the r rolled). two, nergu (e=ai). three, ksirgu. four, gsairgu. five, wargu. six, shturgu. seven, shnergu. eight, kshargu. nine, rbergu. ten, khadrgu. eleven, khdtyi. twelve, khaner (without the final r). thirteen, khasi. 1 As rightly recognized by my learned friend Col. H. Tule, C.B., LL.D., the above-quoted memoir. „„,.,._ a The numerals 1-12 and 20 have been published by Col. H. Tule in his Essay, where he has pointed out their identity with those of the Thochu. 3 Probably ka-ssye. * In the MS. the final gu is gob. 100 fourteen, khasia (sia like Asia). fifteen, khonga. sixteen, khdchou. seventeen, khashner (without the final r). eighteen, khaksha. nineteen, khdrgiie. twenty, nesd or nersa (without the r). 177. A few words, like Shui tang tzai, and a full line written in the note-book of the traveller, show their writing to be the cursive Tibetan umin. The short vocabulary is not uninstructive. The numerals exhibit a superimposition of finals : -gu, which reminds us of the Chinese -ko, and is a class- particle, for which some others may be substituted according to the range of objects enumerated. They are not used — and such is the case with many other languages where similar co-efficients are employed — beyond the first ten numerals ; and their utility is to make up for the difEcufty frequent in little developed minds to find a sufficient rest in a single word which is often a monosyllable. The question is a very important one, and cannot be more than alluded to in the present pages. The first nine numerals of the language under consideration have a final r, which may be simply an adjective final or a former class-particle of a general application, now simply enclitic, should not these numerals be a foreign acquisition. Their similarity is so great with those of the Thochu on the Tibeto- Chinese frontier, 1 which have every one of them the suffix -ri or -re in full, that the probability that they are loan-words is very great. The few other words we know of the language show it to be altogether distinct from Thochu, and not without some relation with the Gyarung. It seems that two sorts of class- prefixes are used, ta- or te- for objects, ko- or ka- for adjectives and verbs. This language is mixed, and the ideology is not apparent. 1 A Thochu vocabulary has been compiled by Mr. Brian H. Hodgson, in On the Tribes of Northern Tibet and Bifan, in Journal of the Bengal Asiatic Society, 1853, vol. xsii. p. 121. , 101 178. The Li-fan Man-tze H # f§| ■?, i.e. the Man- tze of the town of Li- fan, in N.W. Szetohuen, have a dialect which is known only through the exertions of the late Capt. "W. Grill. I extract from the stray leaves of his pocket-book, which he has left with me, the following list of words still unpublished : 1 yes, 2 pai. no, nipa. man, 3 choize, me (the vowel short). woman, chime* boy, chibye. water, 4 tse. mountain, 5 pse. cold, pa (like hat without the t). hot, 6 khsi. eat, 7 gnadze. one, 8 chek (like shirt without -rt, but ch not sh ; -k scarcely perceptible), two, nye (ye not as in die, but like English ye). three, se (very short, like English sir, cut short), four, zshe (same termination), five, kna. six, true (like English true, cut very short), seven, dan, or den, or dun. eight, gyot (g and y joined together, very short, almost like the English yacht). nine, giich (a very slight guttural sound at the end). ten, 9 pchL eleven, pchechek. twelve, pchenye. 1 Excepting the numerals 1-12 and 20, which have been published by Col. Yule. 2 Cf. S'okpa bi. 3 Cf. Manyak chhoh, Tibet, mi. 1 Cf. Tibet. chM, Gyamng lichi. 6 Cf . Thochu spyah. 6 Cf . Gyarung hassi, Manyak eheche. 7 Cf. Manyak gnajen. 8 All the numerals are Tibetan, with slight variants. 9 Cf . written Tibetan Bohu. 102 thirteen, pchese. fourteen, pchizshe. fifteen, pchikna. sixteen, pchetrue. seventeen, pchedan. eighteen, pchegyot. nineteen, pcheguch. twenty, nyeshe. 179. There is no evidence concerning the ideology in the above list, with the exception perhaps of the postposition of the adjective (indice 4). The class-particles or co-efficients so conspicuous in other lists are altogether absent. A few comparisons of words which I have been able to point out in the foot-notes show a relationship with the other dialects of the region, and not exclusively with Tibetan, as the numerals would have suggested. It is a Si-fan dialect Tibetanized. Part VI. Aborigines and Intruders. XX. Pre-Chinese Proper Names. 180. Researches in Onomastics, as a help to modern eth- nology, are generally daDgerous, and, as far as names of tribes are concerned, have better to be left aside ; similarities of sur- names may be mere coincidences of a temporary character, which the respective antecedents of the assimilated names may prove to have been originally dissimilar, while names may survive and be transmitted through a superimposition or a succession of various populations. A race may have disappeared, leaving only some of its names and appellatives. Therefore such researches limited to an investigation of the former ethnology of a country cannot be entertained other- wise than if it is without reference to the modern population. Among geographical names, those of rivers are well known as resisting more than others to disappearance, and as being often interesting survivals. In the case of China they do not fail to satisfy our expectations under this respect. 181. A glance at the map with reference to the names of rivers is at once suggestive of a plurality of ethnic 103 elements in the former population of the country. In the whole basin of the Yellow River, or Huang ho, the latter term ho frf is applied to all or nearly all the rivers, as Lo-ho, Huei-ho, Shu-ho, Wen-ho, Hu-to-ho, Ma-liao-ho, etc., etc. Coming down to the basin of the Yang-tze kiang and the South, we find another term kiang j|£ spread everywhere : Tcheh kiang, Mei kiang, Heng kiang, Kia kiang, Si kiang, Peh kiang, Yu kiang, etc., etc. North-westwards we meet with a third word shui 7JC, properly ' water,' as Tchih shui, Heh shui, Sin shui, Hung shui, etc., etc., which apparently are simply the Chinese transcriptions of the Tibetan tchu. The first two, Ho and Kiang, have now the accepted meaning of ' river ' in Chinese, but they do not belong to the ground-stock of the Chinese language, which possessed but one word and symbol for ' a stream,' tchuen )\\, and none for 'river.' The same fact finds a confirmation in the formation of the characters ffl ho, and %£ kiang, which is a late one, and made of the mute ideogram determinative for ' water,' and a phonetic to suggest the sound. So is the representative of a word cognate to Mongol ghol 'river,' 1 and kiang, formerly kang, kung, kong, is a survival of the same linguistic formation to which belongs the name of the Ganges. This distribution agrees in its broad lines with the general arrangement of the populations which have preceded the Chinese in China, as illustrated all through the present work. 182. "With regard to the names of Pre-Chinese tribes, there are several difficulties to overcome, which are by no means unimportant. The excessive multiplication of small communities, and the apparent minute division based on local appellative names of extensive tribes, are only one of them. I have found about four hundred of such names. Another difficulty consists in the looseness of the names indicated in Chinese sources, or by European travellers. Such names as Lolo, Man-tze, Miao-tze, Si-fan, Tu-y, Tu-jen, Y-jen, Pin-ti-jen, etc., are vaguely used in different places, and may be looked upon with the greatest probability, as 1 To be distinguished from the Turki leul ' lake.' ; 104 loose names, indefinite Chinese expressions, mostly con- temptuous, and altogether devoid of any ethnological signi- ficance. They were not so loose at the beginning, and real distinctions of ethnic stocks were meant by them, but the subsequent interminglings of race, and the ignorance of the Chinese compilers and travellers, have caused the present confusion. 183. Many names are not genuine names of tribes, and consist only in qualificatives of a prominent characteristic in customs, costume, etc., singled out by the original author of the report, or simply of their geographical location. The same tribe described by another traveller has often received another soubriquet, and afterwards has been inscribed in the later works as a different tribe. So that tribes of one and the same stock have received different names in different places from different people, while different tribes of various stocks have received identical surnames. Such an onomastic exemplification is not calculated to facilitate the elucidation of the intricate problems of ethnology involved in the history of the population of the Middle Kingdom. 184. Those of the native names which are genuine are not, however, always recognisable. They appear in a Chinese dress, generally distorted to suit the limited capacities of Chinese orthoepy, and most often have a Sinicised significa- tion imparted contemptuously to them by the composition of the Chinese symbols employed for their transcription. 185. A few names at the beginning of history have escaped the scornful stigma. They are mostly those of the tribes whose power and proximity or interlocation of settle- ments with those of the haughty Chinese, still weak and unimportant, had compelled them to respect, such, for instance, as those of the Jung 3%, whose meanings of ' weapon,' ' war-chariot,' were soon extended to that of ' warrior.' 1 1 In the quaint Eu-wen spelling it is written ^ NGU ' a club,' placed under jrjj, Mou ' an axe ' ? (cf. Min Tsi kih, Luh shu hung, bk. i. f. 11). The same word was written later on with other symbols, somewhat similar in shape to the former ones, but not expressive of sound, and purely ideographical, i.e. EH kia ' a 105 Y |^, tlie ' Great-Bow man,' so translated from the com- position of the symbol, said to be made of ^ ta ' great,' and ^ kung ' a bow ' ; but this composition is not old, and originally the symbol was written differently. 1 These names, like those of the Lai, Lu., Lo, etc., probably one of the native terms for ' man,' written with indifferent ' characters which have no contemptuous meaning, are only a few of a longer list. 186. But still longer is the list of the written names of a contemptuous meaning, generally resulting from the ( notation of the native appellatives, written with a selected Chinese symbol having a meaning of ' animal,' a system objectionable for the Chinese themselves, as exemplified in the following : % Ma ' horse.' HI Man ' ungovernable vermin,' a general name for the southern non-Chinese tribes. HI] Min ' a sort of serpent,' in Fuhkien, etc. "jg Miao 2 * cat,' the central aborigines. shield,' under ;j^ kwo ' a spear,' as explained in the Shwohwen, which refers only to the Siao tchuen or Small Seal characters. Cf. Dr. J. Chalmers, The Structure of Chinese Characters after the Shwohwan, p. 51, and above, § 60, u. 1. 1 In Kuwen it is written _fl Tuong or Shavg 'high,' under p T'i or Shi ' a lying body,' suggesting a word like Tit, Tish, Shit, Shish. Something has remained in theSinico-AttnamjyteRojrnd^^. Another Ku-wen form of the same modern characS^BarcorrespcTdrn^prooably to its actual meaning of ' pacific,' was written p shi, placed over ^ ni, or shi-tti, which perhaps was connected as a collateral antecedent with the modern colloquial expression Sih-nu (Ji, "jfct^ 2 As in Tsing Miao, Seh Miao, etc., i.e. ' Blue Miao,' 'Black Miao,' etc., and als in ' Miao-tze ' 'jjf ^F, the famous ' sons of the soil ' of the ancient Sinologists, whose error still lingers in all the books concerning China. This wrongful interpretation is an interesting instance of graphical mythology not unfrequent in Chinese proper names and historical legends. The ancient interpreters have been misled by the analysis (always dangerous in modern style) of the modern character ■Sr, which they have supposed to have been originally composed of -h* ' grass ' and m ' field,' while they, at the same time, have taken ^p with its meaning of 'son,' 'child.' Both of these handy hypotheses are worthless. Miao "ggj", so written in the Small Seal or Siao tchuen style of writing, was an imitation of an old pictorial character figuring the head of a cat, and meaning ' cat,' now written /[flf . Tue ^f is only an enclitic of appellatives. As to the reason which caused the selection of Miao 'ffif for the rendering of the name of these tribes, we may assume that it was twofold : their strongly vocalic language, which the Chinese could not understand, was to them like the mewing of cats, and they called them- 106 d Pa ' huge serpent,' in Szetchuen. :Hi Shu(k ' silk worm,' also in Szetchuen. * M |H HEhcqE-iP oo Eh -fc ft Eh B Eh E-i W B "A \ ! »"ILb'"53 2-€Lm ■a 1 I S : ;.= = 1 E *1 a -a a Chinese Branch of the Ruenlunic stock (§ 231). 121 206. The greatness of the early Chinese rulers, so-called emperors, and the great extension of their dominion, are purely mythical, as we have had occasion to show in several instances. They were simply civilized chieftains struggling for the good of their followers. Even at the later time of the Tchou dynasty, during the short period of its splendour (1050-770 b.c.) which followed its establishment, the power of the Chinese was still very small indeed. It is true that the founders of the dynasty had acknowledged some eight hundred barons within and without their dominion, but many of these were simply native kings and local chieftains. At the latter date, the Jungs, whom we have mentioned repeatedly, were powerful enough to kill the Chinese ruler and cause the removal of the Chinese capital from Tchang- ngan (mod. Si-ngan fu in Shensi) to Loh (near Ho-nan fu, Honan). The Chinese agglomerations, which formed numerous states under the rule, at first absolute, afterwards nominal, of the kings of Tchou, were much smaller than is generally supposed. Fourteen of them, mentioned over and over again in the Tchun tsiu of Confucius, and in the invaluable chronicle of his disciple Tso Kiu-ming, were the most im- portant. And this small importance may be appreciated from the fact that one of them, the state of "Wei, which may be depended upon as representing an average of their strength, and whose territory covered about the thirty-sixth part of the whole Chinese dominion, had not a population much over 5000 souls, all told, in 660 b.c. Confucius, whose bravery was not his chief virtue, could not speak without awe in 500 b.c. on the N.W. borders of modern Shantung of the distant barbarians, who were simply the Lai tribes of the Shantung peninsula ! 1 207. The list of native and pre-Chinese states or political agglomerations would alone 2 be sufficient to show that the bu ildin g of the present Chinese greatness has been the result of Cfort^ centuries of up-hill work seldom discontinued. She Hwang-ti, of the Ts'in state, IS. W. China, the founder of the » Cf. above, § 191-195. 8 Tso tchuen, Ting kung, tenth year, 2. 122 Chinese Empire, was really the first who began the task 221 B.C., and the results of his efforts and' conquests, jeopardized through the weakness of his unworthy successor, were upheld again by the following Han dynasties (b.c. 206-220 a.d.). The splits which occurred severally in the course of history in the succession of the Chinese government, and resulted in the fragmentation of the dominion between several contem- poraneous dynasties, have greatly helped, as did the inter- necine wars and Tartar conquests, the maintenance and independence of power of the pre- Chinese tribes. For instance, such times happened in 220-280 a.d. between the E. Han and the W. Tsin dynasties, in 420-580 a.d. between the E. Tsin and the Sui dynasties, in 907-960 a.d. between the T'ang and the Sung dynasties, in 1127-1280 a.d. between the Sung and Mongol Yuen dynasties, etc. Provinces once occupied were given up, and could be recovered only a long time afterwards, and others were only conquered, either for the first time or finally, in recent times. 208. The Chinese Empire of She Hwang-ti had ill-defined limits, and did not cover more than two-thirds of the present China proper. Fuhkien, partly occupied for a few years, was given up in 105 b.c, recovered in the fifth and sixth centuries, again the seat of an independent dynasty in the ninth century, and conquered at last only in 939 a.d. During the same ninth century the south of modern Tchihli was abandoned to anarchy by successive emperors, and it was even a question whether modern Shansi was worth recovering. Euangtung^ which had been made a Chinese dependency about the end of the third century B.C., and soon after returned to freedom for some time, was sinicised much later. ' Canton,' said the late Wells "Williams, ' still in the ninth century, and even long after, was comparatively a small place, and the people of that part of the country but little removed from gross barbarism.' Returning north- wards, we see that Kiangsi was only conquered in the tenth century. Kueitchou, Hunan, W. Szetchuen, S.E. Szetchuen, Kuangsi, and Yunnan were not yet subdued in the thirteenth century. The great T'ang dynasty (618-906 a.d.) has done 123 much for the unification of the Empire, which, however, required stronger hands than theirs. It was the work of the energetic Mongol or Yuen dynasty, 1260-1367 a.d., continued by the Mings (1368-1640 a.d.), and on its way to achievement since the conquest of the present Mandshu dynasty, in power since 1644 a.d. XXIII. Other Intruders. 209. Numerous were the tribes and races who, for the same reasons as the Chinese Bak tribes, or attracted by the wealth and civilization of the latter, forced their way into China, imperilling the existence of its government, often superseding it altogether over a part or over the whole of the country, and afterwards disappearing, not however without leaving traces of their sway in the civilization, the language, and the population. The Jungs, who had partly preceded the Chinese, the Teks, the Kiangs, etc., have been already mentioned in this work as having contributed to swell the ranks of the mal- contents and banished Chinese families, 1 as well as those of the aboriginal tribes, in pre-Chinese lands. Now we must refer more particularly to those of the intruders who have exer- cised an influence of some importance either politically or in civilization. 210. The oldest intruders of this class were the Shang "^ , whose name suggests that they were traders, while their traditions indicate a western origin near the Kiien-lun range, and perhaps a parentship with the Jungs. 2 They 1 Some tribes, interspersed with the aborigines and pre-Chinese, claiming to be descendants of Chinese, migrated by their own will as malcontents, or by com- pulsion as prisoners of war or exile. Such, for instance, as the Ts'ai Ma, the hi min tze, the Peh-erh tze, the Tch'e tch'ai Miao, the Ta Lang, and the Sung kia, all generally in Kueitchou. 2 Their oldest female ancestor, Kien-tik, who begot a fatherless child Sieh, belonged to the great state of Sung Jj|j , which according to the Shan ha'i King, bk. xvi., was situated in the wide unknown west. The written character for Sung is the same as that for Jung ajj , with the addition of the determinative for ' woman.' Its Kuwen spelling, however, was different. It was written with two signs TCEung and Man, thus \£ Tcho- or Tehom, read from bottom to top. Cf. the various Ku-wen forms in Min tsi kih, Luh shu t'ung, bk. i. f. 11, where the variants bear only on the inferior character jji , which is exchanged with 1{J , pfi, 124 appear on the N.W. of the Chinese settlements since the beginning of and in the sixteenth century; they upset the Hia dynasty, took possession of the parts of Shensi, Shansi, and Honan then occupied by the Chinese, driving the , Hia towards the coast. The Tchou Jj§J, formerly Tok, who drove away the Shang- Tn dynasty, established their brilliant rule over the Middle Kingdom in 1050 b.c. ; some of them had lingered on the Chinese borders in Shensi for several centuries. They were , most probably Red-haired Kirghizes, and were not apparently without Aryan blood among them. It seems so, from the fact that they were acquainted with some notions derived from the Aryan focus of culture in Kwarism, which they introduced into China, and that several of the explanations added to the Olden texts of the Yh-King by their leader Wen-wang were certainly suggested by the homophony of Aryan words. 8 The Ts'in Jg, or better Tan, as formerly pronounced, formed an important state on the west of the Chinese agglo- meration. It grew from the tenth century to the third B.C., when, having subdued the six other principal states of the confederation, its prince founding the Chinese Empire, declared himself Emperor in 221 B.C. 3 Their nucleus was and -f', all pointing to a hushing or hissing initial. The name Kien-tik has a great similarity with that of the ancestors of the Turks Hiung-nu, variously ■written Kuen-tuh, JTuntik, etc. Sieh or Sie(t J2, the written name of the ancestor of the Shang, means 'great or important writing,' and though this character has probably replaced an older one, which represented a sort of bird, it has given rise to the unwarranted conclusion of the Eev. J. Chalmers, that the art of writing had been only introduced into China by the Shang people. The man so called was according to tradition an officer of Shun. 1 Cf. above, § 193 n. 2 Such is one, and the most important it is, of the explanations which can be put forward scientifically concerning the finding of Aryan names in Chinese. We must, however, declare that affinities of this sort, which have been eagerly pointed out by several writers in China with more zeal than discretion, are tor the greater part spurious or accidental. The narrow range of the Chinese phonesis, and the disregard of scientific method, explain the number of apparent similarities, which have been unwarrantably indicated by numerous writers. Another source of introduction of Aryan words into Chinese is that of the native dialects, which, after having received many Indian words (cf. below, §§ 212, 213), have furnished numerous terms to the Chinese vocabulary. 3 Some fugitives from Ts'in had fled to Corea in the Han country, where they were called Shin-Han. Their language, according to the Hou Han aim, bk. 115, hot e a resemblance to that of the Ts'in dynasty ; they called a kingdom, pang f$ ; a bow, hu gjj ; robbery, hou yg ; to pass wine, hang shang ft jij|£ ; calling each other, they said tu $| ; all words obsolete in the Han period. 125 not Chinese, and made of Jung tribes who absorbed gradually many Chinese families from inside, and also Turko-Tatar' tribes from its outside borders, the limits of which are not well known. This state was a channel through which passed, or a buffer preventing the passage of, any intercourse of the west with the Middle Kingdom. 211. After the foundation of the Empire, there was no longer any barrier to stop foreign intercourse, at least in the North-West, and the central government could itself open relations with the outside. The result was the Mission of Tchang Kien in Central Asia, and by B.C. 115 a regular intercourse with thirty-six states of Turkestan had been established. This signifies the entrance into China of many new items of civilization, new ideas and new words. The Jews as a colony entered into China in the second century of our era; the Nestorians, the Persians, the Ma- hommedans followed in the seventh. The Persians had a considerable intercourse with China, from 723 to 747, when ten envoys reached the Middle Kingdom. Every one of these races brought something of its own civilization, and was also the channel through which a certain amount of western culture was introduced into the Flowery Land. 212. "We know hardly anything of the immigrations which have taken place in the western and south-western regions non-Chinese, in former times. In the fourth century B.C. a native dynasty arose in the country of Shuh, i.e. Szetchuen, and the fourth ruler, who was the first to assume the title of King, is stated to have come from India. This important event was undoubtedly the outcome of the commercial relations which had existed for eight centuries or more between the traders of Shuh and those of India. 1 Many Hindu ideas have penetrated into non-Chinese China through this channel, and from thence partly into Chinese China. To the same time and means may be assigned a curious series of mythological resemblances. The many notions of fabulous ethnology and natural history, which we know, 1 On this trade, cf. Beginnings of Writing, § 1564. 126 from Ctesias, MegastMnes and others, as Indian, and the existence of similar, sometimes identical, notioils in ancient Chinese literature of the same period, especially in the Shan hat King, were due, I think, to the marvellous reports made in both countries by these travelling merchants about the intermediary, unknown, and therefore awful regions through which they had to pass during their journey to and fro. 1 213. The imperfect and embellished tradition of the arrival of a Buddhist missionary Li-fang with seventeen companions, under the reign of the First Emperor, circa 227 B.C., which subsequently served as a pattern 2 for the expeditions sent to India by Han Ming-ti (65 a.d.), and by the Tibetan Srong btsan sgam po (632 a.d.), refers most probably to an introduction of Buddhism from India in Szetchuen. Archaeological remains of great interest in the shape of statues and carved caves with Indian emblems, which are to be met with in Szetchuen, Hunan, Kiangsi and Tchehkiang, extending like a wedge turned eastwards, show another current of influence, if not of immigration, from the South-West. 3 Taoism, at least in its leading features, was introduced into China from the same quarters, but nothing remains to show if the two belong to the same current, and the matter has not as yet been investigated. Buddhism was introduced in an effectual manner through Imperial patronage in 67 a.d. Its great development and evolution as a religion in the country does not concern us here ; on the writing its influence was not unimportant, but its effect on the spoken language has been very small. 214. On the eastern side, otherwise the sea-coast, it was difficult for any immigration to be important enough to have any lasting influence under any respect. 1 Ibid. § 156c n. 3 Ibid. § 91. 3 Mr. E. Colborne Baber has carefully described several of such caves he visited in Szetchuen [Travels and Researches in Western China, pp. 129-141). All that I know of the others I have learned from the Chinese topographies. The curious horse-shoe shape of the Chinese tombs reminds us of the yoni of India, and must, most probably, be attributed to the same Brahmanic influence. 127 In the south-east of Shantung, the city of Lang-nga, founded ahout 500 B.C., which recalls so much to mind the Lanka, Lankapura of the old Ceylon, the Lang-nga of the north coast of Java, and seems to have been a settlement of colonist traders from Asianesia, in a region which was not yet Chinese, was the channel through which so many foreign notions have entered into China, that it deserves the special attention of future inquirers. The Japanese in the Middle Ages made several raids on the coast, leaving behind them nothing else than destruc- tion. The same must be said of the Bisayas of the Philippines, who made a raid on the coasts of Tsiuan-tchou in Fuhkien during the period 1174-1189 a.d. under the Sung dynasty. The Arab traders who frequented the old port of Kanf u, now embedded near Hangtchou, in the ninth century, have introduced many items of civilization. But numerous as they were, they have had no influence on the language, as in the personal statement of the relation of Wahab and Abu Zaid, no Chinaman could speak Arabic in their time. The same phenomenon, which is worthy of the attention of comparative philologists, is still experienced in the present day, as Chinese do not speak Arabic. The latest and most important influence for the future which has ever entered into China by the eastern coasts is that of the Europeans, which promises to be the greatest incitement and help to development which the Middle Kingdom has ever received. 215. The influence of the Turko-Tatar races has been considerable. Several of them, spoken of in the previous pages, belong to olden times. For several centuries after the Han period, ignorant Tatar dynasties have ruled over parts of Northern China. The Sien-pi, cognate to the Coreans, have produced the dynasties of the Former Yen, 303-352 a.d. ; the After Yen, 383-408 a.d. ; 'the Western Yen, 385-394 a.d. ; the Southern Yen, 398-410 a.d. ; the Southern Liang, 397-414 a.d.; the "Wester nTsin, 385-412 a.d. The Hiung-nu Turks have produced the dynasties of 128 Northern Liang, 397-439 a.d., of the Hia, 407-431 a.d. in W. Shensi (to he distinguished from the later Si-Hia), and afterwards the Northern Han, in 951-799 a.d. The Tchao Turks produced the dynasties of the Former Tchao, 304-329 a.d., and After Tchao, 319-352 a.d. The Si-fan have produced the dynasties of Tcheng in Szetchuen, 301-346 a.d. ; of the Former Tsin, 390-395 a.d., After Tsin, 384-417 a.d., hoth in Shensi. The Tobat Tatars, who produced the great dynasty of the Northern "Wei, 386- 532 a.d., belonged to the same group. They were apparently acquainted with the Syriac writing, at least about 476-500 a.d., and they had a court language of their own, in which their ruler Wan-ti at that time (in 486 a.d.) ordered that a translation of the Hiao king or ' Book of filial piety ' should he made. 1 Its use was not abolished before 517 a.d. 216. The rule of the Northern Wei extended over the whole of Northern China, with a few regional exceptions in the proximity of the Yang-tze Kiang. Later on, that of the Mongol dynasty of the K'itan or Liao, 907-1202 a.d., was restricted in the north-east. In the north-west, the Si-Hia or Tangut dynasty ruled from 982 to 1227, until it was swept away by the Mongols. The Meniak (§ 173) are their descendants. The Kin or Jutchih, the ancestors of the present Mandshu dynasty, ruled over a larger area than the N. Wei, from 1115 to 1234 a.d. The Mongol Yuen dynasty established by Kubila'i-Khan in 1271, and which lasted until 1367, was the first to rule over the whole of China; its great power did more for the homo- geneity of the Middle Kingdom than any previous effort. And at last, in 1644, the Mandshu Ta Tsing dynasty established its sway all over the Empire, and is still reigning brilliantly, with all prospect of not coming to an untimely end. 2 217. These various dynasties brought each of them their own language, as their names suggest, and restricted as it was in its use to the court and soldiery, its influence was in 1 Of. my Beginnings of Writing, § 164 and n. 2 All these dynasties had special writings made for them, as recorded at length in my Beginnings of Writing, §§ 101-110, 127-129. - 129 every case limited, though by no means unreal, as shown by the alteration of pronunciation and the introduction of words in the official dialect. "With regard to the present Maudshus, their presence has hurried on the phonetic decay of the Peking Mandarin dialect, now the official language, on the path of hissing and hushing the sounds, where it had entered since the days of the Yuen Mongols. Their small number, and their habit of living somewhat apart from the population, restrict the influence of the soldiery, which is felt only in the proximity of the post-towns over the empire, by the introduction of a few terms in the vernaculars. Part VII. Results and Conclusions. XXIV. General and Historical. 218. The results of our survey, however concise it is in many of its parts, are serious and complex, for the science of language and for history. The importance on Chinese soil in former and recent times of the native and intruding languages, spoken of in the previous pages, is clearly under- stood when considering how numerous and large were the various Pre-Chinese states or political agglomerations of tribes, which existed contemporaneously or superseded one another, over the whole at first, and more or less extensive parts afterwards of China proper. The slow growth of the Chinese from very small beginnings to their present stand- ing, and the corresponding gradual diminution of the non- Chinese states and territories, throw still more light on the whole affair. 219. We shall examine the various results we have come to, with reference to the science of language, under several respects, such as an increase in the number of classified languages, and an alteration of previous arrange- ments, with the formation of an altogether new linguistic group, that of the Tai'-Shan languages; and also serious warnings and teachings about the hybridology of languages, the non-mechanical character of the pronunciation and the formation of tones. But we must also indicate here some important results, 130 for the history of civilization, of our linguistic and other researches. They show that the Chinese greatness from antiquity was simply a fabulous legend, and far from being permanent, is, on the contrary, a modern fact and an im- portant contingent of the future of mankind; that there is no such a thing as a great antiquity and purity of type of the Chinese language, which, on the contrary, is a result of intermingling; that the Chinese civilization is not the result of their self-development, but an importation; and, therefore, that the theories of monosyllabic languages, primitiveness of the tonic linguistic formation, and also the theory of the self-progress of a secluded population, must be deprived of the supposed conclusive supports which have always been sought for them in China. XXV. Additions to Classified Languages. 220. As to the general classification of the languages of the Indo-Pacific and Turano-Scythio stocks, the results obtained in the preceding pages produce several new sub- divisions and groups, and the enlargement of others : the ■whole may be resumed in the following lists. We subjoin to the names the Ideological Indices available, and one or more of the italicised initials of their general characteristic, such as Unmixed, Mixed, .Hybridized, iTybrid, Developed, involved, i.e. transformed without progress, and Regressed. 221. Beginning with the Indo-Pacific stock of languages, INDO-CHINESE division or family (I), we have found a new section a) Mon-Taic including — 1) Pre-Chinese dialects (Unm. and M.) : a. Pang or Pan-hu dial. + . . . . 2 4 6 8 VI. b. Yao-jen dial. «£■ c. Pan-yao dial 2 4 6 8 VI. d. Mo-yao dial 2 4 6 8 VI. e. Ling Kia Miao dial. 2) Pre-Chinese dialects (Hd. and H.) : a. Tung jen dial 14 6 b. Ta-shui Miao-tze dial '4 60 c. Peh Miao dial 2 3 6 131 d. Hua Miao dial 2 J 60 e. Yao-pu Miao dial 2 J 6 0' /. Leng-ky Miao dial 6 g. Min Kia-tze dial. (M. Hd.) . . . 2 4 h. Liao dial. •$> ' J 60 i. Kih-lao dial 1460 j. Heh Miao dial \ \ 6 7c. Yao Min dial 14 6 222. Of the Mon-Khmek family, or section b), we have met with two of its languages, the 1) Cochin-chinese or Annamite (M.) . . 2 4 6 8 VI. 2) Palaong (M.) 2 4 6 8 VI. 223. Of the Taic-Shan family, we have been enabled to recognize several members of great importance, inasmuch as they have shown to us in the most unmistakable manner its formation, rise and growth. The first section is com- posed of the Pre- Chinese, subdivided in three subsections of dialects : A. Undeveloped: a. * Chief dial, of Ts'u. b. •£■ Ngai-Lao dial. c. 4* Nan-tchao dial. B. Unmixed and Mixed : a. Tsing Miao dial 2 4 6 b. Ngan Shun Miao dial 2 4 6 c. Tchung Kia tze or Pu-y dial. . . 2 4 6 d. Tu-jen dial 2 4 6 8 VI. e. Pai-y dial 2460 /. Pah-peh-sih-fu 2 4 6 C. Hybridized and Hybrid : a. Lien-Miao dial. b. Li of Hainan dial 14 6 c. Loi of Hainan dial. d. Hotha Shan dial 14 6 e. Khamti dial 2458 III. 224. The traces of Negritos which were disclosed by us in the course of our investigation were not sufficient to form 132 any positive idea as to their language, and we do not know if they belonged to the Himalaic Negrito- Andaman, to the Indonesian Negrito- Aetas, or to the Mon-Khmer Negrito Kamucks divisions,, though the first of these three is the less, and the third the most, probable. 225. One of the most curious results is the finding of traces on the pre-Chinese soil of an Indonesian occupation which has left in situ no living languages representative of its former standing. These, however, in several disjecta membra, now hybridized, were driven out of the Chinese soil, West, South, and East. Therefore the Interoceanic division of the Indo-Pacific stock, Indonesian Section, pre-Chinese hybrid group, includes : — a. •£ Pre-Chinese Indonesian . . . 13 6 7 IV. ? b. Gyarung or Tchentui (E. Tibet) H. 1 3 5 8 III. c. Toungthu (S. Burma) H. . . . 1 4 6 8 VI. ? d. Tayal (N. Formosa) H 1 % Q 7 V. 226. The relative position of these various additions to our present knowledge of the INDO-PACIFIC STOCK OF LANGUAGES may be seen from the following general scheme of the whole stock in its two divisions : I. INDO-CHINESE. a.) Mon-Taic. 1) Pre-Chinese dialects (Unm. and M.). 2) „ „ (Hd. andH.). b.) Mon-Khmer. 1) Cochin-Chinese or Annamite (M.). 2) Palaong (M.). 3) Talaing or Peguan. 4) Khasi (M.). 5) Khmer and its numerous group (M.). 6) Negrito Kamucks, etc. c. Taic-Shan. 1) Pre-Chinese (Und., Unm., M., Hd., H.). 2) Ahom group (M., Hd.). 3) Shan group (D.). 4) Laocian-Siamese (D.). 133 II. INTER- OCEANIC. a.) Indonesian. 1) Pre-Chinese •{«. 2) Formosan (M., Hd.). 3) Tagalo-Malayan (D., E.). 4) Negrito- Aetas (M.). b.) Micronesian (M.). • c.) Polynesian (E.). d.) Melanesian (M., H.). 227. The great Kiienlunic family of the Turano-Scythian stock of languages was represented among the populations who occupied some parts of China before the Chinese by several groups of tribes speaking languages of the Tibeto- Burmese type, and of the Kareng group. 228. The latter Kareng group is divided into a northern branch in ancient pre-Chinese country, and a southern branch including the present dialects spoken in Burma. It is the existence of the Northern and older branch which has been disclosed in the present work, as follows : — KUENLUNIC, 3) Kareng family, a Northern branch. a) Pre-Chinese Kareng * . - . (1 4 6 8 VI ?) b) T'u Man dial. M 14 6 229. The 4) Tibeto-Burmese family was, and is still, repre- sented by a large number of languages and dialects, thus, the /.) Naga-Kakhyen group includes : — b.) Western Naga group 1. Pre-Chinese Jung >%>. c.) Eastern Naga subgroup. 1. Pre-Chinese Lu-tze, Hd. . . 2 4 5 8 III. 2. Melam, Hd 2 4 5 8 III. 3. Pagny or Ghien. 4. Telu. 5. Remapan. The j) Laka-Lolo group, which is altogether newly recognized, 1 is composed as follows : — ' a. Laka-Lolo (Szetchuen- Yunnan) E. . . 14 5 8 III. 1 Cf. my Beginnings of Writing, i. § 76. 134 b. Y-kia (Yunnan) H 6 c. Liso or Leisu (N.W. Yunnan) M. . . 14 5 8 III. (Moso-Nashi (N.W. Yunnan) I. 14 5 8 III. d. Moso -| Mu-tze (Muang-lim, N. Indo-China) 14 5 8 III. I Musur-Lahu (Shan country) . 14 5 8 III. e. Kouy (Siemlap, JST. Indo-China) M. /. Ka-to, Nopi and Heh Po (S. Yunnan) M. g. Honhi (S. Yunnan) M. h. Ka-kho (Paleo, K Indo-China) M. 230. The k) Sifan group has also received several additions which we note in the following scheme with an asterisk : — 1. Pre- Chinese Kiang ►£■■* 2. Meniak 1458 III. 3. Sung-pan Sifan.* 4. Outside Mantze. * 5. Lifan Mantze.* 6. Thotchu. 7. Horpa. 8. Takpa. This arrangement is provisional, as we know very little ahout these languages, and new information is much re- quired. 231. All these additions and the relative position of the groups to which they belong, are better understood when examining the following general scheme of the TURANO- SCYTHIAN STOCK OF LANGUAGES. I. S.W. ASIATIC. Sumero- Akkadian, etc. + Hd. II. URALIC. 1. TJgro-Finnish, D. 2. Samoyed, E. 3. Yamato-Corean, E. III. ALTAIC. Turko-Tartaric, E. 135 IV. KtJENLTJNIC. 1) Yenissei Kotte, E. 2) Chinese family, H. a. Ancient ►£■. b. Sinico-Annamite dialect. c. Canton dialects. d. Fokien. „ e. Shanghai „ /. Mandarin „ 3) Karen g family, H. a. Northern or Pre-Chinese branch. b. Southern or Burma branch. 4) Tibeto-Burmese family. a. Bhot group. b. Nepal. „ c. Sikkim ,, d. Assam „ e. Kachari-Koch group /. Naga-Kakhyer i » g. Kuki 9) h. Arrakan-Chin » «. Burma M J. Laka-Lolo. ft k. Sifan it Y. HIMALAIC. 1) Dravidian, D. 2) Gangetic, M. E. 3) Kolarian, M. E. 4) Negrito-Andaman, &c, M. E, 5) Australian, R. YI. KUSH-CAUCASIC. 1) N. Caucasian, M. E. 2) Alarodian, M. E. 3) Kushite, &c, M. E. VII. EUSKARIAN, M. E. And other divisions. 136 XXVI. Otheb. results as to Ideology and Phonetics. 232. Most important results for the history of languages have come out from the contacts historical and variously intense, chiefly in Chinese regions, of languages belonging to the Turano-Scythian and to the Indo-Pacific stocks of languages. Both were opposed in ideology, as shown by their respective indices when undisturbed, viz. 13 5 8 III. for the former, and 2 4 6 7 B IV. VI. for the latter. And an alteration or divergence from these standards in a language belonging to one or the other of these two stocks always occurs when the affected language has been engaged in this remarkable linguistic struggle. "We know from history in so many cases that such was the fact, that we are authorized in other cases, concerning which historical testi- mony is lacking, to draw a similar conclusion. A strong negative evidence in favour of these views comes from the fact, most important here, that languages belonging to the two aforesaid stocks, which cannot have come into the social contact alluded to, and therefore have not been parties in the struggle, do not present the same phenomena of divergence and alteration. Their evolution has not been impressed in the same way. 233. As the variations of ideology, temporary, or perma- nent, have been indicated throughout the present memoir, among the aboriginal dialects, we need not go over the same ground again. As a complement, let us recall the altered ideologies of the Chinese 13 6 8 VI., of the Karengs 14 6 8 VI., and of the Tibeto-Burmans 14 5 8 III., instead of the original 13 5 8 III. in the Kiienlunic family. 234. We have seen, then, the undeniable existence, not only of languages mixed in their stock of words, but also of many others hybridized in their grammar, and of some new linguistic formations hybrid altogether in their vocabulary and grammar. I shall not insist here on the importance of the matter, as I have done so in another work on the com- parative ideology of languages. It will be sufficient to call attention to this important fact, which finds exemplification all the world over. 137 235. Another point which requires due consideration is that of pronunciation. The scientific achievements lately obtained in perfection of transcription by several English and German scholars go beyond human looseness. They have reached the high level of the respective idiosyncrasies of the speaker and of the transcriber, above the com- mon average of speech. The activity of man's speaking organs, and also that of his ear-sense, have nowhere the mechanical and permanent precision which their principles and those of the new school of grammarians imply. Uncul- tured populations and uneducated men are not naturally bent in the material of their speech to the yoke of steady precision which is only the result of a training in educated social surroundings through several generations. Audition and articulation of language, except in the higher races, seldom arrive together at some sort of perfection in their effective- ness. For instance, we may quote the well-known fact that the acuity of the ear among the races paying peculiar attention to the colour and pitch of the vowels exists only at the expense of precision in the articulation. 236. Tribes in a rude state of culture have a looseness and uncouthness of pronunciation and hearing, which escapes, in its group's fancies or individual distortions, from any unflinching law of regularity. The cases and causes of variance from analogy, relative easing, symbolical strength- ening or weakening, scorn anything like a formulated law. The segmentation, dispersion, and migration of tribes grown from a homogeneous linguistic stock in that state of un- culture, combined with the complication resulting from the frequent though often unknown superimposition of races and languages in a similar condition or otherwise, imply large divergences of pronunciation apparently inconsistent with their genuine derivation from common parents. And the efforts at reducing the whole of the divergences to regular and somewhat mechanical equivalence cannot lead otherwise than to numerous confusions and misapprehensions. 237. After the disturbance of ideologies, the most im- portant result for all the languages engaged in the struggle, 138 a result produced at the same time by the intermingling of blood, concerns the phonesis. We have called attention to this fact again and again. 1 The difference of phonetic peculiarities between the two great stocks was on a par with the opposition of their ideologies. The Southerners, Mons and Indonesian, were in possession of elliptic tendencies, and, above all, of a characteristic nicety of distinction in vowel sounds. The Northerners or Kiienlunic, on the other hand, had just a reverse tendency to simplify the varieties of the vowel-sounds and to unify those of a word, a process leading straight to contraction and ellipsis. The first case is illus- trated in the present day in the reports of European scholars on the extraordinary sharpness of the Khmers at catching the most delicate nuances of colour in the vocalic sounds. 2 The second is exemplified in the remarkable phenomenon of the vocalic harmonization which exists among many of the Uralo-Alta'ic languages. 3 Such were the conditions of the contest. Neither of the two parties could adopt the preferences and characteristics of the other. These were reciprocally ob- jectionable to their physiological possibilities and tendencies. 238. A compromise became forcibly the natural outlet of the contending phonologies in the languages of the inter- mingled populations. Unable to find, in a difference of colour of the vowel, the compensation required by the natural equilibrium of language for the losses in the phonetic stuff of the words by contraction, ellipsis and otherwise, they have found, as a physical necessity, this compensation in a differ- ence of pitch of the vocalic sound, which pitch is simple or compound according to the peculiar character of the loss 1 For the first time in my Early History of the Chinese Civilization (London, May, 1880), p. 19. Vid. also my Beginnings of Writing, i. §§ 62-53. 2 This is most difficult for European ears, and proves a serious obstacle to those ■who go there. Yid. G. Janneau, Manuel pratique de la langue Cambodgienne (Saigon, 1870), p. v. 3 It was disclosed for the first time by Dr. J. L. Otto Roehrig, at length, in his Researches in Philosophical and Comparative Philology, chiefly with reference to the Languages of Central Asia, in 1849, presented to the Institut de France. Cf. L. Dubeux, Compte Rendu (Paris, 1850), pp. 12-14. And previously in his Eclaircissements sur quelques particularity des langurs tartares et flnnoises (Paris, 1845), pp. 5-6. A complete exposition of the phenomenon has been given by M. Lucien Adam, De V 'harmonic des voyelles dans les langues Ouralo- Altai jites (Paris 1874), pp. 31-76. 139 sustained. This is the simple explanation, which nobody- has hitherto given, 1 of the tonic formation so remarkable in its outlines, as it has affected languages belonging to the two great linguistic stocks we have mentioned ; it does not properly belong to either of the two, and, as already said here, only the opposed languages which have come into social contact have been touched by it. Though the tones of a language are the most variable part of its phonesis, they have come to occupy an important position in the economy of the language. Their use is open to extension by analogy, want of distinction, imitation, or symbolism, and to diversifi- cation for the same reasons, besides the phonetic reaction of the vowel-sound and consonants. As a part of the material of a language they have to answer to its various requirements in the same way as the other parts. 2 And they are greatly responsible for the apparent monosyllabism of the tonic languages, which has so thoroughly deceived the philologists of former times. 3 239. It remains to be noticed that the hold of the tones on languages is in proportion to their stay within the influ- ence of the struggle we have described, and the proportion of intermingling they display in their glossary and ideology. The Chinese dialects have four tones, in some dialects ex- tended to eight by segmentation in a lower and upper class ; the Shan-Siamese have five ; the Annamites, the Karengs, and the Kakhyens six tones ; some of the Miao tribes have eight 1 It is a simple phenomenon of equilibrium, and not the survival of' an hypo- thetic primitive musical language, ' the everlasting song of the soul,' as proposed by L. de Rosny in De Vorigine du langage (Paris, 1869), pp. 36-39. Of. also D. Beaulieu, Memoire sur Vorigine de la Musique (Niort, 1859), pp. .5-8. 2 It has been remarked by Brian Hodgson that those languages which are most given to adding other syllables to the root make the least use of the tones, and vice versd where the tones most prevail, the least recourse is had to determinative syllables.' Cf. his paper On the Tribes of Northern Tibet and Si-fan, 1863. Also E. L. Brandreth, On the Non-Aryan Languages of India, 1878, Jotirn. Soy. Asiat. Soc. ■ and cf . Prof. Dr. Anton Boiler, Die priijix tnit voealischem imd gulturalem Anlaute in den einsilbigen Spraehen (Wien, 1869). 3 The languages of Tibet, Burma, Pegu, Siam, Annam, China, are generally called monosyllabic, and are still erroneously supposed by many to be living illus- trations of the imaginary primitive language of monosyllabic roots. Such mono- syllabism does not and never did exist. In reality there are three sorts of monosyllabism only — one of decay, one of writing, and one of elocution. It is to the last and first that the tongues of south-eastern Asia belong, with the complica- tion of the second in the case of modern Chinese. 140 tones ; the Lolo and the Meniak have three tones ; the Si-fan, Li-so, Mo-so, and Burmese only two ; the Nagas, the old Jungs of the Chinese, have two ; and the Tibetan has hitherto grown two tones. The gradual growth of the tones is an historical fact which we see still at work in the present time as in the last instance. It has been demonstrated to be a fact in Chinese by a native scholar, Twan-yu tsai, of the last century, whose views have proved to be substantially correct. 240. This memoir is the first (and therefore incomplete and imperfect) attempt at grasping the whole of a subject of singular importance in history, though hitherto neglected, and about which hardly anything had been done. Deprived of all the historical and ethnological data which would have made the matter less dry, and easier to comprehend by jus- tifying many an arrangement of these pages, the lin- guistic information compressed here will strike every one by its insufficiency and defective character. Materials, are wanting for the study of fifty out of the fifty-five languages and dialects mentioned therein. My last word cannot b6 less than an appeal for help, and nobody will feel more than I do myself the defects and lapsus of this work. But the importance of the results obtained must be a strong incite- ment to further efforts, and the contempt of the Chinese for the scanty remnants of the former population of their country ought not to continue to blind the Europeans, who have occasion of travelling through China, on the scientific im- portance of these ethnical and philological remains, dilapidated and hybridized as they may be, of a former state of things highly interesting for the elucidation of serious problems of anthropology, linguistics, and philosophy of history. Let us hope that this appeal will not be a vox clamantis in deserto, and that our co-workers in China will turn their attention to these living relics of the past, and gather with due care the proper materials which are required for their scientific study, before the not remote time of their complete disap- pearance under the levelling activity of progressing China. '* 141 Summary of Contents. Part I. The data and their treatment, § § 1-12. I. Data. 1. They embrace a great length of time. 2. Scantiness of information. 3. Little attention paid to the native populations. 4. Linguistic materials are meagre. 6. Dubious character of Vocabularies of Chinese origin. 6. Difficulty of their phonetic reading. 7. It detracts from their importance. II. Methods op Classification. 8. Affinities of Vocabulary and Ideology. 9. Description of Ideology. 10. Grammars mix and change. 11. Points of Ideology. 12. Symbols for their notation. Part II. Aborigines and Chinese. §§13-19. III. Arrival op the Chinese. 13. Arrival of the Chinese from the N.W. 14. Small area of their occupation. 15. Unlike the Aborigines, they were civilized. 16. Preponderance obtained by their culture. IV. Chinese and Aborigines. 17- Chinese and Aborigines. 18. Aborigines and other neve comers. 19. Gradual migration of Aborigines southwards. Part III. The Aboriginal Dialeits in the Chinese Language and Ancient Works. §§20-62. V. The Chinese Language appeoted by Aborigines, 20. Succession of races and transmission of languages. 21. Influence of the aboriginal languages on that of the Chinese. 22. Ideological Indices at present. 23. Temporary Indices 2 3 6 7 IV. 24. Definitive Indices. 25. Phoneais, Morphology, Semasiology. 26. Vocabularies full of loan-words. VI. The Aboriginal Languages in Chinese History. 27. Linguistic data in Chinese literature. 28. The Jung language in the Tso-tchuen. 29. Other languages not mentioned, though certain. 30. Languages of the outside Barbarians. — Interpreters. 31. The dialect of Tsu. 32. A mythological account on two words. 33. They are Taiic-Shan. 34. Dialects of Hu and Yueh. 35. Prefixes ngu and kon. 36. Prefixe wtt. 142 VII. Ancient Chinese Works on the Aboriginal Dialects. 37. Introduction of dialectical words in Chinese leads to a reform of the Chinese writing. 38. The reform did not answer to expectations. 39. Annual collectors of provincialisms. 40. The Erh-ya, the oldest Dictionary. 41. It contains many provincialisms. 42. The Fang-yen, comparative Dictionary of dialectical words. 43. It contains words from 44 regions. 44. 20 Chinese regions. 45. 24 Non-Chinese regions. 46. Variety of their names suggest several periods. 47. The region of Mien. 48. The Kiang-hwai region. 49. Proofs that documents therein are of various dates. 50. Difficulty for the transliteration. 51. Examples of its contents. 52. Relation to modern dialects. 53. Equivalences of sound in the Fang-yen. 54. Equivalences hetween Chinese, Sinico-Annamite and Tunkinese. 55. Equivalences between Chinese and Tai'c. 56. Equivalencies hetween Mandarin and Cantonese. 57. Equivalences between various regional sounds. 58. Dominating influence of the Chinese court dialect. 59. Chronology of equivalences. 60. Another Dictionary, the Shwoh-wen. 61. Its contents and dialectal bearing. 62. Importance of the three works. Part IV. The extinct and surviving Aboriginal Languages and Dialects, §§ 63—144. VIII. Families op Languages. 63. A complete survey of all of them is out of the question. 64. They belong to the Indo-Pacific and Kiienlunic families. 65. Mixed, Hybridized and Hybrids. IX. The Pre-Chinese Aboriginal Mon-Tai Dialects. a) Unmixed and Mixed. 66. The Pong or Pan hu. 67. Belies of their language. 68. The Yao-jen or Fan-k'oh. 69. The Pan-yao or Ting-Pan-yao dialect. 70. The Pan-y shan-tze or Mo-yao dialect. 71. The Ling-Kia Miao or Ling jen dialect. X, The Pre-Chinese Aboriginal M5n-Tai Dialects. b) Hybridized and Hybrids. 72. The T'ung-jen or Tchuang-jen. 73. Relics of their language. 74. Their Mon-Taic character. 75. The Ta-shui Miao-tze dialect. 76. The Peh Miao dialect, hybridized. 77. The Hua Miao dialect, hybridized. 78. The Vaop'u Miao dialect, hybridized. 143 79. The Leng ky Miao dialect. 80. The Min-kia Tze dialect, mixed. 81. The Liao dialect. 82. Relics of vocabulary. 83. It was hybridized. 84. The Kih Lao dialect, hybridized. 85. The Heh Miao dialect. 86. Its hybridized character. 87. The Yao min dialect. 88. Its hybrid character. XI. The Pre-Chinese Aboriginal M5n-Khmer Dialects. 89. The Annamites from Central China. 90. Their traditions. 91. Their ancient history. 92. Two languages in Annam. 93. The Annamese or Cochin-Chinese language. 94. Three writings in Annam. 95. The Palaoung dialect, mixed. XII. The Pre-Chinese Aboriginal Tai-Shan Languages. Undeveloped. 96. Linguistical influence of the State of Ts'u. 97. Its principal language was Mon-Ta'ic. 98. Has helped the Tai-Shan formation. 99. The Ngai Lao. 100. The Kiu-lung legend. 101. Their history. 102. Single relic of their language. 103. The Nan tchao language. 104. Influence of the State. XIII. The Pre-Chinese Aboriginal Tai-Shan Dialects. a) Unmixed and mixed. 105. The Tsing Miao dialect. 106. The Ngan-shun Miao dialect. 107. The Tchung-kia tze or Pu-y. 108. Data of the Tchung tze. 109. Data of the Tchung kia. 110. Data of the Tchung Miao. 111. The Tu-jen language. 112. Grammatical remarks. 113. The Pai-y and Pah-peh sih fu. 114. Recent data. 115. A vocabulary of Chinese source. 116. The Shan-Siamese. 117. Their general characteristic. XIV. The Pre-Chinese Aboriginal Tai-Shan Dialects. b) Hybridised and Eybrids. 118. The Lien Miao dialect. 119. Its hybridized character. 120. The Hotha Shan dialect, hybridized. 121. The Khamti dialect, hybridized. 122. The Li of Hainan dialect, hybridized. 123. Its relationship. 124. Their writing. 125. Another Hainan dialect. 126. Its genuineness unascertained. 144 XV. The Pre-chinese Aboriginal Negritos. 127. Historical Traces in Eastern China. 128. Linguistic characteristics of their race. XVI. The Pre-Chinese Aboriginal Indonesians. 129. Traces of Indonesians in E. Pre-China. 130. The Gyarung of N".E. Tibet. 131. Gyarung compared to Tagala of the Philippines. 132. Hybridized by Tataric influence. 133. Traces of class prefixes. 134. Affinities with the Miaos, Toungthus, and Tagala. 135. It objects to monosyllabism. 136. The Toungthus of Burma. 137. Their name and traditions. 138. Their Gyarung, and other affinities. 139. The Tayal of Formosa. 140. Defective documents. 141. They came from the Pre-Chinese mainland. 142. Connection of their language known to the Chinese. 143. Connection with the Pre-Chinese and Philippines languages. 144. Ideological Indices. Part V. The Pre- Chinese Intruders. Extinct and surviving Kuenlunic Dialects. § § 145 — 179. XVII. The Karengs or Burma and their Pre-Chinese cognate Dialects. 145. The Karengs. 146. Their linguistic evolution. 147. Their formation in Pre-China. 148. Their Pre-Chinese affinities. 149. Surviving tribe, the T'u Man. XVIII. The Jungs, Nagas, and Lolos. 150. The Jung invaders, pre- and post-Chinese. 151. Affinity of their tribal names with those of the Western Nagas. 152. The Laka-Lolo or Lokuei of S. Szetchuen. 153. Some words in Chinese records and vocabularies. 154. Originally from N.E. Tibet. 155. Their Tsuan writing. 156. The T-lda of Szetchuen- Yunnan. 157. The Li-so of N.W. Yunnan. 158. The Mo-so of N."W. Yunnan. 159. Their language and hieroglyphic writing. 160. The Mu-tse of N. Indo-China. 161. The Musurs of Indo-China. 162. Their common language and indices. 163. The Ho-nhi of S. Yunnan. 164. The Kato of S. Yunnan. 166. The Kho of Indo-China. 166. They form a special subgroup. 167. The Lu-tze of the Teks. 168. Their connection with the Kakhiengs, etc. 169. The Melam in S. E. Tibet. 170. Ideological Indices, hybridized. 171. The Lu-tze are Pre-Chinese, not aborigines. 145 XIX. The Si-fan Tibetans. 172. The Kiang, Jungs and Si-fan. 173. TheMeniak. 174. The Sung pan Si-fan. 175. Their Tibetan affinities. 176. The outside Man-tze. 177. Their language is mixed. 178. The Si-fan Man-tze. 179. It is mixed. Part VI. Aborigines and Intruders. XX. Pre-Chinese Proper Names. 180. Researches in proper names. 181. Distribution of l're-Chinese terms for 'river.' 182. Vagueness of names of aboriginal tribes. 183. Causes of their intricacy. 184. Their Chinese garb. 185. The older names are honourable. 186. The later ones are contemptuous. XXI. Gradual retreat of the Pre-Chinese. 187. Aboriginal tribes and infiltrations of the Chinese. 188. Retreat, generally southwards, of the Pre-Chinese. 189. The standing of the Pre-Chinese made not apparent. 190. Aboriginal states and political agglomerations. 191. In the East. 192. In the South East. 193. In the centre and "West. 194. In the South. 195. In the South "West. 196. Their native civilization. XXII. The Chinese Intruders. 197. The Chinese were intruders in their country. 198. Their imported civilization from S. W. Asia. 199. Remarks on the list of borrowings. 200. ' Bale ' their primitive name. 201. It was an ethnic from S. W. Asia. 202. H.i&=Kiitc/ie, another primitive name. 203. The ' Peh Kia sing' shows their absorption of native tribes. 204. Early Chinese language and modern dialects. 205. General scheme of classification. 206. The smallness of the Chinese lasted long. 207. Their divisions and civil wars. 208. Building the Chinese greatness. XXIII. Other Intruders. 209. Many tribes followed the steps of the Chinese Bak tribes. 210. The Shang, the Tchou, the Ts'in. 211. The Jews, Persians, Nestorians, Mahommedans, from the N.W. 212. In the west, in Pre-Chinese lands, from India. 213. Early entrance of Brahmanism and Buddhism. 214. In the east, early and late colonists. 215. Ancient Turko-Tatar dynasties in the north. ' 216. Later dynasties, from the K'itans to the Mandsbus. 217. Their linguistic influence. 10 146 Part VII. Results and, Conclusion. XXIV. Their General .and Historical Character. 218. They are important and complex. 219. For linguistics and history. XXV. Additions to Classified Languages. 220. For the Indo-Pacific languages. 221. Mon-Ta'i. 222. Mon-Khmer. 223. Taic-shan. 224. Negrito. 225. Indonesian. 226. General scheme of the Indo-Pacific stock. 227. For the Kiienlunic family. 228. The Kareng sub-family. 229. The Tibeto-Burmese Naga, Kakhyen, and Laka-Lolo. 230. The Tibeto-Burmese Si-fan group. 231. General scheme of the Turano-Scythian stock. XXVI. Other Results as to Ideology, and Phonetics. 232. Existence of numerous mixed and hybrid languages. 233. Produced by intermingling of conflicting ideologies and vocabularies. 234. Several important instances. 235. Imperfection of pronunciation versus perfection of transcription. 236. Permanent causes of divergences. 237. Causes of the arising of tones in languages. 238. They are a natural phenomenon of compensation. 239. Their inequal repartition proves their formation. 240. Conclusion. 147 ADDENDA ET CORRIGENDA. § 1, 1. 8, read Kueitchou instead of Kueitchon. 4, 1. 2, read 58 instead of 38. 13, 1. 7, reflrf sixteen instead of a, dozen. 17, 1. 9, read extortions instead of exertions. 19, 1. 15, read Gyarungs instead of Gyalungs. 23, n. 1,1. 6, after divination insert : They were arranged for that purpose at the beginning of the Tchou dynasty, with the largely increased addition of words of fate. ibid. 1. 12, read impossibility instead of improbability. 28, n. 2, read N. Hunan instead o/N. Honan. 31, 1. 8, read Honan instead o/Homan. 33, n. 2, after decayed forms insert Wu was a prefix. 38, 1. 7, read dynasty, consists instead of dynasty. It consists. 41, n. 1, 1. 9, after published) insert The Ta-hioh and the Tchung-yung are not included in the list of the thirteen classics because they form respectively the books xxxix. and xxviii. of the Li-Ei. 62, n. 5, add Among the Kacharis Batho is a name of the supreme deity. 93, n. 6, 1. 3, read Truong instead o/Tsuong. 105, 1. 8, read Tai Shan instead of Mon. 117, n. 1, 1. 13, read many Shan affinities instead o/many Chinese affinities. 130, 1. 1, after Gyarung insert or Tchentni. 136, n. 3, 1. 2, read of Kangoon instead of to Eangoon. 143, add the following note : I reproduce here the §§ 101, 102, and part of 103, of my Formosa Notes : — The Gyarung glossary exhibits numerous similarities with the Blue Miao and T'u Man tribes now in Kueitchou, the Toungthus of Burma (in which case they extend to 25 °/ )i with the Tayal of Formosa and with the Tagala of the Philippines. This remarkable connection would some years ago have proved unintelligible, while in the present day we may look upon it almost as not unexpected. I have carefully compared the Tayal glossary with the lists of words available from the Aborigines or Non- Chinese tribes of the Middle Kingdom, and I have found with several of them the following proportion of similarities : Tchung Miao, 33 °/ ; T'u Man, 25 % ! K'ih Lao, 25 °/ \ Loi of Hainan, 25 °/ ; Ngan Shun Miao, 20 % ; Slue Miao, 20 %; Black Miao, 15 %; White Miao, 10 %, etc. These figures, with the exception of those concerning the last two names, which belong to a inore distant group, exhibit an undeniable connection and larger affinity than with the Malayan groups (which ran from 8 to 13 °/ only). 154, n. 1, add cf. below, § 174, n. 1. 169, ii. 2, add for Ghien pronounce Dj'ien. 193, n. 1, 1. 3, read Hou Han instead of Hon Han. ibid, u. 1, 1. 8, read Shang yn instead of Shang-yu. 198, n. 1, 1. 18, read 16, 17, 34, 39, 40, 45, 49, 57, instead of 1, 15, 16, 17, 22, 23, 50, 51. 200, n. 1, 1. 2, read ~lj instead of i^. 219. 1. 18, read civilization of the Chinese instead of Chinese civilization. 230. 1. 10, add M. Hd. ibid. 1. U,addM. Hd. 148 POSTFACE. The scheme of this book was presented to the Philological Society and read as a part of the President's Address at the Annual Meeting, Friday, May 21st, 1886. As a delegate of the same Society to the Seventh International Congress of Orientalists held at Vienna last year, I read in French a resume* which was very favourably received by the fourth or Eastern section at the meeting of the 30th September, 1886. My best thanks are offered to my colleagues the Members of the Philological Society for the publication of this work, which has appeared in full in the Philological Transactions for 1885-1886. It has been made partly with notes from my MS. work China before the Chinese, from my other work on The Beginnings of Chinese Civilization, in preparation, and from my lectures on The Science of Language, chiefly with reference to South-East Asia, which will soon appear. Some of the corrections are important, for instance, § 105, 1. 8, and § 198, note 1, 1. 18. TERRIEN DE LACOUPERIE. University College, London, June, 1887. i/m: ■4K 'l&fc qM&eSvfa m s y