410 181. COTAN
in these Kharosthi documents is everywhere else Khotamna or Khotana. Moreover, we are now in a position to prove definitely that Levi was right. Among the documents in Khotanese which I brought back from the Tunhuang caves, BAILEY found a sort of a bilingual « manual of conversation » in Sanskrit and Khotanese; the Sanskrit name of Khotan is given there as Gaustana and Gârrmstana, both graphic renderings of Gostana. But Hsüan-tsang tells us that this was « the elegant name [used] locally », which I understand to mean that, though it was not the name of Khotan in Khotanese current speech, yet the Khotanese language used Gostana as a more literary name of the country. This is confirmed by the occurrence of Gâmsta (_ *Gaustam) and Gaustamä in a Khotanese document (cf. BAILEY, in BSOS, ix, 522, 541). BAILEY explains Gaustamä as « an inverse spelling of *gaustam, after -ana and -ama had become -am. It may be so, but the solution depends to some extent upon the very history of the name Gostana.
At the same time as Ltvi, WATTERS (On Yuan Chwang's Travels, ii, 299) was also struck by the constant value of ch'ü as go- or gu- in transcriptions. Starting from the other forms Yü-t'ien, Ch'ü-tan, etc., WATTERS supposed that they represented « a word like Go-dan or Gothàn ». «The former », he added, « is found in Turki, and the latter in the Indian vernacular, and they represent the Sanskrit Gosthâna. These words denote a place or station for cattle ». According to WATTERS, it is this Gosthâna which, mispronounced and misunderstood as Kustana, led to « the silly legend invented to account for the name ». It is true that Skr. gosthâna, lit. « cow-station », means a « stable », or an « enclosure for cattle » ; but the Turki word intended by WATTERS, gown, « enclosure for cattle (mainly for sheep) », identical with Mong. gotan > hoto, « enclosure for cattle » and « city », has of course nothing to do with gosthâna, nor with the name of Khotan. Something of WATTERS'S hypothesis may however be retained, and was in fact retained by THOMAS (in Asia Major, II, 261; Tibetan Texts and Documents, I, 18). Skr. sthâna, « station », « country », occurs mainly in the geographical nomenclature of the North-West, and in some cases seems to be due to the much more frequent use of its Iranian equivalent -stâna. I have no doubt that for instance the well-known Skr. Cinasthâna, « China », which is used in a Kharosthi document (RAPSON, Kharosthi Inscriptions, 12, No. 35, where there is once cina and twice cinasthanade), is the Indian recasting of an Iranian *Cinastâna, the original of the Sogdian instan (see « Cin »). The correct form of the name restored by LLvi as « Jagudavardhana » (JA, 1915, I, 84) is Jagudasthâna, Zabulistan. Tibetan texts mention an Aryastana on the Go§trlga hill of Khotan (THOMAS, in Asia Major, ii, 261; Tibetan Texts and Documents, 1, 8, 18, 27, 108, 109). A Gostana, actually formed of go ± Iran. stâna > stana, « country », could of course be interpreted as a Skr. compound of go + stana meaning « earth-teat », and this mistaken interpretation would be at the basis of the whole legend recorded by Hsüan-tsang and in Tibetan texts (I find no trace of the intermediary Prâkrit form «Wothana» mentioned by Ltvi, JA, 1915, I, 110); that Gostana in its turn may very well have become *Gostan at an early date in Khotanese, and be used as such, in « elegant » speech, alongside with what we know to have been the current form of the name in Khotanese in Hsüan-tsang's time. As to the value of the first element go- of Gostana, I shall come to a discussion of it farther on. It may be that a form related to Gostana left traces in Tibetan
IIIF Color HighRes
