国立情報学研究所 - ディジタル・シルクロード・プロジェクト
『東洋文庫所蔵』貴重書デジタルアーカイブ

Besides the temples mentioned above, there is at Alchi a great number of ancient
*mchod-rten* and *mchod-rten* gates with four doors. Several of them contain pictures simi-
lar to those noticed at Basgo and Ubshi, *i.e.*, Buddha-like figures of blue complexion.
They probably represent Nāgas, and these gates and *mchod-rten* may be of Bonpo
origin. But there were so many that it was found impossible to register them all.

Let me add a few words with regard to Tibetan palæography, based on a compari-
son of the Alchi and Tabo inscriptions of the 11th century. Besides the Alchi and Tabo
inscriptions of the 11th century, the following datable inscriptions of ancient Tibet are
known : (1) The Endere manuscripts, excavated by Sir M. A. Stein in Turkestan.
The latest date which can be attributed to them is the year 788 A.D., but several appear
to be earlier. (2) The inscription of Khri-srong-lde-btsan in Lhasa of c. 780 A.D.
discovered and edited by Dr. Waddell.¹ (3) The inscription of king Ral-pa-can in Lhasa,
of c. 810 – 820 A.D.²

The most archaic of the Endere relics have the following peculiarities :—
(1) The *i* vowel sign is often inverted.
(2) In several cases, the final consonant of a syllable is written below the first.
(3) The article is in many cases *pha* and *pho*, instead of the ordinary *pa*
and *po*.
(4) Ordinary *c* and *ts* are in many cases replaced by *ch* and *ths* ; and *ch* and *ths*
both have *g*, *b*, and *d* prefixes attached to them.
(5) When *m* precedes *i* or *e*, *y* intervenes.
(6) Words ending in *r*, *l*, or *n*, are furnished with a *d* suffix, called *drag*.

If we examine the Tabo, Poo and Alchi inscriptions of the 11th century, with re-
gard to the peculiarities of the Endere manuscripts enumerated above, we find that they
exhibit only the two last mentioned peculiarities, *viz.*, *y* intervenes between *m* and *i* or *m*
and *e* ; and *r*, *l*, and *n* are often furnished with a *d* suffix. Thus we see that during the
period from the 8th to the 11th century, the Tibetan orthography has come much nearer
to its present state. The *i* vowel sign, for instance, is no more found inverted, but always
in its present position. At Alchi we found the *i* vowel sign only once inverted and that
was probably due to want of space. From this observation we may conclude that all
those inscriptions which contain inverted *i* vowel signs, may be older than the 11th
century. As regards the position of the *e* and *o* vowel signs on the right or left upper
end, or above the consonant base, it varies with the age of inscriptions. I am of opinion
that the compilation of the *bKa'-agyur* and *bStan-'agyur*, in c. 1300 A.D., marks an
epoch in Tibetan palæography. It probably put an end to the intervening *y* between
*m* and *i* or *e*, and to the suffixed *d*. From the year 1300 A.D. onwards Tibetan ortho-
graphy presumably remained stationary, and the age of an inscription after 1300 A.D.
can be estimated only by the form of its compound letters, and the position of vowel
signs on or above their consonant bases.