412 181. COTAN
v, 289-290; some of the names in the Chinese list are somewhat surprising at such an early date); the same rendering of Hiina with Hsiung-nu occurs in 308 in the P'u-yao ching (BEFEO, Iv, 575-576). But the days of the Hûna in North-Western India were already passed when Hsüan-tsang visited India and Central Asia. In the 7th cent., the Hsiung-nu had been replaced in Central Asia by the T'u-chüeh (Turks), who, rightly or wrongly, called themselves in their inscriptions the descendants and successors of the old Hsiung-nu. On the other hand, Hsüantsang often speaks of the T'u-chüeh, and if he had only wanted to say that Yü-tun was a T'uchüeh form of the name of Khotan, he might have said it in so many words. My impression is that he wanted to convey the idea that Yü-tun was the name of Khotan among the nomads of Central Asia, Mongols as well as Turks. At any rate, there can be no doubt that such was actually the case, since Yü-tun certainly represents what will soon be shown to have been the Turkish and Mongolian name of Khotan in the Middle Ages, to wit Odon.
We have seen above that the f yü (*jiu) of Yü-t'ien and Yü-tun, which is hardly ever used in Buddhist transcriptions, is of rather frequent occurrence in early transcriptions of Hsiung-nu words. Here again, it was for a « Hsiung-nu » term that Hsüan-tsang adopted it. We may therefore conclude that it probably had an initial element which did not occur in Indian speech, perhaps some sort of laryngeal opening, followed in the present case by a labial vowel. In other words, Yü-t'ien would be *'Odan, and Yü-tun would be *'Odon. If we remember the History of Chang Ch'ien's missions, it would be a natural inference to suppose that the name of Khotan came to his knowledge through a Hsiung-nu channel. In agreement with the native form, the Hsiung-nu pronounced it with an -a- in the second syllable, as *'Odan, but at a later period, the initial labial vowel, as has often been the case in Altaic languages, and particularly in Mongolian, labialized the vowel of the second syllable, hence *'Odon. The case is absolutely parallel with that of Mong. qotan, « enclosure for cattle » and « city », which has given qoton > hoton and hoto (the word is already qoton, plur. qotot, in the Secret History, §§ 124, 195, 247, 272, except once with plur. qotat, § 248). In such circumstances, it may not be necessary to resort to a hypothetical pronunciation *giu for -J yü (*jiu) in Han times. The native name must really have begun with a guttural consonant, perhaps sonant or spirant, but we are certain that there was none at the beginning of the Altaic form in the Middle Ages. As the absence of a guttural initial in the Altaic form is of course independent of the hypothetical evolution *giu > *jiu in Chinese, the simplest solution seems to be to conclude that Yü-t'ien already transcribed *'Odan in Han times, as Yü-tun rendered *'Odon in the 7th cent. Perhaps the Hsiung-nu language did not any more admit of an initial g- or y- than the Turkish of T'ang times.
The result of our examination of all these ancient forms is that the name of Khotan must have existed at a very early date, at least in the 2nd cent. B. C., in a form *Godan, which, in the first centuries of our era, was doubled by a form *Gostâna > Gostana. It seems reasonable to suppose that the two parallel forms have the same value in different languages, or at least in different dialects, and that both mean « Place (or Country) of the *Go ». Further speculation would be risky. It may be, but it cannot be proved for the present, that the -dan of *Godan is the same as the -dan of Humdan, the ancient name under which Hsi-an-fu was
IIIF Color HighRes
