98 Journal of the Asiatic Society of Bengal. [March, 1910. 



Barnett. I readily accepted his view. Dr. Waddell has recently 

 repudiated it stating that it was founded on nothing better 

 than the occasional occurrence of the drag in ancient docu- 

 ments. (J.R.A.S. October, 1909 ) No, it was never based on 

 so poor a foundation. 



What induced Dr. Barnett to believe in a possibly earlier 



introduction of the Tibetan script, was the fact that in Tur- 

 kestan at Endere, on the very confines of the Tibetan empire, 

 were found specimens of Tibetan writing in not one but two 

 forms of script, one of which exhibited already traces of great 

 simplification, which can be explained only with the accept- 

 ance of the theory of a long period of use of the same. And 

 the latest date which can be assigned to these documents is 

 c. 780 A.D., about 120 years after the asserted invention of 

 the script by Thonmi. But the Endere relics are not all of 

 the same type, some exhibit a more archaic type of orthogra- 

 phy than the others, and have to be dated considerably earlier 

 than 780. 



Turkestan is exactly the country where a new kind of 

 script and literature could most probably have originated. 

 The Buddhists of Turkestan were more eager than any other 

 to provide people of various tongues with Buddhist literature 

 in their own language. Proof of this are the various MSS. in 

 unknown languages which have come to light there. Turkes- 

 tan was in possession of a form of the Gupta alphabet (the 

 Bower MSS., etc.), and this alphabet impressed its type on the 

 Kashgar Brahmi as well as on the Tibetan script. When I 

 wrote my article on l- The Similarity between the Tibetan and 

 Kashgar Brahmi alphabets,' 5 published by this Society, I might 

 have added a column showing the Gupta characters ! The 

 Gupta alphabet has variants, but its descendants here in the 

 West are all sprung from one and the same type of Gupta. 

 These descendants are the Kashgar Brahmi and Tibetan 

 characters, and the Indian characters used in Ladakh between 

 700 and 900 for writing the Sanskrit formula Ye dharma, etc. 

 This formula was written in Tibetan characters as well, and at 

 first sight, the Indian and the Tibetan variety of this formula 

 can hardly be distinguished; for most of the characters look 

 the same. The difference rests in this, that in the Tibetan 

 version the aspirated mediae are written with an ordinary 

 media furnished with a subjoined "A," whilst in the Indian 

 version gh, dh and bh are written with simple characters* 

 Besides, the Tibetan version has the Tripartite y, whilst the 

 Indian version has a later form of the y. The west (Kashmir, 

 Ladakh, Turkestan) is the country where the Gupta form of 

 characters remained stationary for a longer period than else- 

 where. Here is the probable home of the Tibetan script. It 

 was invented not many centuries before Srong btsan sgampo, 

 possibly one or two centuries before him. The Brahman 



