National Institute of Informatics - Digital Silk Road Project
Digital Archive of Toyo Bunko Rare Books

purport may be often seen even now stamped on the outer edge of silk rolls as exported westwards from Ssû-ch`uan, etc. To the Brahmi inscription would correspond the brief markings in Persian with which Muhammadan traders like my friend Sher `Ali Khan, the Kabuli merchant from Khotan and Yarkand, are accustomed to label the fabrics their caravans carry, whether silks exported from Ssû-chuan or British muslins, etc., brought from Kashmir.

Accepting then gisli >gillh to mean a ' span ' and assuming, which seems reasonable, that the ancient trader's note referred to the complete piece of silk as purchased, we might attempt to determine its approximate original length. It is certain that by the modern Indiangillh is meant a span measured between the extreme tips of the little finger and the thumb. But such a measurement can scarcely have an exactly fixed value nowadays, and we know still less what its accepted value may have been in Central Asia about the time of Christ. So it seems better to base our attempt at a conjectural determination on the assumption, likely a priori, that the standard of length was approximately the same about the beginning of the first century A. D. (T. xv. a. iii. 57) as about its end (T. xv. a. i. 3). There seems distinct support for this assumption in the fact previously noted that the width of the silk exported from China had undergone no change between the beginning of the first century A. D. (T. xv. a. iii. 57) and the third or early fourth century A. D. (L.A. r. 002).3 If, then, we suppose the 46 gisli or spans of the note in Brâhmi script to have been the equivalent of the 40 (Chinese) feet which we find recorded in T. xv. a. i. 3 as the original

length of the silk piece, we arrive at the equation of z gÇc = 22 9 46 X 4°_ t9 9 cm., or close on

8 inches. The result coincides closely with the average span of the hand in India and the Middle East, and thus indirectly offers some support for M. Boyer's interpretation of both gisli and saparisa.

The first word still awaits explanation. Except for the initial aksara, which has suffered through a hole in the silk, but which both Dr. Hoernle and M. Boyer are inclined to take for ai,' the reading [ai}slasya is assured. That the word shows the Sanskrit genitive case ending -sya is certain. The mixture we see here of Sanskrit and Prakrit forms can, as M. Boyer rightly points out, cause no difficulty to any one who is familiar with the language of the Kharosthi documents from the Niya and Lou-lan Sites.b There remains the question of the import of this genitive and of the meaning of [ai]sla itself. No suggestion can be made as to the latter. But keeping in view the purpose of the ' label ' entry, it has occurred to me—and M. Boyer sees no objection to urge against such a view—that we might possibly have here a partitive genitive meant to designate the particular quality or material of the silk contained in the roll. It is possible, however, that the word in the genitive was intended to designate the purchaser or something of the same sort.

But even with . this point left in doubt some observations of archaeological and also historical interest may be founded on the brief record now interpreted. In the first place, it proves that during the period between 6 i B. C. and A.D. 9, roughly comprising the last reigns of the Former Han dynasty, traders accustomed to use an Indian script and language must have already made their way across the Chinese Limes for, the sake of the 'silk of the Seres'. It would be useless in the

Measure of grti or span.

Interpreta- tion of [atislasya.

Indian language of early silk trade record.

3 Cf. above, p. 374.

' M. Boyer, writing on April 4, 1917, remarks : ' Je crois très probable que, dans cette inscription prâkrite, le caractère en question représente un ai, dont la forme est basée non sur celle de e, mais sur celle de a, d'après le procédé graphique des alphabets par vous découverts et publiés par le D* Hoernle [see J.R.A.S., 1911, pp. 451 sqq.]. De cet ai il reste la double courbe spécifiant le phonème et des portions de la

forme base a, savoir : le haut et le bas de la haste de droite et la tête pâteuse (par la rencontre des courbes voyelle) de la partie gauche. Parmi les variantes de forme possibles pour cette dernière partie, celle que la déchirure a fait disparaître ici demeure naturellement indéterminée.'

For some general observations on this point, cf. Ancien! Kholan, i. pp. 364 sqq. ; above, p. 414.