Sec. i] THE EARLY RECORDS AND NAMES OF KHOTAN 153
this assumption 7. Yet owing to the Khotan rulers of this period appearing only under Chinese names in the Tang Annals, it is impossible to attempt the identification of any of them with those bearing Indian names in the Tibetan list. All that can be suggested at present with some probability is that the family name of Wei-ch`ih, which is given to the period by the Tang Annals and Wu-k`ung 8, may be a rendering of the Sanskrit Vijaya-, which appears invariably as the first part of the names recorded in the Tibetan list. The occasional references which we meet to the relations of ` Li-yul ' with other territories besides China afford little or no clue, since the latter can scarcely ever be identified under the strange disguise of their Tibetan names 9.
The difficulties and obscurities here indicated make it impossible to treat the information gathered from the Tibetan accounts in connexion with the strictly historical data supplied by our Chinese records. Nor can I attempt to subject that information to a separate critical analysis. Whatever in the Tibetan notices I have been able to recognize as throwing light on the early legends, sacred sites, and historical relations of Khotan, will be duly noticed hereafter in connexion with other available evidence, while for a complete list of the kings and the main facts noted regarding them in the ` Annals of Li-yul ' I must refer to the synopsis furnished by Mr. Thomas in Appendix E.
An ancient legend connects the name of Khotan with the very origin of the kingdom. The transformations of this name reflect, as it were, the succession of cultural and political influences which the territory has undergone from an early period. Hence our review of the historical notices concerning Khotan may fitly commence with a brief survey of the varying forms of its name. We find the most important of them conveniently enumerated in an interesting passage of the Hsi yü-chi which, omitted by Julien and Beal, has been duly noticed and translated by M. S. Lévi 10. Hsüan-tsang, at the conclusion of his account of Chê-chü-chia, tells us : After marching Boo li, one arrives at Ch`ü-sa-tan-na . -- fii MS. In Chinese this means " breast
of the earth ". This is a popular interpretation. In current speech one says Huan-na I3.
The Hsiung-nu say, Yii-tun 2. The Hu say, Huo-tan n (for ) q. The Yin-tu
[Hindus] say, Ch`ii-tan MA.. Formerly one said, gti
The name Ch`ii-sa-tan-na, which Hsüan-tsang mentions first, is undoubtedly intended to reproduce the Sanskrit Kustana, `breast of the earth ', as recognized by Chézy and Rémusat12. It is subsequently accounted for in the legend which Hsüan-tsang records of the first king of Khotan and his miraculously born son, and which we shall have occasion to discuss presently. The pilgrim himself calls the meaning given to the name ` a popular interpretation ', and it can scarcely be doubted that the form ` Kustana' itself is only the result of a learned ` popular etymology ' which endeavoured to provide an orthodox Sanskrit derivation for the old local
See the notices concerning the kings numbered g, lo, 13 in Mr. Thomas' supplementary list, App. E.
See Chavannes, Turcs occid., p. 1 26 ; also below, p. 173•
' The sanie difficulty is experienced about the identification of Khotan localities mentioned in the Tibetan extracts with names which seem more puzzling even than Chinese transcriptions of local names.
1° See S. Lévi, Notes chinoises sur l'Inde, iv. p. 44. The passage had been previously reproduced by A. Rémusat, Ville de Khotan, p. 35, from a note of the Chinese editor of the Pien i tien, but without an indication of its origin.
" The same forms of the name are enumerated in the passage which opens the Tang Annals' notice on Yii-tien, and which is manifestly an abbreviation of the above passage in the Hsi yii-chi ; see Chavannes, Turcs occid., p. 125. Here the forms Ch`ii-sa-tan-na, Huan-na and Ch`il-tan are mentioned together, without a distinction as to their use. The barbarians of the North (Ti) instead of the Hsiung-nu are credited with the form Yil-tun, and the form Huo-tan
appears in a different graphic form as + ; for the
latter form comp. also L'Itinéraire d' Ou-kong, p. 26. 12 See Rémusat, Ville de Kholan, p. 36, note.