JXa << . i< QJotncll UttlttcraitH Slibtarg Stifata. l^tw ^nrk CHARLES WILLIAM WASON COLLECTION CHINA AND THE CHINESE THE GIFT OF CHARLES WILLIAM WASON CLASS OF 1876 1918 P 211 TasT™"' ""'**"">' '-"'rary ^*^ii"nMl9iii.iSl(.ffi!!iM.J.9.,.=?.n»ra^^ and eas 3 1924 023 345 782 Cornell University Library The original of tiiis book is in tine Cornell University Library. There are no known copyright restrictions in the United States on the use of the text. http://www.archive.org/cletails/cu31924023345782 BEGINNINGS OF WRITING m CENTRAL AND EASTERN ASIA, OR NOTES ON 450 EMBRYO -WRITINGS AND SCRIPTS. TERRIEN DE LApOUPERIE, Ph.D. and lAtUD. (Lovan.J ; Laitreate of the Institut de France f Academic des InscHptions, 1889-1893^; Professor of Indo-Chinese Philology ('late of University Coll. LondJ; Member Mon. Royal Asiatic Society; Author of ^^ The Languages of China before the Chinese" (1887), " The Oldest Book of the Chinese" (1892), ^^British Museum Catalogue of the Ancient Chinese Coins" (1892), " Western Oi-igin of the Early Chinese Civilization " (1894), etc. ; Director of the Bahylonian and Oriental Record^ etc. LONDON: D. NUTT, 270, STEAND 1894. Vj s-sc^a [All rights reserved.) INTEODUCTOEY. The present work is made up of several parts, which have been printed successively since 1885. The pages 1-67 are reprints from the Journal of the Royal Asiatic Society for 1885, vol. xvii. pp. 415-482. Pages 68-144, printed in 1887, have never been published ; and pages 145-189, long left in type, were as much as possible re-cast in January of the present year, thanks to the kindness of Messrs. Stephen Austin and Sons, the long famous Oriental printers of Hertford. The result in the work is a great inequality, which a complete alphabetical index of the writings referred to is intended to obviate to a certain extent. Terrien de Lacodperie. FuLHAM, June, 1894. EEEATA. Page 14, 1. 1, read The Loi 15, 1. 12, 13, » (6, 18, and 20) 18, 1. 8, )) of whom 28, n. S, )) not ID ancient Tamil 31, 1. 13, )) Captain Calder „ 1. 21, )) vid. §266. 38, 1. 13, )) pitch, otherwise " tone," „ 1. 14, )) of the single 41, n. 1, 1. 2, )j p. 207 42, n. 1, 1. 2, jj and Batang 57, n. 3, , )) Bk. 801, i.\v., 81, 1. 26, )) to Loulan, 82, 1. 14, 7» Buddhists after 445 a.d., and one 86, 1. 7, )1 858 to 1337. 94, 1. 6, ?J A Vertical writing in Ceylon or Java 108, 1. 2, J» KangJcu or Samaroand, 114, 1. 8, )J from North-East 121, 1. 14, 3) distinct works. 123, 1. 6, 5) pilgrim had remarked J) ■'■• J) )) at Akni (Yenki= Karashar) Kutche » 1. 10, >J The Suleh and other writings. 124, 1. 29, >» Maverennahar or Transoxian 126, 1. 25, J» In South- East Asia. 127, 1. 7, 1) Emperor ShSng Tsu Jin (K'ang-hi). CONTENTS. Chapter I. Embeto Weitings. §§ 1-12. A. Objects. „ 3-4. a) Used singly ,,5. 5) Strung together ,, 6. c) Fastened on sticks or strings. . . „ 7-12. (?) fixed, carved, or drawn . . . , ,, 13-35. B. Signs. ,, 13-14. a) Used singly ,, 15. h) Beans or pebhles strung or netted ,, 16-22. e) Knotted cords ,, 23-31. d) Notched sticks • ,, 32-35. e) Marks incised or drawn . . . 4 6 6 7 9 11 11 15 21 Chapter II. "WEiiDies Impeded and Decayed. §§ 36-45. "Writing struggles for life. Various instances 25 ,, 46-55. Chinese writing 31 Chapter III §§ Mo-So HlEEOGLTPHICAL WnilING OE TlBETO- China. 56-58. 59-63. 64-73. 74-76. 77-82. 1. History . 2. Description 3. Writing . 4. Linguistic 5. Ethnology 40 42 45 51 53 Chapter IV. Alphabet in Tibet. §§83-100 56 Chapter V. Shoet-Lived Weitings in Centeal Asia. §§101-102. a) The K'itan writing 69 „ 103-107. h) The Tangut or Si-Hia writing . . 70 ,, 108-109. c) The Jutchih writing 74 ,, 110. d) TheBagspa 76 CONTKNTS. PAGE Chapter VI. Deflected Writings. §111 77 §§112-115. a) The Babylonian and its derivates . 77 ,,116-117. J) Various writings of the West . . 80 ,, 118-130. c) The Syriac writing and its Eastern derivates 81 ,,131-133. (i) The Lepcha or Rong writing . . 91 ,,134. e) The Battak of Sumatra .... 93 ,, 135-136. /) A Vertical writing in Ceylon or Java? 94 Chapter VII. Lost, Fokgotten, akd Recovered Writings. §§ 137-142. a) In S.W. Asia 95 143-146. h) Bactrian and Chinese ... .99 147-158. c) India 102 159-168. rf) In Central Asia 119 169-174. «) The Sulek and other writings . . 123 175-178. /) In S.W. Asia 126 179-185. ff) In Indonesia, the islands, etc. . . 130 Chapter VIII. Some More Facts (Supplement). §§ 186-234. Embryo writings 136 ,, 235-253. Writings of Central Asia and the Pseudo-Runic 157 „ 254-271. Nineteen writings of S.W. China . 175 ,,272-275. Easter Island inscriptions . . . 185 ,,277-280. Art of Printing 189 ,,281. Number of writings 191 Additions and Emendations 193 Alphabetical Index 203 BEGINOTNGS OF WRITma I. — Embryo Writings. 1. Man writes, as he speaks, by a special aptness of his nature. As a consequence he has used all sorts of methods and devices which are now in practice, more or less, for the transmission of thought by images, symbols, or arbitrary signs. Rude systems of writing are found everywhere in use, survival, or tradition. Many more have totally dis- appeared in course of time, superseded by some preferable system, either m.ore advanced or better fitted to the surrounding circumstances. It is not a necessity of nature that these low means of communication should always be pictorial. Conventional marks used alone or in connection with figures play quite as great a part as images among these embryonic writings. And combina- tions of material and conventional symbols are frequently met. From the embryo writings, consisting either of material or symbolical objects carved, delineated, and drawn, or of symbolical or conventional signs like marks, strokes, and lines, the genealogical development oi phonetic writing begins by the substitution of the sign of one idea for that of another whose sound is nearly or quite the same. Such were the early graphic systems of Egypt and of Babylonia, and such 4 BEGINNINGS OF WRITING AEOUND TIBET. substantially to-day is that of the Chinese. Above this stands syllabic writing, evolved from the use of the sign of one idea to represent the sound of its first syllable and from the neglect and oblivion of its idea which entails a corruption of its original graphic shape. Such were those of the Cypriotes, of the Ancient Persians (Cuneif.), of the Japanese, and the semi-syllabic signs of the old Semitic Alphabet. The perfect result of these various attempts is the invention of a true alphabet in which a definite sign corresponds to a definite elementary sound. The Corean is the only real alphabet of Eastern Asia. 2. The many devices made use of for these low means of communication or embryonic writings, may be classified into a double division : of (A) material objects or symbols, and of (B) symbolical or conventional signs. (A) The material or symbolical objects are either {a) used singly ; (6) strung together ; (c) fastened on knotted sticks or knotted cords ; {d) fixed on a board ; or {e) carved, delineated, or drawn. (B) The symbolical or conventional signs consist of I. {a) twigs, reeds, pebbles, goats' dung variously placed ; {h) netted beans (like the wampum in N. America) ; (cj knotted cords (like the Quippus of Peru) ; II. {d) notched sticks ; (e) marks on stones, like cup-marks, lines, etc. ; (/) strokes and lines of all sorts. A. Objects : (a) used singly. 3. Material objects sent singly are, of course, the most handy system for low-cultured tribes to communicate out of BEGINNINGS OF WEITING AEOUND TIBET. 5 sight and ear. Sucli, for instaace, was the system in use among the Scythians at the time of Darius's campaign against them.i Pherecydes of Heros relates that Idanthuras the Scythian King, when Darius had crossed the Ister, threatened him with war, sending him not a letter, hut a symbol, which was a mouse, a frog, a bird, an arrow, and a plough. When there was (not unnaturally) much doubt concerning the meaning of this message, Orontopagas, the Chiliarch, main- tained that it was a surrender of the empire ; for he conjectured the mouse to mean their dwellings, the frog their waters, the bird their air, the arrow their arms, and the plough their country. But Xiphodres interpreted it differently, for he explained it thus : — " Unless like birds we fly aloft, or like mice burrow under ground, or like frogs take ourselves to the water, we shall never escape their weapons ; for we are not masters of their country." Herodotus tells another version of the same story .^ 4. On the Tibeto-Chinese frontier, the Lu-tze and the Li-su have still means of communication of the same kind. The Lu-tze, being unable to read or write, have arranged with the Chinese a sort of code of signals or tokens, by which important messages are carried to and fro between them. 1 In Clem. Ales. Stromat, v. pp. 671-672 (ed. Potter, Venice, 1757), quoted in G. EawUnson, History of Herodotus, 3rd edit. vol. iii. pp. 105-106, n. " I extract the following sections of bk. iv. (transl. Eawlinson) : " The Scythians had willingly exposed some of their cattle to be seized by the Persians, in order to attack them in a trap. "131. This they did several times, until at last Darius was at his wits' end ; hereon the Scythian princes, understanding how matters stood, despatched a herald to the Persian camp with presents for the King : these were a bird, a mouse, a frog, and five arrows. The Persians asked the bearer to teU them what these gifts might mean, but he made answer that he had no orders except to deliver them, and return again with all speed. If the Persians were wise, he added, they would find out the meaning for themselves. So when they heard this, they held a council to consider the matter. " 132. Darius gave it as his opinion, that the Scyths intended a surrender of themselves and their country, both land and water, into his hands. This he con- ceived to be the meaning oi the gifts, because the mouse is an inhabitant of the earth, and eats the same food as man, while the frog passes his life in the water ; the bird bears a great resemblance to the horse, and the arrows might signify the surrender of all their power. To the explanation of Darius, Gobryas, one of the seven conspirators against the Magus, opposed another, which was as follows : — ' Unless, Persians, ye can turn into birds and fly up into the sky, or become mice and burrow under the ground, or make youi'selves frogs, and take refuge in the fens, ye will never make escape from this land, but die pierced by our arrows.' Such were the meanings which the Persians assigned to the gifts." 6 BEGINNINGS OF WEITING AROUND TIBET. For example, a piece of chicken liver, three pieces of chicken fat, and a chili, wrapped in red paper, means, " Prepare to fight at once." ^ Among the Tihetans themselves, a system of the same kind existed formerly. In the Chinese description of the seventh century, speaking of an earlier period, it is reported that : " for collecting warriors they use gold arrows. They use a gold arrow seven inches long as a sign of office. There is a post-station every hundred li. If the war be important, the courier carries also on his breast a silver hawk ; if of urgent importance, several of these hawks." ^ (6) Strung together. 6. Of material or symbolical objects strung together, I do not know any instance in Tibet or the neighbouring regions, but the practice is now current among the negroes of Gambia and Guinea, on the "Western Coast of Africa. There, a log, a stone, a feather, or other things, are strung together, and sent as messages.^ The somewhat similar, though diflferent custom of the Kakhyens, on the South-Eastern borders of Tibet, is rather of a higher standard.* They hang on strings, stretching across the pathway to their villages, small stars of split rattan and other emblems.* (c) On sticks or strings. 6. The following instance in the same region of Tibeto- China is of a mixed character and belongs to our A (c) class. The use of material objects is combined with that of ' T. T. Cooper, Travels of a Pioneer of Oommene, p. 310. 2 Bushell, The Early History of Tibet, pp. 440-1 (J.M.A.8. Vol. XII. 1880). 3 Capt. C. A. Moloney, C.M.G., of Bathurst (Gambia), has collected some very valuable information on this custom. — Vid. also an interesting paper in Journal of the Anthropological Institute. * Vid. Shway Toe, The Kachyens — See below § 72. s The custom of wearing symbolical objects in a necklace, which are seen on the figures of Assyrian kings, is perhaps a superstitious revival of this early system. Anyhow, it is interesting to see many, if not all of these signs and emblems among the zodiacal ? signs of the laud-mark stones, a dozen of which are in the British Museum. And it is very suggestive to meet them among the written-in-relief Hittite hieroglyphics. Vid. below, § 7. BEGINNINGS OF "WRITING AROUND TIBET. 7 notched sticks. When the Li-su are minded to rebel, they send to the Mo-so chief (who rules them on hehalf of the Chinese Government) what the Chinese call a Muh- k'i^ and the Tibetans a Shing-tchram,^ It is a stick with knife-cut notches. Some symbols are fastened to it, such, for instance, as a feather, calcined wood, a little fish, etc., etc. The bearer must explain the meaning of the notches and symbols. The notches may indicate the number of hundreds or thousands of soldiers who are coming; the feather shows that they arrive with the swiftness of a bird ; the burnt wood, that they will set fire to everything on their way ; the fish, that they will throw everybody into the water, etc., etc. This custom is largely used among all the savage tribes of the region. It is also the usual manner in which chiefs transmit their orders.' (d, e) Fixed, carved, or drawn. 7. The fixing of the objects, material or symbolical, in nature or in figure, on a board, is the stepping-stone to the more advanced systems. Carving hieroglyphics, i.e. the above objects in relief, or delineating them on a board, are the intermediary systems which in a tangible way lead to the drawing or painting of a hieroglyphical writing. I do not know any historical record describing the practice as I have put it; but it is almost impossible that something of the kind should not have been naturally in use as a consequence of the more simple systems. Objects fastened on notched sticks or knotted cords could not do for long records. The system of writing in relief which appears among the oldest specimens hitherto known of the hieroglyphics of Egypt seems to me a survival of this old process. And the relief system in the Hittite inscriptions, — whatever may be their early connection by derivation or imitation from the Egyptian writing, — are perhaps also survivals of a native process in > Muh-Jc'i 7f: ^, ^ Shing = piece of wood. 5 Cf. Zes aauvages Lysaous du Lou-tze Kiang, par I'Abbe Dubemard, in Bulletin de la Societe de Geographic de Pari), 1875, t. x. pp. 55-66. 8 BEGmNINGS OP -WRITING ABOUND TIBET. earlier times. So, too, the relief of the Himyaritic or Sabean inscriptions, which is not explained away by any influence of casting, may be also a survival from a writing, perhaps similar to the Hittite system, which may have preceded the adoption of the Semitic writing in which they are written. All this is speculation, but perhaps not idle, and may help to the solution of moot questions, as we know not what future discoveries may disclose, on these interesting problems of origin. 8. We have only to register here hieroglyphics incised or drawn. Some hieroglyphical graffitti have been discovered on rocks above Tomsk, on the right bank of the Tom river in Siberia.^ They are incised at a height of more than twenty feet. They are very rude, and somewhat like the famous Livre de Sauvages of merry fame in palseography. Quadrupeds, men, heads, all roughly drawn, and some indistinct lines, are all that can be seen. It looks more like the pictorial figures which can be used as a means of notation by ignorant people at any moment, than like an historical beginning of some writing. There is not the slightest appearance of any sort of regularity or conventional arrangement in them. 9. No genuinely historical hieroglyphics have hitherto been found in China. This interesting peculiarity is not, however, surprising in face of recent researches which show that the Chinese Bak tribes brought the knowledge of writing with them when they migrated into China, and that this writ- ing possessed at the time an historical antiquity of some two thousand years.^ The Chinese did not stretch eastwards to the sea-shore till four or five centuries after their entry into N.W. China. The region towards the sea remained for many centuries afterwards still sporadically occupied by the former aborigines.^ ^ J. Spassky, De Antiquis quihusdam sculpturis et inscriptionibus in Siberia repartis, Petropoli, 1822 ; L. de Eosny, Archives Paliographiques, p. 144, pi. xiii.— Vid. our remarks on other graffitti of Siberia, below § 32. 2 Cf. below ^ § 46, 47, 49. ' It is not improbable that one or another of the aboriginal group of tribes possessed a rough kind of writing, at the pictorial stage, such as is found every- where, and that something of this writing may have crept-into the more perfect BEGINNINGS OF ■WRITING AROUND TIBET. 9 10. It is apparently to the art of the aboriginal non-Chinese that the following inscription belongs, should it be proved to be primitive; and it is the only precise mention I have ever found of the kind in my researches. Outside of Li-tch'eng (in N. Shantung^), at some 500 li on the west towards the north, is a stone cliff mountain,^ on the upper part of which may be seen marks and lines representing animals and horses. They are numerous and well drawn, like a picture.' 11. Hieroglyphical inscriptions in Japan are mentioned by several authors,* but I am not aware that any copy or facsimile of them has been published anywhere. In a Japanese work, the Qiji Hen,^ some facsimiles are given of characters and inscriptions of fanciful forms, among which some might be considered as pictorial, though I dare say the whole lack sufficient indications of genuine antiquity. 12. From these vague and unsatisfactory legends, we come now to more tangible matter. It is the existence now-a- days of a hieroglyphical writing, preserved by the sorcerers of the Mo-so, a tribe of Tibeto-China. It might be a late invention. It might be an ancient one. Nobody knows. But it is not unlikely that the truth is between the two. The writing apparently contains survivals of an ancient and undeveloped system of communication by written hiero- glyphics. As the third section of this paper deals with this writing, we must leave the matter for the present. B. Signs : I. {a) — used singly. 13. In the second category of the embryonic means of system brought by tbe early Chinese rulers. We know that some of these tribes did use knotted cords and notched sticks, but we have no tidings of any other sort of writing than these besides the cup-marks on the river cliffs, which seem to have been found in China by the new comers. — Vid. below § 33. 1 Lih-Uheng, ^ ^, lat. 36° 40', long. 117° 01'. Vid. Addenda. 3 Shui-Mng 7]^ J^ comm. ; Tai Ping yii Ian, bk. 50, f. 7. * L. de Eosny {Archives paliographiques, p. 233) possesses a fac-simUe of an old inscription in hieroglyphics from Japan. — Leon Meutchnikoff, L'Mmpire Japonais, p. 200. 5 Hirata Atsutane, Giji Sen (1819, 8vo.), ff. 10 BEGINNINGS OF WRITING AROUND TIBET. communication we have put first the use of twigs, reeds, stones, heans, goats' dung, etc., variously placed. The Tang-hiang in the N.W. of Tibet had no written characters in the sixth century, and only arranged reeds and pieces of wood to remember the seasons. They boast with the Tang-tchang and Poh-Lang of being descendants of a monkey.^ One of their six most important tribes bore the famous surname of Tu-p6t, and it is from the same stock that came Tupot Fanni,^ the conqueror of Tibet, as we shall see below. ^ Down to 402 a.d. the Ju-Juan people of Tartary, not with- out relation with the Tang-hiang, used to take goats' dung, and dispose it in a certain fashion on the ground to indicate what they wanted to record.* 14. A recent account of Formosa states that the aborigines ignore writing. " They have not even any means of keeping time, and when they have made an appointment for any date, their only means of keeping a check on the days as they pass is by means of a tally of stones or grass, one stone or one knot in the grass representing a day." ^ In Timor Island, according to the Chinese records in 1618,* the people had no writing; when they wanted to record something, they did it with flat stones, and a thousand stones were represented by a string.' Before the time of their acquaintance with the Quippus, the Peruvians used in the same way pebbles or maize-beans of various colours.^ Such a practice was not unknown in Europe in the pre- historic period.^ 1 Sui Shu, or Annals of the Sui Dynasty (a.d. 581-618) ; Tai Ping yii Lan (Cyclopaedia of 983 a.d.), bk. 795, f. 3. They were the ancestors of the Tangut. Vid. also S. W. Bushell, The Early Sistory of Tibet, loc. cit. p. 528. 2 Tang Shu, ibid. * Vid. below § 86. ' De Guignes, Sistoire des Buns, vol. i. part 2, pp. 337-8. 6 A. R. Colquhoun, J. H. Stewart Lockhart, A Sketch of Formosa (1884, Hong Kong), Eicei'pta, p. 203. ' Tung si yang hao, bk. iv. ' Groenevelt, Malay Archipelago, p. 117. Vid. below § 39, n. 4. ^ H. Wnttke, Die Entstehung der Schrift. 9 Dans certains endroits on a remarque parmi les alluvions quatemaires, a cote d'armes de pierre de travail humain et de cailloux perfores pom former des grains BEGINNINGS OF WIIITING AROUND TIBET. H B. IT. (b) Beans or Pebbles strung or netted. 15. Strung or netted beans and pebbles (£. b) are not used, as far as I know, among tbe rude systems once in practice around Tibet. Unless we understand, as meaning something of the kind, the records in gold and turquoises in which the sages of Tibet are reported to have glorified their first King.i The gaionnS, garthona, or garsuenda of the Red Skin Americans belongs to that class. ^ So too the wampum belts of the Iroquois.^ B. II. (c) Knotted Cords. 16. Knotted cords were originally used in Tibet, but we have no information about their system of using them. The bare statement comes from the Chinese annals.* 17. It is commonly reported that the ancient Chinese used, de coUiers en de bracelets et servir de parures, des groupes d'autres cailloux remarquables par leur formes bizarres, leurs couleurs variees, certains hazards de mesure. Ces groupes ont ete formes intentionnellement par la main de I'liomrae, on n'en saurait douter quand on les trouTe en place, et d'un autre c6te les cailloux qui les composent n'ont ete utilises ni comme instruments ni comme parures." Vid. Fr. Lenormant, Histoin ancienne, 9th edit. vol. i. p. 401. 1 See below § 88. ' They consisted in necklaces of beans, the differences of which were suggestive of the intended meaning. ' " ' This belt preserves my words ' was a common remark of an Iroquois chief in councO. He then delivered the belt as the evidence of what he had said. Several such belts would be given in the course of a negooiation to the opposite party. In the reply of the latter, a belt would be returned for each proposition accepted. The Iroquois experienced the necessity for an exact record of some kind of pro- position involving their faith and honour in its execution, and they devised this method to place it beyond dispute." — Morgan, Ancient Society, p. 139. ' ' Among other things, the ancient wampum belts, into which the structure and principles of the confederacy ' had been talked,' to use their expression, were pro- duced and read, or interpreted for the instruction of the newly inducted sachem. A wise man, not necessarily one of the sachems, read from them the facts which they recorded. According to the Indian conception,- these belts can tell, by means of an interpreter, the exact rule, provision, or transaction talked into them at the time, and of which they were the exclusive record. A strand of wampum consisting of strings of purple and white shell beads, or a belt woven with figures formed of beads of different colours, operated on the principle of associating a particular fact with a particular string or figure ; thus giving a serial arrangement to the facts as well as fidelity to the memory. These strands and belts of wampum were the only visible records of the Iroquois ; but they required trained interpreters who could draw from their strings and figures the records locked up in their remembrance." — Ibid. p. 143. * Yid. Tang Shu in Bushell, The Early Miatory of Tibet, p. 400. And below 12 BEGINNINGS OF WEITING AEOITND TIBET. before the iuvention of writing, and previous to that of the Kwas by Fuh-hi, a system of knotted cords invented by a ruler of the mythical period named Sui-jin.^ 'Now it results from my researches that this tradition, which crept lately into Chinese compilations,^ is no historical truth, so far as Sui-jin and the ancient Chinese are concerned. By ancient Chinese I mean the Bak tribes who brought the knowledge of written characters with them into the Flowery Land, and not the aboriginal tribes non-Chinese, which had apparently the said custom of knotted cords. 18. The name of Sui-jin is not connected with the invention of knotted cords in the various traditions collected about this person and his invention of fire-drill, in the Tai ping yii Ian published in a.d. 983.' Neither was it in existence when Sze-ma Tcheng (circ. a.d. 720) wrote his introduction to the She Ki of Sze-ma Tsien.* We only find in the latter that Fuh-hi invented writing instead of the knotted cords in former use. The oldest statement about the subject is that which we find in the great appendix to the Yh-King, commonly at- tributed to Confucius, but which is certainly not the work of the Sage and has a flavour of later conjecture.^ There we read : " In the highest antiquity knotted cords were used for the administration of government. In subsequent ages the sages substituted, for these, writing by notches." ^ There is no name quoted for this change in this statement. And we ' H. Wuttke, Die Entstehung der Schrift, p. 243.— L. de Eosny, La Oivilisation Japonaise (Paris, 1883, 18mo.), pp. 130-131. 2 Tung Kien Kang muh (circ. 1180 a.d.), De MaUla, Sistoire gmeraU de. la Chine, i. i.—Kang Kim y tehi luh ty Wu-sMng Kiuen (1711 a.d). — Kana Kien tcheng she yoh (1737 a.d.), i. f. 3. ' " 3 Cf. bk. 78, fi. 2-3. * San Hwang pun hi, i. 1, where the substitution of shu-k'i to knotted cords by Fuh-hi is mentioned without reference to Sui-jin. » Yh-King ; hi tze, ii. 23. It is also found in the Tao teh King. « i.e. ShuK'i ^ ^ said to mean: "written contracts." On the inter- pretation of this expression cf. Tai-ping yii-lan, bk. 747, fi. 1, .5 ; and Yuen Kien luy han, bk. 325, f. 16 ; where an explanation by Shin tze (400 B.C.) is quoted. This is a forced interpretation, as K'i is nothing else than "notches " Cf. The Six Scripls, a translation by L. C. Hopkins (Araoy, 1881), p. 6. Mr T. Watters, in the second of his valuable Essays on the Chinese Language, trs^as- lates it by " mdentures," which is half-way between the original meaning and the moral sense afterwards imputed to the K'i. BEGINNINGS OF "WRITING AROUND TIBET. 13 find it in a section of tlie above appendix, where the writer has fancied inventions and progresses of all sorts as being suggested by an examination of the Kwas of the Yh-King. Now the Kwas are these very writing-notches which were substituted for the knotted cords of former times ; so that they could not suggest their own invention. 19. This shows the childishness of the speculations attempted by the Chinese author ; it recalls to mind a similar attempt in one of the latest additional sections of the Shan Sdi King, the sacred book of the mountains and seas, where the statements do not agree with those of the above appendix. It is, however, highly interesting for history to find such allusions at so early a date, as the author could not have spoken of realities as knotted cords and notched sticks should he not have heard of them being in use at some time and somewhere. But is the tradition referring to the Pre- Chinese Bak tribes previous to their migration eastwards and before they learned writing ? or is it a combination made by the writer, based upon his knowledge or hear-say of such customs among the aboriginal tribes of China ? The answer to the question would entail an inquiry of such length that we had better leave it as it stands until a further opportunity. 20. The Yang tung, south of Khoten and consequently north of Tibet, who first communicated with China in a.d. 641, had no written characters ; they only cut notches in sticks and tied knots in strings for records.^ The Bratyki and Buriats of Siberia are credited with the use of knotted cords.^ The Japanese are also reputed to have employed knots on strings or bind-weeds for records.' 1 Tung Hen by Tu yu (850 a.d.) in Tai ping yii Ian, bk. 798, f. 7 v. Also BusheU, The Early Sistory of Tibet, loc. cit. p. 527. ^ A. Maury, Origine de VEeriture, in Journal des SavamU, Avril, 1875, p. 217. Vuttke, Die Entstehmg der Schrift, p. 143, where other examples are quoted. And also Eemusat, Eeeherches sur les languea Tartares, i. 65-6. 3 L^on Mentohnikofi, L' Empire Japonais, p. 200. — M. Leon de Eosny, Etudes Asiatigues de Geographie et d'Eiatoire, p. 4, mentions the knotted cords and notched sticks in Japan. I do not find any reference to this custom in the exhaustive and scholarly introduction of Mr. Basil Hall Chamberlain to his careful translation of the Eo-ji-Ki (Yokohama, 1883, 8vo.). Mr. C. Satow, in his 14 BEGINNINGS OF 'WRITING AROUND TIBET. The Li of Hainan, of whom we shall have to speak further on, being unacquainted with writing, according to Chinese sources, use knotted cords or notched sticks in place of bonds or agreements.^ 21. Such is also the case with the Sonthals of Bengal. "Their accounts are either notches on a stick, like those formerly used by the rustics for keeping scores at cricket matches in country villages in England, or knots on a piece of grass string, or a number of bits of straw tied together." ^ " I weU remember my astonishment ^ while trying my first case between a grasping Mahajun and a Sonthal, when I ordered them to produce their accounts. . . . The Sonthal produced from his back hair — where it had been kept, I suppose, for ornament — a dirty bit of knotted grass string, and threw it on the table, requesting the court to count that, as it had got too long for him. Each knot represented a rupee, a longer space between two knots represented the lapse of a year."* 22. In the first half of the present century, cord-records were still generally used in the Indian Archipelago and Polynesia proper. The tax-gatherers in the Island of Hawaii by this means kept accounts of all the articles collected by them from the inhabitants. A rope 400 fathoms long was used as a revenue book. It was divided into numerous portions corresponding to the various districts of the island; the portions were under the care of the tax-gatherers, who with the aid of loops, knots and tufts of diflferent shapes, colours and sizes, were enabled to keep an accurate account of the essay on the Transliteration of the Japanese Syllabary, in Transactions of the Asiatic Society of Japan (vol. vii. pp. 226-60), has collected and discussed all the authentic information concerning early Japanese writing. ^ Kiung shan Men tehi ; Notes and Queries on China and Japan, vol. i. p. 83. J. Moura, Ze Eoyaume du Gamhodge (Paris, 1882, vol. ii. Svo.), i. p. 512. ^ Sonthalia and the Sonthals, hy E. J. Man, late Assistant Commissioner, Sonthal Pergunnahs (Calcutta, 1867, 8vo.), p. 42. 3 Ibid, p. 75. — " Some method of calculation by means of knotted cords exists among the Sonthals of Bengal, and is mentioned in the Report on the Census for 1872." — Herbert R. Giles, Historic China and other Sketches (London, 1882, 8vo.), p. 3. * Darius (Herodot. iv. 98) made something of the kind, when he took a thong, and tying sixty knots in it, gave it to the Ionian chiefs, that they might untie a knot every day, and go back to their own land if he had not returned when all the knots were undone. BEGINNINGS OF "WRITING AEOUND TIBET. 15 togs, pigs, and pieces of sandal' wood, etc., at which each person was taxed.' Polynesia was the way through which apparently the custom of knotted-cord records reached the new world. The remarkable instance of dissemination we have to quote further on about the Easter Island inscriptions is highly sug- gestive of such a fact. It is by the Peruvians that the cord system of mnemonics was carried to the greatest perfection,^ and the name of quippus they gave to them might be taken as a generic appellative for the system. B. II. {d) Notched sticks. 23. We have had occasion in former paragraphs (11, 17, and 20) to mention the use of notched sticks. It is useless to repeat here the statements quoted above on this 1 Cf. Wuttke, op. cit. p. 143.— C. F. Keary, The Dawn of History, p. 181. ^ ' ' The quippu was a cord aljout two feet long, composed of different colonred ttreads tightly twisted together, from which a quantity of smaller threads were suspended in the manner oJt fringe. The threads were of different colours and were tied into knots. The word quipu, indeed, signifies a knot. The colours denoted sensible objects ; as, for instance, white represented silver, and yellow gold. They sometimes also stood for abstract ideas. Thus, white signified peace, and red war. But the quipus were chiefly used for arithmetical purposes. The knots served instead of ciphers, and could be combined in such a manner as to represent numbers to any amount required. By means of these they went through their calcu- lations with great rapidity, and the Spaniards who first Tisited the country bear testimony to their accuracy."- — Vid. Prescott, History of the Conquest of Peru, vol. i. chap. iv. " On the quipus devoted to population, the coloured strings on which the number of men in each town and viJlage was recorded had depending from them little strings for the widowers, and no doubt the widows and the old maids had their little strings from the coloured cord that denoted women. One knot meant ten ; a double knot one hundred ; two singles side by side twenty ; two doubles two hundred ; and the position of the knots on their string and their form were also of immense importance, each subject having its proper place on the quipus and its proper form of knot. The art of learning to read quipus must have been difficult to acquire ; it was practised by special functionaries, called quipucamayocuna, or knot officers, who, however, seemed only able to expound meir own records ; for when a quipu was sent from a distant province to the capital, its own guardian had to travel with it to explain it." — Cf. C. F. Keary, The Dawn of History, pp. 181, 182. Sometimes instead of knots the little strings of various colours were of the most elaborate character ; they represented all sorts of objects — suns, stars, waxing and waning moons, rainbows, birds, animals, lizards, fruits, and even pandean pipes. — Vide illustrations, p. 20, in L. de Eosny, Les Ecritures Figu/ratives et Hieroglyphiqaes. — On the quippus, vide H. "Wuttke, Die Mtstehung der Sehrift, pp. 179-190. " The messages from the Inca were indicated by a characteristic red string. At the end of the last century the Spaniards, advised in proper time, prevented a general insurrection of their Peruvian subjects ; the intended rebels had commiinicated between themselves by quippus, the date, orders, and details of the rising." — Julien Vinson, in Dic- tionnaire d' Anthropologie (Paris, 1885, 8vo.), p. 407. 16 BEGINNINGS OF WRITING AROUND TIBET. custom among the ancient Chinese, the Li of Hai-nan, the Yang-tung (N. of Tibet), and the Li-so of Tibeto-China. The current expression which connects the notched sticks and knotted cords in the Chinese descriptions of uncivilized tribes, looks like one of those stereotyped compound idioms so numerous in their language. The hearsay of one of the two processes may have been sometimes a sufficient reason to employ the well-known expression. Perhaps we must not accept, without some reserve for the said reason, the Chinese statement of the use in ancient Tibet of knotted cords and notched sticks,^ which is given in the very terms we put in suspicion. One of the two statements is certainly exact, but we dare not be so confident about the double assertion. 24. Marco Polo, in his account of the province of Zar- dandan (Western Yunnan), relates that, "When these people have any business transactions with one another, they take a piece of stick, round or square, and split it, each taking half. And on either half they cut two or three notches. And when the account is settled, the debtor receives back the other half of the stick from the creditor." ^ Dr. John Anderson, in his Report on the Expedition to Western Yunan vid Bhamo, gives some interesting in- formation, taken in situ, on the same subject : " The use of tallies to which the great traveller ^ refers is stiU prevalent among the Kakhyens and Shans. A slip of bamboo, about eight inches long, is fractured at intervals. The fractures are simple, and do not separate from each other." And further on he speaks of a Momein messenger who was anxious to get away :".... he was continually referring to the small bamboo tally on which he had marked off the days as they had passed. It was the same as in use among the Kakhyens, a thin strip of bamboo broken across at intervals." * ' Bushell, The Early History of Tibet, I.e. p. 440. 2 The Boole of Ser Marco Polo, ed. Yule, 1874, vol. ii. p. 74. ■' Marco Polo. < Calcutta, 1874, 8to. pp. 36, 270. BEGINNINGS OF WRITING AROUND TIBET. 17 25. Sir Arthur Phayre ias related a curious instance of the same custom : ^ "In the year 1863 the Tsaubwa (or Prince) of a Shan province adjoining Yunnan was in rebellion against the Burmese Grovernment. He wished to enter into com- munication with the British Government. He sent a messenger to a British officer with a letter tendering his allegiance, and, accompanying his letter, was a piece of bamboo, about five inches long. This had been split down the middle, so that the two pieces fitted closely together, forming a tube in the original shape of the bamboo. A notch at one end included the edges of both pieces, showing that they were a pair. The messenger said that if the reply was favourable, one of the pieces was to be returned and the other kept. I need hardly say the messenger received no written reply, and both pieces of bamboo were retained." 26. The custom of notching sticks was prevalent among the aboriginal tribes of China. A Chinese writer of the middle of the seventeenth century describes it as follows : — "The notches of the Muh-k'i, like the checks (of the Chinese), are used for fixing covenants. The Miao-jen, though having written characters, are not all competent to write ; so that when they make a business contract, they notch a stick as a proof of their respective good faith." ^ The Chinese annals of the Tang Dynasty,' describing in the seventh century the Sie and other aboriginal tribes of a large region (of which the west of modern Kueitchou province may be considered as the centre), say of the Sie, that they had no written characters, and that they use to make notches on wood for their contracts.* About the other tribes the state- ment that they had no written characters does not appear,* but that of the notched sticks stands. ' Quoted from a MS. note by Col. Tule in his noble edition of the Venetian traveller, vol. ii. pp. 78-9. ^ Tung KH sien tchi, by Luh Tze-yun, f. 13 v. [Shwoh ling ooUeotion, bk. 29). 3 Tai ping yii Ian, bk. 788, f. 3v. ' Ibid. Bk. 791, f. 9, v.— Of. the rather loose statements of Ma Tuanlin in the translation by Marq. d'Hervey de St.-Denys, Sthnographie dea peuplea elrangers, yol. ii. pp. 81, 91, 139. 2 18 BEGINNINGS OF WRITING AROUND TIBET. The Black Lolos of Yunnan, the Tsing tchung Kia and the Tung Kia Miao tribes of Kueitchou province, are said, in the absence of any written language, or of ia regular calendar, to still use pieces of carved or notched wood for records of events.^ Now it may be interesting to say here that the two latter tribes belong to the same stock as the Tai-Shans.^ 27. The practice was also known in Fpper Asia.^ For in- stance, in 402, the Ju-Juan, of which we have already spoken, were taught by their chief, Tu-lun, to make use of notched sticks in substitution of their former mode of notation.* In the eleventh century, the chiefs of the Ju-tchi, cognate with the preceding, were still issuing their orders by the old device of an arrow with a notch in it, while matters of urgency were distinguished by three notches.^ Before making a writing for themselves in 920 a.d., the Khitans or Liao used to keep their records by means of wooden tallies.^ 28. The oldest remains of notched sticks in literature are most probably the Kwas of the Yh-King.' These symbolical marks, made of lines, broken or entire, are now arranged, on a basis of eight sets of three, in sixty-four rows of six or double-three, each of which is placed at the head of a chapter of the above book. They are, in my opinion, nothing else than a survival of the notches formerly made ' E. C. Bridgman, Sketches of the Miau-tsze, § § 3, 9 ; in J.N.C.B.Jt.A.S. for 1859. — G. w. Clarke, A Manuscript Account of the Kwei-chau, Miao-tzu, § § 8, 49. App. to A. E. Colquhoun, Across Chryd, vol. ii. pp. 368, 383. ''■ Vide China before the Chinrse. ^ Above § 4, 13. * De Guignes, Mistoire des Huns, vol. i. (2), p. 338. "• Ma Tuanlin, Wen Men tung X'ao, hi. 327.— French translation, vol. ii. p. 440. ^ Al. "Wylie, Translation of the Ts'inff wan K'e mung, a Chinese Grammar of the Mandohu Tartar Language (8vo. Shanghai, 1855), p. xviii. — H. H. Howorth, The Northern Frontagers of China, V. The Ehitai or Khitans {J.R.A.S. 1881), excerpt, pp. 23-48. — G. Deveria, jExamen de la stele de Yen t'ai, in Zevue de V Extreme Orient, 1882, vol. i. p. 178, n. ' I have found that the early text of the Th-King, which has never been understood, will never be so, because the majority of the chapters axe made with fragments of an old dictionary, somewhat like the so-called syllabaries of Assyro- Babylonia; the other chapters are very old documents on various subjects, dating, as the others, from the introduction of writing in China. This solution, which is now accepted by nearly all the Sinologists who have scientifically studied the question since my paper has appeared, is established in The Oldest Book of the Chinese (J.R.A.S., 1882-3). Of. J.S.A.S. 1884, Vol. XVI. p. 460. BEGINNINGS OF "WRITING AROUND TIBET. 19 on the eight arrows of divination.^ They are the forms given in writing to the notches on wood, the broken lines representing the vertical, and the entire lines the horizontal lines, or the reverse. The Chinese traditional legends claim the Kwas (§ 17) as a regular means of intercommunication, which replaced a still older system of knotted cords. It may be that such a substitution did really take place, and that the Kwas are a survival of it, kept for purposes of divination. 29. The fu 1^ or check of the Chinese was nothing else than a tally or notched stick. As the Shwoh Wen has it, the description cannot be mistaken. Intended to secure faith between two parties, it was formerly made of bamboo, about six inches long, which, being cut in two, each party held a portion, in order to see whether they agreed at any future time. The use and meaning of the word was extended to tallies of all sorts and to the symbols of office, of which a part matching to the other always remained in the hands of the superior. Legendary history ascribes the practice to the rulers of the semi-historical period for the appointment of their officers.^ There are various historical instances of the use of the fu? It was in practice in the half-Chinese (or, better, non-Chinese) state of Tsu, on the banks of the Tang-tze Kiang, in the sixth century B.C. During the Han dynasty they were frequently employed, and the most interesting occasion was during the revolt of the " Eed Eyebrows," * circd a.d. 30, in N.W. China, which the Chinese connect with the Siamese.^ The fu was composed of two fragments, both fitting one another. The K'i K'iuen ^ |^ were another sort of binding device. Like the notched sticks, they were used ' The eight wands, or arrows of fate, are marked on many Babylonian cylinders as held in the hand of Marduk (Lajard, Culte de Mithra, pi. xxxii. No. 2 ; Ut. a. No. 5) or of Istar [ibid. pi. xxxvii. No. 1). 2 Of. K'anghi Tze-tien, 118 + 5. ff. 8-9. 3 Tai-Pitig yii-lan, bk. 598, fol. 1-2. * Teh'ih mei 'gfe Jg, so called because their leader, Fan Ts'ung, with his whole army, adopted the practice of dyeing the eyebrows blood-colour, in order to increase the terror that their appearance inspired. * Suh Wen fiien tung E'ao. — Ming y tung tchi. — Turn Kien luy han, bk. 334, fol. 6-7. 20 BEGINNINGS OF WHITING AEOtJND TIBET. double, each bearing corresponding marks. They were extensively used under the Han period and downwards. Sometimes they were made of silver ; more often of iron when dealing with aboriginal or foreign tribes — for instance, by the Emperor Min when dealing with the Ti-Kiang or Tibetans.' But the usual material was slips of wood or the bark of trees. ^ 30. The use of notched sticks is not confined to the Far East. It was formerly greatly practised in the West, and it still lingers in some countries. It will be sufficient for our illustrative purposes to remind our readers of the huch- stahen of the Germans, the hok-stafir of the Scandinavians, the coelhren of the Welsh, and of the old statement by Tacitus of the use in Grermany of notched sticks for divination. Every one is aware that the Chancellor of the Exchequer used (till 1834) formerly to present his budget (or bag) con- taining the tallies with which he checked his accounts ; and that the burning of the discarded tallies caused the fire and destruction of the building in the place of which were built the present Houses of Parliament. Notched sticks were, and are perhaps still, in use among the colliers in Scotland ; ' and they are still used by bakers in various parts of England and France,* in the danterbury hop- gardens, locally in some other trades, at Constantinople, etc.^ 31. The notches on sticks do not seem in the East and in the West to have ever been any other thing than a simple mnemonic process of numbers or objects. There is, however, one exception ; a singular instance of their development into a regular alphabet occurred in the case of the Oghams of Wales and Ireland. This was done under the influence of a previous alphabet about which there are two opinions. According to the most probable explanation, the Ogham writing was simply an adaptation of the runes to xylographic ^ Tain tchung king shu ; Tai-Ping yil-lan, bk. 598, fol. 6v. ^ On the K'i K'iuen at large, vid. Tai-Ping yii-lan, bk. 598, fol. 3-7. ' Where Colonel Yule saw them. Vid. his note in Tlie Book of Ser Marco Polo, Tol. ii. p. 78. * I saw those used by bakers in Normandy. ° Notes and Queries, ser. 1, vol. x. p. 485.— Edelestand du Meril, Esaai sur I'originet des Runes {8vo. Paris, 1844), p. 29. BEGINNINGS OF WRITING AROUND TIBBT. 21 conTenience, notches cut with a knife on the edge of a squared staff being substituted for the ordinary runes.^ Another and later opinion maintains their phonetic values from the Latin of the classical period.^ B. II. {e,f) Symbolical marks incised or drawn. 32. Grraffitti, not properly speaking inscriptions, have been found in Siberia, but they are not the expected primitive remains of ancient writings.* Some * are purely Tartar, being written in Mongolian and Kalmuck. Others, obviously the work of common people, may be Arabic ; while some others found on the left bank of the Jenissei river are much more interesting. They seem to me to be badly written in Syriac from right to left horizontally, before the time of the adaptation of this writing to the Uigur and Mongol. The characters are still separated one from the other. On one of these graffitti found at the same place, several Chinese characters, as written by common people, are recognizable ; for instance, J% ^ cyclical characters.^ The last we have to speak -of are quite peculiar and altogether different from the others. The signs are painted in red. They are made of straight lines disposed like draw- ings of lattices and window shades, and also like the tree characters of the Arabs and like the runes. They are met with near the Irtisch river on a rock over the stream Smolank.^ 33. The early rulers of the Pre-Chinese Bak tribes 1 Rev. Isaac Taylor, Greeks and Goths; a Study on the Sunes (London, 1879), pp. 108-139; The Alphabet, vol. ii. pp. 225-227. 2 M. D'Arbois de Jubainville, V Alphabet Irlandais primitif et le dieu Ogmios, pp. 20-26 in Aeademie Inser. et Bell.-Lettr. Comptes Eendus, 1881, vol. ix. 3 We neglect here, of course, the obviously Tataric graffitti and inscriptions of Mongol and Kalmuck characters. * P. J. von Strahlenberg : JDeseription of Bitssia, Siberia, and Great Tartary, etc., "with plates. 4to. London, 1738.— Greg. Spassky: JDe Antiquis quibusdam sculpturis et inaeriptionibus in Sibiria reperlis ; 4to. Petropoli, 1822. — An abstract of the latter, with plates, under this title : He quelques inscriptions deeouvertes en Siberie in L. de llosny : Archives Paldographiques de I' Orient et de I'Amerique; 8vo. pp. 143-6, Paris, 1872. -Meiners: tie Antiquis Monum. in Sibir. Auslrali extantibus ; in Comment. Soc. Reg. Getting, vol. xiii. 1799. Pallas, Neue nordische Beytraege, torn. v. St. Petersb. 1781. * Vid. above § 8, on a hieroglyphical one. * G. Spassky, op. cit. Archives, p. 145, pi. xix. Vid. Addenda, 22 BEGINNINGS OF WRITING AROUND TIBET. once migrated in China, found on tlie banks of rivers, near places of worship of the Aborigines, some curious marks, which I understood to be of the pattern known as cup- marks.i Such findings are attributed to the great Yii, to Yao, to Huang- ti, the first emperor, and to the legendaryFuh-hi himself. The legends of the findings are mostly centred about the Ho t'u,^ presented to Fuh-hi and the Lo shu^ offered to Yii, The regular and numerical disposition of these cup-marks, as understood in the Chinese traditions, reminded me of the groups of cup-marks as found by an Indian archaeolo- gist, Mr. Rivett Carnac, on the southern slopes of the Himalaya.* This assimilation was accepted, and further researches by Prof. R. K. Douglas^ and by myself have shown that cup-marks are met with in China in several places, in Shantung, Ngan-hui, Hupeh, Szetchuen and Kwangtung provinces. There is no doubt that these cup- marks were a mnemonic means of notation used by some tribes of the pre-Chinese population of the Middle Kingdom.^ ' The Oldest Book of the Chinese, § 28. ^ Or Map of the Ho river. , ^ Or writing from the Lo river. On these, vid. Mayers, Chinese Reader's Manual, pp. 56-9. * Journal of the Royal Asiatic Society of Bengal, 1877, vol. xlvi. Mr. J. H. Rivett Carnac had the kindness to write me from Ghazipur, and to send me his following papers; Pre^ Historic Remains in Central India, Calcutta, 1879; On Stone Implements from the North-Western Provinces of India, Calcutta, 1883; A.rchceological Notes on Ancient Sculpturings on Rocks in Kumaon, India, similar to those found on Monoliths and Pocks in Europe, 8vo. Calcutta, 1883. * Cup-marks, in Saturday Review, Nov. 24, 1883. ^ Some marks, straight lines and circles, on rocks were found along the southern coast of Hawai. Cf. A. Bastian, Sprachvergleichen Studien .... der Indo- Chinesischen Sprachen, p. 104 (8vo. Leipzig, 1870). I extract the following note from a contemporary (Feb. 13, 1883) : " In many parts of Switzerland, writes our Geneva correspondent, are often found smooth flat stones, evidently hand- polished, and covered with dots, lines, circles, and half-circles. The origin and use of these stones, known among country people as Schalensteine, has long been a moot point among the learned. Some have thought they were charms, others that they were meant to commemorate the dead, or that the signs on them were undecipherable hieroglyphics ; hut it has been reserved for Herr Rodiger, of Bellach, in Solothurn, to throw a new light on these mysterious relics of the past, and suggest a theory concerning them which seems to meet aE the necessities of the case. The Schalensteine, he says, are neither more nor less than topographical charts, as a comparison of them with any modem map of the districts in which they are found will show. The engraved dots correspond with existing towns and villages, the lines with roads. Even the fords and moimtain passes are indicated. Herr Rodiger has examined many of these stones from various parts of the country, and he possesses a collection, picked up in Solothurn, which form together a map of the entire canton. Another significant cii-curastance is that the Schalensteine are mostly found at intervals of about two hours (say, six miles) BEGINNINGS OF WEITING AROUND TIBET. 23 34. There is in Chinese books a curious legend about a very ancient writing from Yueh-tchang. This country was in the Indo-Chinese peninsula, and the region which formerly bore that name was half in Tong King and half in Cochinchina. A mission is said to have come from there to the Chinese Court in the middle of the eleventh century B.C.' Another and much earlier mission from the same country, mentioned however, and as far as I know, only in historical compilations of late date,^ is attributed to the time ofYao (2145-2043 B.C.). In the fifth year of his reign, a mission from Yueh-tchang came to his Court and presented a tortoise, a thousand years old and three feet in size ; on its back were characters of the Ko-teu (or oldest) style recording what had happened since the beginning of the world. Yao ordered it to be transcribed and called it the " Tortoise Annals " (or, better, ephemeridis).^ This second tradition has apparently developed from that dating from the Tchou dynasty, i.e. the mission of the eleventh century. The uncritical Chinese compilers of later times are very fond of embellishing their records with repetitions of events which might enhance the glory of their sages of antiquity.* The number of legends engendered in this way is by no means inconsiderable. from each other, and at spots where several roads meet. The former Herr Eodiger calls "headstones" (Sauptateine) , the latter he denominates "hy- stones " ( NebensteineJ . If he he right in his hypothesis, the places where these stones are met with possessed considerable populations long before the dawn of history ; even the villages shown on the Sehalensteine must be far older than the Christian era. Herr Eodiger considers the Swiss map stones to be of the same origin as the similar stones which are found in Germany, Scandinavia, India, and further Asia, and sees in them another proof of the high antiquity and common origin of the Indo-Germanic races, and the existence among the latter, in an indefinitely remote age, of civilized habits, organized trade, and more culture than is generally supposed." ' Vid. below § 44. '^ It was apparently not yet current in the tenth century, if we may so infer from the silence of the Tai-ping yu Ian, bk. 785, f. 2, which mentions only the second mission. ' K'ang Kien tcheng she yoh, 1737, bk. i. f. 9.—Kang Kien y tchi luh, 1711 ; Medhurst, Ancient China, p. 330. — T'ung Kien Eang muh (twelfth cent.), bk. vii. f. 18. * On these embellishments and their subsequent development, cf. some remarks in my paper on The Old Numerals, the Counting Rods and the Swan-Pan in China, London, 1884, p. 1 (excerpt Numismatic Chronicle, vol. iii. pp. 297-340). 24 BEGINNINGS OF WRITING AROUND TIBET. 35. The use of conventional marks belongs to all times and to all countries, and it is needless to insist upon such a well- known fact. Disagreement of opinion between scholars may- happen in some particular cases. For instance, the striae found on a piece of reindeer's horn at Cro-Magnon of the Stone age are explained by some as mnemonic marks.^ And so too are the striae found under similar circumstances. Such was the opinion of Lartet, Christy, Broca, Lenormant and others. Another explanation, however, is given by different authorities ; they maintain that such notches were only made on the handles of the implements in order to prevent their slipping out of the hand.^ Many instances could be put forward here of the use of marks everywhere, on the rude stone monuments of Europe,^ as well as in other countries, but it would lengthen uselessly the present section. The two great lines of evolution which we have attempted to explain in the preceding sections, in numerous examples of embryo-writings, though occasionally intercrossed, keep pretty well their individuality up to the point we have now reached. They both converge to regular writing, and we do not think that any system of writing, deserving to be so called, has ever been framed without the co-operation of these two great sources of notation.* This remark does not apply to the numerous cryptograms which are based on a previous knowledge of alphabetism ; such, for instance, as the Anaitsi and other characters of Japan,^ the tree-alphabets El ^ Vid. Fr. Lenormant, Sistoire ancienne del' Orient, 9tli edit. vol. i. p. 399; Dr. Broca, Sur les Troglodytes de la Vezere. * G. de Mortillet, Le Prehistorique, p. 408 (Paris, 1883). ' Such, for example, as the marks on the stones of the dolmen of Mane-Lud. Sir J. Simpson has collected all these marks in his work, Archaic Sculpturings of Cups, Ciicles, etc., on Stones and Roehs in Scotland, England, and other Countries. 8to. Edinburgh, 1867. * A French scholar, Mr. C. Scboebel, published in 1882 a learned but not convincing paper, Memoire sur les Origines de VEeriture Alphabetique (in Actes de la Soeiete Philologique AsVecns, pp. 137-213), where he denies the evolution towards alphabet from a hieroglyphical basis. ' Cf. Sanaki-bara yoshi-no : Bun-gei rui-san, bk. i. f. 22. — Hirata Atsutane, Giji hen, ff. 14, 16, 16, 17. BEGINNINGS OF "WRITING AROUND TIBET. 25 Mushajjar and El Shnjari of the Arabs,^ or the numerous examples published in the curious work of Abubekr ben "Wahshih,^ etc. II. Wkitings Impeded and Decayed. 36. The struggle for life is a condition of existence for a writing as for other things. The best fitted resists and lasts. But the better fitting is purely objective ; it depends upon the surrounding circumstances much more than upon the intrinsic perfection and high standard of the writing concerned. I do not think this interesting problem of evolution has ever been considered from this important standpoint ; while opposite statements, as if they were a matter of course, are often met with in learned works. Alphabets and phonetic writings once acquired have been lost because they were too far advanced for their surroundings. In some cases they have either disappeared, in others they have dropped their too much advanced capabilities. Of such cases there are not a few. Let us consider some instances of the two classes. 37. " The Ainos have the custom of inscribing all their objects with signs which vary according to the owner. These signs are made of curved and straight lines. Trees in the forests and points of bamboo arrows are marked in the same way." ^ Fac-similes of these signs have been published by the learned author of this statement.* Ifow an examina- tion of them has satisfied me, without leaving any doubt in my mind, that the people who use these marks were once ac- quainted with the alphabetic writing still used in Corea and not unknown in Japan. Several of these marks can be still resolved into their alphabetic elements, and consequently ' " Constructed out of the Arabic alphabet, after the Arabs had come into contact with the Varangians in the ninth century." — Isaac Taylor, T/ie Alphabet, ii. 226n. * Ancient Alphabets and Hieroglyphic Characters Explained, .... in the Arabic language, by Ahmad ben Abubekr ben Wahshih. Transl. by Joseph Hammer. Svo. London, 1806, pp. 38, 45. ^ Heinrich Siebold, Ethnohgische studien fiber die Ainos auf der Jnsel Tesso (BerUn, 1881, Svo.) s. 19. < Hid. Taff. II. 26 BEGINNINGS OF WRITING AEOUND TIBET. easily read ; for instance, oku, sao, sno, us, yes, are easily read ; some more may be deciphered with perseverance, while others are corrupted beyond hope. Therefore we are justified in assuming that the present people in using these marks employ the groups bodily, and know no more about their composition and Talue. Since I have examined the specimens published by Herr H. von Siebold, an Aino inscription has been published by Dr. B. Schube of Kioto. ^ There again the oblivion of the old alphabetic writing is obvious, and the characters, many of which are still recognizable, are mixed with hieroglyphic and symbolic signs. As to the time when the A'inos were made acquainted with this Corean or Onmun alphabet, we have no information. They may have obtained it direct from the Coreans, as the Japanese do not seem to have been seriously acquainted with or to have ever used it, except in modern forgeries.^ Besides that, the peculiar combinations made for the rendering of Japanese sounds, which appear in the Japanese form of the alphabet,^ and which obviously bear a modern face, do not appear in the Aino use of the alphabet. 38. The Lolos of Szetchuen have an alphabetic writing,* 1 Die A'inos, Taff. vii. Inschrift zweifelhaften TTrsprungs bei Oturanai. Mittheilungen der Seutschen Gesellschaft far Natur- und Vollcerkunde Oslasiens, 26 stes Helft. Februar, 1882, Band iii. pp. 220-256. Vid. Addenda. ' M. L^on de Eosny thought he had found proofs of the early use of this writing in Japan, and he communicated the fact to the Academie des Inscriptions et Belles-Lettrea. Vid. Zfs sources des plus anciennes de I'histoire de Japan, pp. 105-116, and L'ecriture sacrie et les inscriptions de I'antiquite Japonaise, pp. 170-177 of Comptes nndus A.I.B.-L., 1882, t. ix. But hi supposed discovery turned out to be a misconception according to the severe criticism of Mr. Basil Hall Chamberlain, On two Questions of Japanese Archceoloqy, pp. 315-332 of J.S.A.S 1883, Vol. XV. 3 Vid. the specimens dating from 1477 a.d., in the Bwn-gei rui-san (1878), bk. i. S.. 14-15. * Vid. my paper On a Lolo Manuscript written on Satin in J.S.A.S. Vol. XIV. pp. 1 19-123. _ The great interest of this writing lies in its bearing on the history of Indian writing. _ The oldest specimen of writing hitherto known in India is a stone seal found in some ruins at Harapa, near Lahore, upon which General Cunningham writes as follows : — " Its age is of course quite uncertain, but I do not think its date can be later than 500 to 400 e.o. I now think it may be archaic Indian letters of as early an age as Buddha himself." — Corpus Inscriptionum Indiearum, p 60. Now these characters ake the same wEiTi.s'G AS THAT NOW POSSESSED BY THE LoLos, as I have showu in the above pamphlet by juxtaposition of the two writings. BEGINNINGS OF "WRITING AROUND TIBET. 27 connected (by common descent according to my views ') with the South Indian Alphabet of A^oka. But they no more understand it as such, and are unable to compose new groups or disentangle the old ones. They often use each word or group as an ideo- or phonogram. This is shown by the extreme corruption of many of these groups as exhibited in the latest documents we have received of the same writing. The clustering of the characters in groups, which, as in Chinese, is a characteristic of this writing, has helped to the obliteration of the characters individually ; they can no more be used otherwise than in these groups of ancient make representing the old pronunciation. And these groups are now considered as inseparable ideograms, and used accordingly without knowledge of the respective values of their compo- nent characters.^ This writing, once phonetic, is returning to the ideographic stage. Its phonetic practice entailed more mental work than the common capacities of the people would permit. It was hence fated to drop its higher capability.* 39. Among the several writings which were used in Borneo two have left interesting relics and survivals.* The Dayaks engrave as ornaments some signs which they obviously understand no more. Some bamboo objects ex- hibited at the India Museum, London, bear these marks. They are apparently the survival of an alphabetical writing 1 Vid. my remarks in E. C. Baber's Travels and Researches in Western China, pp. 142-143 (London, 1882, E.G.S., suppl. pap. toI. i.), and also in J.R.A.S. 1882, Vol. XIV. pp. 802-803; T. de L., On the Bistory of Archaic Chinese Writing and Texts, p. 8 (London, 1882, 8vo.) ; The Academy, July 2, 1881. ^ These tribes are reported to bave a 'written language of OTer 600 characters. These, he says, are symbols of sounds and not of things, as the Chinese characters. Such was the saying of a Mandarin of Kueitchou province, to Deka, Notes and Queries on China and Japan, vol. i. p. 104. ' Up till now we bave five texts in this writing. Three fac-similes were published in E. C. Baber's Travels and Researches. The fourth is the MS. written on satin described in my pamphlet On a Lolo MS. above quoted. The fifth consists of three pages sent by the missionaries MM. Gourdin and V. CrabouiUet to my friend Mr. E. C. Baber, who gave them to me. They will appear in my book China before the Chinese. A sixth exists in the Library of the Asiatic Society of Shanghai ; it has been described in Revue de V Extreme Orient, 1884, vol. ii. pp. 682-83. * There are many teaces in some parts of Borneo of Chinese influence, shape and ornamentation of roofs, etc. Important Chinese colonies were formerly settled there from the fifteenth century. A Dayak tribe in the interior claim to be descendants of Chinese. Cf. W. P. Groeneveldt, Notes on the Malay Archipelago and Malacca, compiled from Chinese sources, pp. 101-115 (8vo. Batavia, 1876). 28 BEGINNINGS OF WRITING AROUND TIBET. anciently known there, and afterwards forgotten.^ We find a similar writing on an earthenware vase from the same island belonging to the Ethnographical Museum of Dresden. This vase, as far as I remember from a sketch communicated to me by Mr. A. W. Franks, is ornamented with two figures of the Chinese dragon, but not of Chinese make.^ Dr. Kern has published some inscriptions found at Koutei in the same island, which are written in this character of Eastern India the Vengi-Chalukya in Kalinga, the same that was carried to Cambodia, to "Western Java, and elsewhere, as we shall presently see.^ 40. This same character of the Vengi-Chalukya was also carried to North-Celebes Islands. The people have not remained at the level required for the practical use of a phonetic writing. It is no more used as an alphabet. Curiously enough, it is employed as pictorial ornaments on the MSS. they now write in a pictographic style of the lowest scale. This I have seen on the facsimile* published by Dr. A. B. Meyer, of Dresden, in his splendid album on the writings of this region. The finding was extremely welcome because it is a partial confirmation, and anyhow a stepping- stone helping to understand another palseographical discovery of mine of rather a startling character.' 41. In the Easter Island, or Va'ihu, some fourteen inscrip- 1 I was acquainted -with this inscription through a fac-simile sent to ray learned friends, Col. H. Yule and Dr. R. Rost, by Dr. A. B. Meyer, Keeper of the Museum. This writing is not without some apparent connection with one of the writings of Sumatra. On the other hand, it presents no less a curious semblance with an inscription dug out in Japan at Jlsu/ci ^ ^^ ^j^ in Fiuffa Euni |p] ^ in 1821 (Bem-sei, fourth year). Cf. the fac-simile in Ta Jlhpon ti Ming, f. 189». ^ J do not know in what part of Borneo the vase was found. It is, however cui-ious to point out that the Chinese annals of the Ming dynasty (1368-1643), Ming she, bk. 323, in the notice about Bandjermasin, speaking of the people, say that " they very much like earthen jars with dragons outside." 3 Over de Opschriftcn uit Eoeter in Verband met de Geschwdenis van het schrift in den Indisclien Arehipel. 8vo. Amsterdam, 1882, p. 18. — Also K. V. Holle, Tabd van Oud- en Nienw-Lidische Alphabetlen. Bijdrage tot de palaeugraphie van Nederlandsch- Indie {?,m. Batavia, 1882), n. 80-1. * Bilderschriften des Ostindischen Archipels, pi. I. 1. 11. ^ It is interesting here to note that a bronze bell bearing an inscription in ancient Tamil characters, has been discovered at Wangarei. — Cf. T. R. Taylor Te Ika a Maui ; or, New Zealand and its Inhabitants, p. 34. —A. de Quatrefages, Bommes fossilea et hommes sauvages (Paris, 1884, 8vo.), p. 476. BEGINNINGS OP WRITING AROUND TIBET. 29 tions have been found incised on wooden boards, perhaps of drift wood. The characters are peculiar. Most of them display strange shapes in which with a little imagination forms of men, fishes, trees, birds, and many other things have been fancied. A curious characteristic is that the upper part of the signs are shaped somewhat like the head of the herronia or albatros. A pictorial tendency is obvious in all of these. Some persons in Europe have taken them for hieroglyphics, and have ventured to find a connection with the flora and fauna of the island. The knowledge of this writing is now lost ; and it is not sure that the few priests and other men of the last generation who boasted of being able to read them, could do so thoroughly. Anyhow, in 1770 some chiefs were still able to write down their names on a deed of gift^ when the island was taken in the name of Carlos III. of Spain. 42. Now, an enthusiastic archaeologist of this country, Mr. J. Park Harrison, has spared no pains to bring together all the possible information on these inscriptions. He has put himself in communication with persons acquainted with the island, and he has published with views of his own all that he could get, and also fac-similes of two sides of one of the above inscriptions on drift wood. He had also another one photographed, and had the great kindness to communi- cate to me his materials some three years ago.^ 43. In examining carefully the characters, I was struck by the forked heads of many of them, which reminded me of the forked matras of the Yengi-Chalukya inscriptions. ' Sen. Gonzalez de la Eoza has exhibited a tracing of these signatures, -which were reproduced by Mr. Park Harrison, op. cit. pp. 14, 15, pi. 27. — Journal of the Anthropological Institute, -vol. iii. pi. ixvii. For instance, one of these signatures is a big monogram Inu, which was apparently used as an ideogram, while the other signatures are written in a concise form of the characters and monograms of the inscriptions. 2 J. Park Harrison, The Sieroglyphus of Hosier Island (8vo. London, 1874), p. 16, with five plates. — Jfoie on Five Hieroglyphic Tablets from Faster Island, p. 2. — rW. also mature, Sept. 17, 1874. — J.Linton Palmer, iJa™ or Easter Island (with plate) in Proceedings of the Literary and Fhilosophieal Society of Liverpool, 1875, xxix. pp. 215-97 ; On some Tablets found in Faster Island [-mfii plate), ibid. 1876, xxx. pp. 255-63. — A small inscription enlarged is given in Meyer's Bilderschriften des Ostindisehen Archipels, pi. vi. — Also one in Tour du monde, 2° sem. 1882, A. Pinart, L'ile de Faques. 30 BEGINNINGS OF WRITING AROUND TIBET. A closer comparison with plates i. to viii. of the Elements of South Indian Palceography ^ soon showed me that I was on the right track. And a further study of the Vdihu characters and their analysis hy comparing the small differences (vocalic notation) existing between several of them, convinced me that they are nothing else than a decayed form of the above writing of Southern India returning to the hieroglyphical stage. With this clue, the inscriptions of Easter Island are no more a sealed text. They can easily be read after a little training. Their language is Polynesian, and I can say that the vocabulary of the Samoan dialect has proved very useful to me for the purpose.^ It is useless to dwell on the importance of this little palseographical discovery for the history of civilization, and its dissemination eastwards. 44. The Giao-tchi or Annamites had once a phonetic writing, which they have lost under the influence of the Chinese.^ A great antiquity is claimed for such knowledge among them ; but we have no means for verifying such an early date as the eleventh century B.C. And though this writing might have been a derivation from the phonetically used Chinese characters, we are not inclined to accept, with- out reserve, the genuineness of so early a claim.* Dr. A. Bastian was given by a Shan of the Yuns who live at Kiintun, near the frontier of Yunnan, an alphabet, " which may probably resemble that of the Quanto, the ancestors of the Tunkinese."^ P. Montrouzies, a missionary there, says that 1 A. C. Bumell, Elements of South Indian Paleography, from the fourth to the seventeenth century A.D., being An Introduction to the Study of South Indian Inscriptions and MSS. 2nd edit. 4to. London and Mangalore, 1878. — Plates i. vii. viii. are specially interesting for the forked matras. ^ I had vrithheld the publication of this discovery with the hope of spending more time over it, and giving a transcription and translation of several of these inscriptions. Though it is stUl my intention, I find my hands so full for a time to come, that I venture to print the above information for the use of some scholar conversant with these matters. ' It is said in their records that their writing was in use at the time of their sending a mission to the Chinese Eng Tching of Tchou (viz. U09 or 1039 B.C.), and that it required translation into Chinese. Vid. P. J. B. Truong-vinh-ky, Cours d'hisioire Annamite (Saigon, 1875), vol. i. pp. 11. * The question is discussed at length in China before the Chinese. * Remarks on the Indo-Chinese Alphabets, in J.R.A.S., 1868, Vol, III. pp. 65-80. BEGINNINGS OF "WRITING ABOUND TIBET. 31 the phonetic writing of the Giao-tchi is still in use nowadays among the mountainous tribes of the province of Nghe-An, in Tongking.^ This is the modern name of the region formerly called Yueh-tchang or Viet-thuong, from whence the above mission to China is said to have come.^ The same missionary thinks that he has found some specimens of it in inscriptions of a grotto of the basin of the Sang- Gianh. In 186 A.D. Si-nhip, Chinese viceroy of Annam, formally introduced the Chinese writing, forbidding most expressly the employment of the former phonetic writing, which the Annamites have entirely lost since that time.^ 45. In the island of Hainan " Captain Culver found what appeared to be characters of some kind scrawled on the walls of a Li village near Yu-lin-Kan, but none of the villagers could tell him anything about them, except that they had probably been written by a medicine man, who was, however, not forthcoming. The characters have a resemblance to a kind of mixture of Chinese and Malay, something like what one might imagine these characters written on the surface of rippling water. Yid. Addenda. "The Li, who are non-Chinese rude tribes, occupiers of the island, do not seem to have any form of writing, none certainly that is generally known to the people ; they do not seem to have even any recognized symbols for numbers." * 46. Another instance of great moment is that of the Chinese writing. Its history, instead of being that of an evolution towards progress (as we understand it), is, on the contrary, that of a slow decay and adaptation to surrounding circumstances of a lower standard than those of its former focus in the west. The writing of China was not born in the Middle 1 Dutreuil de Ehins, Kotiee sur le Tong-King, in Bulletin de la Society de Oeographie de Paris, Avril, 1880, p. 311. * Vid. the official geography of Aimam, puhlished in 1829, Eoang viet dia du cM, vol. i. pp. 1, 9; vol. ii. p. 31. The Yueh-tehamg or Viet-thiiong region is now covered by the provinces of NghS-An, Thuan-hoa and Quang-nam. 3 Vid. Truong Vinh Ky, ibid. p. 27. * James George Scott, France and Tonghing (London, 1885), pp. 353-354. 32 BEGINNINGS OF WIIITING AROUND TIBET. kingdom,! ^^^^ there are no reasons to doubt tlie tradition that the leaders of the Bak tribes, when they entered into China, were acquainted with it, while there are proofs of all sorts in its favour. In such a country of conservatism and preservation carried to the utmost, if the writing had followed a regular course of development, and been evolved from the limbo of any embryonic means of communication, or of a rude system of hieroglyphics, traces of this forcibly slow growth would certainly be found. But it is the reverse that we discover. This writing is not of indigenous growth. It is an importation from the west. 47. The oldest palseographical remains which have been preserved prove the writing to be from the beginning in China, a scholarly writing comprising ideograms and phono- grams (former ideograms turned meaningless and phonetic) susceptible of a great power of phonetic combinations.^ Many signs had reached purely conventional shapes. A proportion by no means small of the ideograms had still some pictorial features, survival of their former hiero- glyphical stage, which the Chinese scribes eventually turned into profit at the expense of the phonetism. They insisted on the pictorial features, and probably increased their number, in order to facilitate the understanding of the writing to the various dialect-speaking people. But they were not aware that their characters had passed through several phases of evolution before reaching them ; they did not know that their writing, written in columns, must be derived from an older one written horizontally, since, as a natural consequence, many characters have to be laid down before ' This unexpected conclusion, to which I was long adverse, has cost me many years of patient work before I could conceive and understand it. But it is now so clear that any scholar conversant with the matter who examines without prejudiced views the facts I put forward in my publications, cannot fail to be convinced. 1 am now myself rather ashamed not to have found it out at the beginning of my researches. Vid. §§9, 17, 19, 33, 49. ^ In favour of this result of my studies 1 find the following views : Tai Timg, a renowned palaeographer of the thirteenth century, author of the Zuh Shu Kn, arrived at the conclusion that hieroglyphics do not constitute the only original ground of the Chinese writing. Vid. J. Nacken, A Chinese Webster, in China Beview, 1873, vol. ii. p. 176. M. Leon de Eosny, Les £critures Jiguratives et hierogh/phiques des differents peuples aneiens et modernea (2nd edit. Paris, 1870, 4to.), pp. 3 and 4, expresses a similar opinion. BEGINNINGS OP WRITING AROUND TIBET. 33 the survival of hieroglyphieal features remaining in them can be recognized. This characteristic, which we ascertain to- day, was never guessed by them.^ It has for us been a help to find its antecedent writing in that of south-west Asia. 48. The rude pictorial characters which appear in many European books as representatives of the primitive writing of China cannot be accepted as anything of the kind. As a matter of fact, they are not old. They are taken from inscriptions on vases which are forgeries.^ Imbued with the idea that the rougher the writing the older and con- sequently the higher priced would be their vases, the forgers (generally of the Sung dynasty downwards to the present day) have drawn their inscriptions accordingly. The proof of this statement is easy to give. The Chinese are noted forgers, but, like those of Europe, they seldom escape detection, which in this case comes from comparison of their own works. We find, for instance, some inscriptions of the most common kind repeated over and over again on genuinely antique vases and correctly written ; while in the forgeries they are distorted pictorially, and gradually turned ruder in the successive imitations made by unskilled hands. And, happily enough, in manipulating these inscriptions, which they do not understand, they display the most curious blunders.' An examination of such inscriptions disposed in ' On its curious effect on the phonetic reading of the ancient groups cf. The Oldest Book of the Chinese, § 23, and also a valuable article in The Times, Aug. 26, 1884, Further Progress in Chinese Studies. ^ Such is also the opinion of Rev. J. Chalmers, The Origin of the Chinese ; an attempt to trace the connection of the Chinese with western nations in their religion, superstitions, arts, language, and traditions (8vo. London, 1868), p. 60. A rather indifferent work, which does not repose on a sulficiently extensive knowledge for the ground it seeks to cover. ' For instance, in the first volume of his great work on China, the well-known German traveller, F. von Richthofeu (C7»«a, vol. i. p. 371), has given a sketch of a bronze yase with its inscription as a specimen of the oldest bronze industry under the Tchou dynasty, 1000 e c. Now the inscription proves that the vase is a forgery. The founder, to escape detection, has dropped the first or left column- line, and the last character of every other five column-line of an old inscription of six lines, which is known from two ancient objects on which it was inscribed. Fac-similes have been published in the Ku yii t'u, bk. iii. ff. 7-8, and bk. xvii. S. 14-5. A splendid bronze vase brought back by Gen. Malcolm contains an inscription in twelve characters, copied simply from the 16th, 17th, and 18th columns of the well-known Sun she pan inscription. Mr. J. Drury Fortnum, of Stanmore, has in his collection a beautiful vase, containing the half of an old inscription, etc. 34 BEGINNINGS OF WfilTIXG AROUND TIBET. rows is highly instructive : we intend publishing several of them some day. Is it necessary to add that these fanciful characters have no connection whatever with the oldest forms preserved from antiquity ? 49. The Chinese have some traditions, enveloped in a mist of fiction, that their writing comes from the west. Though entangled with secondary myths of later growth, under the late influences of national reasoning and of foreign ideas, it is not impossible to trace out of them a few reminiscences which underlie their whole fabric. It would be trespassing on the purpose of the present paper to do so here, and it will be sufficient to mention only that which relates to the shape of the primary characters learned by the early Chinese Bak tribes before they migrated to the Flowery Land. This writing (according to the traditions and archaeo- logical evidence) was neither scratched nor painted, but was cut deep into soft material ; many signs represented (or were supposed to represent) all sorts of things and objects; some strokes of the writing were thick at one end and thin at the other; they were imitated from "prints of birds' claws on clay," and sometimes they were compared to " tongues of fire," or to " drops of rain freezing when falling," These legendary peculiarities precisely enough remind us of the cuneiform writing of Babylon and the surrounding countries. liow it happens that among the oldest specimens of characters kept traditionally in the best Chinese paleeo- graphical works, there are in a few characters survivals of a wedge-writing appearance. The few thousand written words which formed the whole material of the ancient Chinese books rest on a basis of some five hundred difierent signs (tradition says 540) ; the oldest forms of the half of this quantity have been preserved. Compared with the few hundred characters which, in the same way, form the basis of the ancient writing of Babylon, they display the most remarkable likenesses. Taking into account the allowance to be made for the difference of material used for writing (which has caused the wedge to vanish), their identity is indisputable. This identity goes much beyond the mere shapes, sounds and BEGINNINGS OF WEITING AROUND TIBET. 35 meanings, as many other peculiarities are common to the two writings ; the Chinese presenting an imitation, somewhat im- perfect, of the other. Comparative researches on a scientific footing show, beyond any doubt, that the elements of the early civilization of the Chinese, and the bases of their knowledge and institutions, were borrowed from a region or people con- nected with the old culture fostered in Babylon.^ The greatest probability is that the borrowing was eflfected through prac- tical intercourse with Susiana or the country of Elam. And in the history of the Chinese this is the first of the sixteen sources which I find to have contributed to the formation of their civilization. 50. In the oldest style the writing was the faithful ex- pression of the spoken language. By the phonetic reading of the oldest Ku-wen characters we find the ancient spoken forms of the words. Taking into account the wear and tear, many of them can still be identified with the same words of common descent in cognate languages. We cannot here give all the necessary details which show how closely connected are the oldest means of phonetic expression of this writing with that of Babylonia (a connexion corroborated by the shape of the characters) in polyphony, phonetic com- plements, and phonetic combination, etc. The oldest phonetic order in the Chinese phonetic groups is from left to right, and also, curiously enough, from bottom to top ; ^ the latter ' This is established by a formidable array of facts. Some of them are mnst convincing. "We have learned from the Chinese palaeographers, for characters identified with ancient Babylonian, some meanings still unknown, and deciphered afterwards by Assyriologists. On the other hand, in May, 1880, 1 was able to announce that the Chinese signs of the cardinal points, similar to those of Babylon, exhibit a shifting then unexplained ; three years afterwards Mr. T. G. Pinches made a corresponding discovery on the Assyro-Babylonian side, cf. T. de 1/., Early History of Christian Civilization, 1880, p. 29 ; T. G. Pinches, in Proceedings of the Society of Biblical Archceology, 6 Feb. 1883. I read myself at the Eoyai Asiatic Society, on April 30th of the same year, a paper On the Shifting of the Cardinal Points, as an illustration of the Chaldieo-Pabyloman culture borrowed by the Early Chinese, which will soon appear in my book on the Sources of Chinese Cinlizaiion, where it is shown that it gives a clue to the road followed by the civilizers of China. And I had the satisfaction at the Eoyal Asiatic Society's meeting, on March Hth, 1884, to hear Sir Henry Eawlinson, the founder and the most prominent member of the Assyriological school in this country, state his belief (after three years of opposed opinion) iu my discoveries of the derivation of the early Chinese civilization from that of Assyrio-Babylonia through the intermediary of Susiana. ' Vid. some instances. The Oldest Book of the Chinese, § 23, and notes. 36 BEGINNINGS OF WRITING AROUND TIBET. resulted, at first, from the turning up of groups originally- written horizontally; these processes were imitated after- wards, and sometimes only for convenience with some special characters. The phonetic-comhination groups (where a closed syllable, for instance, is written with two signs, one for the initial, the other for the final) are more frequently met with in the oldest specimens than afterwards, when they were repeated only by tradition. On the other hand, the polysyllabic-combination groups (where each part has a meaning and a sound by itself), as a result of the evolution of the language and the oblivion of the old principles, are found later on. I have been able to give elsewhere a curious instance of them explaining an unintelligible passage of the Shu-King at the beginning of the Tchou dynasty.^ 51. The Chinese ofiicial scribes clung to the old principles as much, and as long, as they could. But with the gradual differentiation and decay of pronunciation caused by enlarge- ment of the nation and absorption of foreign native tribes, they eventually found themselves outstripped by a gap between their traditional rendering of the ofiicial language and the dialectal varieties. Successive transcriptions of the same words make us assist at the struggle that took place. We see how the scribes ventured to render, by the old means of phonetic composition, then fading away (as shown in the preceding section), the new words introduced into the lan- guage. But the ideograms which could be understood notwith- standing the dialectal discrepancies proved stronger in the long run. With the growing independency of the various principalities, ideographism was more extensively used in some states, while in some others the ancient principles of phonetic rendering adapted to the regional pronunciation were still adhered to. The inscriptions show that it is specially in the western quarters of the Chinese dominion that the old phonetic orthography was maintained. 52. In order to understand the sequel of the evolution, and see bow the Chinese were led to the threshold of the 1 Vid. The Oldest Boole of the Chinese, § 23«. /. ij. A. S. Vol. XIT. p. 800, BEGINNINGS OF WEITIXG AROUND TIBET. 37 alphabetic principles without grasping them, and even dropped what they were taught originally, we must consider a few phenomena of the history of their language. The Chinese B.ak tribes, when they reached the Flowery Land, spoke a language presenting (so far as comparison may he allowed at three thousand years' distance) more affinities with the TJgro-Finnish than with the Turko-Tartar lan- guages. But it was characterized, like the Akkadian, ^ by a marked tendency to agglutination, somewhat, though in a smaller degree, like the polysynthetic or holophrastic lan- guages of North America. The words dropped a part of their significative elements in case of association in phrase groups. The Bak tribes found China occupied by popula- tions of several races speaking at least two classes of languages of different ideology, and at variance with that of the Chinese.^ One of these classes, that of the M^n, originating from India, possessed an ideology opposed to the Chinese, elliptic tendencies, and a characteristic nicety of distinction in vowel sounds. The intermingling of these languages with that of the new comers has produced the Chinese language and its dialects, and the cognate and special characteristics of the Tibeto-Burman and Tai-Shan languages. The phonetic consequences of this intermingling, con- sidered here specially in that which concerns the Chinese languages, were peculiar ; so we must remember that nearly everything here explained applies also to the above lan- guages, but on a different scale.^ 53. The archaic language of the Bak tribes possessed the special phonetic feature called the harmonization of vowels, though thematic only, confined, as in Akkadian, to the signi- ' Cf. Franijois Lenorraaut, La langue primitive de la Chaldie et les idiomea Touraniens (8to. Paris, 1875), pp. 278-279. ^ The purpose of my forthcoming work, China before the Chinese, Is to disen- tangle the nexus of all the non-Chinese tribes which have occupied the country before the Chinese, in order to permit me to consider exclusively the Chinese origin in a following work. A short abstract of a part of it forms my paper, The Cradle of the Shan Race, introduction to A. R. Colquhoun's book, Amongst the Shans (8vo. London, 1885). ^ Vide a clear classiiication of this phenomenon, Lucien Adam, De Vharmonie des voyelles (8to. Paris, 1 874). 38 BEGINNINGS OF "WRITING AROUND TIBET. ficative parts of the words.' This was the weakest point, and the first which yielded to the native tendencies of ellipse and contraction. A bisyllabic word, for instance, of which the syllabic Towels belong to one class only, is very easily contracted by the dropping of one of the vowels;^ the aggre- gate consonants which remain are gradually resolved into a single one, and sometimes eventually dropped altogether. The natural equivalent required to compensate the dropped elements of a crippled word reduced to one expiration or puflf of breath, could not be obtained by a stress on one part (the crippled one), and indifference on the other.^ It was sought for by the genius of the language in a difference of pitch,* simple or compound, according to the place of the lost part, in the utterance of the whole puff-of-hreath word. 5i. The gradual curtailing or crippling of the words entailed in speech the addition of a completing word, suggestive of class, as in the native dialects, or of a synonym, the selection of which was, of course, a matter of dialectic preference. Often, words borrowed from the native languages were used to facilitate understanding among the mixed population. This is a phenomenon of frequent occurrence in the history of languages. The rendering of such additional words was, of course, for the scribes, a matter of perplexity. In the canonical books they were often expressed by additional phonetic- combination sub-groups, but the prevailing growth of ideo- graphism led to the selection of ideograms, which in the ' Francois Lenormant, ChaUcean Magic, its origin and development (8vo. London, 1878), p. 271. - Vid. my Early Sistory of the Chinese Civilisation (London, 1880), p. 29. But these affinities date from a period very remote, when we may safely assume that the present Ugro-Altaic and Turko-Tartar groups were not yet bent by their surrounding circumstances to their present distract course of evolution. This dis- tinction of the two groups is by no means clearly established, as shown by the late contention about the classification of the Magyar (cf. A Magyarok eredete irta Vambery Armin, Budapest, 1882. Also a valuable article by Count Geza Kuun on Les Origines Song raises in Revue Internationale, Mai, 1884, pp. 465-495). ' Cf . on the syllabic pronunciation, Henry Sweet, A Handbook of Phonetics, pp. 87, ««?.— Prof. A. H. Sayce, Introduction to the Science of Language, vol. i. pp. 286, seq. * Otherwise " tone." — On Tone and Accentuation, vide my book, Du Langage, lEssaisur la nature et Velude des Mots et des Langues (8vo. Paris, 1867), §§ 22-30. BEGIKNINGS OF WRITING AROUND TIBET. 39 dialects could be read according to their idioms and preferences. The part played by these ideograms became larger, and, from a small starting-point which belonged to the ancient period of the writing, it gradually assumed a prominent rank. The parent-writing of Assyria presented the same phenomenon of small beginnings and eventually large use of ideograms. 55. The Chinese were led by the evolution of their language and writing to the threshold of alphabetism, but their surrounding circumstances did not permit them to perceive it.^ When in phonetic composition groups, the last phonetic element stood no more, but for the final consonant of a word which was deprived by decay of the sound or sounds formerly following this consonant, the alphabetic isolation was reached.^ Besides that, the harmonization of vowels, which assimilated the vocal sound of an added word expressed by an ideogram, rendered indifferent the original vowel of this ideogram. But the facility of grasping the sense carried by ideograms, notwithstanding the regional pronunciation and dialectal words,- was too great a political advantage to be neglected. It had imposed itself quite in a natural way, by mere exigency of the surrounding circum- stances, and it required only an ofacial systematization to be the law of the land. This systematization, as we have said elsewhere, was carried on in 827 B.C. during a temporary revival of the Tchou dynasty's power. The breach between the spoken and the written language of China was an accomplished fact. The possibility of an alphabet had been lost, and a cumbrous system of writing— which expresses the succession of ideas and is not a language, which has proved 1 For some examples, vid. J.S.A.S. Vol. XIV. p. 799, n.s. 2 The curiomi Chinese system of indicating the pronunciation of a word by the initial of one word and the final of another, called ih,fcm-ts>eh, was appaxeiitly 40 BEGINNINGS OF WHITING AROUND TIBET. most -noxious to philologists because of tlie wrong views they took from it on the spoken language — was obtained. But the phonetic advantage was not lost everywhere. III. Mo-So HiEROGLYPHICAL WRITING OF TiBETO-ChINA. 1. History. 56. The Mo-so g ^ i or Na-shi form the bulk of the popu- lation betwixt the Lu-tze Kiang, the Lan-tsang Kiang and the Kin-sha Kiang between the 27th and 30th parallels,^ in the prefecture of Li-Kiang ^ (N".W. Yunnan). The name of Mo-so,* by which they are known among the Chinese, was at first objectionable to them, and it is only in late years that they have accepted and used it.* We suppose that they objected only to the Chinese transcription of the appellative as written with two characters meaning respectively " small" and "little." Yet this was nothing else than a form of an appellative apparently of native or Tibetan make ; Mo for Mu is the name of their once leading family, and so for sa ^ is perhaps the Tibetan word sa for place or country. On the other hand, let us not forget that zo in their own language means "son," like the Chinese tze, which is probably its antecedent. Muso means "hunter" in Shan and in Li-so, a sister-language of the Mo-so, and may be derived from the very name of that people, so that it does not preclude any other origin for the appellative, as we shall see below. Their full name Mo-so Man in Chinese records is nothing else than " southern Barbarians Mo-so." 57. They are mentioned several times in history. First about 796 a.d., in the Annals of the T'ang Dynasty, I-mou- ^ The character & must be read here as ^ so, and not sie, its otherwise common pronunciation. The K^anghi tze tien giyes this equivalent and quotes the names of Lo-so (Lhassa), Mo-so and Tu-kuang-so as examples of this special pronunciation. Vid. ii. 5, f. 30«. ^ Vid. F. Gamier, Voyage d'Exphration en Indo-Chine, vol. i. p. 520. ^ Their own native name is Na-shi. * " Le nom de Mosso, tout injurieux qu'il etait dit-on dans le principe a fini par dominer tellement que les Mosso actuels I'acceptent sans repugnance et s'appellent eux-memes Mosso." * The Wan-y tehi (quoted in the Tai Ping yii Ian cyclopedia (a.d. 977), Kiv. 789, f. 6) locates them above and below the Iron bridge (in the N.W. of Li-Kiang) . ' Sa in Tibetan means : earth, country, place, spot, ground, etc. Vid. H. A. Jaeschke, Tibetan-English Dictionary, p. 569J (London, 1881, 8vo.). BEGINNINGS OF "WRITING AROUND TIBET. 41 siin, the King of the Nan-tchao, whose title had just heen recognized by the Emperor of China, led a campaign against the Tibetans, and afterwards conquered several small States of his neighbourhood, viz. those of the Mo-so Man (our Mo-so),^ Long-tong Man, and Mo-tchang Man? They recovered their independence when the kingdom of the Nan-tchao collapsed at the death of Shun-hua (899 a.d.), the last prince of the dynasty established by Mung-she, which had ruled over Yunnan for eight centuries.^ The state of Nan-tchao (since 860 a.d. called Ta-li), which had begun in 629, ceased to exist, and his successors were not able to recover the power of the former kingdom. In 937 the Tuan dynasty began to rule, and gave to its dominion the former name of Kingdom of Ta-li, which, however, was changed several times, and was not definitively used before its recognition by China in 1115 a.d. It lasted till its sub- mission by the Mongols in 1354^1357, who put an end to the Tuan hereditary power. 58. At the outset of his campaign against Yunnan, Kublai Khan in 1253 first met before him in the north-west of Yunnan the very State of the Mo-&o Man, which he subdued without delay. The power of the Yuen or Mongol dynasty of China was never firmly established in Western Yunnan ; the submission of the native small States was nominal only ; and the Mongol princes who ruled from father to son over Yunnan, and who lasted a long while after their brethren had been expelled from China, had their authority restricted to the centre and east of the province till their destruction by the Ming dynasty of China, and the suicide of the last of them in 1381 a.d. The Imperial troops were unable to make a regular conquest of the region occupied by the Mo-so, but the supreme suzerainty of the Chinese 1 The passage; -without quotation as usual, is reproduced by Ma Tuanliu. Yid. Ethmogrwphie des peuples etrangcrs a la Chine, vol. ii. p. 297 (transl. d'Hervey St.-Denys), where the name is inadvertently transpribed Mo-sie-man. ' The Mo here i-s written with the same character as that of Mo-so, and suggests apparently a relationship between the two populations. * Their name is not quoted, nor any other, in the Historical Documents con- cerning Yunnan, where it is stated only that the country fell into anarchy after the death of Shun-hua, and that several mbes threw over the yoke. Of. E. Rooher, La province Chinoiae du Yunnan (Paris, 1879), vol. i. p. 166. 42 BEGINNINGS OF WRITING AROUND TIBET. Emperor was recognized by them, and this shadow of authority was sufficient to guard the susceptibility of the central power (1384 a.d.). The paternal administration of Wu San-kuei, and his final revolt at the beginning of the present dynasty, had loosened the reins to such an extent that in the end the Chinese Government was compelled to conquer the region, which was finally subdued in 1775 a.d. Ba-t'ang Li-tang and Tchung-tien,^ the latter in the south near Li-Kiang, indicate the extension of the newly incor- porated country, including the Mo-so dominion, a part of which was then again under Tibetan rule. 2. Description. 59. From inquiries and a long habitat on the spot, the renowned Pere A. Desgodins was able to collect some valuable facts and information on this interesting and once powerful tribe. Formerly they had a flourishing empire covering a large area. It extended northwards to Dzo-gong on the Ngu-kiu or Ou-kio river, to the salt-mines near and above Terkalo on the Lan-tsang Kiang, and to Dzong-ngu in the Kin-sha Kiang basin.^ The above missionary has often seen, in his travels on the borders of the Lan-tsang Kiang and on those of the Lu-tze Kiang, numerous ruins of houses and forts, remains of Mo-so military posts and colonies, built by them after their conquest of the salt-mines' region from the Tibetans three or four hundred years ago.' The capital was Li-Kiang, which the Tibetans and the native chiefs call Sadam,* and the name of their king was Mu Tien Wang {i.e. the Celestial King Mu), whose family ^ is still in existence and numerous. 1 On Mr. E. 0. Baljer's map, Bjedam is written under the name of Tchung- tien, half-way between Li-Kiang and Bat'ang (Vid. Travels and Researches, p. 93) ; but as he does not give lus authority, we do not know which of the two conflicting statements is genuine. 2 Desgodins, Za Mission du Tibet, p. 258. 3 P6re A. Desgodins, Notes Ethnographiques sur le Thibet, in Annates de V Extreme Orient, JmUet, 1879, pp. 10-12. 4 Under the Han dynasty it was called ^ ^ Ting-tao {Of. Playfair, The Cities and Towns of China, n. 4147). Sadam is perhaps the same two syllables inverted with a dialectal pronunciation. The case would not be isolated in that part of China, as we know several other instances of the same phenomenon. Cf. Teng-yueh, which was also Tueh-tan. 5 From p. 21 we are informed that the members of the tribe in the prefecture BEGINNINGS OF WRITING AEOUND TIBET. 43 60. Though destroyed as an independent power, and with- out any chance of ever recovering it by themselves, the Mo-so are still preponderant in the country of their former greatness. Nearly all the indigenous chiefs are Mo-so, hold- ing appointments for life from the Chinese Government under the rule of Chinese mandarins. The natives occupy only the secondary posts, yet, as their authority has a family character and is hereditary, they have a good deal of influence on the people. It is specially so in the territories of Aten-tze and Wei-si, going southwards from Yerkalo on the borders of the Lan-tsang, that the greater number of the oflB.cers are Mo-so. 61. They are deeply despised by the Tibetans, who call them Djiong} P. Desgodins says that to call somebody Guiong-god, i.e. Mo-so-head, is an insult in frequent use, which he explains by the low character of the Mo-so communities ^ in Tibet, probably of mixed blood. A great traveller and charming writer, Mr. E. Colborne Baber, has seen at Tatsienlu, a popular Tibetan poem, rather epic than lyric, in the world-wide metre of " Twinkle, twinkle, little star," recounting the invasion of part of Tibet by the Mo-so. " The epic is styled Djiung Ling (Mo-so Division), and is only one of three parts of a very extensive work known as the Djriung Yi, or " Story-book." Another part, called Hor Ling (Hor Division), recounts the con- quest of the Hor (Turk tribes) by the Tibetans, and conveys much historical information in a tale of magic and marvel. The third part, Bjia Ling (Chinese Division), of Li-Kiang are all surnamed So ^[J , but intermarriages are not on this account forbidden (as would be the case according to Chinese customs), so that there are at least two great-subtribes among the Mo-so, the Mu and the So. 1 Guiong = Djiung . The latter orthography is that of Mr. Colborne Baber, Travels and Researches in Western China, p. 88«. In H. A. Jaeschte's Tibetan Dictionary, p. 75, 1 find a word yyony-, meaning, "rough, rude, impolite," but I do not find anything like god, with the meaning of ' head,' in the same work. The latter word is apparently a dialectal form corresponding to the ancient Chinese GET, modem Met, hieh. * ' ' Aux vices des Tibetains dont j'ai parle aUleurs, dit le courageux missionaire, il faut ajouter un esprit chicanier, querelleux, ladre et aimant les proc&s ; ajoutez a cela I'ivrognerie et vous aurez une idee de ces etres degrades. Quant aux traits physiques Us sont bien alteres et ne representent plus le vrai tj'pe Mosso, cepeudant en pent le reconnaitre S. certains caracteres ; front plus f uyant, nez plus aqniUn, les deux os maxiUaires inferieurs moins ecartes, menton plus fuyant que chez le Thibetain . . . ." Notes JEthnographiques sur le Thibet, I.e. 44 BEGINNINGS OF WEITING AEOtJND TIBET. narrates a contest of unknown date between the Tibetans and tbe Chinese." i 62. A Chinese notice gives some details of their customs and habits which are worth reproducing here, and concern mainly those who inhabit the prefecture of Li-Kiang : "Their houses are huts built of boards, with walled doors. As to their clothing, the men pierce their ears, wearing pendants made of a green stone. Their hair is arranged in a twist under a black cap. They wear coats with long collars and wide sleeves, fastened either with a red girdle or a green flowered sash. The women wear short jackets, pointed caps, and cylindrical skirts, finely plaited and fastened with a gay-coloured embroidered girdle. Over all they wear a sheepskin, cloak. On the death of a parent, neither coffin nor shell is used, but the body is burned and the bones scattered in a deserted place. A half-burnt log is brought home, and to it sacrifices are offered. The prevailing religion is Budd- hism, and Lamas are held in great respect. But they have also other ceremonies. On New Tear's Day the members of each family bum incense and take a ceremonial bath. Then with incense in their hands, and carrying rice on their backs, they all repair to the buUding containing their family altar. The priestess of their rites is respectfully entreated to offer prayers and sacrifices on their behalf. These ceremonies, which last for eleven days, are called ' Days of sacrifice,' and are intended to ensure a happy year. Again, in the 6th and 11th moons, the priestess, at their instance, plants a branch of chestnut, as ' branch for the gods to roost on,' and offers ancestral sacrifices. Their land being too cold to grow rice, they live on barley, darnel, and various other grains." '^ 63. With the preceding remarks it is useful to compare the following description, from the Travels of a Pioneer of Com- merce,^ more recent in date, and showing the gradual sinicisa- tion of the tribe. "The Mosos . . . are quite Chinese in appearance, the men wearing the common blue cotton jacket and short wide trousers of ^ Tid. E. Colbome Baber, Travels and Researches in Western China, p. 88. 2 G. M. H. Playlair, I'he Miao-tzu of Kweichau and Yunnan, ii. 21. A. 108, 63, gives only a part of the same text. Vid. China before the Chinese, s.v. ^ Tra/eels of a Pioneer of Commerce in Pigtails wnd Petticoats ; or. An Overland Journey from China towards India, by T. T. Cooper, London, 1871, 8vo. pp 312-315. BEGINNINGS OF WRITING AEOUND TIBET. 45 China, shaving their heads and growing the pig-tail. The custom of the women is fantastic, but graceful. It consists of a very becoming little cap of red and black cloth, with pendant tassel, jauntily worn on the top of the head, inclining a little to one side ; a short loose jacket, vnth long wide sleeves, over a tight-fitting cotton bodice, covering the breasts ; with a kilt-like petticoat of home-made cotton stufi, reaching from the waist to the knee, and gathered in longitudinal plaits. Instead of stockings, their finely- shaped limbs are swathed from the ankle to the knee with white or blue cotton cloth, while leather shoes, turned up in a sharp point at the toe, complete the chaussure of the Moso ladies, who, though not quite so fair as the Chinese, are generally well-proportioned and good-looking, and unembarrassed by the shy reserve of the fair Celestials. As ornament they wear huge silver earrings (resembling in shape the handle of a common key), silver rings and bracelets, and bead necklaces."^ At Ya-tse (the J^-tche of P. Desgodins ^) the women often substitute for the little cap " a red cloth hood thickly braided with cowrie shells." The Moso chief of Ya-tse is the most powerful ruler of all the tribes along the Lan-tsang Kiang. He rules over the Moso and the Liso tribes, and collects the taxes in the name of the Chinese government represented by the Mandarin in residence at Wei-si. 3. Writing. 64. As to their language and writing, what we know on the matter comes, again, mainly from the indefatigable missionary, Pere Desgodins. He was able, in 1867, to make a copy of several pages from a manuscript written in hiero- glyphics, and belonging to a Tom-ba or Tong-ba, a medicine man among the Mo-sos. These pages, eleven in number, were sent to his family, and much later on, in 1879, I received eight of them through the kindness of the learned 1 T. T. Cooper, ibid, p. 314. This traveller has mistaken the names of Ya-tse and Mooquor (Mukua) as appeUatives of tribes instead of localities. ' Mots principaux de certaines triius qui hahitent les bords du Lan-tsang Kiang, du Zou-tze- Kiang et Irrawaddy, par I'Abb^ Desgodins, missionnaire an Thibet (Yerkalo, 26 Mai, 1872) ; Bulletin de la Society de Q4ographie de Faris, vi. serie, t. iv. 46 BEGINNINGS OF WEITING ABOUND TIBET. Mr. Grerard de Rialle, and I kept these sheets with the in- tention of publishing them, as soon as more information should come forward. 65. In the mean time a genuine manuscript of the same writing had reached the British Museum. The lamented Capt. "W. Gill, during his famous journey with his faithful companion, Mr. Mesny, had been able to secure three of these MSS. while staying at Ku-deu, on the Tibeto-Chinese frontier east of Li-t'ang.^ Two of them were sent to Jersey, the seat of Mr. Mesny's family ; and an enquiry from me there remained unanswered.^ The other MS. was presented by Capt. Gill himself to the British Museum.' The donor was not aware of their exact nationality ; he had obtained them from two men of rather European and specially French appearance, who offered them for sale. Apparently these men were Lolos, who had obtained possession of the MSS. by plunder of Mo-so people, as the latter would never have parted willingly with books so scarce and valuable to them. 66. However, the MS. in the Oriental Department was described only as a " Hieroglyphic Book of Prayers from the Mountains between Burma and China." It had remained there for a while when, attracted by the above title, I examined it and recognised the Mo-so writing, with the general features of which I was familiar since the copies from Pere Desgodins had reached me. Col. Yule was informed at once of this discovery, and inserted in his remarkable Geographical Introduction (then in type) to Capt. Gill's The River of Golden Sand,"^ a few notes embodying the result of my first examination. 67. In an amiable letter from Darjiling, April 25th, 1882, Pere Desgodins had the extreme kindness to send me some more details about this writing in answer to several ques- 1 The fact is not recorded in the published record of his journey, but I have it from the traveller himself. ^ Capt. Gill was not more successful than myself in a similar attempt. 3 Add. MSS. Or. 2162. Published in fac-simile on Plates I. II. herewith. * The Biver of Golden Sand ; being the Narrative of a Journey through China and Eastern Tibet to Burma ; London, 1880, 2 vols. 8vo. Introd. pp. 90-2, EEaiNNINGS OF "WEITING AROUND TIBET. 47 tions of mine. They do not match exactly with what we expected from the former rather vague statements, viz. that this writing, now obsolete, was formerly current among the people. However, they are extremely important for the general theory of writing, inasmuch as they do not pretend to show in that peculiar hieroglyphical writing any survival of former times. According to these views, should they prove correct, it was apparently made up for the purpose by the tombas or medicine-men. This would explain, perhaps, the anomalous mixture of imperfect and bad imitations of ancient seal characters of China, pictorial figures of animals and men, bodies and their parts, with several Tibetan and Indian characters and Buddhist emblems. The superfetation and addition of the Chinese, Tibetan, Indian, and Buddhist signs are obvious, while the pictorial ground of the writing with peculiarities of its own is no less visible. The tails of animals, caps of men, etc., are modified according to the sentence ; on the other hand, these occasional additions are also used independently. This feature deserves more atten- tion than would be supposed ; should we get a phonetic rendering and a translation of these texts, those appendices might turn to be phonetic complements. 68. Now let us return to the simple facts and statements as they were communicated to us by this energetic and intelligent missionary. "These hieroglyphics," according to his letter, which we translate verbatim, " are not, properly speaking, a writing, still less the current writing of the tribe. The sorcerers or Tong-bas alone use it when invited by the people to recite these so-called prayers, accompanied with ceremonies and sacrifices, and also to put some spells on somebody, a speciality of their own. They alone know how to read them and understand their meaning ; they alone are acquainted with the value of these signs, combined with the numbers of the dice and other implements of divination which they use in their witchcraft. Therefore these hiero- glyphics are nothing else than signs more or less symbolical and arbitrary, known to a small number of initiated, who transmit their knowledge to their eldest son and successor 48 BEGINNINGS OF 'WRITING AROUND TIBET. in their profession of sorcerers. Such is the exact value of the Mo-so manuscripts ; they are not a current and common writing ; they are hardly a sacred writing in the limits indicated above." ^ 69. Yet we cannot help thinking that this sacred writing embodies survivals of the pictorial stage of notation indepen- dent of synchronieal dates and progresses elsewhere, which seems (within their limited area of self-progress) to be proper to all races of mankind, the white race with exceptions. The latter was more often satisfied with systems of notation more symbolical and conventional; simple combinations of dots and strokes, straight, curved or spiral lines, round and deep as cup-marks or angular and square, were sufficient for them, while the inferior races have' always wanted, and have made a more material and eye-speaking system of nota- tion. We may be sure that pictorial writings have crept up everywhere, though very few have survived in the struggle of civilizations ; the long period required for their passing through the pictographic and ideographic phases was not allowed to them, and the untimely intrusion of an older and more perfected system, or another one better fitted to the surrounding circumstances, superseded them altogether. 70. The only possible life for still-born writings of that description is that which lingers in obscure corners of super- stition and witchcraft. "We think that this Mo-so writing may be an instance of the fact, excepting the modifications in- troduced in the mean time for the purpose of those who use it. And we can support this view by the reproduction, in the learned work of Emil von Schlagintweit, On Buddhism in Thibet, of charms found by his brother Robert v. S. during his journeys there. On some of these charms are drawn hieroglyphical signs, which are not without analogy with those of the MSS. drawn by the Mo-so sorcerers. Now, the Schlagintweits 1 Letter from PSre Desgodins to the author, dated from Dariiling, 21 Avril, 1882. In forwarding this letter, the amiable Mr. Desgodms, of Nancy, brother of the Missionary, had the kindness to send me two pages and a half more of MS., completing the whole of the copy made by P^re Desgodins from the Tong-ba's MS. Published in fac-simile on Plate III. herewith. BEGINNINGS OF "WRITING AROUND TIBET. 49 did reach Tibet, but only the -western part of the country, and the finding there of specimens not unconnected with that writing indicates for it, or at least for the rougher ground script from which it has been evolved, a much larger area than could otherwise be supposed. As the Mo-so have not taught the western Tibetans any more than they did the eastern, from whom, on the contrary, they have learned so much, we must understand the ground of this writing to be of Tibetan origin of unknown date. The inference is plain and cannot be impugned. It requires the attention of future explorers of Tibet, when this forbidden land is open to scientific researches. 71. Up to Pere Desgodins' discovery nobody had ever heard of any writing among the Mo-sos. The Chinese documents do not mention this accomplishment of theirs. And the late Mr. Cooper, to whom we are indebted for not a few details on the same people, positively states that they have no writing of their own.^ If the existence of this acquirement has remained unheard of so long, through the secrecy kept by the very few men acquainted with it, we must expect the same difficulty in Tibet. The notice of the latter country in the annals of the T'ang Dynasty,^ states that in the sixth century the Tibetans had no written characters, and used notched sticks and knotted strings in their covenants. We know how very little is proved by negative testimonies of this kind, in a case like that which we put forward. The above-quoted instances are cases to the point. And we must leave the matter as a moot question to be elucidated by further researches and new materials for study. 72. It is not uninteresting to remark here that a kind of embryo picture-writing, understood by none but the mee- tway or Toomsah, i.e. priest, has been pointed out among the Kakhyens of Upper Burma. The description is interesting to quote. ' Travels of a Fioneer of Commerce, vid. above, § 8. ^ Vid. the able translation of this notice with annotations by Dr. S. W. Bushell, The Early Sistory of Thibet, p. 440 in J.S.A.S. 1880, Vol. XII. ; and also Dr. K. GanzenmuUer, Tibet, Stuttgart, 1878, p. 103. 50 BEGINNINGS OF "WRITING AROUND TIBET. "We extract the following from a paper on these people, published in a periodical ^ in 1882 : " A formal avenue always exists as the entrance to a Kaehyen village .... On each side of the hroad grassy pathway are a number of hamhoo posts, four feet high or thereabouts, and every ten paces or so, taller ones, with strings stretching across the path, supporting small stars of split rattan and other emblems. There are also certain hieroglyphics which may constitute a kind of embryo picture-writing, but are understood by none but the meetway or priest." At the following page (p. 471), we hear of the meetway or toomsah. In the latter we have, perhaps, a cognate appellative of the Tom-ba of the Mo-so; ba being probably the Tibetan suffix. There is no apparent connection between the system of hanging up symbols and hieroglyphics and that of the Mo-so writing, excepting that in the latter there are signs and symbols which might be compared to those of the Kakhyens. From the proximity of the two peoples, and the higher standard of the Mo-so writing, it might be supposed that the Kakhyen toomsahs are the pupils of the Mo-so tombas. 73. Pere Desgodins has inquired if, previously to their absorption by China, the Mo-so had a writing of their own, besides these hieroglyphics ; but he was unable to obtain any information on the matter. In these days the Mo-so of the south, i.e. those who reside about Li-Kiang on the borders of the Kin-sha Kiang, and about Wei-si and Aten-tze on the Lan-tsang Kiang, being Chinese subjects, use exclusively the Chinese characters. Those of the north, between 29'' and 30° N. lat., in the region conquered by them from the Tibetans, use Tibetan characters. Notwithstanding this difference of writing and influences, their language has been preserved, and they use it between themselves with a mixture of Chinese or of Tibetan according to the region which obliges them to speak also either Chinese or Tibetan. 1 Cornhill Magazine (Oct. 1882, pp. 466-476) by Shway Toe (Mr. J. 6. Scott, once resident in Burma, and the author of the best book ever written on its subject, The Burmmi, London, 1882, 2 vols. 8vo.). BEGINNINGS OP "WRITING AROUND TIBET. 51 4. Linguistic. 74. We have no grammar nor current text of their language. A short vocabulary of some 200 words has been published by Pere Desgodins from notes taken by his col- leagues, PP. Q. Biet, F. Biet, and J. Dubernard,^ as follows : 1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 19 20 100 1000 moon star day month earth world wood djre lu stone lou gnilu iron chou se lu siher ngou lolu gold ha ■ ngoa lu copper heu {red), eu tohoa lu {yellow) che lu air heu ho lu body goumo ngo lu head koulu tse lu eyes men tse djre lu ears h^ tze tse gni lu nose gni ma tse se lu mouth krouhe tse lo lu hand la tsi ngoa lu Idly deu men tse tchoa lu foot kheu tse eta lu eat dze tse ho lu drinle tchre tse ngo lu food ha gni tse lu rice tchoa djre chi tea le tong tchra flesh chi mou butter marpeur gni me salt tse he me tze tobacco yo kheu father aba-aou gni mother am^ he brother heze khou sister mehe mou deu son zo dzom hou ling daughter mi guid master daha s^ chief officer su-mouquoi-aqua 1 Mots principaux de certains tribus .... loo. cit. 52 BEGINNINGS OF WRITING AROUND TIBET. servant guieu ZQ plain pa tze lo house gui^ da river i bi guii door kto field mou deu window kho-ka-go tclira grass field ko khou roof kia kou rock ha hitehen range koua cloth bala stable tso bou girdle bouke animal goghe boots za horse joa knife jete ox Isghe sword dapia-dapre cow ghe me soul oua he dog khe to love obi to ba cat ha le to think choun drou lird a to le mou fowl am^ to have guiou sheep iu to will djra-djro da goat tsi to do fou red tu lu to speak chado white pe sa to rejoice ba Hue lie le pain chou djrou guiem green guiong ko ma ba black na me rich 'h.h la gni mountain guieu kh.ou poor ma ha 75. The words must be read as French. Though several errors have crept in, through the MS. copies and the print- ing, the aflB.nities of the language are obvious. The common words with the Li-so are more than fifty per cent., and the proportion of Tibetan and Burmese is considerable. The gram- mar seems to be in accordance with the dictionary, and shows, like that of the Li-so dialect, Burmese features, so that the classification of the Mo-so language must be easy. It must be put down as a member of the specially Burmese division of the great Tibeto-Burmese family. As materials for the study of the language, in addition to the above vocabulary, we have one word and a single phrase which have been collected by the late Francis Gamier. ^ The word is hantse "manger," for which "meaning we find in P^re Desgodins' vocabulary dzi. The phrase is Khipa khe ' Voyage ff exploration en Indo-Chine, vol. i. p. 520». BEGINNINGS OF WBITING AROUND TIBET. 53 tche ma seu, lit. " Chinese I do not know the language," or " I do not know the Chinese language." 76. With several other languages of the same region and some of southern Yunnan and Indo-China, they form a group or subdivision weU delimitated (to which, however, several more will have to be added in the sequel) in the great Tibeto-Burmese family of languages, with a special connection to the Burmese and some Tai-Shan ingredients, as follows : Living Languages. Brandies, Dialects oftheLaka or Lokuei(Szetchueii-Tuniian) \ ) -nt +>, „ „ Li-so or Leisu (N. W. Yunnan) ... ! ] J^»"l- „ „ Mo-so or Na-sM (N. W. Yunnan) ) ) ^™- „ ,, Mu-tse(MuongLini,N.Indo-China) i ") „ „ Kouy (Siemlap, N. Indo-CMna)... J I }■ Laka. „ Ka-to (S. Yunnan) 1 v. South- „ „ Ho-nhi (S. Yunnan) \ ^™- ,, ,, Ka-kho (Paleo, N. Indo-China)... , 5. Ethnology. 77. Ethnologically the position of the Mo-so cannot be ascertained otherwise than in its main lines. They belong, in their underlying original type, to the same group of popula- tions that the Nung=^Njung^=Jung tribes which appeared on the west borders of China as early as the sixteenth century before our era. The name of Djiung, by which they are known to the Tibetans, belongs undoubtedly to them as a survival of the general name quoted above, which is also represented in that of the A-nung or Lu-tze, their cognates in Eastern Tibet. In the other name they give to them- selves, Na-shi,'^ it is not at all unlikely that we have a modern appearance of an old name of the same group, which the Chinese once in former times rendered by the punning transcription of Niu-tze ^ •^, alluding to some gynecocratic customs of their own. We know so little of the Mo-so habits and traditional institutions that we are still unable to cor- roborate this probability by any survival of the kind other than the priestess mentioned above (§ 62). ' Is this HO the same as the Tibetan nag, ' woman,' the old Chinese nok ? 54 BEGINNINGS OF WRITING AROUND TIBET. 78. Unless we venture to see another survival underlying the very name of Mo-so given to them by the Chinese, the vocabulary from Pere Desgodins does not contain the word for ' woman.' Should it be mo, as in Tibetan, or mu, ' mother,' as in Chinese, combined with zo, which is their very word for ' son ' (Chinese fee), we could translate Mo-zo as ' woman son,' corresponding exactly to the old Niu-tze or (?) Na-shi, and now objectionable to them because they have dropped a long while ago the gynecocratic institutions which once justified it for foreigners. "We give this suggestion for what it is worth, and shall not dwell more upon it, as ample confirmation is found of the existence of such institutions in the history and customs of other cognate tribes, such as the Lakas or Lolos. 79. The connection of the Mo-so is narrower with the Li-so than with any other tribe. Their parentage is openly admitted in a legend of the country which says that the Li-so and Mo-so, of Burmese origin, were driven away on elephants from the latter's country.^ The genuineness of the story requires confirmation, as the same legend is told by the Pwons near the third defile (Kyandwen) of the Irawady.^ Does this indicate a parentage between them ? It is not at all unlikely ; the Pwons do not speak a Tai language, and they are connected with the Kadus, who belong to the same stock as the Mo-so and Li-so. We must see probably in this curious tradition embodying the souvenir of a small historical fact, the popular and widespread expression of a conscious knowledge of original parentage with the main stock of the Burmese, upheld and maintained because of the high estate of the latter. The historical fact is worth considering, since it might help to the elucidation of a point of Burmese history. The current tradition says that, about the commencement of the religious era, or partly during Gautama's lifetime, the town of Tagaung (built on the left side of the Irawady, 300 years before the lirth of Gautama, by Abhiraja, who had 1 Mentioned by P6re Desgodins in his paper, Mots prineipaux, etc., he. cit. 2 Vid. Ney Elias, Sketch of the History of the Sham, Calcutta, 1876, 8to. p. 12. — And also below, § 100, Addenda. BEGINNINGS OF WRITING AROUND TIBET. 55 migrated from Kapilavastu) ^ was taken by an invasion of Taruk or Taret tribes coming from a country to the east, called G-andalarit, in the land of Tsin or Sin, which cor- responds with Yunnan. 80. 'Now, the only tidings we have, from the Chinese side, connected with an event of that kind, goes back to the middle of the fourth century B.C. After his accession to the throne, in 338 B.C., the King Wei of the Teru or Tsu State ^ despatched an army to the south-west, and his general, Tchwang Kiao, subdued part of Yunnan, and Tchwang Hao his son (?) became King of Token or Tsen ^ (now written Tien \%). This name, which, by the way, is extremely interesting, as it became that by which China was known to the southern traders from the west, was selected because of the large central water of the region, now the lake of Yunnan fu ; the word Tchen^ in the language of the region meaning "water." Now we have here the Tsin of the Burmese legend without doubt, and in that of the mother-country of the prince- family, the State of Teru (mod. Tsu), of which the new State was the real offshoot, we might find the very name which appears in the legend under the double form of Taruk or Taret. 81. To conciliate the statement of Burmese history with the date of the Chinese record, we must admit a rectification of the chronology of the southern Buddhists as proposed by some scholars, and premise the great probability that the establishment of the new State of Tsen in Yunnan had a direct influence on some more meridional regions. Either by a migration southwards of tribes dislodged by the new kingdom, or by, what is more in accordance with the spirit of the legend, the new ruler himself pushing further south his attempts at conquest. In the latter probability he 1 Cf. Ney Elias, loo. cit. ; Major H. R. Spearman, British Surma Gazetteer, vol. i. p. 236 ; Sir Artliur Phayre, History of Burma (London, 1883, 8vo.), p. 8. 2 More is said of this once famous and important state in my introduction on The Cradle of the Shan Kaee, to A. E. Colquhoun's book, Amongst the Hhans (London, 1885, 8vo.). Excerpt, pp. 27-8. 3 We find this word connected with the following which will have the same meaning : Tchung Miao iws ; Singpho ntbin ; Kakhyen intzin ; Munnipuri ISHING, etc. 56 BEGINNINGS OF WRITING AEOITND TIBET. had to utilize some intermediary tribes as elephant-drivers, a fact whicli is part of the tradition kept by the Pwons, and partly still impressed in that of the Mo-so and Li-so. 82. The wave of migration of all these tribes, we might say of the race altogether, is strongly marked from north to south ; the occasional retrogression of a tribe or two under peculiar circmnstances has nothing to do with the general movement which, for ages and ages, let us say forty centu- ries at the least, has been proper to the various populations which can be traced back to the Kuen-lun range as to their cradle within historic times. IV. Alphabet in Tibet. 83. Tibet enjoys now the privilege of being T/ie Forbidden Land for Europeans. It is not long ago that another land could boast of the same position, but the barriers are now removed, and since last year Corea is open to Western influence. It will before long be the same with Tibet. The jealous monks who hold the country do not yet allow, under any conditions, Europeans to overstep its frontiers. And it could be considered as nearly certain still lately that the supreme authority of the Chinese government would not be respected, should it yield to pressure and authorize the entrance of Euro- peans into this abode and refuge of decayed and corrupted Buddhism. But, we may be sure that from one side or the other, the opening of the country must take place at no distant date, apparently through India by mutual interest of trade (Yid. Addenda). Yet many years must elapse before any scientific exploration can be made. And even then a long time will pass before the country shall have revealed its secrets. Indeed, we cannot foresee what archaeological researches, regularly made, may disclose in monuments, rock inscriptions, etc. So far as our present subject is concerned, we have already from Chinese, Tibetan, and other sources, some valuable information, which, in my opinion, will prove interesting when put together. The country by itself is not BEGINNINGS OF 'WRITING AROUND TIBET. 57 ricli. It is thinly inhabited. It offers by itself little incite- ment to civilization, which has to come from the outside. Hence the information bears more on the outskirts than on the interior of the country. Its central position in Asia has made it witness, all around its borders, to several evolutions, derivations, progresses and decays of importance for the history of writing in the East as seen throughout this memoir. 84. The hitherto scattered and independent groups of tribes occupying the country now known as Tibet proper were submitted and organized into a regular government (towards 434 A.D.) by Tupot Fanni,^ who gave his family surname to his new country.^ This surname meant "Prince of the Land." ^ He was a scion of the Tartar dynasty known as Southern Liang, which ruled in Kansuh on the N.W. of China, with its capital at Liang-tchou, from 397 to 415 a.d.* The Tupot are more generally known among scholars from the modern pronunciation of the Chinese transcriptions of their name : Tu-fah, Tu-poh, To-poh, all interchangeable, and from the form To-pa adopted by De Guignes in his ■ wonderful work Histoire des Huns.^ 85. We have seen above that Tu-pot Fanni was a scion of the Southern Liang dynasty ; he belonged to the same stock as the famous Wei dynasty which ruled over the north of China from 386 to 535 a.d. The family name of this dynasty was Tu-pat, like that of the Southern Liang ; there is only a slight difference in the Chinese transcriptions f£ ^ and ^ ^, both pointing to the same sounds and not much diversified since then. 86. The To-pat were a division of the Sien-pi^ race, according ^ Cf. a valuable paper by Dr. S. "W. BusheU of Peking, T/ie Early Sistory of Tibet, from Chinese Sources, in J.R.A.S. 1880, Vol. XII. pp. 435-541. Vid. p. 440. Since, a very promising scholar, Mr. W. AVoodviUe EockhiU, has compiled from Tibetan sources The Early History of Bod-yul (^Tibet), pp. 203-229 of his valuable work The Life of the Buddha (London, 1884). * I have treated at length this question in a special paper : Ttbet : why so called ? * Annals of the Wei dynasty or Weishu in Tai-ping yO-lan, bk. 800, f. Iv., and bk. 101, f. 1. * Cf. Li-tai Ti-vimg-nien-piao, E. Tsin, ffi. 11-13. ' Vid. Histoire des Huns, vol. i. part i. pp. 197-198. * The Coreans are still called Sien-pi by the Japanese. 58 BEGINNINGS OF WEITING AROUND TIBET. to the Chinese annalists, and so far were connected with the Coreans and the Mandchus Tungus. We know several branches of them, almost all great and powerful. Besides the ruling families of the mighty Wei, the petty southern Liang dynasties of China, and the dynasty of Tibet, they formed one of the six chief tribes of the Tang-hiang, the To-po #•? ^> the descendants of which tribe established the kingdom of Si-Hia, ruling over Tangut and the north-west of China from 982 to 1227 a.d.i A subdivision of the To-pat Jg^ has formed the nucleus of the J§ ^ Ju-Ju, or Ju-Juan^ a separate tribe of the Hiung- nu according to the Chinese annals. They dwelt originally in the country of Kalka, on the frontier of Siberia, and con- stituted a regular and dynastic power with 18 rulers from 402 to 554 A.D.^ Destroyed by the Turks, they fled westwards and passed the Volga in 555 a.d. ; they were the same as the Ouar-k' umi, the ancestors of the (language of the) now much mixed Avars of the Caucasus.* ^ Several tribes of the same stock, under the name of Tu-p6t ^ '^j are mentioned at thirteen days' journey north of the Uighurs, and three of them had their dwellings south of the Baikal Lake. 87. The Bods or early occupiers of Tibet were still in a state of barbarism when some occasional refugees from India, and especially Nepal, reached them. And it is doubtful if they had arrived at any higher standard when they 1 Cf. Tang shu, in T.P.y.L., Bk. 795 f. 3 v. ; where it is -written ^g |^, a variant wMcli leaves no doubt as to the equivalence of all these forms. * Also read Juan-Juan and Jan-Jan. 3 Nan she fg ^ (420-589 a.d.), monography of the Y meh ^ ^ %■ in E'anghi tze tien, 142 + 14, f. 64.— Vid. above J § 13, 27. * They were the first who are known to have used the title of K'an. Vid. Deguignes, Histoire des Suns, vol. i. (1) p. 188. 5 Their chief relationship was with the Jurtchi, Dzhurtshit, Tshurtshit, Zhudzhi, otherwise Nyudzhi, Neu-chin, Nio-Uhi, and TchorUhog (Uighur orthography), Jurjeh, Jurji (Persian orthography), also Soh-shin, Nuhtchi (in- older Chinese transcriptions), of which the names of the Tchachounhe and Nakhtchusi are perhaps survivals in the Caucasus. I have found myself many afiiaities between the Awar and the Mandchu languages in vocabulary and ideology. The Caucasian affinities of languages of the far East have been pointed out by Klaproth, Latham, Norris, Logan, Hodgson, Charencey, Schiefner, etc., but have not yet been established on a scientific footing. BEGIKNINGS OF "WRITING AROUND TIBET. 59 were conquered at the beginning of the fifth century a.d. by Tupot Fanni. They were in that simple stage of civilization where the absence of needs and no incitement from the outside leave man at liberty to forget and drop any earlier acquire- ments hence useless and not required for the satisfaction of his daily wants.^ From the low stage of culture which the ancient Tibetans occupied, we are not free to infer that such had always been their social state. We do not know from whence they came, nor what was the degree of civilization of the stock from which they had separated. Tribes reduced by surrounding circumstances strictly to the satisfaction of the wants of nature, cannot fail to lose any previous knowledge and arts which are henceforth of no use to them. They must adapt themselves to their new circumstances of life. And if any high acquirement comes abruptly within their reach, they are unable to understand and grasp it. 88. Before the reign of their famous king, Srong htsan sgam-po (629-698 a.d.), the Tibetans had no writing. Notched sticks and knotted cords were their means of com- munication, but we have no information on these processes, nor on their likeness or non-resemblance to similar devices in use among neighbouring nations.^ We have the bare statement of the fact in the Chinese Annals of the T'ang dynasty. But in the Tibetan traditions, with the excep- tion of the following, there is no known reference to these rude means of communication. In E. Schlagintweit's Konige von Tibet, we read that " the five principal sages of the country glorified the (first) king^ in records in gold and turquoises," a statement which may be taken as an allu- sion to the former use of a sort of quippos or wampums. ^ Cf. my remarks in my paper on The Cradle of the Shan Race, in excerpt, pp. 6-9. It was printed before the valuable article of Prof. Max MiiUer, The Savage, appeared in the JS'ineteenth Century, Januaiy, 1885. 2 Vid.TAe Harly History of Tibet, from Chinese Sources, by Dr. S. "W. BusheU, in J.R.2.S. Vol. XII. 1880, pp. 435-641. Cf. p. 440. s This first king was Gnya-Khri btsan-po, who ruled five hundred years before the birth of the King Thotori, according to Csoma's Tibetan Grammar, p. 194. 60 BEGINJflNGS OF 'WEITING AROUND TIBET. We have said above all that cau be said on the subject. Cf. §§ 4, 15.1 89. One of the first occupations of Srong-btsan was, apparently, to obtain a writing for the Tibetan language. Tradition says that Buddhist books had appeared in the country five generations previously ; ^ should it not be altogether spurious, we might premise this appearance to be the incitement, which became ripe under the reign of the above king. He seems to be the first ruler with real power who was enabled to turn his attention to the welfare and advancement of his subjects. Srong-btsan soon after the beginning of his reign sent a mission of seven nobles to India for that purpose, but they were unable to find a route, and so returned without having accomplished their object.^ Such a failure does not prove much in favour of former and regular relations between the two countries. 90. The King, however, was not at all disheartened, and in the third year of his reign (632 a.d.) he sent Tongmi Samb'ota, son of Anu, with sixteen companions, to study carefully the Sanskrit language and thereby obtain access to the sacred literature of the Indian Buddhists. He also instructed them to devise means for the invention of a written language for Tibet by adapting the Sanskrit alphabet to the phonetic peculiarities of the Tibetan dialect. So we are told, but we must cut down a good deal of the device as an afterthought of the compilers of traditions. Such a scheme could not have been thought of as here stated, without a previous knowledge in Tibet of Sanskrit and Buddhist notions, which would lead one to suppose some relations with India, and consequently makes more unintelligible the previous 1 Cf . The Life of the Buddha and the Early History of his Order, derived from Tibetan works, in the Bkah-hgjnir and Bstan-hgyur. Followed by notices on The Marly History of Tibet and Khoten. Translated by W. "WoodviUe EockhiU, Second Secretary U.S. Legation in China. London 1884. Svo. p. 208. 2 A copy of the Za-mag-ihog bhod-pai mdo or Karandavyaha s&tra, an almsbowl [palra), the six essential syllables {Om mani padme hum), a golden tchaitya, and a clay image of the ohintamani, are said to have fallen from Heaven in the royal palace (Woodville Rockhill, The Life of the Buddha, p. 210). 2 Bodhimur, in Sanang Setsen, p. 327, edit. Schmidt. BEGINNINGS OF "WRITING AROUND TIBET. 61 statement about the first mission, and its inability to find a route to go there. ^ 91. Another feature in the report of the second expedition suggests some doubts as to the veracity of the story. The envoy has sixteen companions with him. Now let us re- member that the fabulous tradition about the introduction of Buddhism in China in 217 B.C. mentions also seventeen as the number of the missionaries who were imder the guidance of Li-fang.^ And when Han Ming Ti of China despatched, in a.d. 65, Ts'ai-yn to India for inquiries about Buddha and his religion, the Imperial Messenger was sent with seventeen companions.^ The latter expedition is an historical event, and its record may have suggested the similar number of envoys in remodelling the other traditions. The silence of the original reports concerning that poiut was supplemented, apparently, by a little imagination from the recorders. But should it be the case, it is not calculated to inspire great confidence in the other features of these tradi- tions. One of them, that concerning the mission of Li-fang, is considered by many, as spurious from beginning to end, though we are not inclined to accept so severe a verdict, ' I am indebted to Dr. Ventzel, a fervent pupil of the late Dr. Jaeschte, for the following note : — "A detailed description of the introduction of writing into Tihet is contained in the tenth chapter of the Gyalrabs {rgyal-rais), a history of the kings of Tibet, made use of by Jaeschke for compiling his Dictionary. (Another copy of this work is in the library of the Petersburg Academy, N. 433« in the catalogue in the Bullet. Mst.phil. 1851 ; a third is mentioned in Schlagint- weit's DieKonige von Tibet, p. 19 of the separate edition from the Abhandl. d. Kgl. Bayr. Ak. i. cl. x. iii.) In Jaeschke's copy, the tenth chapter reaches from the end of fol. 29 to the beginning of 34. The substance of it is translated into German (from Bodhimor, the Mongolian Torsion of the work) by Schmidt, in the annotations to his edition of Sanang Setzen, Geschichte des Ost-Mongolen, p. 327 sy. There it is said (fol. 31* 3, Schmidt, p. 328) that Thomni Sambhota formed the square writing I>bu-djan out of the characters of the gods Lanelui, and the cursive writing (here zmdjan ' the angular,' properly the Aa(/'-cursive, which itself then was developed to the more current dbu-min) out of the characters of the Nagas, T'artula. What Indian alphabet may have had this last name is not known. Devalipi and nagalipi occur also side by side among the 64 alpha- bets that Siddharta is instructed in (Lai. 144, 2, of the Calcutta edition)." " Of. Prof. E. K. Douglas, China, London, 1882, p. 318.— Eev. Prof. Samuel Beal, Abstract of Four Lectures on Buddhist Literature in China, London, 1882, pp. 1-2; Buddhism in China, London, 1884, pp. 47-48. The Fosielun, by Fa-lin, where the tradition is reported, was written between a.d. 624-640, according to Bunyiu Naniio, A Catalogue of the Chinese Translation of the Buddhist Tripitaka, Oxford, 1863, col. 331, «. 1500. ' Cf. Mayers's Chinese Reader's Manual, 340, 754. 62 BEGINNINGS OF WEITING AROUND TIBET. because we think that, after all, the legend might have some foundation. 92. As to the Tibetan expedition, there is no apparent reason to doubt it, with the exception of the additions and embellishments which have been added by the historians. Let us remember that we have no contemporary records nor annals of the time, and that all the knowledge we have from the Tibetan history is derived from native compilations, if not of a late date, at least made many centuries after the events they purpose to record. 93. The Tibetan king furnished the members of the mission with a large quantity of gold to make presents to their Indian professors. After having had to overcome great difficulties on their road, they safely reached their destina- tion in Aryavarta {i.e. abode of the Aryas or the whole central region between the Himalaya and Vindya mountains). So says the Baboo Sarat Chandra Das in his Contributions on Tibet, from native sources.^ The Bodhimur or Mongolian version of the rgyal-rabs? a native history of the kings of Tibet, states that the mission was sent to (Enoed Kok^ to learn the writing of the country. This last name seems to me to be a corrupted form of the Chinese appellation for India, Tntu-Kuok or country of Yn-tu f P ^ . The same work, as we have it, through I. J. Schmidt,* says that it was in Southern India, which I understand to be an indication of the position of India with reference to Tibet, and not at all as the southern part of India. 94. Tong-mi Samb'ota made himself acquainted with the Indian characters, or, as Baboo Sarat Chandra Das says, he 1 Baboo Sarat Chandra Das, Contributions on the Religion, History, etc., of Tibet, in the Journal of the Asiatic Society of Bengal, vol. L 1881, pp. 187-251 • vol. H. 1882, pp. 1-75, 87-128, of. p. 219. ''■ The title of this important Tibetan work is Egyal-rabs-Jcyi gsal-bai me-long or "A bright mirror of the history of kings" (of. Jaesehke Tib. Engl. Diet., p. 417). It was compiled by the fifth Gyalwa-Rinpochhe, or Great Lama (Sarat Chandra Das, op. cit. p. 212). ' The name slightly altered was still used in the last century, and figures in D'Anville's map as Anonkek or Anongen. * Translated by Schmidt in the Annotations to his edition of Sanang Setzen, Geschichle der Ost-Mongolen, p. 327 sq. Ssanang Setzen Khung taidshi com- pleted in 1662 his work entitled Mongol Khadwn Toghudji or " A History of the Mongol Khans " (H. H. Howorth, Sistory of the Mongols, i. xvi.). BEGINNINGS OF "WRITING AROUND TIBET. 63 acquired a thorough knowledge of the Sanskrit and of sixty- four different characters known in the Arya land. Surely this is an exaggeration, though the number is reiterated from the sixty-four alphabets that Siddharta is instructed in, as reported in the Lalita Vistara. The envoy and his companions were taught by a Brahman called Li-hyin (i.e. Lipikara or Livikara, which means merely " a scribe ") all the intricacies of the language and writing, while the pundit Devavid Singha, or Singha-ghosha, called Tcengnhn-arkhagan or Arsalan in the Bodhimur, instructed them in the Buddhist books and precepts. After returning to Tibet, Tong-mi Samb'ota framed the system of Tibetan characters in two styles, the yi-ge dbu-djan, and the yi-ge zur-djan} 95. The yi-ge dhu-djan, or, as the name means, the letters furnished with heads, also called rom-yig ^ or " thick letters " in the Western provinces, and commonly called TT-djan by a sim- plification of the first name in speech, are the ordinary Tibetan characters commonly used now in printing.^ The legendary 1 Shang dchoh Khantouktou in the preface to his Tihetan-Mongol Dictionary, states that Taomisamkandra, when hack in India framed two Tibetan alphahets, the tsab from the Landza, and the char or kchar from the Vardo, cf. Journal Asialique, 1822, p. 331. * Som = ' thick, big, stout,' whence rom-yig as a distinction from the p'ra yig or cursive writing, where p'ra, means ' thin, fine, minute,' cf. Jaeschke, Tibet Engl. Diet. pp. 353, 536. s They are said to have retained faithfully the primitive forms which were cut on wooden blocks for printing in the seventh century, soon after their introduction into Tibet (cf. H. "Wuttke, JDie Entstehung der Schrift, p. 471). Printing was introduced from China, where the art was flourishing, especially in the west, on the boi-derland of Tibet. It began by the habit, still in use, of taking rubbings of engraved stones {i.e. of blackening, with a pad, paper squeezed on the inscribed stone, so that the deepened marks appear white on black ground). Such rubbings werein circulation under Han Tchang-ti( A. D. 76), andTsin"Wu-ti(A.D. 265). The engraving of the texts of the sacred books on stone, in a.d. 175, by Tsai-yung, and in a.d. 240-9. afforded facilities for such rubbings. The art was improved in the region of Shuh, i.e. Szetchuen, and much used by the Buddhists for the pro- pagation of their texts and images of Buddha. But we do not see it adopted by the Chinese government before the year 693 a.d. for printing the pictures, autographs and neglected texts. Printing on blocks was carried to Korea and Japan, where it was in use in a.d. 764. Two specimens of the latter date printing are in the British Museum. As to the printing with moveable types, the art was known or invented in China circ. 1041-1049 by Pi-shing and improved by Tch'en Knob (circ. 1080) and Yang K'oh; most of their types were in clay. A century afterwards printing moveable types in copper were made in Korea. Copies of books so printed later (in 1317) are still in existence. 64 BEGINNINGS OF WRITING AROUND TIBET. account says that they were derived from the holy writing the Ihai yi-ge or Landza. Now so far as the Landza characters are those that we know in Nepal, the derivation of the JJ-djan from these characters is not borne out by comparison. The connection between the two writings is that of a family parentage and not at all that of a derivation. The similarity presented by the U-djan to the characters of the Grupta inscriptions at Allahabad is, on the other hand, remarkable, and in order to conciliate the tradition with the material evidence, we ought to surmise that the latter characters were at that time considered also as Landza. But we have no proof -of such a fact, and the whole confusion arose from the embellishments and magnifying details added in later ages. 96. The ancient tradition of the appearance in the country five generations before the king Srong-tsang, of a Buddhist book, which was specially venerated in Nepal, '^ has appa- rently suggested the precise minutiae given about a connec- tion of the Dhu-djan with the Landza of the latter country, while the only truth is that the Landza was a beautified and ornamented style of writing down the same Indian characters which were used as a pattern for those of Tibet. The Ihai-yi-ge name in the Tibetan tradition may have the same meaning as Landza, but it has not necessarily the very same characters, and it corresponds undoubtedly to the devalipi, "Divine writing," which figures among the sixty-four writings of the list given in the Lalita Vistara, which devalipi is not the Landza. The perfect likeness in form exhibited by the Ica-p'reng dhu-djan ^ to the inscriptions of the Gupta dynasty at Allahabad represent fairly, we have no doubt, the monumental writing, which was used in all the religious monuments at that time, and is quite sufficient to prove that they were imitated from the latter. The Ka-smad sum-dju Ka-li,^ i.e. the writing in thirty 1 According to Woodville RockhiU, op. eit. p. 210. It is mentioned by Brian H. Hodgson, Estaya on the Language and Literature of Tibet, vol. i. pp. 17, 37. ^ Ka-p'reng dbu-djan, i.e. the capital alphabet, the same as the yi-ge dbu- djan, or more simply u djan. ^ /S'«ff»-rf('«= thirty. Cf. Jaeschke, Tibetan-English Dictionary, p. 426. BEGINNINGS OF WEITIXG AROUND TIBET. 65 characters, a general name of the Tibetan writing, includes the rgya-la med-pai yi-ge drug,^ or the six letters which are not existing in Sanskrit. 97. The second of the styles of writing found by Tong-mi Samb'ota was called, as we have seen above, ye-ge zur-djan, i.e. letters furnished with an angle, or cornered letters, so called from the fact that the upper part or head of the letters were not so regular as in the dbu-djan style. He is reported to have derived them from the EJui-yi-ge or Vartula characters, the two words being considered as equivalent, notwithstanding their respective meanings, which are different. The statement is translated incorrectly by some, as intended to show the derivation of this second writing from that of the Nagas, or simply from the Nagari character, which is another way of escaping the difficulty.^ Now Klu in the Tibetan dictionary of Jaeschke means "serpent demon," and corresponds to the Sanskrit Naga. But in the com- pounds it is written Klui, and does not seem to carry on this extreme meaning. For instance, Klui-skad, in which sltad is "language," means the Pracrit language, i.e. the vernacular dialect, in contradistinction to the rgya-gar shad or "Sanskrit language," properly the "language of India," in which rgya-gar, Utt. " white plain," stands for India.' In Klui-skad we have seen Klui corresponding to a peculiar denomination of " that is in common use." It is obviously in the same acceptation that it must be taken in the words Klui-yi-ge, which are used for naming the vartula character. 98. The latter is the Sanskrit word, meaning "round, circular," which, applied to a writing, is suggestive of the rounded shapes of the cursive characters, in opposition to the angular and straight forms of the monumental or lapidary style. Now, to complete the parallelism, the nagalipi occurs side by side with the demlipi, among the sixty-four alphabets 1 Jaeschke, ibid, p. 418. ' Baboo Sarat Chandra Das ananges it still otherwise, and reads Wurlu m his text, while in his note thereon he says, " Wartu is probably the langoage of the people of Kafiristan and Bactria." Cf. his Contribuiiom, loc. cit. p. 2. But the reading vartula is qnite plain, though arranged after the fashion of the Tibetan lexicographs vartu-la. s Jaeschke, Tib. Diet. p. 105. 5 66 BEGINNINGS OF "WRITING AROUND TIBET. of the Lalita Vistara,^ as in tlie Tibetan record the Klui-yi-ge and the Ihai-'yi-ge appear together. Indianists do not kaow what writing was denominated vartula, and was the ante- cedent of the zur-djan of Tong-mi Samb'ota. From the great resemblance between the characters of the latter style as they are drawn in its immediate derivative the dhu-med, with those of the first style dhw-djan, of which they differ only by the thick strokes of the heads, which are absent, and some looseness in the shapes, which are less tight and want regularity, it is quite clear that the two styles, the Indian antecedents of the two styles of Tibet, were one and the same writing, one drawn or incised on the monuments, the other in use in daily life and for common purposes. 99. The appearance repeatedly of the name of Naga in connection with the current and common writing remains to be explained.^ It is quite clear that the intended meaning was not a special use and acceptation of the word naga, but the proper meaning of this word, viz. " serpent," which, however, could not be separated from its usual surroundings of super- stition and demon-character. Its translation by the Tibetan Idui, instead of its transcription as in the case of vartula, shows it plainly. It seems to me that this qualification was given to that current writing because of the curved and snahe-like forms of its characters in opposition to the stiffness of the monumental or to the ornamented forms of the sacred style of writing. 100. The yi-ge zur-djan writing, properly the half-cursive, was developed into the more current dhu-med or head-less characters, also called fa yig in the western provinces. Of the dhu-med, commonly pronounced Tims, there are various kinds : the dpi-yig (from rfpe=pattern, model), the more distinct and careful, used in copying books ; the 'k'yug-yig, lit. the running writing, the cursive and often rather illegible style used in writing letters ; and the 'bam-yig, the ' Vid. above, § 90«. As to the date of the Lalita Vistara it is not known. It waa first translated into Chinese during the Shuh Han dynasty, a.d. 221-263 (cf. Bunyiu Nanjio, A Catalogue of the Buddhist Tripiiaka, n. 159). The date of A.D. 76 first given by Stanislas Julien was the result of a mistake. 2 No satisfactory explanation has hitherto been given of the name Nagari though four hypotheses were put forward. Cf. A. Bumell, Elements of South Indian Palaeography y 2nd edit. p. 52«. BEGINNINGS OF "WEITING AROUND TIBET. 67 very large and regular style inyented for the use of elementary writing schools.^ The monumental writing of Tibet was that from which was derived, for the special use of the Mongols in the thir- teenth century, the short-lived writing known as Bagspa, from the name of the Lama who worked it out, as we shall see in a further section of the present book.^ 1 On all these names, vid. Jaesclike, Titet. Engl. Diet. pp. 60, 327, 392, 508. For specimens of the writings see the plates in Csoma de Aorbsi's Grammar of the Tibetan Language, Calcutta, 1S31, Ito. 2 Cf. below, § 109. PART II. BEGINNINGS OF WRITING. 69 V. Short-lived "Writings in Central Asia. a) The K'itan Writing. 101. In 910 A.D., on a rock near the Lan river, and in 915, on a stela south of Tsingtsung, Apaoki, first Emperor of the K'itans, had memorials of his victories engraved ^ in a sort of writing on which we have no information. The only thing we may safely infer is, that the writing was foreign to them, since they wanted, later on, to have one of their own, and until that time used notched sticks for their contracts and agreements. In 920 Apaoki employed several Chinese scribes in framing, on the pattern of the Chinese Kiai and Li writings,^ several thousands of characters for the K'itan language. The latter were called Ta tze or Great K'itan characters.^ Five characters of the Ta tze, preserved in a Chinese work, have been published by Mr. A. Wylie. In 924 a stone tablet was engraved in K'itan, Uigur and Chinese characters.* 102. Among the K'itans, badges were worn in the shape ^ H. H. Howorth, The Northern Frontagers of China, Part V. The Khitai or Khitans, pp. 32, 43. * Liao she, K. 2, f. 2, Pen ki ; G. Deveria, Examen de la stele de Ten-t'ai, p. 178. {Sevue de V Extreme Orient, 1882, vol. i.) ' Shu she hui yao ; A. Wylie, Ancient Buddhist Inscriptions at Keu-yung Ewan, J.R.A.S. 1871. Vol. V. Mr. 6. Deveria, Examen de la stele de Ten- t'ai, has also reproduced them. Vid. his plate Nos. 57-61. * Cf. Ma Tuanlin, Wen hien t'ung K'ao, bk. 34S. Suh wen hien t'ung K'ao, bk. 184, f. 31. A. WyUe, Introd. to his Translation of the Ts'ing wan k'e mung (Shanghai, 1855, 8vo.), p. xviii. 70 BEGINNINGS OF WHITING. of a fish.i The King's relatives as far as the fourth degree wore a Jade fish, which in the latter case was made only in gold, while the relatives of a more distant rank were only- entitled to a silver one.^ Dr. Bushell, of Peking, has in his possession a small fish, 2| inch in length, ^ inch in width, one side convex, ornamented with scales inlaid with silver, the other flat, with an engraved inscription. He has kindly sent me a rubbing of the latter.^ The characters, which he calls Jutchih, are similar to the inscription of Lang-Eiun, dated 1134 a.d., and published by Mr. A. Wylie. This identification aflbrds a clue to the somewhat intricate problem about the origin of subsequent writings, said to be derived from that of the K'itan, but which did not bear any resemblance to the five characters Ta-tze. The characters of the fish badge must have been those of the small K'itan characters, which were severally spoken of afterwards, but whose date of invention or making has not been kept on record.* b) The Tangut Writing. 103. From 982 to 1227 a.d., betwixt China and Tibet oQ their northern borders, stood a powerful kingdom which was swept away by the Mongols, and has left hardly any trace of its existence.^ The Si-Hia or Tangut, such was its name, did possess a writing which was only an adaptation of the 1 A somewhat similar custom existed among the Peh-hai, which were hordering the K'itans on the north-east. Vid. Ethnographie de Matouanlin, trad. d'Hervey de St. Denys, vol. ii. i)p. 366-367. ^ Kin She, quoted in Khang-hi Tue-hm, pu 196. "Wells Williams, Syllabic Dictionary, p. 1119 : " Fei kin yii, a prince royal among the Khitans, hecause he wore a fish made of gold." 3 MS. note, 27 May, 1881. * During the Kin dynasty scholars were famous for their knowledge of the Kitan characters large and small; and these smaU characters, with the required alteration, were in use among the Jutchih or Kin as an intermediary transcriptive translation from the Chinese previous to the transcription into the great Jutchih character, for instance, in 1164 a.d. (cf. G. Deveria, Examen de la Stele de Yen-t'ii'i, I.e. p. 179). They remained the sole writing after 1191 a.d. ' It is not at all unlikely that the Meniaks (Mi-nyag) of N.E. Tibet are the descendants of this ancient people. The Tangut or Si-Hia country was also called Knshi, which the Chinese rendered by Ho-si, and which has survived in the modem name KhatcJie in N. Tibet. BEGINNINGS OF WEITING. 71 small K'itan characters to the requirements of the lan- guage.^ In 1030 the king Tchao Te-ming, or more probably in 1037, his successor Tchao Yuen-hao, otherwise Weili, who had married a K'itan princess, is reputed to have invented the Si-Hia characters,^ of which several specimens remain on coins and inscriptions. The Chinese sources describe them as made of numerous strokes, somewhat like the cabalistic symbols of the Taoists, angular like the Pah-fen, and the ends often curved as in the Seal character. 104. The numismatic evidence shows that from 1049 to 1120 A.D., when their names or period of years or Nien-hao were mostly of several words, while, as in the Chinese names, the others are generally made of two words only, the Hia Kings did not issue currency with Chinese legends, as they did before and after the above dates. During these seventy-one years, comprising eleven nien-hao, they issued money with legends in their own character. Though specimens of this mintage are known in numismatic collec- tions since 1149 a.d.,' we do not know if they have or not issued a new type at each change of the periods named. Three only are known hitherto, and they remain still undeciphered. The symbols belong without doubt to the same style as those of the hexaglotte inscriptions of Kiu-yung Kuan, which hitherto were supposed to be Jutchih. On the other hand, Chinese numismatists are most positive in their statement that these coins were issued by the Hia emperors, and they give the following information.* 1 Besides tlie numismatic evidence given below, as to the intimate connection of these writings, may be quoted the tact that a rubbing of an inscription from Shend, in the possession of Dr. Bushell of Peking, is marked as written in Si-Hia characters ; and the characters of this inscription are similar to those of Kiu-yung Kwan. Cf. G. Deveria, ibid. p. 185, and § 109 below. 2 -St Bia shu she, K. 12, f. 9 ; Sanq she, K. 485, f. 12 ; Shu she hut yao, K. 8, f. 11 ; Liao slie, K. 115, f. 4; G. Deveria, I.e. p. 185. ' Hung Tsung in his Tsiuen tchi, published in 1149 a.d., gave a woodcut of one of these coins, but he did not know its origin. In the Tamba collection I have found two specimens now in the British Museum. Dr. BusheU, of Peking, has also two specimens in his collection. Li Tso-hien, a learned Chinese numis- matist, has published in 1864 two rubbings of such coins in his Ktt Tiimn hut, Li, K. XV. f . 8. * Ei kin so kien luh, Ku tsiuen hui, loo. cit. 72 BEGINNINGS OF "WRITING. 105. Once several urns containing old money were un- earthed^ at Liang-tchou,^ in Kansuh. These pieces of money- belonged to the T'ang, Sung, Liao or K'itan dynasties, and to the Western Hia or Tangut, all in Chinese character, besides these pieces with legends in an unknown writing. Now, in the temple of Ta-yun at Liang-tchou was an old stela with an inscription in the same writing on the obverse, while the reverse was inscribed in Chinese Kiai shu and dated the fifth year of Tien-yu min 'an? This is one of the nien-hao of Si- Hia, during which no money with Chinese legends is known to have been issued by them. And the Chinese numismatists do not admit any doubt as to this identification of these speci- mens of currency. On the other hand, some considerations might be put forward to strengthen it. One of them is that they could hardly belong to another people. The characters of the legends have no resemblance to those of the Jutchihs, and cannot be disintegrated into two parts with the same facility as the latter, all known mintage of which bears legends in the Chinese character.* As to the K'itans, whose Great character was invented some 117 years before the Hia writ- ing, they have never used it on their mintage. Ruling over the northern boundaries of China and struggling to extend their sway southwards, they used the Chinese symbols in view of having their currency accepted beyond their dominion.* "With the Hia, who were settled on the outskirts in the north- west, the necessity was not so close, and, as we have seen, they tried for seventy-one years to follow their own way, but the requirements of trade were after all the stronger, and they were compelled to give up and return to their former use of Chinese characters on their money. 106. In his narrative W. de Eubruquis speaks of two • The date of the find is not given. 2 Lat. 37° 59', long. 102° 48'. ' i.e. 1095 A.D. * It is, however, ominous that the oldest specimen known with certainty of their currency bears Vne Nien hao of Tchang-lung, i.e. 1166-1160 a.d., and that no specimens are known of their previous eight periods of years. Pieces of a currency bearing the legends Ta kin tung pao. Kin she tung pm, and She kao tuna pap, in Chinese characters, perhaps belong to that period of their mintage. 6 They began to cast money as early as their first nieu-hao, viz. Shen ts'eh 916-921 A.D. ' BEGINNINGS OF WRITING. 73 writings. He says : ' — " The Thibet people write as we do,^ and their letters bear a strong resemblance to ours. The Tangut people write from right to left like the Arabs, and their lines advance upwards. The first statement agrees with the well-known Tibetan writing derived from India,' and written as usual for the writings of this provenance from left to right. The second, concerning that of Tangut, is more difficult to understand. The empire of Tangut, other- wise called Si-JSia, disappeared in 1227, and nothing was left of its former power and civilization, as everything was swept away by its mighty conqueror Genghis Khan. Twenty-six years had passed away since this complete destruction when the Minorite Friar made his celebrated j ourney.* Therefore he cannot have spoken of the complicated writing of the Si-Hia, the more so that it was written like the Chinese from top to bottom, and in columns from right to left. His description does not apply better to the writing of the Uigur, which also was called the Tangut character. Rubruquis had himself in another passage described this writing,^ borrowed "by the Tartares from the lugures," written from top to bottom, in columns, arranged from left to right.^ This description is quite precise, and it is not unlikely that such a disposition was in use in thef west ' previous to the spread eastwards of ' This passage was pointed out to me by Colonel Yule. ' t.e. from left to right. 3 See above, §§ 83-100. * Major G. H. Eaverty, On the Turks, Tartars, and Muffhals, p. 119, from Muhammadau sources, in Congres des Orietitalistes, St. Petersburg, 1876, toI. ii. speaks of a compact which was drawn up between Kabal the great-grandfather of Genghis Khan, and his brother Ka-jCile, by order of their father T(imn&-i, in order to ensure the sovereignty to Kabal and his descendants. "A compact to this effect was drawn up in the 'I-ghiiri language — which is said to be the same as was used in Tibbat, and called Tiingfit characters, according to some writers signed by them both, and deposited in the treasury." ' Composed of fourteen consonants and three vowels. Cf . Ahmed ben Arabshah who wrote about 1440 A.D., and who was made known first by Le Eoux des Hauterayes {Encyelopedie Petity, 1767, t. iii. p. 551). See, also, Abel Eemusat, Eeeherehes sur les langues tartarea, p. 29 sq. ' Harris's Collection of Voyages, London, 1744, vol. i. Same statement made by the Monk Bacon. Tid. J. Klaproth, VerzeicAniss der Chinesischm mid mand- ahuischen Sticker der i. Bibliotek zu Berlin, Paris, 1822, fol. H. Yambery, Spraehmoimmmte und das KudatJeu Bilik, Innsbruck, 1870. The Kaoudat-Kou- biMk had been previously described only by Amedee Jaubert, Notice d' un man merit turk en caracteres ouighours, Paris, 1826 ; Journal Asiatique, vol. vi. pp. 39, 78; and by A. L. Davids, Oraminaire Turke, pp. xixii-xxivii. ' Cf. Dr. Isaac Taylor, The Alphabet, vol. i. p. 305, who has collected the 74 BEGINNINGS OF "WRITING. the Semitic alphabet cognate with the Syriac, Estranghelo, Mandaite, which became the prototype of the Uigur, Mongol, and Kalmuk writings. 107. We must leave the writing of Tangut described by Eubruquis without identification, however precise his state- ments appear to be. As no other writing of Tangut existed besides the complicated symbols of Chinese descent, and the Uigur character, the problem will not be easily solved. We may venture to suppose that the explanation might be found in the simple fact that the Tangutan scribes writing in Uigur character, were accustomed to the Chinese writing, and perhaps wrote the two characters sometimes in parallel columns. When tracing the Uigur character from right to left, though read from top to bottom, they were compelled, in order, to match with the Chinese text, to put the latter top-lines on the right-hand side. The result of this process was to place the following Chinese lines, and consequently the parallel required lines in Uigur character upwards, as described by Eubruquis. c) The Jutchih Writing. 108. The Kin or Jutchih,'^ ancestors of the present Mand- shu dynasty of China, established their sway and ruled from 1115 to 1234 A.D. over the northern part of China proper. Several sorts of characters were used by them,^ but two only were specially adapted to the requirements of their language. reasons wMch favour tliis probability, against the suggestion of Abel Remusat, who had attributed the vertical writing to an imitation of the Chinese practice. 1 On this name cf. § 86, u. 5 above. Prof. C. de Harlez has made a transla- tion of their history, Aisin gurun-i suduri Hike, from the Mandshu. Some extracts have appeared in the Mmdon, 1886, vol. v. pp. 379-390, 636-651. ^ On these writings cf. A. Wylie, Translation of the Ts'ing wan k'e mung, a Chinese Grammar of the Manchu Tartar Language, with Introductory Notes on Manchu Literature (Shanghai, 1865, 8vo.), p. xix. On an Ancient Inscription in the Neuehih Language {Journ. Roy. Asiat. Soo. 1860, vol. xvii. pp. 331-345). L. de Eosny, Zes Niutckih, leur langue et leur Scriture (Archives paleographiques, 1872, 8vo.), pp. 179-189. A. WyUe, On an Ancient Buddhist Insenption at Keu-yung Kwm, in North China (Journ. Roy. Asiat. Soc. 1871, vol. v. pp. 14-44, and plates). G. Deveria, Examen de la Stele de Yen-t'cA, dissertation sur les caracteres d'ecriture employes par les Tartares Jou-tchen (Jievue de I' Extreme Orient, 1882, vol. i.), pp. 173-183. Previous to their adoption of any written character, they used symbolical means of communication. Cf. above, § 27. BEGINNINGS OF WRITING. 75 At first, when they came into contact with the K'itan and the Chinese, they made use of their respective writings. But in 1129 a.d. Tai-tsu Aguda commissioned Wan-yen Hiyn, styled Kuh-shen, to make characters on the same principle as those of the K'itans, but suitable for expressing the sounds of the Jutchih language, and bearing a re- semblance to the Chinese Pattern hand or Kiai shu. This order was complied with, and in 1120 and 1126 a.d. the characters were put in circulation, they being termed Ta-tze ' great characters,' otherwise Uncials. These characters were based upon those (also Ta-tse) of the K'itans, and were for that reason occasionally called K'itan characters among the Jutchih, who kept them in use until 1191 for official purposes. The inscription of Yen-t'ai (Kai fung fu, Hon an), which has been published by Mr. G. Deveria, is most probably a specimen of these characters. 109. In 1138 A.D. the Kin^ Emperor Hi-tsung framed himself, in imitation of the small characters of the K'itan, a set of characters which accordingly were called small characters, and were first brought into use in 1145. This writing is that which is used on the Liang Kiun inscription of 1134 A.D., and on that of the Kiu-yung gate of 1345, both published by Mr. A. Wylie.^ The official teaching of these latter characters was only abolished in 1658 at Peking.^ • Kin 'gold,' is the CHnese translation of the native name of this people calling themselves A i-sitij "while Jutchih, etc., was the name of their original country. It is interesting to point ont that the Lang- Kiun inscription begins in Chinese bv Ta kill 'great Eiu,' and in the Jutchih text by two characters apparently derived from the Chinese y^ [Ij Ai shan. ' In the Jourii. Boy. Asiat. Soc. for 1860 and 1871, as indicated above. M. G. Deveria, in his paper on the stele of Ten-t'ai, above quoted, has collected a great deal of information on these matters. * In the Annals of the Ming is recorded the establishment of the Sze-t/ Kiuin or ' translatorial office,' in connection with the national collegiate institute in 1407 A.D. The object of this office was to facilitate the transaction of diplomatic correspondence with foreign nations, and the incumbents were charged with the study of eight different foreign languages. One of these was the Jutchih, with seven interpreters, increased to eight in 1483, and niae afterwards. The other languages were the Mongolian, Tibetan, Sanskrit, Bokharan, Uigur, Birman, and Siamese. In 1470 a iixed number of interpreters were appointed for each of the nations above named. On the accession of the present Mandshu dynasty in 1644, they re-established this translatorial office the same year, adding two sections in addition to the eight previously existing. These were the Peh-y and Pa-peh, 76 BEGINNINGS OF WHITING. d) The Bagspa Writing. 110. The history of this alphabet framed for the use of the Mongols on special order of Khubilai Khan by the Tibetan lama Bagspa in 1269 a.d., and derived from the Tibetan characters, has been written several times,^ so that we need not enter into the details of the fact. We know that it was in use for about eighty-five years, though not in general practice. The only disclosure made about this writing by more recent research is the occasional use which was made of it on his coins during the years 696 to 703 of the Hijreh (1295-1304 A.D.) by the Ilkhan Ghazan Mahmud of Persia.^ Regulations were enacted by the Mongols of China pro- scribing the use of this writing in a.d. 1269, 1272, 1275, 1282, 1284, 1307, etc., but without great avail, as they could not eradicate the habit of using the Uigur-Mongol characters, which eventually recovered their former general use.^ The Bagspa characters were 41 in number, namely, 22 two nations on the south-west of China (as we shall see belo'w, § 175). A presi- dent was appointed, and fifty-six professors ; and tie following year another section was added, consisting of thirty interpreters for the tributary nations. In 1659 the section devoted to the Jutchih language was suppressed, as also that for the Mongol, A. WyUe, On an Ancient Inscription in the Neu-ehih Language, I.e. pp. 335-336. A. Bemusat, Secherches sur les langues Tartares,pp. 218-220. G. Deveria, Examen de la stile de Yen-t'ai, pp. 180-181. The latter scholar has in preparation an extensiTe history of this ' College des Interpretes.' ' The most important notices are the following: A. Wylie, Ts'ing wan A'« mung, introd. pp. xxiv-xxv ; On an Ancient Inscription in Chinese and Mongol, from a stone tablet at Shanghai, in Transactions of the China Sranch of the Moyal Asiatic Society, part v. 1855, Hongkong, 8vo. pp. 65-81 ; Sur tme inscription Mongole en caracteres Pa-sse-pa, in Journal Asiatique, Juin, 1862, pp. 461-471 ; On an Ancient Buddhist Inscription at Keu-yung Kwan, in North China (in J.R.A.S. 1871, vol. V. pp. 14-44), pp. 25-31. G. Pauthier, Rapport sur deux medailles en cuivrejaune trouvees a Sourabaya [tie de Java) in Journal Asiatique, Avril-Mai, 1860, pp. 321-338 ; De I'alphabet de Fd-sse-pa et de la tentative par Khubilai Khan au Xllle sieele de notre ire pour transcrire la langue figurative des Chinois au moyen d'tme icriture alphabetique, ibid. Janvier, 1862, pp. 6-47. J. Edkins, An Account of Sanskrit and Mongolian Characters found in Chinese Books, in Transactions of the China Branch of the Royal Asiatic Society, Hong- Kong, 1855, pp. 101-108. V. Grigorieff, Lettre adressee d la Sociiti Asiatique de Raris sur Vorigine et les monuments de I'ecriture carree dont I'invention est attribute au Pagpa Lama, in Journal Asiatique, Juin, 1861, pp. 522-558. Conon von der Gabelentz, Zeilschrift fiir die Kunde des Morgenlandes, vol. ii. p. 1. ^ Of. my note On a Bagspa Legend on Coins of Ghazan Mahmud, in Stanley Lane Poole, The Coins of the Mongols in the British Museum (London, 1881, 8vo.), pp. l-Ui. 3 CI below, § 126. BEGINNINGS OP "WRITING. 77 consonants, 10 vowels, 8 final syllables, and one initial sign, and the total of their combinations was a thousand.' VI. Deflected Writings. 111. The deflected writings are those which cannot be traced back to their origin without, either in their derivation, or in the course of their evolution, having undergone a change of direction either from right to left instead of from left to right, or horizontally instead of vertically, and so forth. But the deflection takes place on the whole number of the characters, alphabetic or syllabic, and not '^ on a few of them only, with a single deflection in case of immediate derivation, and if these are different deflections well established, they indicate that the derivation is remote, and through several intermediary stages, which require to be authenticated before any safe conclusion may be obtained. The latter cases, which, in the study of the genealogy of phonetic writings must not be accepted without most cogent reasons, and as isolate exceptions, have not the same importance as pictorial writings, where the design of certain symbols, or their size, may have exercised an influence on their later position. a) The Babylonian Writing and its Derivates. 112. The oldest instances of deflected writing are those (stiU, however, uncleared) which have occurred in the evolution of the Babylonian writing. Pictorial in its ' Cf. Abel Eemusat, Recherches sur les langues tartares, p. 345. " The neglect of this unwritten principle of comparatiTe palseography has led to many unsuccessful attempts at connecting writing of different origin. For instances of late years : J. Hal^vy, Simmi d'un Mimoire sur Vorigine des eeritures iiidiennes, in C. R. Acad. Inscript. B.-L., 1884, vol. xii. pp. 214-223 (cf. helow, §§ 161 sq. of the present work). Park Harrison, Phoenician Charaeiers from Sumatra, 1874, Jour. Anthrop. Inst. vol. iv. pp. 387-8. Eejang Manuscripts on Bamboo, 4to. with plates ; Characters Tattooed on a Motu TFbmaM (London, 8to.). Dr. Hyde Clarke, LoU and Vei Characters, in The Athenteum, 16 Sept., 1882, and my rejoinder, ibid. 23 Sept., when I recounted the well-known circum- stances of the recent invention of the Yei characters about 1834 a.d., and the improbability of their connection with the ancient Lolo characters (§§ 38, 1). The only remark which can be added here is that, judging from the similarity of a few characters, the inventors of the Yei syllabary ha3 most likely seen, but not learned, the Abyssinian writing. 78 BEGINNINGS OF WRITING. beginnings, and derived most probably from the same source as the Egyptian hieroglyphics, and not from them, it was carried to Babylonia perhaps by the Kushites.'^ Anyhow, it is to that early period that belong the coincidences which are detected between some Babylonian, Egyptian,'' and Hittite symbols. The characters were then arranged, apparently according to shape and figure, as with the Egyptian and Hittite, and, probably likewise, in boustrophedon. No relics have been preserved of this early phase, which, however, may be restored hypothetically by palseographical necessity. 113. The next step in the evolution of the Babylonian writing seems to have been the abandonment of the boustro- phedon system and the adoption of a serious modification of the symbols. The shape of the characters, in cases of isolated or compound symbols easy to disintegrate, was ' I shall examine this question at length in the third part of the present work. I have adTOcated the existence of the writing in Bahylonia previous to the settle- ment of the Akkado-Sumerians, and given my reasons for it as early as 1883 (cf . my paper on The Oldest Booh of ike Chinese, § 111 and notes, in Jouro. Hoy. Asiat. See. 1883, vol. xv. pp. 278-279, and my note on Th^ Fre- Akkadian Writing in The Academy, 758, Nov. 13, 1886, p. 331). It has heen argued (by Mr. G. Bertin, ilid. Ibl, 6 Nov., p. 313) that Hincks and Norris had already pointed out that the Babylonian hieroglyphics were derived from those of Egypt, and therefore that my above-mentioned (fiscovery had been made previously, since this derivation would imply the pre-existence of the Babylonian writing to the Akkado-Sumerian period. But this suggestion of Hincks, adopted by Norris, was a mere shot in the air, and was not, more in their time than in ours, borne out by facts ; the decipherment of the Babylonian characters, and the recovery of their ancient forms, were still unripe, and did not permit these great scholars to adduce any serious fact in support of their view, except the coincidence of two or three characters Babylonian and Egyptian, a coincidence which receives another explanation. Anyhow this view of Hincks and Norris was so little accepted, and was so far from being public property, that Mr. Bertin himself, in his paper on The Pre-Akkadiam Semites, had not made any allusion to it. There- fore I remain the first who may claim to have given palseographical reasons for the existence of the Babylonian writing previously to the Aliado-Sumerians, who were generally attributed the inventors of this writing (cf. Journ. Roy. Asiat. 8oc. 1886, Vol. XVIII. p. 848). ' I have begun to draw up a list of these coincidences, and my friend Pro- fessor P. Hommel has also begun the same work independently. Professor Hommel thinks that the Egyptian hieroglyphics were derived from those of Babylon, while I think that both, as well as the Hittite hieroglyphics, are derived from one common stem, of which the symbols preserved on the land-mark stones of Babylonia and Susiana were perhaps an early derivation (cf . above, § 5, n. 5 and § 7). A derivation of the Babylonian symbols from the Egyptian is altogether out of the question. The figures of Babylonian pictographs, known from an ancient tablet from Nimroud, which has been published several times (notably by W. Houghton, On the Sieroglyphic or Picture Origin of the Characters of the Assyrian Syllabary, in Trans. Soc. Bibl. Arch. 1879, p. 465), are much more rude than the Egyptian hieroglyphics. BEGINNINGS OF "WRITING. 79 rectified in order to have them facing to the reader (a feature common to this writing and to its derivate the Chinese ^) in contradiction to the profile feature of the Egyptian and Hittite hieroglyphics. The direction selected was from left to right, most likely because it was previously more often used, as was the case in Egyptian.^ Then the characters were strictly disposed in perpendicular lines, a disposition which may have been introduced by the Akkado-Sumerians in analogy to the knotted cords and notched sticks which, in common with the other Scythian populations, they probably made use of previously as means of communication ; and these perpendicular lines were arranged from right to left.* The system has been long preserved on the seals, as may be seen in any of the publica- tions which have been made of them.* 114. It is from this period of the writing that dates the eastward expansion which permitted the leaders of the Bak tribes, between the Hiudu-Kush and the Caspian Sea, to be acquainted with it by practical intercourse, in same time as with the other elements of civilization which they carried in their migration to the N.W. borders of China.^ The oldest writing of the Flowery land, of which some relics have been religiously preserved in this country of servile traditions, was derived from it ; and as the wedge-strokes, though not unknown in the early traditions of the Chinese, were not kept by them because of their writing on vegetable materials, ' The latter has occasionally preserved in the forms of its oldest period the three shapes of some symbols, left, right, and face. 2 Cf. what I have said about this matter in Jbwn. Roy. Asiat. Soe. 1883, vol. XV. p. 279 n. Many shapes and dispositions of these characters, concealed by the wedge-strokes, are explained away by their derivates in early Chinese, where the wedge is not adhered to. The ancient direction from left to right is shown by the early compounds. ' I was mistaken in attributing this change to the Chinese themselves, while they have servilely copied their masters, as said in my communication on Babylonia and China, in The Academy, 7 Aug., 1886. Cf. also my note on The Affinity of (he Ten Steins of the Chinese Cycle of Ten with the Akkadian Numerals (ibid. 1 Sept., 1883). * Cf. Theo. G. Pinches, The Babylonian and Assyrian Cylinder-Seals of the British Museum, London, 1885, 8vo. And the great work of M. Joachim Menant, La Glyptique Orientate, vol. i. Paris, 1884, 8vo. ' Cf. above, § 49 and notes; also the comparison of ancient Babylonian and Chinese characters, plate iv. of the present work. 80 BEGINNINGS OF WRITING. these early characters contain some precious information when compared with the Babylonian unclear wedged symbols.^ 115. The next phase in the evolution of the Babylonian writing was the arrangement of the characters in horizontal lines, and their reading from left to right. This twofold result was obtained by a single turning of the former per- pendicular lines to the left, as shown by the following diagram : 4 3 2 1 before after 1 > 2 > 3 > 4 > V V V V Then the spread of this writing gave birth to varieties in Nineveh and Susiana; and afterwards the Assyro-Babylonian, Vannic, Medo-Scythic, and Persian Achsemenid syllabaries evolved from it, all of them keeping, besides their purely phonetic symbols, the use of a few ideographs as survivals of the former stage of their mother-writing. b) Various Writings of the West. 116. The nearest and best-known instance is that of our Latin writing and alphabet, from left to right, derived from the Phoenician characters, which were written from right to Mi? 117. The characters of the Lybian inscriptions and the Tamashek or Tifinag alphabet, are most probably, in my opinion, derived remotely from the Hymaritic or Musnad characters through some intermediary like the Safa writing, as they exhibit an irregular and apparent succession of de- flections. 1 Cf. my Early Siatory of the Cldnese Civilization, with plate of early Chinese and Babylonian characters (London, 1880, 8to.), and above, §§ 46-50. * This elementary fact has suggested a most ingenious and demonstrative device to the author of a little work on ■writing, where a transcription made by him of the Lord's Prayer in Phoenician characters on transparent paper can be read, the same paper turned over, as if in our writing, with a little practice. Cf. J. Van Drival, He Vorigine de I'icriture (Paris, 1879, 8vo.), p. 20. BEGINNINGS OF WRITING. g^ c) The Synac Writing and its Eastern Derivatives. 118. The Syriac writing lias played a great part in Central Asia, but it has not kept, in its descendants, its original arrangement, from right to left in horizontal lines from top to bottom. It was first carried eastwards in 435 a.d., when Barsumus, a doctor of Edessa, who was a zealous partisan of Nestorius, having been ejected from his chair, took refuge in Persia, and became Bishop of Nisibis, where he founded a school of theology in rivalry to that of Edessa.^ From the school of Nisibis proceeded those bands of missionaries who, during the following centuries, spread with or without the Nestorian tenets the knowledge of this writing.^ 119. It is not one of the less curious features in the eventful history of the Syriac characters that, driven away from their original country by " theological disputes which turned on the most subtle metaphysical distinctions," they became the Uigur writing employed in translations of the Buddhist books, and also of the sacred literature of China. It is an extremely curious fact, pointed out for the first time by Mr. A. Wylie,^ that the greater part of the Buddhist books must have been translated into Uigur. In the Chinese history of celebrated Buddhists, published under the Sung dynasty in 988, it is recorded that "when the Sutras and Vinayas from India were taken to Kutche, on the north of the Tsung-ling mountain, to Lenlan, to Kharashar, to Khoten, and to Khashgar, the natives not understanding the lan- guage of India, the books were translated for them into the barbarian languages which they spoke." * This implies countries speaking Uigur and using the Uigur character, which is an adaptation of the Syriac. ' Cf. Isaac Taylor, The Alphabet, an Account of the Origin and Development of Letters (London, 1S83), vol. i. p. 291. 2 "Wei Tsih, a Chinese author of the ninth century, enumerating 56 different kinds of writing that have been used in China, mentions as the 64th the Wat Kuioh Hu shu, ' the writing of foreign nations,' which he says was introduced by the prince A-ma-kwei-me. Cf. Suh wen hien tuiig k'ao, bk. 134, f. 25. A. Wylie, Ts'inff tvan ic'e mung, p. xx. * Ancient Baddhist Inscription at Kcu-yung Kwan, I.e. p. 32. * Cf. Stan. Julien, Melanges de Geographic Asiatique, pp. 223-224. 82 BEGINNINGS OF WRITING. 120. The exact date of the adoption of the Syriac characters by the Uigurs is not recorded, to my knowledge ; but in the middle of the fifth century, consequently after the advance of this writing eastwards, we find the Uigur, as reported by Chinese authorities, in possession of a writing of their own, and translations from the Chinese of Mao's Shi King or ' Book of Odes,' Confucius's Lun-yu or ' Analects,' the Siao King or ' Book of Filial Piety,' besides some poets and historians.^ The Kiu-sze in 478 a.d., the Wei of northern China circ. 480, and the Tuh-kiueh hun in 518, were in possession of a writing which we can almost with certainty say to have been the, or derived from the, Syriac- Uigur characters.^ The Wei of northern China were staunch Buddhists from the beginning, and one of the Emperors of their dynasty, Hiao Wu, had a collection made of Buddhist books in a.d. 532,' in Chinese, after they had relinquished their former language and writing.* 121. We do not know how long after their establishment at Nisibis, the Nestorians went to Balk (Bactra), but it is from the latter seat that a Raban^ and his companions went to China in. a.d. 635,® where they set up at Si-ngan the celebrated inscription which was accidentally dug out in 1625, and is the most important monument of ancient • Bemusat, Eechenhes sur Us langues Tartares. p. 284. A. Wylie, O.C. p. 31. ' Cf. below, §§ 162-164. The unknown writing, which the Uigurs were supposed to have possessed previously to that derived from the Syriac, is a myth ; the hypothesis arose from not knowing of the early advance of the Nestoriau eastwards. 3 Bunyiu Nanjio, A Catalogue of the Chinese Translations of the Buddhist Tripitaka (Oxford, 1883, 4to.), p. xvii. * Probahly when they removed their capital city from Tchung-shan (Tchihli) to Lo-yang (Honan), a.d. 493. » Haban is rendered into the Chinese clumsy orthoepy hy 'Olopen. 6 It has been wrongly stated that Nestorians had reached China as early as a.b. 505. Ebedjesus Sobiensis remarks that "the Catholicos Salibazacha created the metropolitan sects of Sina and Samarcand, though some say they were constituted by Achans and Silas." Silas was Patriarch of the Nestorians from a.d. 505 to 520 ; and Achseus was Archbishop at Seleucia in 415. The Metropolitan Bishop of Sina is also mentioned in a list of those subject to this Patriarch, published by Anno, and it is placed in the list after that of India, according to the priority of foundation, (wells Williams, The Middle Kingdom, vol. ii. p. 276.) But this Sina near Samarcand was not China. It is a well-known region in Persian Uteratmre near Samarcand, afterwards Tchiniztan, and probably connected with Mount Tchino. Cf. my article on The Sinim of ladiah, not the Chinese, in the Babylonian and Oriental Record, January, 1887. BEGINNINGS OF WIIITING. §3 Christianity in Eastern Asia. In the Chinese Imperial decree^ of 638 authorizing their estatlishment at Tchang-'an, they were denominated Persian priests Po-sze seng, but subsequently in 745 another decree ^ had this name altered into that of Ta-Ts'in, and accordingly on the aforesaid inscription, they called themselves from Ta-Ts'in, the Eoman Orient.* The ^reat persecution of a.d. 845 was against the Nestorians at the same time as against the Buddhists; and by Imperial decree the priests of Ta-Ts'in and those of Mohufut (Mohommed) were laicised like the others. The wholesale destruction of temples and monasteries involved that of the Nestorians, so that the celebrated tablet of Si-ngan fu was buried in the ruins until its recovery by the Chinese them- selves in the seventeenth century.* It stands now among ' Quoted in the inscription with the alteration of Ta Ts'in for Po-sze, which occurred in the, original text as published in the T'aiig huei-yao of a.d. 961, bk. 49. - Published in the Tsih-fti-yuen-ktiei, a cyclopedia of a.d. 1013 in 1000 books. ' This carefully-studied conclusion of P. GaubU, P. Yisdelou, Deguignes iunior, Abel de Remusat, G. Pauthier, and others, has been established again by ±)r. P. Hirtb, China and the Roman Orient, Shanghai, 1885. * The number of papers and works written about it would make a library ; cf . their titles in Henri Cordier's Btblioteca Sniira, vol. i. The most important papers are those of G. Pauthier, Be I' Auihentieite de I' Imet-iption ycstoi-ieiine de 6i-ngan fou (Extr. of A}iiiolts de Fhilosophie Chretienne vols sr. xvi), Paris, 1857; L' inscription Syro-Chinoise de Si-ii'jan fou, monument Xestoi-ien eleie en Chine fan 781 de noire ere et dieouvert en 1625 (Paris, 1858, 8to.). Party spirit has several times blinded the mind of some writers on the matter, but the authen- ticity of the tablet is now beyond any possible doubt. The following notes, extracted from The Tinus in 1886, are the most recent contributions to the matter. From The Times, i February, 1886 : Sir, — I beg to say a few words about the Xestorian tablet at Si-ngan fa as a protest to the letter signed " G. VT.," in Th* Times of yesterday, which I find to be untrue iu facts and unfair in suggestions. AU the possible objections to the reliability of this inscription have been carefully sifted by several competent scholars, and the result of their careful researches is that "the weight of evidence, both internal and external, leaves no doubt regarding its verity. These are the very words used by the late Dr. ■VTeUs "Williams in his work " The Middle Kingdom" (revised edition, London, 1SS3, vol. ii. p. 277), where he has reprodnced in fuU the best translation ever made of the inscription, by Mr. A. Wylie, lat«ly of the London Misson at Shanghai, "who has gone over the whole subject with a fulness and care which leaves little to be desired." Dr. "Wall had certainly not made himself acquainted with the conditions of the case when he wrote in the work quoted by your correspondent his objections to the genuineness of the tablet. The doubts put forward have no foundation whatever, as I will try to show. The inscription was mentioned in the eleventh century by Min K'ieu, author of a description of Ch'ang-ngan, name of a part of Si-ng»n. Far from being lost, the original stone, after its recovery in 1625, was carefully preserved, and the 84 BEGINNINGS OF WRITING. the ruins brought out by the late Mahommedan rebellion of 1859, as a witness of an early advance of Christianity in the great Imperial geography published at Pekin in 1744 states that it was kept in the Kin-shing monastery. As to the objectionable references in the text, they have proTed to be mis- translations of the early translator, P. Boym (" Briefve relation de la convention notable des personnes royales et de la religion Chrestienne en la Chine, ' ' Paris, 1654), whose inexact work was reproduced in P. Kircher's "China lUustrata." In order to see how far the translation of P. Boym was defective, it will be sufficient to compare it with that of Mr. A. Wylie, the best Chinese scholar who has ever lived. The style of the inscription is very terse, and the exact meaning not easily perceived even by learned natives. An exact translation required a scholar much more competent than P. Boym. It is not astonishing that Dr. Wall should have reproduced the ancient objection, based upon the lack of archaism in the characters of the Chinese text. 'i his was said when the history of the Chinese writing was not known, and ought not to be repeated now-a-days. The current writing of China (heng-shu), as every Sinologist knows, has hardly been modified since the latter part of the fifth century. A decisive proof that the writing of the Nestorian tablet has not been tampered with, and is genuinely that of the eighth century, is happily at hand for every one anxious to verify the point. A facsimile of this famous inscription, erected in 781 a.d., is reproduced from a rubbing in Colonel Vule's monumental work, "The Book of Ser Marco Polo," second edition, vol. ii. p. 22. On the other hand facsimiles from rubbings of a Chinese and Tibetan inscription set up at Lhasa in 822 a.d. — viz. 39 years only after the Chinese and Syriac tablet of Si-ngan Fu — are reproduced in Dr. S. W. Bushell's paper on " The Early History of Tibet," in the Journal of the Royal Asiatic Society, new series, vol. xii. 1880, p. 534. A comparison of the Chinese characters in the two inscriptions proves them exactly similar, at large and in their small variants with the writing of modem times. The Nestorian tablet was discovered in 162.5, when Tsen Tsing-chang, prefect of Si-ngan Fu, had a tomb made, south of the Tsung-jin monastery, for his son just deceased. And it was only three years afterwards, in 1628, that it was seen for the first time by the Jesuit Alvarez Semedo. The habit of making casts and rubbings of their antiquities and inscriptions, for the antiquaries and students, is very old in China. The process commends itself, since the accuracy of these reproductions is complete, as there is in it no possibility of any personal fancies or unknowing, nor of any alteration of characters or texts. The making of a copy of the Nestorian tablet after its discovery, to be shown away, was in accordance with the usual practice. The imputation made seriously in your correspondent's letter, that the Chinese part of the inscripton was altered by some Jesuits in favour of their creed, cannot be read without protest and contempt. The facts are against its possibility. On the other hand, the learned and distinguished Jesuits who carried the standai'd of Christianity into China were men of high character and morality, utterly incapable of conceiving or helping any base action. We have only too often occasion of seeing how often the prejudices of a missionary mind hinder the scientific independence and criticism of men otherwise distinguished and learned. But there is an abyss between interpreting and explaining, with a partial mind, an inscription or text difficult to understand, and committing a forgery such as that which is ^atuitously supposed. The earnestness and candour with which the Jesuit missionaries have laboured in China is patent to any one who has read their works. I hope that the eloquent appeal of Mr. Frederic H. Balfour (The Times, January 21, 1886) will not remain unheard, and that steps will be taken by the Foreign Office to obtain the stone of Si-ngan Fu. It is a most precious monument, imique in its kind, which ought to take a place among the Measures of the British Museum. — T. de L. BEGINNINGS OF WRITING. 85 Middle Kingdom. The Syriac characters are inscribed on it in perpendicular columns from left to right. 122. Some important finds of inscribed tombstones in the province of Semiretch, north of Khokand, the ancient region From The Times, 1 September, 1886; — I have read with great pleasure your letter of January 30 in The Times upon the subject of the Nestoriau tablet at Si-ngau Fu, in Shensi, and fully concur with you in the conviction that beyond doubt it is genuine. It has been my privilege to travel a great deal in North China. I have twice visited Si-ngan Fu, and have cai-efuUy examined the tablet in question. The stone itself is remarkable, unlike that of any other I could find in the neighbourhood, and must have been brought from some considerable distance. There need he no difficulty in obtaining possession of the stone. No value is set upon it by the natives or by the Government, and in a few years it vrill be again lost or broken up. It is standing a sUent and solitary witness of God's truth, upon a heap of the ruins of temples of many generations, without a covering of any kind, exposed to every wind and storm that blows. — A. G. Parrott. Si-ngan Fu, Shensi (China), June 14. — • Coming up the Han river some two months ago, I read your interesting letter in The Times Weekly Edition of February 5 on the Nestoriau tablet. My missionary duties bringing me to this historic part of China, I made my 'pilgrimage' to the Nestorian tablet the day before yesterday, before even entering the city. It is, as you know, situated outside the west gate, in the grounds of the great Tsung-jin Monastery. This monastery was burnt about twenty-three years ago, and only a few rooms remain of what must have been a magnificent series of buildings. My feelings of awe and delight, in the presence of this famous memorial to the labours of devoted men, were soon replaced with an indescribable feeling of horror and shame that this precious monument should be thus rudely and recklessly exposed to the elements and any stray ' rough ' who cared to damage it. The old man in the only remaining building of the Tsung-jin Monastery told me it had been so exposed ever since the destruction of the monastery in 1863. So fai' it is in a remarkably good condition, considering this exposure, though there are unmistakable marks of decay and ill-usage. I have procured a recent rubbiu" and take the liberty of asking your Mnd acceptance of it. It will show fairly the present condition of the tablet. The two side bits represent the thick- ness of the stone, and, of course, the engravings also. Tou have every character and letter in the rubbing. From a cursory examination I am inclined to think that it is not stone, but moulded of a very fine and peculisir kind of earth and well fired after the manner of the best Chinese bricks. I tried to cut it at the back with a knife, and found it soft, which rather confirmed this theory. It stands on a stone tortoise. It is, moreover, so isolated that any person might go and batter it to pieces without risk of the sound reaching any habitation. That this world-famous tablet should be exposed as it is to either gradual or sudden destruction is a crying shame and a disgrace t^o the nineteenth century. I most heartily and sincerely hope with you that our rulers will see the wisdom and the urgent necessity for bringing their good offices to bear on the high authorities in Pekin, to induce them to let this ancient and unique monument pass from exposure in the open air to wind, rain, sun, frost, and vandalism into the safe keeping of the British Museum. If this cannot be done, surely something might be done by the distinguished representatives of Western nations in Pekiii, to induce the Chinese Government to put up a house to cover the tablet, and otherwise make adequate provision for its preservation. It would be deplorable if this valuable memorial of Nestorian Christianity in China should, through sheer neglect and carelessness, be lost to futme ages. — JoHX W. SiEVENSoif, Chiua Inland Mission. 86 BEGINNINGS OF 'WEITING. of Ferganah/ which, studied by Dr. Chwolson ^ of St. Peters- burg, have proved to be those of Christian Nestorians. The inscriptions in the Syro-Nestorian characters, written in perpendicular and also in horizontal lines, so as to make the four sides of a square around the flowered pattern of the Nestorian cross, are dated from the years 1169 to 1648 of the Seleucidian era, corresponding to a.d. 858 to 1237. These dates show a persistence of activity of this writing in these regions much later than was hitherto supposed. 123. This twofold arrangement of the Syriac characters in perpendicular lines from left to right, and in horizontal lines, was known of old by the Syriac scribes. As the grammar- ians inform us, they were accustomed to write from the top to the bottom of the page, the writing being turned round into the usual position in order to be read. This habit was not relinquished by them before the thirteenth century. The fact is stated, but the reason is not given by those ' who have investigated the matter. It may have arisen from the • The following, which I extract from a daily paper, refers to it (2 Sept. 1886) : The Vossisohe Zeitung says : "At the beginning of this year a discovery was made at Semirjetsche, in the district of Fergana, west of the Chinese frontier at Kuldja and north-east of Kokand, which is of great interest to historians and Orientalists. It was near the ruins of a fortress called Burana, buUt of bricks, and situated on the spurs of the Alexander Mountain chain, that Dr. Porjakow found two old decayed graveyards with numerous gravestones, many of which bore a cross and others cmselled inscriptions. They could not be deciphered on the spot, both characters and language appearing to be quite unknown. The Catholic mission- aries in Kuldja believed the characters to be Uigurian, whUe the Lama of the Kalmucks believed he recognized in one of the inscriptions the Buddhist form of prayer, om ma horn. It was determined to take several tombstones, fourteen photographs, and eight rather imperfect copies of the inscriptions, to St. Peters- burg, where they were handed to the Councillor of State Chwolson, who soon declared the character to be Syriac-Nestorian. It was difficult to decipher, for, besides the age and decay of the stones and inscriptions, they contained new italics not usual in Old Syriac, and also quite foreign letters, probably taken from the Arabian alphabet, and, very unexpectedly, some Turkish words and proper names. This having been found out, the deciphering proceeded rapidly, and Professor Chwolson sent a report to the Eoyal Academy of Sciences in St. Petersburg. The inscriptions have primarily a palaiographio mterest, and are further remarkable on account of their peculiar grammatical forms and turns of speech. ' ' '' Syrische Orabinschriften aus Semirjetschie, by Dr. Chwolson {Memoires de . V Aeademie Imper. des Sciences de St. Petersbourg, 1886, Vile ser. tom. xxxiv., No. 4), 30 pp. and plate. Dr. Chwolson has read at the Seventh Oriental Congress at Vienna, 1 Oct., 1886, a paper TJeber die NordostUeh von Kokand {Fergana) gefundenen syrisch nestorianischen Grabschriften, and exhibited several of these tombstones. ' As for instance Dr. Isaac Taylor, quoting Bayer, in his work The Alphabet, vol. i. p. 305. BEGINNINGS OF "WRITING. 87 necessity of writing Greek and Syriac together. The first running from left to right, and the second from right to left, the only means of making their letters correspond, as for instance in the case of proper names, was to write the Syriac in columns; this process could not be objectionable to the habits of people reading from right to left, to the same extent as putting these letters contrariwise to their usual arrangement. An instance which occurs in a MS. of the sixth century, and now in the British Museum, and reproduced in Dr. I. Taylor's book on the Alphabet, would agree with this explanation. The disposition of the writing in columns from left to right, occasionally on the funeral inscriptions of Fergana, and regularly on the tablet of Si-ngan fu, shows that this habit was in current use, and could not have come from Chinese influence, as was formerly supposed.^ The following diagrams explain the matter : Greek. Original Sjrriao. Later Syfiao. 12 3 4 Chinese. 4 3 2 1 1- 2- 3- 4- -> ^ -1 -2 -3 -4 124. The Jacobite Christians^ of Aleppo have preserved a survival of the ancient Syriac alphabet, but they have adopted some special signs in place of the primitive Syriac system of points. These signs are no less than superscribed Greek vowels which now appear to be laid aside, simply because at the time of their adoption the Syriac letters were written in perpendicular lines.* 1 Dr. Taylor, in his above-qnoted work (toI. i. p. 304), has shown the impos- sihiHty of this explanation of Bemusat, which would hare required the succession of the columns ot writing from right to left as in Chinese, while it is the reverse with the Sjniac writing. * Their alphabet differs little from that of the Maronite of the Lebanon, and goes by the names of Modern Syriac, Peshito {i.e. current), or Serta {i.e. Enear). The latter name was probably so called (as suggested by Hoffmann, says I. Taylor, T7ii' Alphaiel, toI. i. 286) because of the characteristic horizontal line or ligature which unites the lower portions of the letters, and thus distinguishes it from the Estranghelo or " rounded characters." It must be remembered, however, that this ligature or straight line is found already on the Xestoriau tablet of Si-ngan fu in A.D. 781, as well as on the tombstones of Semirjetche of a.d. 858. s Cf. I. Taylor, T/ie Alphabet, vol. i. pp. 290, 296, and 306. 88 BEGINNINGS OF WRITING 125. The Syriac alphabet, through its immediate derivate the Uigur, where the letters useless to the limited phonesis of the latter language, such as the distinctions between surds and sonants, have been neglected, has given birth to the Mongol, Mandchu, and Kalmuk writings. The history of their successive systematizations has been written several times, and we need r^ot give here more than a short sketch. 126. About A.D. 1200,' the Mongols, under the leadership of Jingis Khan, after having defeated the Naimans near the Yenissei river, were taught the TJigur writing by Tatatonggo, formerly Secretary of the Naiman prince, who was a Uigur by birth. The practice continued to prevail during the supremacy of Qgdai, Guyu, and Manggu, the three successors of Jingis Khan. Ruybruck, who went on an embassy to the Mongolian court, from France, in 1263, speaking of the Uigur character, says : " Mangu Khan has sent to your Majesty (Louis IX.) letters written in the Mongol or Tartar language, but in Uigur characters." ^ Under the succeeding monarch, Kublai Khan, the inadequacy of the Uigur character fully to express the articulations and vocables of the Mongol language, were felt to be a source of embarrass- ment and difficulty. 127. A venerable Tibetan Lama Sadja Pandita, who had been invited,^ in 1244, to come to the court, where he arrived in 1247, and remained until his death in 1251, is attributed by the Mongol Sagas, the invention of their first alphabet. They tell us that while he was pondering over the matter, he one night had a dream, in which he was told to fashion the letters after the first object he should see after he awoke. This happened to be a woman carrying a notched stick over her shoulder. He thereupon constructed an alphabet and 1 Of. H. Howorth, Sistory of the Mongols, vol. i. p. 65. ^ Of. J. Klaproth, Abhandlwng iiber die Sprache und Schrift der Tliguren, p. 56. A. Wylie, Ancient Buddhist Inscriptions, I.e. p. 33. ' By Donda, a brother of Kublai Khan, sent on a mission to Tibet. Klaproth {I.e.) says that Sadja Pandita occupied the office of High Priest of Lamaism for seven years at the Mongolian Court, counting from 1244 to 1251, but he has forgotten that the Lama was three years on his journey. BEGINNINGS OF ■WRITING. 89 formed a set of perpendicular lineal characters^ for the letters a, e, i; na, ne, ni; ha, he, hi; Ma, he, Id; ga, ge, gi; ma, me, mi ; la, le, Ji ; ra, re, ri ; sa, se, si ; da, de, di ; ta, te, ti ; ya, ye, yi ; tsa, fse, chi ; dsa, dse, gi ; tea, ice? The system was very defective, though Sadja Pandita had preserved the fourteen consonants ; but he died before having completed the work in which he was still engaged. None of the religions writings could be translated with that system/ and the ancient Uigur translations of the Buddhist books were stni resorted to. It is this unsuccess which caused the formation of the Bagspa character which we have recorded.^ 128. In 1304, Eldja'itu, successor to Kublai, commissioned the priest Tsordji Osir, a relative of Sadja Pandita, to trans- late the Tibetan sacred books in the Mongolian language and the Bagspa characters. But the difficulties proved insuper- able, and Tsordji Osir was compelled to revert to the un- finished work of Sadja Pandita. He made such additions to it as were needful for his purpose, and wrote out with it a Mongol translation of the Tibetan work Bangcha Rakcha, but found it necessary to express a great many words in the Tibetan characters. The two letters addressed to Philip the Fair of France, the first by Argun, the Mongol prince of Persia in 1289, and the second by Eldjai'tu, his successor in 1305, were written at these stages of the writing. Under the direction of Kaisun KuHuk, the succeeding Emperor, who reigned from 1307 to 1311, Tsordji Osir made still further improvements in the writing. His various additions consisted of the vowels o, u, 6 and ji, the consonants sha, sa, dsi, and pu and the finals », p, k, I, m, r, t, i, u and ng. He retained the Tibetan numeral figures unchanged. And at last a proper system for transcribing the Mongol sounds was definitively obtained.' The number of simple ' Cf. J. Eaproth, VerzHchniss der chinesisehen und mandshuischen Bueher der Konigliehm Bibliotheh :u Berlin (Paris, 1882, fol.) ; Archives Faleographiques ; De I'eeiiture Oiiigoure, vol. i. p. 87 (Paris, 1872). - Cf. Howorth, History of the Mongols, vol. i. p. 506. 5 But it appears on the coins of Abaga of Persia in a.d. 1279. « Cf. above §110. * All these details are given not only in the Chinese history of the Yuen dynasty, and also in a Mongol work, Brulba Seuidja Bandida yia gargaksan Mongol usuk, published in China in 1730. 90 BEGINNINGS OF WRITING. characters was 30, making 127 compounds arranged in 12 classes according to their terminations; they consist of 7 vowels, 6 diphthongs, and 17 consonants, each of them having final, medial, and initial forms. 129. The Mongol alphabet has produced two derivates, the Mandshu and the Kalmuck. In 1599 A.D. Tai Tsu Dergi, the first Mandshu chieftain who assumed the dignity of Emperor, and the predecessor but one of Ichishsu dasan or Shitsu eltembughe, the founder of the present dynasty of China, commissioned Erdeni Bakshi, a scholar well skilled in the Mongolian literature, and Gagai Dsharguzi, the high minister, to invent a system of writing for the Mandshus. Both demurred from the difficulty of the task, objecting that the Mongol characters were too old to be altered, and that inventing new characters was not possible. The emperor overruled the first objection by remarking : " There is no difficulty ; merely apply the Mongolian characters to the sounds of our language, and connect the syllables together to form words ; thus the written characters will show the meaning." Soon after Gagai was beheaded for political ofiences, and Erdeni alone remained in charge of the work of forming the characters ; when this task was accomplished, the emperor caused the alphabet constructed to be put into circulation, from which time the native character dates its origin. After the death of Erdeni Bakshi the superintendence of this writing was committed to Dakai Bakshi, a literary officer of the Blue banner, who, together with Kurshen and others, removed difficulties and made additions, arranging the whole under twelve classes.^ The earliest coins of the Mandshus (period T'ien-ming, 1616-1626) bear a superscription in that stage of their writing. From 1632 to 1641, Dakai undertook a thorough revision of the characters,^ added points and dia- critical marks as well as supplementary classes of characters.' 1 Of. Man-tcTiou ming tehin tchuen, Bk. 8, f. 28 : A. Wylie, Translation of the Ts'ing wan Tcimimg; pp. xxvi-xxix. * Of. Man-tchou ming tchin tchuen, Bk. 3, f, 21 ; ihid. p. xxix. ' The Emperor, addressing Dakai, said: "Hitherto our twelve classes of characters have not been marked with rings or points, and the natural divisions BEGINNINGS OF WRITING. 9I Thus augmented the writing consisted of 38 simple characters, 29 compounds forming nearly 1300 syllables, and from that time to the present there has been no change of importance in the form of the characters. In 1699, the K'ang-hi Emperor caused a stone tablet to be erected to the memory of Dakai Bakshi, on which a record of his labours is set forth. 130. The Kalmiicks or Eleut in their migration westwards from their old home in Sungaria to the banks of the lower Volga in 1616 a.d., brought with them a simplified form of the Mongol alphabet, composed of 14 consonants and seven vowels, making together 105 sound-syllables. ^ The great majority of the people migrated again eastwards, returning to China in 1770 a.d. ; ^ but some of them remained in the region of Astrakan, where the aforesaid characters are still employed for the preservation of some fragmentary remains of Buddhist literature. All these writings are written in vertical columns from top to bottom, arranged from left to right. That disposition arose from the original arrangement of the Syriac characters turned upwards, as shown by the diagrams of the original and later Syriac given above (§ 123). d) The Lepcha or Rang Writing. 131. The Lepcha of Sikhim, east of Nepaul and south of Tibet, have an alphabet whose character Csoma de Koros have been run together without distinction. Now a youth who has studied this writing, when he merely meets with ordinary words in it, understands them without difficulty ; hut when proper names occur, he is sure to faU into error. Do you now invent a system of rings and points, in order to make the distinctions, so that the characters may clearly express the sense and sound, and the study may be attended with greater advantage." Dakai thereupon applied himself to follow out the emperor's views, and arranged the points and guttural marks now in use. Besides this, finding the characters of the twelve classes insufficient to express all the sounds of the Chinese, he added a number extra. StUl even with these additions it was found that some words could not be accurately represented ; whereupon he resorted to the expedient of running two syllables mto one, after the manner of the Chinese fan-tsieh syllabic spelling, and this was found to be more efficient for the Mandshu than the Chinese. A. WyKe, ibid. ' Cf. B. Bergmann, Nomadisehe Streifereien unter der Ealmiickm. The syllabary is reproduced in Faulmann's Illuatrirte Geschiehte der Schrift, p. 327. ^ On these migrations westwards and eastwards, ef. H. Howorth, History of the Mongols, vol. i. pp. 561-579. And also M. Gueluy, La Chine oceidentale, Mceurs et histoire par un voyageur, traduit du chinois, in le Musfion, AvrU, 1886, vol. V. pp. 238-246. 92 BEGINNINGS OF WEITING. pronounced to be not Tibetan. Its Indian deriTation cannot be doubtful, though its genealogy has yet to be established. The arrangement and phonetic peculiarities are Nagari, but the shapes of the characters seem peculiar. I have carefully examined it and think that its strangeness will disappear before any observer as it did to me, when I considered the characters as being sideways, i.e. laid down on the left. This shows that the writing which is now drawn from left to right, was formerly written from top to bottom in the Chinese and Kalmuk-Mongol-Mandshu fashion. The dia- critical marks are in their right place according to the Indian standard ; but the characters individually have remained as they were when written one under the other. The prototype seems to be the same as that of the Tibetan,' viz. the Nagari of the fifth century, as shown by the Allahabad inscriptions. 132. This peculiarity is not unimportant for the history of writing in that country, and may lead to further disclosures. Applied to its language, the Lepcha writing is phonetic ; there is no important discrepancy between the transcription and pronunciation, no gap between the written and spoken language, as in the case of the Tibetan, Burmese, and Siamese, of which we have ancient written texts. There is nothing conventional in the reading of the characters, so that there is sufficient ground to infer that the people have not long been in possession of the writing, unless a complete reform should have taken place at a particular time : for instance, when the writing was written horizontally instead of in its once perpendicular direction, still preserved by, and shown in the lying down to the left, of the shapes of the characters, otherwise similar to the Indian standard from which they were derived. Now there is no indication of any such literally revolution among them, and the small number of traditions collected by Dr. Campbell and Col. Mainwaring ^ does not justify any idea that the people ever 1 Vid. I 96 atoTe. 2 Dr. Campbell, Journ. Seng. Asiat. Soo. 1840, vol. ix. Col. 6. B. Main- waring, A Grammar of the Bong (^Lepcha) Language, as it exists in the Darjeling and Sikim Mills (Calcutta, 1876, 4to.). BEGINNINGS OF -WIIITING. 93 was in so high a state of culture that such a thing could have happened. They must have obtained their writing in recent times, and I should say that the five hundred years given as the most remote date of their national existence are more than required under the circumstances of the case. 133. They cannot have obtained it direct from the Tibetan. The former transitory condition of the writing precludes such a possibility, though it undoubtedly belongs to the same stem. They must therefore have learnt it previous to their migration to Sikkim, at the time when they were wandering or established in the south-east of Tibet. Future travellers will at no distant date be able to find out whether any writing of this description is still lingering in these parts, and to discover traces of its former existence. e) The Battak of Sumatra. 134. The Battaks, the lettered cannibals of Sumatra, have a system of writing disposed in vertical columns from bottom to top and then from left to right.^ This peculiarity gives a strange appearance to the writing, which is undoubtedly a degraded type of Indian descent, through the old Kawi of Java. But if we replacp the Battak writing in the original position of its graphic ascendants, namely, from /I\ /^ /^ /|\ 1 > present arrangement 2 to 3 ]^ turned up to the which is the same up lelt, 4 > 12 3 4 the peculiarities disappear, and their reason for having adopted their present curious process is explained by the material they use to write upon. It consists of long strips of bamboo, welded by beating one to the other, and then folded in a fan- like manner ; they place the latter between two wooden boards and bind the whole together with a string of woven rushes. So that a writing on their strips of bamboo from • There are four vaiieties : MandeUng, Angkola, Toba, Dasri, in K. F. HoUe, Tahel van Oud- and Nieuw-Indisehe Alphahetten (Batavia, 1882, 8vo.), Nos. 130, 131, 132, 133. Other names are also sometimes quoted, such as Pakpak, Ziagkal, by Prof. A. H. Keane, in A. E. Wallace's Auatralasia, p. 630. 94 BEGINNINGS OF WRITING. left to right, eacli line under the other as long as is necessary or as the strip goes, looks when the strip is folded as if dis- posed in columns to be read from top to bottom, and then from left to right. This is apparently the original explana- tion of the above peculiarity.^ f) A Vertical Writing in Ceylon ? 135. Diodorus Siculus, who was a contemporary with Augustus, has recorded in his Biblioteca historica a some- what marvellous account of a journey in the East undertaken by lambulus, son of a Greek merchant. Amongst other descriptions, there is a lengthy one of the manners and customs of the inhabitants of an island in the south : " The inhabitants study all sciences and especially astrology ; their alphabet is made up of seven characters whose value is equivalent to twenty-eight letters, as each primitive character can be modified in four different ways ; their writing is drawn, not horizontally like ours, but vertically from top to bottom."^ The substratum of truth which underlies this statement is difficult to determine. The identification of the island spoken of is not ascertained definitively ; there are many good reasons for belief that it is Ceylon, and the authorities in favour of this view are numerous.' On the other hand, Lassen understood it as being Bali,* and the minor ■authority of Wilford is in favour of Sumatra.^ Tennent said that in the pretended account of the island, he could not discover a single attribute sufficient to identify it with Ceylon, the existence of which had been known to the Greeks three hundred years before.® ^ " The present mode of -writing the Battak is hy scratching it with a kreis on a hamhoo, beginning at the bottom and writing perpendicularly upwards, in direct opposition to the Chinese mode. Marsden has published the radicals, but by mistake placed them sideways horizontally, instead of perpendicular." " Cf. Diodor. Sicul. Biblioteca historica, Hb. ii. cap. 57. ^ Cf. Vincent's Periplus of the Erythrean Sea. * Indische Alterthumsktmde, vol. lii. p. 270. * Essay on the Sacred Isles of the West, in Asiatic Researches, toI. x. p. 150 ; Eamusio, vol. i. p. 176, held the same opinion. * " On the contrary, the traits which he narrates of the country and its inhabit- ants, when they are not manifest inventions, are obviously borrowed from the descriptions of the continent of India given by Ctesias and Megaathenes,' ' — Ceylon, by Sir James Emerson Tennant, 2nd edit. vol. i. p. 631-532, n. BEGINNINGS OF WEITING. 95 136. But if there are still some doubts as to the island spoken of, the information about the writing is certainly- genuine, and concerns a system of Indian origia. Prinsep has said, " It would be difficult to describe the condition of the Indian alphabetical system more accurately than lam- bulus has done in this short summary, which proves to be not only true in the general sense of the classification of letters, but exact as to the origin and formation of the symbols." ^ The numbers, however, do not agree, except if we take the seven letters as representing the seven vargas or divisions of the Nagari alphabet ; the four modifications are the vocalic marks ; but the direction again disagrees, as all the Indian writings run from left to right, so that it may be a writing derived from the Indian stem, but simplified to the fewest possible symbols, and modified in its arrangement by some occasional adventure, as has happened with the Battak writing. YII. Lost, Foegotten, and Recovered Writings. a) In 8.W. Asia. 137. The decipherment of the hieroglyphics of Egypt, and those of the various writings in wedge characters of Babylonia, Assyria, Van, Suza, Media, Persia, which achieved successively since the beginning of the present century, have enlarged beyond any suspected limits the field of history, are likely to be followed, before we reach the beginning of the following century, by some more decipherments of other hieroglyphico-phonetic writings from America, and also in S.W. Asia. The Hittite characters in IT-W. Syria are the most important item of hieroglyphics of the old world now under the considera- tion of a few scholars, and still unread. They were the monu- mental and official means of expression of a once powerful 1 J. Prinsep, in Joum. Asiat. Soc. Bengal, 1837, vol. ri. p. 476. The lamented Dr. Bumell (Elemmts of South Indian Pnloioffraphy, 2nd edit. London, 18,8, 4to.), says, p. 48, n. 2: "I pass over the statement of lambulus, as it is im- possible to explain it by any Indian alphabet as yet known." 96 BEGINNINGS OF "WEITING. empire 1 or agglomeration of states, which were once a match for the Egyptian power ; but the exact dates of their beginnings are not yet known. The great philologist of Oxford, Prof. A. H. Sayce, is the only scholar who hitherto has scientifically attacked their decipherment, and who has been able to recover the value of a few symbols with the help of the Cypriote syllabary, which he thinks, rightly in my opinion, to have been derived from these hieroglyphics. Any great advance in the decipherment is dependent on the finding of some more important bilingual texts than those which have hitherto come to light.^ There are discrepancies between the monuments known hitherto which probably correspond to some difference in the language and a branch- ing off in the writing.' 138. On a small stone inscription of Rim-aku, circa 2200 B.c ,* about the middle of the reverse of the stone, are some curious graphitti, which the late Dr. Birch supposed to be Chinese. I cannot, however, by any possibility identify the combination of characters they represent. But the family resemblance is undeniable, and is easily explained away by my former discovery, that the oldest Chinese writing was derived from a current form of the Babylonian wedge- writing, minus the wedge. 139. The Asianic syllabary is one of the reconstructions still unachieved, due to the genius and acuteness of the 1 The first sketoh was disclosed by Prof. A. H. Sayce, A Forgotten Empire in Asia Minor, Aug. 1880 {Eraser's Magazine, pp. 223-233). The late Fran^pis Lenormant was engaged in a thorough investigation of their historical and geographical remains when an untimely death stopped his pen. Cf. his Origines de Vhistoire, vol. ii. part ii. Paris, 1884, pp. 267-379. * Cf. A. H. Sayce, The Monuments of the Sittites, in Trans. Soc. Bihl. Arch. vol. vii. pp. 248-293 ; The Bilingual Hiiliie and Cuneiform Inscriptions of Tar- knndemos, ibid. pp. 294-308 ; and many notes in the Academy. WUHam "Wright, The Empire of the Sittites (London, 1885, 2nd edit.), where Prof. A. H. Sayce has contributed an important chapter on the decipherment, and Mr. "W. H. Eylands a nearly complete set of the inscriptions. Cf. also A. Amiaud, Simple coup d'aeil sur la tulle de lovanoff et sur les inscriptions h^t^ennes (in Zeitschrifl far Assyriologie, August, 1886, Leipzig), pp. 274-288. ' A new haematite seal, dug out at Yuz'aSd in Asia Minor, has just been acquired by the British Museum. Should its genuineness be proved, it exhibits at the same time similarities and divergences with the Hittite characters hitherto known. A description of it has appeared in The Times (9 Nov. 1886), and an. autotype facsimile published by the Soc. Bibl. Arch., Proc. 2 Nov., 1886, p. 27. 4 T. G. Pinches, MS. note, 8th May, 1885. BEGINNINGS OF WEITING. 97 late Fran9ois Lenormant and of Professor A. H. Sayce. The lamented French scholar had sagaciously suggested ^ that the Cypriote syllabary and the non-Hellenic elements in the Lycian and Carian alphabets would ultimately prove to be derived from an ancient graphic system common to the peoples of Cyprus and Asia Minor. The Cappadocian and Pamphylian are some more of these mixed alphabets, of Greek and Asianic characters ; ^ the latter being all derived from a common source, a syllabic writing evidently of im- mense antiquity, which prevailed throughout Asia Minor. Prof. Sayce has shown reasons to believe that the characters inscribed on the Trojan whorls found in the lower stratum at Hissarlik belong to the same syllabary.' And it may be stated that he has established beyond any doubt, by means of a number of characteristic instances, that the Asianic syllabary was derived from the hieroglyphic writing of the Hittites.* 140. An unknown writing, which seems to fill up a much- wanted gap, has been found' in 1882 on a contract clay- tablet of Babylon, dated in the 23rd year of Artakshatsu (Artaxerxes). Besides twenty- three lines of cuneiform writ- ing, it contains one line and several strips of unknown characters, more or less clearly traced, some forty in number. Their characteristics are some rounded forms and visible combinations of signs, two traits similar to the Indo-Bactrian alphabet ^ in contradistinction to the pointed forms and in- * Essai sur la propagation de V alphabet Fhenicien dans I'Ancien monde, Tol. i. p. 107. 2 Dr. Isaac Taylor, T/ie Alphabet, vol. ii. p. 115. * Appendix to Dr. Schliemann's Ilios. 1880. * Of. A. H. Sayce, in Trans. Soc. Siil. Archesol. vol. v., P)-ineiples of Com- parative Philology (3rd edit. London, 1886), The Asianic Syllabary, App. it. ; and the Comparative Table in Taylor's The Alphabet, vol. ii. p. 123 ; also Prof. A. H. Sayce, in Dr. W. Wright's Umpire of ilte Eittites (London, 1885, 8vo.), pp. 168-188. 5 By Mr. T. G. Pinches, of the British Museum, who then showed it to me. He has since published a facsimile of the whole tablet, and a translation of the Cuneiform text, in Troe. Soe. Bibl. ArehtBology, 3 April, 1883. ^ It is worth noticing that the practice of monograms, where the principle of combination is carried to the utmost, by the pouring in of several characters into one, did not begin in Greece before the Persian period. I am indebted to Prof. Percy Gardner, of Cambridge and the British Museum, for the following note : "The earliest monogram, so far as 1 know, to be found on a Greek coin is 98 BEGINNINGS OP WEITING. dividual isolation of the characters in the Aramsean alphabets. Some of the characters on the Babylonian tablet can be easily enough connected with the Indo-Bactrian letters and com- binations.^ From an attempt at decipherment which I made in March, 1882, with the two-fold aid of the Aramaean and Indo-Bactrian letters, and which disclosed notably words corresponding to several of the names of the witnesses men- tioned in the cuneiform text, I thought and still think ^ that further investigation will show that this unknown writing furnishes the wanted link between the recognized Semitic basis of the Indo-Bactrian writing and its western antecedent. 141. A round and cursive writing, which reminds one of the Indo-Bactrian, and of which neither the ancestry nor the progeny are known, is that which occurs on the Sabsean coins from the fifth century B.C. downwards.^ These coins are trilingual, Greek, Musnad, and the writing unknown. Putting aside the Greek letters A E, which are reproduced in imitation of the Athenian originals, there remain two genuine legends.* The Musnad characters, also called Him- yaritic, give two words, y-n-p and %-l-d, which I have recovered in the unknown letters with the help of the Indo- Bactrian alphabet, to which they are evidently related. The ZE (= E A) on a coin of Edessa, of about B.C. 480 (see Cat. Gr. Coins, Maeedm, p. 37). But for a century after this they are very rare, and do not become usual until the time of Alexander the Great." 1 The late Dr. A. Burnell, in his last days, was made acquainted with this un- known writing by a rude tracing, which he was unable to study properly, and from which he fancied a connection with the Indo-Pali characters (The Aeademii, No. 528, 17 June, 1882), an opinion in which I cannot concur. "We shall have to study the question more fully below. 2 Cf. The Oldest Book of the Chinese and its Authors, § 25 u.. Jour. Soy. Jsiat. Soc. Oct. 1882, Vol. XIV. p. 803. * Cf. Barclay V. Head, Eimyarite and other Arabian Imitations of Athenian Coins [Numismatic Chronicle, N.s. 1878, vol. xviii. pp. 273-284) ; On a Mim- yaritic Tetradrachm and the Triaor de Sand. (ibid. 1880, vol. XX. pp. 303-310). TV. P. Prideaux, On some Recent Discoveries in South-Western Arabia {Trans. Soc. Bibl. Arch. vol. i. 1873). G. Schlumberger, Le Tresor de Sand (Monnaies himyaritiques), Paris, 1880, 4to. The most important work on the history of the region is that of Dr. D. H. Miiller, of Vienna, Buryen und Schlosser SUdarabiens parts i. and ii. * Dr. Mordtmann, in his Neue himyarische MUnzen, p. 299 {Numismatische Zeitsohrift, Wien, 1881), has proposed a decipherment with the aid of the early Pehlvi in reading the legend topsy-turvy ! Besides the incongruity of this pro- cess, there is an insuperable objection to read an inscription with a writing of a distant country seven centuries afterwards. BEGINNINGS OF WRITING. 99 matter requires more extended explanations and a distinct enquiry, which space here precludes.^ 142. In a manuscript of 1535 a.d., twenty-one letters from a forgotten writing once proper to the Albanians have lately been discovered by Dr. N. Karamianz,^ who has identified them with as many letters of the Minuscules Armenian. b) Bactrian and Chinese. 143. An interesting alliance of writings is that which is exhibited by the bilingual Bactrian- Chinese mintage,' of which a specimen, or rather two specimens (round pieces), were brought back by Sir Douglas Forsyth, from his ex- pedition to Yarkand.* These copper coins he procured in the town of Khotan, together with some gold rings of Indian type. The meeting of the Indo-Bactrian writing of Aramasan descent with the Chinese symbols, on the same money, and in certain conditions, would involve a larger study than space can permit here. "We know that the Chinese had extended their political power in the basin of the Tarym, and west of the Tsung-ling range over a large number of small states, towards the beginning of the first century b.c. But this power and influence do not seem to have been great enough to justify anywhere an issue of bilingual money. And should this influence have been strong enough, the Chinese residents could not have introduced another design of money than that which was then the current and standard money of the Han dynasty, namely, the one-inch, round, square-holed pieces, with the legend Wu tchu, i.e. "five tchus." The ' M. J. Halevj {Journal Asiatique, 14 Jan. ISSl) has remarked that the coins must be called Sabean and not Himyaritic, as they belong to a period previous to the Himyaritic d5-nasty. Canssin de Perceval {Mssai sur Thistoire dts Aiabes aeaiit I'islmiiisme, vol. i. p. 63) says that this dynasty began in the first century before the Christian era. Cf. alsoFr. Lenormaut, Lettres Assyriologiqites, vol. ii. p. 13. * N. Karamiauz, Mnwidzwauzig Buthstaien eines verlorenen Alphabets (Zil- achtiftf. D. Moigen. Ges. 1886, xl. pp. 325-319). 3 Prof. Percy Gardner, Coins /rum Kashgar [Xumismatie Chronicle, 1879, vol. XLS. pp. 274-281). . . , J * It was Ion? supposed that the principal one of these two coins, the larger and the only bilingual, was made of iron ; but I had it tested at the British Museum, and there is no doubt that it is of copper. 100 BEGINNINGS OF WRITING. latter, to state it roughly, had superseded p'an Hang, or half- ounce, another legend which on pieces of various sizes had been in use since the days of the Ts'in dynasty (b.c. 221-202). 144. But the Chinese legends on the above two pieces have nothing to do with one or the other of the above legends.^ That of the smaller piece ^ is some debased Chinese characters, like a barbarous imitation of ^ ^p'cm kin 'half-kin,' a design which is familiar to us on several mintages current in Northern China about 300 B.C.' The legend of the other piece is of a more elaborate character. In the centre is the Chinese symbol _^ pei for ' money or precious ' and formerly cauris. The exergue is so much worn as to be difficult to read ; * it seems to me to be ^ ? g — p| J^.- The first two symbols are not the rendering of a proper name, but simply the denomination of the coin ; the others are the weight, i.e. "weighing one ounce and four tchus," and correspond to a well-known design of mintage current also in the third century b.c.,^ but chiefly in the westernmost Chinese state, that of Ts'in, then fighting for the Empire. 145. These two legends must have been borrowed by some neighbouring people, who by intercourse had acquired some- thing of the Chinese civilization, writing and currency. This people can be none other than the Yueh-ti, who in the third century were very flourishing on the north-west of the said Ts'in state and conterminous with it.^ The Yueh-ti formed five tribes, one of which, the Kuei-shwang, afterwards assumed the lead over the other four tribes. But they were attacked in 201 B.C. by the Turks Hiung-nu, and about B.C. ' Cf. T. de Lacouperie, The Coins and Medals of China in the British Museum, etc., vol. i. Nos. 174-302. 2 This small piece is only known to me by the woodcut published by Prof. Percy Gardner in his above-named paper, p. 3, No. 2. ' Vid. the illustrations and decipherments in my work, The Coins and Medals of China in the British Museum, etc., vol. i. pp. 18-30. * The piece was kindly lent to me for examination by the late Mr. Edward Thomas. 5 Cf. The Coins and Medals of China, p. 321. " "When Ts'in subverted the other six states (of the Chinese agglomeration), the Tueh-ti were in a flourishing condition." Taien Han shu, bk. 94 ; A. Wylie, History of the Eiung-nu in their relations with China, 1873 {Journ. Anthrop. Institute). BEGINNINGS OP WEITING. 101 165 compelled to flee westwards,^ carrying with them their knowledge of Chinese writing and currency. In 126 b.c. the Chinese ambassador Tchang Men, who was sent after them in order to secure their alliance against the Hiung-nu, finds them settled north of the Oxus ; and a century later K'iu-tsiu-kio, or Kadphises I., of the Kuei-shwang tribe, the Kushans of the Persians, unites the five tribes, invades^ Amih (Parthia) and annihilates Puk-ta (Bactria) and Ki-pin (Kophene).* They came thus into contact with the Greeks, and Hermseus, the last of the Greek rulers of Cabul, was for a while confederate or allied with Kadphises I. ; they had a common bilingual mintage, Greek and Indo-Bactrian, the legend of the Kushan King being that written in the latter characters.* 146. The solution of the chronological and geographical problem involved by the above mintage, Bactrian-Chinese in its legend, is nearly given by the decipherment of the Bactrian legend. It is on one side so extremely corroded that it will not divulge its complete secret, and unhappily the less decipherable part is that which contains the name of the King. The words visible are Maharajasa Rajadirajasa . . . sa . . . mayasa,^ which fit exactly to the legend of Hermseus as exhibited in the Grseco-Bactrian specimens. Therefore, the date of this Bactrian-Chinese mintage would be most probably about B.C. 26, and its origin Yueh-ti-Kushan. While the smaller specimen, uniliagual, in debased Chinese, would be also Yueh-ti, but previous to their coming into contact with the Bactrians, examplifying another hitherto un- known extension of the Chinese writing westwards through another channel than that of the Chinese themselves. 1 Mr. E. Specht {Journal Asiatiqiie, 1883, vol. ii. pp. 317-350) has collected clironologically the texts concerning them, and has been able to throw much light on their movements. 3 About B.C. 31 Phraates, with the help of a Scythian army, expelled Tiridates from the government of Parthia. 3 Bou Man shu {Yuen kien lei han, bk, 237, fol. 43). ■• Cf. Prof. Percy Gardner, The Corns of the (rreek and Seythie Kings of Baetria and India in the British Museum (London, 1886. 8vo.), pp. xxxi. sq., 120 sq. * Prof. P. Gardner {Coins from Kiuhgai; p. 2) had read : Muharajasa raja yasa, which he proposed to complete by . . . dirajasa tradtitasa Serama , in agieement with some well-known legends ot coins of Hermirus. 102 BEGINNINGS OF "WRITING. c) India. 147. The Lalita-vistara, ch. x., puts in the young B6dhi- sattva's mouth, when he was led to the writing-school Kpisala, an interesting list of 64 writings. "We reproduce the frag- ment from the recent translation of Baboo Rajendra Lai Mitra^ and of Mr. Ph. Ed. Foucaux:^ — "Now Bodhisattva, taking up a tablet made of Uragas4ra sandal-wood and excellent ink, with a golden pen mounted with jewels, thus addressed the tutor Visvamitra : ' "Which is the writing, sir, which you wish to teach me ? 1) Is it the Brahmi writing ? 2) or theKharosti?^ ' 3) or the Pushkarasari P 4) or the writing of Anga ? 5) or that of Banga ? 6) or that of Magadha ? 7) or the M^ngalya writing ? * 8) or the Manushya writing ? 9) or the Finger writing (Anguli) ? 10) or the Qakari writing ? ^ 11) or the Brahmavalli writing (Tib. Yavana) ? ^ 12) or the Dravida writing ? 13) or the Kinari writing (Tib. Kiratas) P ' ' Bibliotheca Indica, N.s. No. 473. ^ Ph. Ed. Foucanx, Rgya-Tcher-Rol-pa, ou diveloppement des Jeux, contenant I'histoire du Bouddha (J)akya-Muni, traduit sur la version Tibetaine du Bkali Hgyour et revu sur 1' original Sanscrit (Lalita-vistara) , 2 vols. Paris, 1847, 4to. ; vol. ii. pp. 122-123. Le Lalita ViUara, traduit du Sanscrit en francjais (in Annales du Musde Guimet, vol. vi. 1884), pp. 114-115. I have compared also the list with those of the i^p Chinese versions, viz. in the P'u yao king, ui. 7, fol. 5, and in the Fang kwang ta ichwang yen king, bk. iv. fol. 6 (edit., of the Tripitaka, Nos. 159, 160), and I have noticed the variants of names when im- portant, marking them A and B for these two works. 3 Chin. A: {^ ^ Eiu-liu; B : f^ ^ gj, Jg Kiu-lu-she-ti. The first syllable is also casually written ■{J[| Ma. * Chin. A : ^ ^ An-kiu. 5 A 13 : Tsih-kien. 8 A 10 : F'an, ^ ; B Te-pa-ni, both transcriptions of Tavanas. Professor Albrecht Weber, of Berlin, in his Indische Studien, xxi. 399-400, has mentioned that in Jain books an enumeration is given of writings known to them. The first is the Labamkht or ' holy writing,' then comes the Tavan&nt. ' A 12 : Kin-dzu, mod. Ki'n-yu. BEGINNINGS OP WRITING. 103 14) or that of Dakshioa P i 16) or that of Ugra ? » 16) or the Figure writing (Sankhya) ? 17) or the Cursive writing (Anuloma) ? ' 18) or the Half-bow shaped writing (Ard hadanus) ? 19) or that of Darada ? 20) or that of Khosya ? 21) or thatof Tchlna?* 22) or that of Huna ? * 23) or that in which the letters are most in the middle (Madhyakshara-vistara) ? 24) or the Flowery writing (Pushpa) ? 25) or the writing of the DSvas ? 26) or that of the Mgas ? 27) or that of the Takshas ? 28) or that of the Gandharbas ? 29) or that of the Kinnaras ? 30) or that of the Mahoragas ? 31) or that of the Asuras ? 32) or that of the Grarudas ? 33) or that of the Mriga-ugra ? 34) or the circular writing (Tchakra) ? 35) or that of the Vayumaruts ? 36) or that of the Bhaumad^vas ? 37) or that of the Antarikshadevas ? 38) or that of the TJttarakuru ? « 39) ' or that of the Purvavideha ? * 40) or the perpendicular writing (TJtkshapa) P 41) or the pendulous writing (Nikshepa) ? 42) or the scattered writing (Vikshepa) ? 1 A 7: Tatan, afterwards Ta Ts'in ^ ^; B 12: To-tso-m ^ i|| HJ. 2 A 14: Bzi-dik-tak ^ Jjit ^, mod. I-tih-seh. ' A 16 : ^ j§ , Eang-kiii, mod. Samarcand. Vid. below, § 150, n. 1. * A (20) : ^ Ta'in; B 19: j^ gg Tchi-na. 6 A 21 : Bimg-nu ; B 20 : Hu-na. 5 CMn. A 37 : Fehfang t'ien Ida, i.e. northern regions. ' Foucaux's text has Aparag&iidiiiii after 38, and before Purvavideha. In Chinese A 38 : Eiu-ge-ni for Gaudani. ^ The first of the four dvipas, the eastern lands. In Chinese A 39 : Tung fang tUeii hill, the eastern regions. 104 BEGINNINGS OF WRITING. 43) or the disordered writing (Prakshepa) ? 44) or the ocean writing (Sagara) ? 45) or the thunderbolt writing (Vajera) ? 46) or the cross writing (Lekha-pratilekha) ? 47) or the slow writing (Anudruta) ? 48) or the Sastravarta writing ? 49) or the Ganavarta writing ? 60) or the Utkshepavanta writing ? 61) or the Nikshepavanta writing ? 52) or the Padalikhita writing ? 53) or the Dviruttarapadasandhi writing ? 64) or the Tavaddasuttarapadasandhi writing ? 65) or the Adhydhanini writing ? 66) or the Sarvarutasangrahani writing ? 57) or the Vidyanuloma writing ? 58) or the Vimisrita writing ? 59) or the Rishitapastaptan writing ? 60) or the Gaganaprekshani writing ? ^ 61) or the Rochamanandharani-prekshana writing? 62) or the Sarvanshadhinisyandan writing ? 63) or the SarvasarasaiSgrahani writing ? 64) or the Sarvabhutarutagrahani writing ? ^ ^ This is missing in Foucaux's text. ^ It is interesting to compare with this list that which is given in the Tihetau version. I quote from M. Ed. Foucaux's French translation, vol. ii. pp. 122-123, and I subjoin the numbers which correspond to the text of the BM. Indie, as above: " Ecriture de 1) Brahma; 2) Khar6sti; 3) de 1' essence du lotus ; 4) d' Afiga ; 5) de Manga ; 6) du pays de Magadha ; 7) de ceux qui ont la benediction; 8) de ceux qui ont des doigts ; 10) de (Jatani; de Tavana ; de BaglSpa; de Paroncha ; des etus volante ; 13) des Kiratas ; li) de I'horizon du sud ; 15) d'Ougra; des nombres SankhyS, ; a tete renversee ; reguliSre ; 19) de Darada; 20) de Eoncja; 21) de Tohina; de Ph'onna ; 22) de Houna ; 23) moyenne ; grosse ; 24) de Penchya ; 25) des Dieux ; 26) des Nagas ; 27) des Takchas; 28) des Gandharbas ; 29) des Kinnaras ; 30) des Mahoragas ; 31) des Asouras ; 32) des Garandas ; des bStes fauves ; 34) du cerele ; de ceux qui eaveut le langage des comeilles ; des dieux qui president h, la terre ; 35) des dieux de 1' atmosphere ; 38) des regions d'Outtarakouran ; d'AparagddSni ; 39) de Ponrva- videha; 40) d'OutkohSpa; 41) de Nikchfpa; 42) de Vikchepa ; 43) de Prak- ohSpa; 44) de r Ocean ; 45) de la fondre ; d'une lettre d'avis et de la reponse ; cursive ; possee ; 48) tournante des Castras ; tournante des calculs ; 50) touruante d'Outkchepa; 51) tournante de Nikchepa ; 52) tracee aveo le pied; 53) des sandhi (liaison) d'un mot repete deux fois ; 54) du sandhi d'un mot repete dix fois ; 55) de MadhyS,hariiii ; de tons les sons reunis ; de la science methodique ; de la science confuse ; 69) des Eichis livres h. I'^xercice de la penitence ; oerfcaine des dieux ; visible de la terre ; visible du ciel ; par ordi-e de tous les remfedes ; de la coUectiou complete de toutes les essences ; de la reunion de la voii de tous les etres." BEGINNINGS OP 'WIIITING. 105 Out of these sixty-four kinds which is it, sir, that you wish to teach me ? '" 148. This complete list, from the LaUta-vistara, is in itself valueless for scientific purposes, but a few of the names derive interest, as far as we can ascertain their precise reference to some known writings, from the date of its com- position.1 Therefore the vexed question as to the date of the work is important. The Sanskrit text now known is almost identical with that which was translated into Chinese in 685 A.D. under the title of Fang kwang ta tchwang yen king, and later on into Tibetan Rgya-toher-rol-pa. The difierences are unimportant, as for instance in the case of this list, which amounts to 65 in these Chinese and Tibetan versions. The said Chinese version covers about 540 pages of text.^ An earlier text of probably the same work had been pre- viously translated three times into Chinese, in circa 250, 308 and 427 a.d., under the title of P'u-yao king, but the first and third of these were already lost in 730 a.d.' The second, which is still in existence, covers about 360 pages of text, namely, one- third less than the version of 685 a.d. in the same edition of the Buddhist TripitS,ka.* The third century ^ is therefore the earliest period which from the translations can be assigned to the ground-work of the Lalita-vistara. Any earlier date must be sought for from internal evidence. Under this respect some statements from the chapter on writings are not without interest. The list in the P'u-yao king is one of 64, as in the known Sanskrit text, but while many names are the same, either > The Lalita-vistara is neither a poem nor a prose text. M. Ph. Ed. Foucaux evaluates the proportion of prose to three-fiftli of the whole work. The other two-fifths are some sort of gathas, or popular verses interspersed with the prose text, and always introduced in support of the latter. The distinction of prose and verse is carefully preserved in the Chinese translations. ^ In the Japanese edition of the Chinese Tripitaka at the India Office Library. ^ Bunyiu Nanjio, Catalogue of the Chinese Tiipitalca, Nos. 159, 160. * M. E. Senart, La Legende du Bouddha, p. 497, n. 4, has given an accurate summary of this work, says S. Beal, The Buddhist Tripitdka, p. 18. * That of 65 a.d., which is often repeated from the late Stan. Julien, was a confusion made by this scholar with another work. Cf. B. Nanjio, Catalogue, col. 3S0. 106 BEGINNINGS OF WRITING. translated or transcribed as usual in Buddhist- Chinese ver- sions, some of them are altogether different, and others are neither more nor less than rough-and-ready identifications by the translator. I have given the identification in all cases of interest. 149. The above list of 64 writings bears internal evidence of successive composition, the names being arranged in several broken series of cognates or classes, which, however, find some additions scattered in the sequel. The names may be arranged under five heads : a) Referring to certain writings of importance, such as those indicated by the numbers 1, 2, 3. Thus 1) is that of Brahma,'^ otherwise the Sanskrit writing. 2) That of Khardsti,^ a ' (Semitic) writing running from right to left in contradis- tinction to the preceding, as explained by a tradition preserved in China, and referring to the Indian, Kia-liu, and Chinese writing, all differing in the direction of their course. 3) That of Pushkarasari is difficult to identify, though the name stands somewhat isolated along with the two others at the ' The Fa wan shu lin, Enoyclopsedia of Extracts from the Tripitaka, compiled in 668 A.D. by Tao-shi, reports a tradition attributing the invention of writing to three individuals : 1) Fan (or Brahma), who wrote from left to right ; 2) Eia-lu (or Kia-lu-she-io for Kharflshta), who wrote from right to left ; and 3) Tsang- hieh (the Chaldean Dungi, as I have shown elsewhere), who wrote from top to bottom, and is the legendary inventor of the Chinese characters. Now the Kia-lu appears with the Fan writing, quoted apart as equally important, in the seventh chapter of the P\i yao Teing, and also at the head of the list of 64 writings of the same chapter, which corresponds to the tenth of the Lalita-vistara Sanskrit text, and of the later Chinese version, where, however, this list is one of 65. The date of 308 A.D. for the P'u-yao king version, still in existence, and that of circa 260 A.D. for the first version of the same work, are the only precise data concerning the existence of the Lalita-vistara, or the non-completed text of this work. 2 It was transcribed variously, as seen above, § 147, n. 3, but explained by >® ^) ^•^' ' *^^' ^'P*'' ■"'^'•''i i^ ^^^ meaning of Kharoati, which name can be easily recognized in the complete Chinese rendering Kia-lu-she-ti. Kharfisti is the name of the man to whom the legend ascribes the introduction of astronomy into Chaldea, according to Armenian authorities. The accuracy of the Chinese transcription and explanation does not permit our acceptance of the suggestion of Prof. Alb. "Weber {Indian Literature, p. 248) assimilating the name to that of Krashutuki, reputed in the Atharva Paris. [Lit. C. Bl. 1869, p. 1497), to have arranged the consteUatione in the order beginning with Krittika. From the aforesaid association of the two names of the Fan and Eia-lu writings on the same footing, we may infer that this association belongs to the time when the two writings known as Aryan-Pali and Bactro-Pali were used in N. India ; their respective directions from left to right and from right to left, as preserved in the Chinese legend, support this inference, which is perhaps suggestive in other respects.. BEGINNINGS OP WEITIN6. 107 head of the list. The name is that of the seventh dvipa in the mythical geography of the Vishnu Purana,i and does not admit there of any reduction to positive geography. Another tradition makes Pushkara a son of Bharata, and elsewhere this same name has been connected with a district north of Peshawar.'' 150. The other heads are the following : b) Referring to writings of particular people or countries, in and outside India ; such as (4) Anga/ (5) Bang^,* (6) Magadha ; ^ (12) Dravida ; (13) of the Kinnari, the Kiratas 6 of the Tibetan version ; (14) of Dakshina, the Ta Tsin of the older and the To-tso-na of the later Chinese version ; (19) Darada, the Dards of the Hindu Kush ; (20) Khosya, the Khasiyas of N.E. Bengal ; (21) Tchina,'' the Shins of Dardistan ; (22) Huna, the Hiung-nu, Turkish tribes, of the older Chinese version ; (10) Cakari, probably for the Sakas in the N.W.; (11) Brahmavalli, the Yavanas^ of the Tibetan ^ Vishnu Purana, ii. 1 and 4 ; transl. H. H. "Wilson, edit. Fitzedward Hall (London, 1865), yol. ii. pp. 101, 201. ^ 8. Beal, Buddhist Records of the Western World, vol. i. p. 109. PTishkarS- vati was the old capital of GandMra. Cf. Vishnu Purana, transl. Wilson, vol. iii. p. 319. ^ The country about Bhagulpoor, of which Champa was the capital, mentioned with the following, in the Mahabharata, Bhisma Parvan, si. 317-378. * i.e. Eastern Bengal. ^ The cradle of Buddhism near S. Behar. ^ The Kiratee of Ptolemy, assimilated to the Kiranti of the present day, N. Bengal ; also mentioned in the Mahabharata. ' Written Tsin ^ , and ^ BjJ Tchi na, in the Chinese versions ; must be distinguished from ^ BR Tohe-na, the Buddhist name of China. On the origin of the latter vid. above, f 80, and Col. H. Tule, Glossary of Anglo-Indian Words s.v. China. In the topographical lists, people and countries in the Mahabharata, there are two Chinas quoted: "Northern and other fierce Mlechchas (or barba- rians), Yavanas, Chinas, Kamboias ; ferocious and uncivilized races, Sakridgrahas, Kulatthas, Hunas, and Parasikas ; also Eomanas, Chinas, Dasamahkas, etc." The Eomanas or Eomans and the Chinas or Chinese, here enumerated apart from the others, were the western and eastern great civilized nations of the time. The Chinas mentioned before were undoubtedly the Shinas of Dardistan, on whom see Major Biddulph, Tribes of the Hindu Kush, pts. iii. and xiv. This identifica- tion, proposed by Sir Henry Eawlinson, was opposed by the late G. Pauthier rather injudiciously, in ignorance of the two Chinas in the enumeration. ^ The Greeks of Bactria, under the successors of Alexander the Great ; this name was also applied in a rather loose way for any western people. The writing there indicated was certainly that of the Greeks. The late Franijois Lenormant, in his unfinished and remarkable work Les origines de Phistoire d'npres la Bible et les traditions des peuples orientaux, vol. ii. (2), pp. 1-28, has thoroughly investi- gated the question. 108 BEGINNINGS OF WRITING. and both Chinese versions; (17) Anuloma or cursive of Kankin ^ or Samarcand, in the older Chinese version. c) Referring to writings of fabulous beings : such as (15) Ugra; (25) Devas; (26) Mgas; (27) Yakshas; (28) Gand- harvas; (29) Kinnaras ; (30) Mahoragas; (31)A8uras; (32) Grarudas; (35) Vayumaruts; (36) Bhaumad^vas, etc. d) Referring to writings of semi-mythical lands : (38) Utta- rakuru, one of the four continents north of Mount Meru ; (39) Purvavideha, another of the four continents east of the same sacred mountain. e) Implying peculiarities in the configuration of letters. f) Cryptic or imaginative forms ; ^ the thirty-five other writings must be classified under one or the other of these two divisions. 151. In the short preamble which, in the later Chinese version of the Lalita-vistara,^ precedes the enumeration of the 64 writings, P'u-sat ^ jj^,* i.e. Bodhisattva, is made to take in hand the tablet of red sandal wood employed for the T'ien-shu or Divine writing, i.e. the Dfivanagari.^ In the older Chinese version,^ which is that of an earlier text, this preamble is somewhat longer and especially interesting for the case in point. The P'u-sat takes in hand the golden pencil and the red sandal wood tablet ; then the master (Vi9vamitra) states that there are two writings, that of Fan ' or Brahma, and that of K'u-liu, both equally good and not ^ J^ ^ ) -2''««? ^ciu, also -written Fang only, and afterwards Sa-mo-leien, i.e. Samarcand. Bunyiu Nanjio, in his Catalogue of the Chinese Buddhist Tripi- taha, cols. 383, 386, 390, has mistaken a modern Chinese name of a part of Tibet for this ancient name of Samarcand, and in consequence attributed wrongly a Tibetan origin to three Sramanas from Samarcand (8, 14, 21). 2 Babu Eajendra Lai (/oc. cit.) had acknowledged four, viz. h, d, e, f, of the above divisions, and indicated some ten of the named writings in reference to them. ' Fan Kwang ta tehwang yen Idn, bk. iv. fol. 5{!. * Now P'u-sa, commonly applied in China for all kinds of gods and idols. 6 On this name vid. supra, §§ 97-99. 6 P'm yao king, bk. iii. f. iv. ' ^ -^"^ (formerly Bam) for Brahma, as found by Stan. Julien {MMhode pour dichiffrer les noms Sanskrits dans les livres Ghinois, N. 281), who, however, has not understood the ingenious formation of this Chinese symbol of Buddhist make. The under character was Bam, and the upper one, lin, was lam for ram ; the two making Barn-ram, or better B-ram, according to the practice followed usually by the Buddhists in their Chinese transliteration of complex groups, which required the dropping of the last pait of the first of two rhyming syllables. BEGINNINGS OF WRITING. 109 differing, but the P'u-sat retorts that there are in fact 64 different writings, and that he does not see why the master advocates only two sorts. As we have seen above. Fan was for the Brahman South Indian writing running from left to right. K'u-Un was for Kia-lu-she-ti or Kharosti, i.e. 'Ass' lips,' the writing running from right to left. These two names are here the designations of the Indo-Pali and Bactro- Pali ; and their association on an equal footing goes far to show that the redaction of the text, where it occurs, belongs to the period when they were both in use, as from the time of or before Agoka to the first century of our era. 152. But in this name of K'u-liu and its meaning is hidden, I think, another valuable information concerning the primeval origin of this writing for the Hindus. Let us keep in mind the meaning 'ass' lips,' of its full form Kharosti,^ and premise that this name, foreign to the Sanskrit Onomas- tics, is a folk etymology, an Indian significant adaptation obtained in the striving after meaning of a foreign name, which name was probably no other than that of Kyros, the Kuras and Kurush of the Cuneiform inscriptions, the Khusrau of Persian history. The great conqueror had vanquished Krsesus (554 B.C.), in fulfilment of the following prediction of the Pythoness of Delphi, whose oracle the Lydian King had just consulted : " Wleu Media's King shall be a mule, Soft-footed Lydian, by tbe pool Of pebbly Hermos fly, nor stay Nor dread the coward's name that day." " l^ow as Kyros was then looked upon as a mule, being the offspring of a Persian and a Median wife,^ the accomplish- ment of the prediction was certainly spread far and wide, in ' Khara, in Skr. ' ass.' Cf. my article, Did Gyrus introduce the Writing into InAa'i in The Babylonian and Oriental Record, February, 1887. * Herodot. i. 55, transl. Sayce. 3 G. Maspero, Histoire ancienne, p. 516. Prof. A. H. Sayce (Museon, 1882, ToL i. p. 550) says : Cyrus was a Mardian, whose father Athadates was a brigand, and whose mother Argoste was a shepherdess of goats. Mirkhond, History of the Early Kings of Fersia, transl. D. Shea, makes Khusrau son of Ferangiz, daughter of Afrasiab, the Turkish King, and of the Persian SiyS-wesh son of Kai Kaoos. Vid. pp. 226, 233. 110 BEGINNINGS OF VRITING. order to magnify his fame and inculcate a just apprehension of his advance as a conqueror. It was previously to 540 B.C. that Kyros extended his sway in the east, north and south, and subjected to his rule the regions bordering India on the north-west, and it is there- fore from that date that the introduction into India of notions of civilization from the west became possible through the channel of the well-managed Persian monarchy.^ The knowledge and use of alphabetic writing, though limited as it may have been, because of the small demand for it in N.W. India, were certainly among the then introduced notions ; and there is a great probability that in the name of the Khar6sti writing, running from right to left, as was that of the Persians, we find concealed the name of Kyros, intro- ducer of the writing in India, and traces of his fabulous capacity of a mule or ass ; the motive of the former having been lost on the way, or found objectionable for the play which was made on the name of the Persian conqueror. 153. We shall return directly to the peculiarities of this Persian writing, but we must first examine with reference to the above explanation of an historical problem, the bearing of the Armenian tradition already reported (§ 149, n. 1). Kharosti is said to have been the introducer of astronomy into Chaldaea, a statement which, with the large allowance due to Persian exaggeration, is by no means antagonistic to the identification of Khar6sti with Kyros. There is a growing feeling among Assyriologists, increasing with the progress of their science, that the ancient knowledge of astronomy in Chaldaea was not worthy of its wide repute, and did not extend much beycind a notation of astronomical events. More sensible notions and better knowledge appear on the Cuneiform tablets of later times only. This would agree with an importation of astronomical lore by the Persian conquest. We are well aware that Kwarism on the east of the Caspian Sea was an old focus of astronomy, and that important progress in the science was made there. An ' The conquests of Kyros the Great were achieved preyious to his conquest of Babylon, which took place on the 16th of June, 538 B.C. BEGINNINGS OF WEITING. HI adaptation of twenty-eight lunar mansions with the solar year was one of them, and it is most likely there also that the zodiacal signs were regulated. The Chorasmians used to reckon from the beginning of the colonization of their country, 980 years before Alexander,^ i.e. 1304 B.C. Kyros had subdued their country and incorporated it in his dominions, previous to his conquest of Babylon,^ and there- fore their advanced knowledge in astronomy must have been carried there afterwards. The great Persian conqueror was an enlightened monarch, and was no doubt interested in the progress of science, if not himself acquainted especially with the science of Kwarism, inasmuch as the Chorasmian tradi- tions report that his father Siyawesh had ruled over their country.^ 154. The Persians employed a Cuneiform writing from the time of Darius (521-485 B.C.) down to Artaxerxes Okhos (362-339 B.C.) * for monumental purposes. They had derived it ^ from the wedge characters of Assyro-Babylonia, and this derivation and adaptation to the requirements of their language has been rightly looked upon as an accessory proof of their having an earlier system of phonetic writing.^ These older characters were traced on skins.' A tradition preserved by a very early and learned Arabic writer, Ibn-el- Nadim, who had unusually good means of information as to genuine Persian traditions, ascribes the invention of Persian writing to Jamshid, the son of Vivengham (who, with the Zoroastrians, was the Eponym of the Persian race), and adds that he, Jamshid, dwelt at Assan, one of the districts of 1 Altiruni, The Chronology of Ancient Nations, transl. Sachau, pp. 40, 57, 172, 173, etc. ^ The same date occurs in India, that of the Jyotisha observation of the Colures, still in use in the Vedic rituals of India. — Edward Thomas, Comments on Recent Fehlvi Decipherments, p. 18. ' Albiruni, I.e. * Prof. A. H. Sayce, The Ancient Empires of the Mast, pp. 438, 483. 5 It was neither an alphabet nor a syllabary, as some of the signs were syllabic. The total of the characters was 38 or more. Cf. Dr. J. Oppert, Expedition scien- tijique en Mesopotamie, vol. ii. 1858, p. 12 ; Melanges perses in Sevue de LinguisUqm, iv. pp. 205-206. « By Prof. C. de Harlez, Introduction a I' etude de VAvesta et de la religion Masdienne, p. 54. ' Vid. the authorities in C. de Harlez, ibid. 112 BEGINNINGS OF WEITING. Tuster (modern Shuster).^ We can dismiss this legend without much concern, because to Jamshid, the sixth ancestor sovereign of Kyros, is attributed the authorship of everything useful and good in the Persian histories, such as that of Mirkhond. It might simply mean that the use of writing among the Persians was much older than the time of Kyros, i.e. some 150 years previously, or about 700 b.c. 155. No specimen of this ancient writing is known to be in existence, but we have : a) nearly in situ the Indo-Bactrian alphabet in the third century B.C., illustrated by the cele- brated edict of Ac6ka at Kapurdigirhi in the district of Peshdwar, and by the legends of coins of the Greek and Scythic kings of Bactria and India ;^ b) an instance of a cursive writing at Babylon under the Persian rule, 443 B.C., quite distinct from the Phoenician dockets found in Nineveh and Babylon.' As I have pointed out already, the latter is not without presenting some serious resemblances to the former, of which it may be looked upon as an antecedent ; c) the specimen of the cursive writing which appears on the Sabsean coins, and belonging to the same Aramaic origin as the preceding, was most likely brought there by the Persian influence at the same time as the habit of mono- grams.* The Semitic ancestry of the Indo-Bactrian alphabet is an admitted fact,^ and it has been rightly remarked that the derivation, remote as it may be, cannot be earlier than the seventh century, as it exhibits an Aramaic feature, such as the opening of the loops of the closed heads of Phoenician letters, which was not developed before that time.^ The immediate ancestor of this alphabet ^ was not known, but 1 Eital-el-Fihrist, p. 12, 1. 22; Sir Heniy Rawlinson, Notes on a newly-dis- covered Clay Oylinder of Cyrus the Great, p. 77 of Journ. Roy. Asiat. Soc. 1880, Vol. XII. pp. 70-97. ' Which cf . in Prof. Percy Gardner, The Cows of the Greek and Scythic Kings of Bactria and India in the British Museum, 1886. * Vid. supra, § 140. * Cf. above, ^ 140, 141. ^ Since its decipherment in 1837. « Cf. Dr. Isaac Taylor, The Alphabet, vol. ii. pp. 234, 236. ' The Indo-Bactrian is represented by a dozen inscriptions of various lengths, besides the legends on coins. Cf. J. Dowson, On a newly-discovered Bactriam- Pali Inscription; and on other Inscriptions in the Bactrian-Fali Character {Journ. BEGINNINGS OF WRITING. H3 now it is exceedingly probable tbat tbe aforesaid remains b and c are tbe only known representatives of that ancestor, wbicb was the writing of tbe ancient Persians. But tbe use of tbe Bactro-Pali was restricted to tbe extreme nortb of India, and no traces of its extension soutb- wards bas been found. Sucb would not have been tbe case should tbe country have been deprived of any sort of writing. This writing was tbe Indo-Pali and its antecedent, upon wbicb tbe following remarks are necessary. 156. It would be more than singular that India, being in the vicinity of and having relations with China, should have remained without any knowledge of writing, fragmentary and imperfect as it may have been, resulting as it were from a mere rebound by necessities of trade. Writing in China was a very old institution, the knowledge of which the early civilizers of tbe Chinese, leaders of tbe immigrating Bak tribes, bad brought with them from Western Asia into tbe Flowery Land. Confined at tbe outset to tbe basin of tbe Yellow River, their influence and their civilization spread more quickly and further than their political dominion, wbicb bad a very slow extension. In the west of China, namely, in Sze-tchuen, the old Shuh region, were established some well- gifted populations of an initiative spirit, who did not shrink from being on friendly terms with tbe civilized new-comers, and avail themselves of tbe benefit wbicb an acquaintance with their scientific notions and arts could procure. There are all sorts of reasons for believing tbat this anciently-acquired knowledge was kept up by a frequent intercourse witb the , Chinese states bordering their respective territories. When tbe state of Ts'in, in the sixth century B.C., wanted to extend its power south of tbe Kiu-lung range, tbe native and non- Chinese populations there, cbiefly Shan,^ were so w.ell ac- quainted witb tbe art of writing tbat a written compact could be made witb them. Roy. Asiat. Soc, 1863, Vol. xx.). These various inscriptions and legends call for a remark hitherto unobserved ; it is the great looseness of the shape of many characters. Does it suggest a worn-out writing, or one which was not sufficiently known ? The former suggestion seems improbable. ' Vid. my paper, The Cradle of the Shan Race. 114 BEGINNINGS OF WRITING. 1566. From the earliest ages relations have existed between India and Ctina. Traders used to run between the Chinese agglomeration, especially the border states and India on one side, and Indo-China on the other.^ Mercantile expeditions, which had reached the middle kingdom under the reign of Tcheng Wang (1044-1007 b.c.) of the Tchou dynasty, are recorded in literature as embassies, and from intrinsic evidence one of them, the second, had come from N.W, ' I have collected in a special paper, read before the Eoyal Asiatic Society, 16 June, 1884, all the available information and possible inferences on the matter: " Three Embassies from Indo-China to the Middle Kingdom, and the Trade Routes thither 3000 years ago," of which the following is a summary. During the first years of the reign of Tch'eng, the second King of the Tchou dynasty (about 1030 B.C.), three so-called embassies came to the Chinese Court from Indo-China, a region vpbich at that time was widely separated from China, whose power was not yet strongly established south of the Yangtze Kiang. These embassies were not political missions, but were only parties of merchants travelling in pursuit of their own business ; they had heard of the newly-established dynasty and its wealth from the indigenous tribes of Western and Southern China, who had helped the Tchou to overthrow the preceding dynasty. The original records of the arrival of these embassies have been lost in one or other of the five great bibliotheeal catastrophes in which the greater part of the historical literature of China has been destroyed. Only a few scraps of information have survived about them, and these fragments have suffered from the various causes of alteration which have impaired the credibility of so many ancient traditions of the country. The most important of these causes being the curious and rather unfurtunate influence of the two schools of Confucius and Lao-tze, which have acted diiferently, though in both cases iu obedience to the national tendency towards the exaltation of ancestors ; the Confucianists, who had no other object than to put them forward as patterns of virtue, neglected all that did not tally with their views ; while the Taoists,on the other hand, desired to bring into prominence their super- human qualities and divine character ; the traditions have accordingly been amended and completed by the addition of moral padding and marvellous circum- stances ; the most common process being the attribution to the palmy days of the early rulers, of felicitous coincidences and glorious events similar to those which occurred at late periods. Conspicuous among these coincidences was the arrival at the court of foreigners from distant regions ; for instance, the arrival of one of the so-called embassies from Indo-China was supposed to have happened also under the reign of Yao, and a detailed statement to that effect was developed accordingly from some obscure and brief entry in the ancient records. The three embassies of about 1030 b c. were the successive arrival of 1st, merchants from Nili, or the Norai country North of Burmah, through the Bamo road; ind, merchants from the S.W. of Yunnan carrying monkeys from the Kudang country, the geographical location and details concerning which prove the existence at that time of several Karen tribes in Northern Burmah and of Dravidians on the north-eastern parts of India ; 3rd, merchants from Yueh-shang or Cochin-China, said to have been sent back with several south-pointing chariots, which, however, do not seem to have been invented until several centuries later. The concluding part of the paper consisted of a review of six ancient trade routes between India and Indo-China with China, and of their respective antiquity down to the beginning of the Christian era. The two most important of these routes were, firstlyj one through Assam to India ; and secondly, tTie one to Tong- King by the Red Eiver. It was by the latter of these that the sea traders at Kattigara (Hanoi) heard of the import 'ut trading state of Tsen in Yunnan, the name of which became the antecedent of that of China. BEGINNINGS OF WRITING. 115 India. Later on, the chief object of the Princes of Ts'in, in the fourth century B.C., advancing in Szetchuen, was to get hold of the head of the trade route which the merchants of Shuh used to take westwards on their way to India. The relations between the country of Shuh or Szetchuen and India were sufficient to permit towards the end of the afore- said century, at Tchengtu, still now the capital city of Sze- tchuen, the establishment of a dynasty of Indian, origin. ^ And it was because the merchants of Shuh stuck to their old prerogatives and refused to let their trade pass into other hands or suffer from competition, that the Chinese ruler, enlightened by the revelations of Tchang kien, endeavoured unsuccessfully towards the end of the second century B.C. to find other routes of communication with India in passing through Yunnan. The trade from China, i.e. Szetchuen, to Patna in India, spoken of by Ptolemy, was that very same trade made with the m.erchants of Shuh which had been in existence for many centuries. 156c. Traders cannot do easily without the art of writing, and it is not admissible that the diligent merchants of Shuh, acquainted as they were in this country with writing, should have neglected to make use of it in the interests of their traffic, which it could not fail to help. Therefore, considering that this trade was a regular one, and was in existence for many centuries previous to the time of Acoka, it would be a matter of surprise, with the extension that this trade had obtained unto the north-western provinces of India, that nothing should have been known in India of the writing proper to these merchants. As the reverse is rather incredible, remains of the casual use of this writing would 1 On this dynasty vid. Tang hiung (bo. 53--A.D. 18), himself a native of Shuh, who wrote a history of his country, Shuh viang pen tsi, in the lai ping yii Ian cyclopedia, hk. 166, ff. 3. lOc. Many Hindu ideas had penetrated into China, about the time of that short-lived dynasty. And to the same time may be reported a curious series of mythological resemblances. The many notions of a fabulous ethnology and natural history which we know from Ctesias, Megasthfenes, and others, as Indian, and the existence of similar, sometimes identical, notions in ancient Chinese literature of the same period, especially in the SItan 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 flirough which they had to pass during their journey to and fro. 116 BEGINNIXGS OF WRITIXG. have been found, should it have been superseded by a more perfected one. And when the time came for the systematiza- tion of the Indian writings, it is possible that Palibothra (Patna), where A9oka ruled in his time, and which was an important centre in the country, perhaps because it was the arrival point of the eastern trade, did not put forth the rights which long usage gave to its writing. I need not enter here into the particulars which the earlier Greek writers on India have given about the narrow know- ledge of writing existing in their time, and I shall confine myself to such necessary remarks which have not as yet been put forward. What the writing of the merchants of Shuh was we can easily know. It was simply an adaptation of the Chinese characters to the transcription of foreign words and names. Recent investigations in the history of the writing of the Middle Kingdom have shown that a system of syllabic spelling was mixed up from the beginning with the use of ideograms, and had gradually faded away through the exigencies of the intelligibility required, notwithstanding the regional variations of dialects, though preserved for the transcription of foreign words, and in such parts of the country as those of Ts'i and Shuh, which were inhabited by well- gifted populations. The simplification which the Chinese characters underwent after the time of Confucius could not but be favourable to such a state of things with regard to foreign relations. And tangible proofs of that former state of writing in the country of Shuh are still now found in the two surviving writings of the Lolos^ and of the Shui-kia tribes,^ both derived from simplified Chinese characters ; these populations belonging, still now in the case of the first, and formerly in the case of the second, to the ancient inhabitants of Szetchuen. 157. Now let us return to what we find in India. The oldest specimens known of the writing are the stone seals ' It is not impossible that the -writing of the Lolos should have been improved from its former ruder state, under some Buddhist influence, which at a certain early time was rife in the country. 2 Cf. §§ 176, 227-232, of the present work. BEGINNINGS OF WRITING. II7 found at Harappa near Lahore.^ They belong to the same kind of writing of Chinese derivation ^ which is now repre- sented by the writing of the Lolos. Next to these in point of date are the inscriptions of A9oka, in that writing which is called the Lat Pali, Indo-Pali, or Acoka characters. It rests undoubtedly on a different basis to that of the northern writing or Bactro-Pali of the same period, though the Semitic prejudice put forth since 1856 has led investigators to minimise the differences, and seek for a Semitic source other than for the first. I think that the seals of Harappa, and a better kno.wledge of the history of writing in the East, cannot fail to dispel the illusion under which many Indianists have laboured in that respect. As a matter of fact, the similarities of shape which have been pointed out between the Sabean characters and the Indo-Pali are insuflScient in number * to be satisfactory ; they do not go much beyond the similarities of the same number which could be pointed out with the Cypriote and other character^. This insuffi- ciency is made up, among the Semitists at any rate, by the supposition that many characters have been developed and varied from the former nucleus. And while admitting the possibility of the case in several instances, which however are not proved, we must state that the others are wholly unjustified. Now for the sake of the south Semitic hypo- thesis, those who upheld it are compelled to suppose four improbabilities : 1) that the Sabean characters existed a long while previously to the monuments hitherto discovered ; 2) that in this earlier period they were framed in the same rigid and square monumental style; 3) that this monumental or lapicide style, which elsewhere is generally used only for inscriptions, was also employed for their current use by the traders who are supposed to have carried it to the Indian coast ; and 4) that these merchants, contrary to what ex- perience shows to have been the case elsewhere, would have 1 Cf. «§38n.,and226. 2 Cf . the analogy I haTe shown on the plate of my paper, On a Lolo Manuscript Written on Satin (London, 1882, 870.). * The number has been eked out mth some shapes taken from a writing derived from the Musnad and younger by several centuries. 118 BEGINNINGS OP WRITING. preferred this stiff writing to the cursive one which, was also employed in their country, as shown by numismatic evidence,^ and which seems to have had a wide extension. I think that these four questions, which are resolved negatively against the Semitic hypothesis, have a great weight in favour of another solution of the problem by themselves, and also still more when coupled with the remarks above. The Semitic hypothesis, which has lately found some skilful advocates,^ who have been unable to show any new and satisfactory proofs more convincing than before, is decidedly below the mark, and an ingenious scholar * was led by their want of success to start a new theory of his own, which, however, is against the full weight of historical history and palseographical principles. The systematization of the Indo-Pali alphabet cannot have been made independently of that of the Bactro-Pali, and it is most probable that both took place at one and the same time, as the parallelism of the two shows obvious evidence of a reciprocal influence. And as I have remarked elsewhere, the vocalic notation arose most probably from the habit of making monograms, a Persian contrivance, leading to em- bodying the vowel characters into a typical consonant, and thus extended to the other consonants. 158. Put face to face one with the other, the lame proposal of a Semitic origin, and the probabilities of an Eastern origin for the groundscript of the Indo-Pali characters,* stand no more on one and the same footing ; the former hypothesis yields certainly to the latter. The solution of the question is given by palaeography, as shown by the seals of Harappa, and the comparative table given on Plate VI., where the characters of A9oka are put in parallel with the Siao-tchuen 1 Cf. § 141, above. 2 Eev. Dr. Isaac Taylor and Dr. R. N. Oust. Prof. Dr. Buhler, of Vienna, thougli still in favour of a Semitic origin, has declared himself not satisfied with the supposed proofs hitherto put forward. 2 Prof. J. Halevy, whose paper has been quoted above, § 111 n. * I have advocated this solution for several years ; and before the Royal Asiatic Society, 20 June, 1881, I read a paper On the Sinico-Iniian Origin of the Indo- JPAli Characters, of which I withhefd the publication, in the hope that newdis- closures would come forward and facilitate the solution of the problem. They have come in favour of my views and against the Semitic theory. BEGINNINGS OF WRITING. HQ characters,! which I look upon as their antecedents.'' All this requires to be explained with more length and precision in a special paper. d) In Central Asia. 159. The Tibetan sources record^ that the writing of Li-yul, otherwise Khotan, was invented by the Arya Vairotchoma, some 165 years after the establishment of the kingdom by Kusthana. And as the date of the latter event is stated to be 234 years after the Nirvana, which here means the Buddhist era 367 b.c.,* the introduction of Indian writing into the Tarym basin occurred not many years (thirty-two) after the Christian era. When Hiuen Tsang, in the seventh century, went there, he remarked, " their written characters and their mode of forming their sentences resemble the Indian model ; the forms of the letters differ somewhat ; the differences, how- ever, are slight. The spoken language also differs from that of other countries." ^ 160. At Yarkand, when Sung-yun, the Buddhist mis- sionary, passed there in 519 a.d., the customs and spoken language were like those of the people of Khotan,^ and the written characters in use were those of the Brahmans. The statement was partly confirmed the following century by Hiuen Tsang, saying, " the letters are the same as those of Khotan, but the spoken language is different." ' ' An apparent objection, which might be pnt forward, can be easily disposed of. It may be said that the comparatively small number of the Indian characters cannot fail to meet similarities among the many Chinese symbols. Such is not the fact ; the Chinese symbols, simple in shape and strokes, are few, and their whole niimber does not reach a hundred. ^ The signs are extracted from the Tchuen tsse wei, and from the SJiwoh wen, both dictionaries of the writing employed during the centuries immediately pre- ceding the Han period (206 B.C.). They are syllabic, and were pronounced according to the phonetic peculiarities of S.W. China. Some remains of a rude syUabie system are found in some of the Indo-Chinese writings, as well as some of the co-called cerebral series, whose home was in S."W. China and Indo-China, as shown by their great number in the several phonesis of these regions. * Gf . Woodville RockhiU, The Marly History of Li-yul {Ehoteii) in his learned work, The Life of the Buddha (from Tibetan sources), p. 237. * Prof. Rhys Davids, Ancient Coins and Measures of Ceylon (Appen.) ; Budd- hism, p. 213. ' S. Beal, Buddhist Records, vol. ii. p. 309. ' S. Beal, Buddhist Records, vol. i. p. Ixxxix. ' Buddhist Records, vol. ii. p. 308. 120 BEGINNINGS OF WRITING. 161. According to Chinese sources, in An-sih, i.e. Parthia/ about the first century b.c, they used to write on skins ^ in horizontal lines, by which means they kept their records.' We know from other sources that this writing was Greek, and recent numismatic research has made it quite evident that the Greek letters and the Greek language were generally understood in Northern India and in Kabul as late as the second century of our era.* It was so at least and at first in the Court circles, seeing that the coins of the period are generally bilingual Greek and Bactro-Pali until the time of Kanerkes, circa 75 A.D., one of the Kushan kings of Yueh-ti origin, when the sole writing used in the exergues is a debased Greek.^ 162. About 478 a.d. the Turks Kiu-s%e ^ gfH® were acquainted with the Chinese characters, and they had a writing of their own. They were then resident in Turfan, , on the North-West of China, previous to the Eao-tchang.'' 163. Sung-yun, the Buddhist missionary, who went to India about 518 a.d., passing west of the Moving sands, meets the Tuh-Kiueh 'hun, the Turks Hun, and observes that their written characters were nearly the same as those of the Wei.8 The second denomination looks like antagonistic to the first, as it is later on commonly attributed to the Indian or Nagari character. But it is not so in the present occurrence, 1 So called by the Chinese from its founder Araakes, B.C. 250, who had estab- lished his first citadel at 'Ao-a^x (Asak-abad) . On the latter cf. Isidore of Charax, ch. xi. ; J.RtA.S. 1871, p. 445. ^ The Arab historian Masudi (tenth century), in' his apparently authentic account of the Zendavesta, states that it was written on 12,000 cowhides, in a character invented by Zartusht (Zoroaster). 3 Tsien Han shu, bk. 96 ; A. "Wylie, Ifotes on the Western Regions, in Jour, Anthrop. Inst. 1880 ; repr. p. 21. * Prof. Percy Gardner, The Cqins of the Greek and Scythic Kings of Bactria and India, p. liii. ^ On debased Greek characters on Parthian Coins, cf. Edward Thomas, Parthian and Indo-Sassanian Coins, p. 9 (regr. Jour. Roy. Asiat. Soo. 1883). ^ A. L. Davids {Grammaire Turke, London, 1836, 4to. p. xii) has remarked that this name resembles that of the Ghauzz or 'JDuzz tribe from whom the Osmanlis claim to be the descendants. ' Abel Eemnsat, Recherches sur les lamgues Tartares, in A. L. Davids, op. cit, p. xix. * S. Beal, Buddhist Records, vol. i. p. Ixxxv, " The written character of the country is nearly the same as that of the Wei.' ' BEGINNINGS OF -WEITING. 121 inasmucli as the Hu shu was indeed the writing used by the Buddhists in some parts of Central Asia, as we shall see below.^ 164. There are some reasons to believe that the Topa Tartars or Northern Wei, who ruled in the north of China from A.D. 386 to 534, had a writing of their own,^ at least about A.D. 476-500, under the reign of Wen-ti. And it is probable that this writing was the, or derived from the, Uigur writing.' 165. The Chinese annals of the Sui dynasty (581-618 A.D.) contain a special section, presenting a general survey of the existing literature as collected in the Imperial library.* The division concerning Buddhism includes some 1950 distinct words. Many of the titles are given, and among them are not a few which treat of the mode of writing, by alphabetic symbols used' in the kingdoms from whence Buddhism came. The first alphabet that was thus introduced appears to have been one of fourteen symbols. It is called Si-yii hu shu "Foreign writing of the Western countries," and also Po-lo-men-shu " Brah- manical writing." * We know what is the common mean- ing of the word hu, and in- this case we are inclined to accept it with its occasional substitution of Uigur. Inasmuch as the number of fourteen is that of the con- sonants in that writing, which was known to some Arab 1 Of. below, §§ 165, 170. 2 "We cannot speak with certainty," says Mr. A. Wylie (Journ. Eoy. Asiat. Soc. 1860, vol. xvii. p. 333), " in this matter about the achievements of the Topa- Wei Tartars, while they held rule in the northern parts of China, as the Northern Wei dynasty, during the fifth century of the Christian era, though there is ground to believe that they had a particular character for writing their original language. Thus we read in the history of the Suy dynasty (bk. 32, fol. 18) that when the Wei Tartars took possession of Lo-yang, they were unacquainted with the Chinese language ; and their Emperor, Wan-te, gave orders to How Fo-how and Ko Seih-ling to translate the Biao Xing or ' Book of Filial Piety ' into the Court language." In the bibliographical section of the same work (bk. 32, fol. 22) the names of nine other books relative to the language of the people are given. The names of three books on the language of the Sieu-pei, a neighbouring tribe, are also preserved in the same place. s Cf. § 120, above. * A. Wylie, Ifotes on Chinese Literature, pp. vii, viii. ' J, Edkins, Chinese Buddhimn, p. 112. Sui shu, bk, 35. 122 BEGINNINGS OF VEITING. historians by this peculiarity,^ and which was derived from the Syriac.^ 166. In the same part of the Annals of the Sui dynasty, a Chinese Buddhist missionary in Central Asia, at the time of T'ai Wu Ti (a.b. 424-451) of the Wei dynasty, was taught a writing composed of thirteen characters only.' 167. In Tokharistan (Tuholo), about 630 a.d., according to Hiuen Tsang, they possessed extensive literary records.* " Their language differs somewhat from that of the other countries. The number of radical letters in their language is twenty-five ; by combining these they express aU objects around them. Their writing is across the page, and they read from left to right." ^ At Bamian, Kapisha, and Shambhi, about the same time, the writing in use was the same as that of the Tuholo.^ "What this writing was is not clear.' The direction from left to right agrees with any Indian writing from the Nagari class, but the number of characters is too small, and Hiuen Tsang would not have failed to indicate the fact of the Indian derivation as he did in other instances. It may have been a writing of special adaptation to the exigencies of the languages by the skilful Buddhist missionaries.* 168. At Kashgar {Kie-sha), in the seventh century, the written characters were imitated from those of India, and • Ahmed ben Arahshah, in his history of Timur. J. Klaproth, Verzeichniss der Ghinesischen and mandehuischen Biocher der Koniglichen Bihliotek zu Mtrlin, Paris, 1822, fol. 2 Cf. § 120. 3 Sui shu, hk. 35, fol. 21-22. * Which were reputed to exceed those of Sulek. ' Cf. S. Beal, Buddhist Secords, vol. i. p. 38. 8 Idid. Yol. i. pp. 50, 54 ; toI. ii. p. 296. On the Tuholo, which must be dis- tinguished from the Tueh-ti, cf. S. Beal, Gleanings from the Si-yu-ld, p. 253 \jour. Roy. Asiat. Soc. Vol. XVI. 1884). Mr. G. de Vasconcellos-Abreu {Be Vorigine probable des Toulchares, Louvain, 1883) finds in the Tuholo the Greek Teuchri. Anyhow they were no more there at the time of Hiuen Tsang, who is careful to say that he speaks of the old country of the Tuholo. ' M. G6rard de Eialle [Memoire sur I'Asie Oentrale, 2nd edit. Paris, 1875, 8vo.) rightly remarked (p. 44) that it could not be the Zend writing, which runs from right to left, but more probably an Indian writing. 8 The late Mr. Edward Thomas has pointed out Hindi legends on coins of the region. Cf. his papers : Parthian and Indo-Sassanian Coins, 1883, p. 24, and he thinks that the debased Southern Hindi characters were those alluded to by Hiuen Tsang. Cf. also from the same scholar, Indo-Set/thian Coins with Hindi Legends {Indian Antiquary, 1883). BEGINNINGS OF WRITING. 123 although somewhat mutilated, yet they were essentially the same in form.' The statement comes from Hiuen Tsang, who also says that at Och (TJsha) and at Khavandha^ {K'ie-p'an-fo) the writing was much like that of Kashgar.' In the Tarym basin, coming from China, the same missionary had remarked * that, at Akin, Kutche (formerly Kuei-tze^) and Oksu, the written characters were Indian with some differences.* At Kumo, between Kashgar and Kutche, the literature was the same as at the latter place.' e) The Suleh Writing. 169. Hiuen Tsang, on his way to Samarkand by a route south of the Tien shan,® reached the Su-yep or Tchu river, and there he found a writing differing from those of Indian descent, which he had hitherto described. " From the town of the Su-yep river, as far as the Ki-shwang-na (Kesh) ' country the land is called Sulek, and the people are called by the same name. The written characters and the spoken language are likewise so called. The primary characters are not many ; in the beginning, they were thirty-two * in number : the words are composed by the combination of these : these combinations have produced a large and varied vocabulary. They have some literature, which the common sort read together ; their mode of writing is handed down from one master to another without interruption, and is thus preserved." ^ ' Hiuen Tsang, circa 630 a d. Cf. Seal's Buddhist Mecords, vol. ii. p. 307. => Ibid. vol. a pp. 299, 304. = Otherwise Sarikhul. • Ibid. vol. i. pp. 18, 19, 24. In my copy of the Ta Tang si yu hi the words used are ^ gl] fp g. ' In the Annals of the T'ang dynasty (618-905 a.d.) Turn Men lei han, bk. 237, fol. 12. In 601 one of the two Indian translators in Chinese of the Saddhartnapundarika-sutra states, in the Preface, that he has seen two texts of the work ; one written on the palm leaves (from India), and the other in the letters of Kwei-tze. Cf. Bimyiu Nanjio, A Catalogue of the Buddhist Tripitaka, p. 46. * Cf. E. Bretsohneider, Notices of the Medi