Trans-Himalayan Dialects.

Before we proceed to trace the southern ramifications of the Bhotîya class in India, it may be well to cast a glance on what may be called the Trans-Himalayan dialects of Tibet.

The Trans-Himalayan members of the Bhotîya class of languages do not properly lie within the limits of Indian phonology. They are mostly, however, dialects of that language which forms the type or norm of the whole Bhotîya class, the literary language of Tibet, and as such they have a certain importance for a study of the whole class. It should be borne in mind, that what we call a literary language, is, after all, only one out of many dialects, which politically may have been more successful than the rest, but which linguistically has no more right to be considered the sole representative of one body of living speech than any other of its dialect. Nay, in many cases, though literary dialects may be richer in words, they have been shown to be more reduced in grammar than their less cultivated sisters; and comparative philology has elicited more secrets from the lips of vulgar idioms than from classical writings of literary celebrities. Besides, with regard to Tibet, Mr. Hodgson tells us, that what we, after Csoma de Cörös, consider as the standard of Tibetan grammar, is positively repudiated by the people of Tibet (1853, p. 125.), so that any new collateral light on this subject will be useful and important.

Tibet, bounded in the north by the Kuenlun, in the south by the Himalaya, is divided again by a third prominent chain, which Mr. Hodgson calls the "Nyenchhen-thangla." This chain, which is partially indicated by Ritter's Nian tsin tangla, is considerably extended by Hodgson, and forms, according to him, the barrier between the north and south, or between the nomadic and civilised portions of Tibet. Between this range and the northern borders of Tibet, we find three large nomad races, the Horpa in the west, the Sokpa in the east, and the Drok pa in the central portion. The Horpa (Ritter's Khor) reach into Little Bucharia and Songaria, where they call themselves Ighurs; the Sokpa extend as far as

the Kokonur and Tangut, and their country is called Sokyeul.* Besides the Drokpa (Brogpa), remain other nomadic tribes known by the names of Kazzak and Chakpa. The general name of these nomads of Tibet is Horsok, in contradistinction to the settled inhabitants of the southern provinces, who are known by the general name of Bodpa.

Some of these nomadic tribes coming into frequent contact or collision with the south, speak the pure Tibetan; others speak dialects. Mr. Hodgson gives a specimen of the Horpa in the west, which is a Bhotîya dialect; while the Sokpa in the east speak a Mongolic idiom. The language of the central Drokpa is not yet known.

Another tract of language, first explored by Mr. Hodgson, extends from the Sokpa on the north-eastern frontier of Tibet, along the confines of Tibet and China, toward the south, as far as Yunan. We have here the Amdoans, the Thochu, Gyarung, and Manyak. The first speak simply Tibetan; the other three speak dialects first collected by Mr. Hodgson. Another language, equally a Tibetan dialect, is spoken by the Takpa (Ritter's Gakpo, Gangpo, and Dakpo), not, however, on the eastern frontier of Tibet, but west of Kwombo, in the central province of Tibet. These tribes, with the exception of the last, are known in Chinese by the general name of Sifan, or western aliens. Finally Mr. Hodgson gives us one dialect spoken in the immediate neighbourhood of the Sifan, the Gyami; and this is no longer Tibetan, but Chinese.

The information which we possess regarding these languages is as yet extremely scanty, particularly with respect to their grammar. The vocabularies published by Mr. Hodgson are here less trustworthy than in other tribes. He says so himself, particularly with regard to the Sokpa and Gyami vocables. Still the linguistic evidence, incomplete as it is, is sufficient to warrant the classification of the Sokpa with the Mongolian dialects. The identity in the

*Of the two lists of words, respectively ascribed to the Sokpa and Horpa, the Sokpa words are Mongolian, the Horpa, Bhotîya. Sok and Sok-bo is the usual Tibetan name for Mongolian tribes; those who live in Northern Tibet and Tangut, nay all Mongolians between Tibet and the towns of Little Bucharia, call themselves Sharaigol, and are sometimes called Chor by the Tibetans, Chor being given as a synonyme of Tata (i. e. Mongol) in the Chinese-Tibetan dictionary of Peking.

numerals is surprising; but in the absence of a complete set of pronouns or pronominal affixes, it is impossible to enter into details. The Horpa language is more ambiguous. By its pronouns and numerals it is Bhotîya, and I have accordingly ranged it with the Trans-Himalayan Bhotîya dialects: forming their most western branch. Mr. Hodgson, however, refers the Horpa to the Turkish family, and he derives his argument "not only from the vocables but from the complex structure of Horpa verbs." The plural termination also, which is riggi in Horpa, sounds like Tataric grammar; for instance, the Horpa gna, I, gnariggi, we; as compared with the Tataric ben, I, and bisigi, we. Nay, the Thochu also, with its plural termination lar, the Manyak with its dur, and the Takpa with its ra, have at first sight a Tataric appearance. But we must wait until Mr. Hodgson will give us all his materials, before these Trans-Himalayan dialects can be classified with anything like safety, and I therefore give my own classification only as provisional and open to correction. Of the eastern languages of Tibet, that of the Amdoans is said to be purely Tibetan. The Thochu, Gyarung, and Manyak dialects are also connected with Tibetan; but again the scantiness of linguistic evidence is such as to make further identifications extremely problematic. The Gyarung, for instance, to judge from occasional instances given by Mr. Hodgson, seems far to surpass the literary Tibetan in grammatical forms. The Gyarung clearly possesses predicative pronominal prefixes, which in the Bhotîya class are confined to some of the Nâga dialects. They are used on the principle of composition represented in my table by ẞa, and, therefore, find analogies in the Caucasus, the Dekhan, and in Sanskrit. Mr. Hodgson points out himself the striking similarity between the Circassian and the Gyarung in the use of these pronominal prefixes, and he contrasts the

Circassian sara (I), wara (thou), ui (he),

s-ab (my father), w-ab (thy father), t-ab (his father),

with

Gyarung nga (I), nanre (thou), watu (he),

nga-pe (my father), na-pe (thy father), wa-pe (his father).

Mr. Hodgson maintains that the same principle prevails in the

Hayu, Kuswar, Kiranti, and Limbu languages of the Himalaya, and in the Uraon, Ho, Sontal, and Gondi tongues of Tamulian India! Unfortunately, he has not published his grammatical outlines of these idioms, which no doubt would throw more light on the intricate problem of the exact relationship of these tongues than pages and pages of mere vocables. As far as our information of these Indian dialects goes at present, I should feel inclined to doubt any connexion between the Gyarung and such languages as the Ho. There is a grammar, by Philipps, of the Sontal language, but it could not be procured for the present Essay. With regard to the Ho language, in which, according to Mr. Hodgson, similar possessive prefixes exist, I can only say, that in Tickell's account of this dialect I looked for them in vain. Tickell gives possessive pronouns, but no possessive pronominal prefixes.

But there are other features in the Gyarung grammar, to which I remember nothing parallel in Tibetan or any other Bhotîya dialects. I give the forms, as well as their explanation, on Mr. Hodgson's authority, who occasionally quotes them in his notes. A verbal root admits of a number of prefixes without any change of meaning. Thus, to go is not only ching, but also yaching, kaching, daching, taching, and naching. These are all used in a present sense. The past is formed by putting ta between the prefix and the root. Thus we get ya-ta-ching, ka-ta-ching, da-taching, tataching, nataching, all in the sense of "I went." Causal verbs are formed by putting sa between the prefix and the root. For instance, zo, to eat; ta-sa-zo, to feed. By using ma instead of the first prefix, we get a negative verb. Thus, ma-ta-ching, I went not; ma sa zo, I did not feed. Sometimes, we are told, two or three indifferent prefixes may be used, for instance, da-na-ra-gy uk, instead of simple gyuk, to run. The causal form of this would be again. da-na-ra-sa-gyuk, to cause to run; and from this again the negative, ma-da-na-ra-sa-gyuk, not to cause to run. This a kind of gram

matical mosaic of which one should hardly have expected a Bhotîya language to be capable. But, on the other hand, it cannot be said to be Turkic; because there the verbal root always maintains its place at the beginning, and though it allows a number of suffixes, in some cases even the same as those in Gyarung, at the end of

words, on the contrary, it excludes most rigorously any prefixes. The same applies to Burmese and its cognate tongues. They are full of modifying verbal suffixes; but the only verbal form which admits of a prefix is the negative, formed by ma. Thus, in Burmese, thwa, to go, forms its causal, thwa-za, cause to go; its past, thwabhu-the, I went. Its negative, however, is, ma-thwa-bhu, he goes not.

A more complete grammatical analysis of the Sifan and Horsok tongues will be invaluable for determining the frontiers between Chinese, Mongolic, Tataric and Bhotîya dialects; and there is reason to hope that Mr. Hodgson will continue his researches in this direction. With the present evidence we must be satisfied to know that, besides the literary Tibetan, many dialects continue to be spoken, particularly in the north of Tibet, which in their vocables are related to Tibetan, and through it to the Sub-Himalayan idioms. The Sokpa dialect, however, seems to form an exception, for we can hardly be mistaken is treating it as a Mongolic dialect introduced into Tibet from Mongolia by nomadic tribes.

FIFTH SECTION.

Further Extension of the Bhotiya Class, and its Subdivision into Sub-Himalayan (Gangetic) and Lohitic Dialects.

AFTER this somewhat unsatisfactory survey of the northern members of the Bhotîya family, we return to India, to trace there the further spreading of the same speech south of the Himalaya. One imaginary barrier, which seemed to separate the languages of the second and third regions of the Sub-Himalayans, and which constituted the Kocch and Dhimal Tamulian, in contradistinction to the Tibetan immigrants, such as Limbu, Murmi, &c., has already been removed. These two groups of dialects once comprehended by one general title (Bhotîya), it will be easier to advance another step, and to include within the same class, many of the tribes of Asam

BB