barely observe, that the first king of the Hyumnu's or Huns began his reign, according to VISDELOU, about three thousand five hundred and fixty years ago, not long after the time fixed in my former discourses for the first regular establishments of the Hindus and Arabs in their several countries.

I. Our first inquiry, concerning the languages and letters of the Tartars, prefents us with a deplorable void, or with a profpect as barren and dreary as that of their deferts. The Tartars, in general, had no literature: (in this point all authorities appear to concur) the Turcs had no letters: the Huns, according to PROCOPIUS, had not even heard of them: the magnificent CHENGIZ, whofe Empire included an area of near eighty fquare degrees, could find none of his own Mongals, as the best authors inform us, able to write his difpatches; and Taimur, a savage of strong natural parts and paffionately fond of hearing hiftories read to him, could himfelf neither write nor read. It is true, that IBNU ARABSHAH mentions a set of characters called Dilberjin, which were used in Khátà: ‹ he had ' seen them, he says, and found them to confift of forty-one letters, a distinct fymbol being ap

propriated to each long and fhort vowel, and to each confonant hard or foft, or otherwise ' varied in pronunciation;' but Khátà was in fouthern Tartary on the confines of India ; and,

from his defcription of the characters there in ufe, we cannot but fufpect them to have been thofe of Tibet, which are manifeftly Indian, bearing a greater resemblance to those of Bengal than to Devanagari. The learned and eloquent Arab adds, that the Tatars of Khátà write, in the Dilberjin letters, all their tales and hiftories, their journals, poems, and mifcellanies, their diplomas, records of state and juftice, the laws of CHENGIZ, their publick regifters and their compofitions of every species:' if this be true, the people of Khátà must have been a polished and even a lettered nation; and it may be true, without affecting the general pofition, that the Tartars were illiterate; but IBNU ARABSHA'H was a profeffed rhetorician, and it is impoffible to read the original paffage, without full conviction that his object in writing it, was to display his power of words in a flowing and modulated period. He fays further, that in Jaghatáï the people of Oighùr, as he calls them, have a fyftem of fourteen letters only, denominated from themselves Oighúrì ;' and those are the characters, which the Mongals are fuppofed by most authors to have borrowed: ABU'L'GHAZI' tells us only, that CHENGIZ employed the natives of Eighùr as excellent pen→ men; but the Chinese affert, that he was forced to employ them, because he had no writers at all

among his natural-born fubjects; and we are affured by many, that KUBLAIKHA'N ordered letters to be invented for his nation by a Tibetian, whom he rewarded with the dignity of chief Lama. The small number of Eighúrì letters might induce us to believe, that they were Zend or Pahlavi, which muft have been current in that country, when it was governed by the fons of FERIDU'N; and, if the alphabet afcribed to the Eighurians by M. DES HAUTESRAYES be correct, we may fafely decide, that in many of its letters it refembles both the Zend and the Syriack, with a remarkable difference in the mode of connecting them; but, as we can fcarce hope to fee a genuine fpecimen of them, our doubt must remain in regard to their form and origin: the page, exhibited by HYDE as Khatayan writing, is evidently a fort of broken Cúfick; and the fine manuscript at Oxford, from which it was taken, is more probably a Mendean work on fome religious fubject than, as he imagined, a code of Tartarian laws. That very

learned man appears to have made a worfe miftake in giving us for Mongal characters a page of writing, which has the appearance of Japanefe, or mutilated Chinefe, letters.

If the Tartars in general, as we have every reason to believe, had no written memorials, it cannot be thought wonderful, that their lan

guages, like those of America, fhould have been in perpetual fluctuation, and that more than fifty dialects, as HYDE had been credibly informed, fhould be spoken between Mofcow and China, by the many kindred tribes or their feveral branches, which are enumerated by ABU'LGHA'ZI. What thofe dialects are, and whether they really sprang from a common ftock, we fhall probably learn from Mr. PALLAS, and other indefatigable men employed by the Ruffian court; and it is from the Ruffians, that we must expect the most accurate information concerning their Afiatick fubjects I perfuade myself, that, if their inquiries be judicioufly made and faithfully reported, the refult of them will prove, that all the languages properly Tartarian arofe from one common fource; excepting always the jargons of fuch wanderers or mountaineers, as, having long been divided from the main body of the nation, muft in a courfe of ages have framed feparate idioms for themselves. The only Tartarian language, of which I have any knowledge, is the Turkish of Conftantinople, which is however fo copious, that whoever shall know it perfectly, will easily understand, as we are affured by intelligent authors, the dialects of Tátáristàn; and we may collect from ABU'LGHA'ZI', that he would find little difficulty in the Calmac and the Mogul: I will not offend your ears by a dry ca

talogue of fimilar words in thofe different languages; but a careful investigation has convinced me, that, as the Indian and Arabian tongues are feverally defcended from a common parent, fo thofe of Tartary might be traced to one ancient ftem effentially differing from the two others. It appears, indeed, from a story told by ABU'LGHA'ZI', that the Viràts and the Mongals could not understand each other; but no more can the Danes and the English, yet their dialects beyond a doubt are branches of the fame Gothick tree. The dialect of the Moguls, in which some histories of TAIMU'R and his defcendants were originally compofed, is called in India, where a learned native fet me right when I used another - word, Turci; not that it is precifely the fame with the Turkish of the Othmánlu's, but the two idioms differ, perhaps, lefs then Swedish and German, or Spanish and Portuguese, and certainly lefs than Welch and Irish: in hope of afcertaining this point, I have long fearched in vain for the original works afcribed to TAIMUR and BA BER; but all the Moguls, with whom I have conversed in this country, resemble the crow in one of their popular fables, who, having long affected to walk like a pheafant, was unable after all to acquire the gracefulness of that elegant bird, and in the mean time unlearned his