Changes

very well aware of the fact that the speech of the people of Fergana was Iranian, for he stated in his report, that, although there were different dialects in the tract of land stretching from Fergana westward as far as Parthia (An-si), yet their resemblance was so great that the people could make themselves intelligible to each other.1 <ref>Strabo (XV. II, 8) observes, "The name of Ariana is extended so as to include some part of Persia, Media, and the north of Bactria and Sogdiana; for these peoples speak nearly the same language."</ref> This is a plain allusion to the differentiation and at the same time the unity of Iranian speech; 2 <ref>Emphasized by R. Gauthiot in his posthumous work Trois Mémoires sur l'unité linguistique des parlers iraniens (reprinted from the ''Mémoires de la Société de Linguistique de Paris'', Vol. XX, 1916).</ref> and if the Ferganians were able to understand the Parthians, I do not see in what other language than Iranian they could have conversed. Certainly they did not speak Greek or Turkish, as some prejudiced theorists are inclined to imagine.

The word brought back by Can Čaṅ K'ien for the designation of alfalfa, and still used everywhere in China for this plant, was ''mu-su H fit'' <>, consisting of two plain phonetic elements, 8 <ref>The two characters are thus indeed written without the classifiers in the Han Annals. The writings <> *muk-suk of Kwo P'o and <> *muk-swok of Lo Yüan, author of the ''Er ya i'' (simply inspired by attempts at reading certain meanings into the characters), have the same phonetic value. In Annamese it is ''muk-tuk''.</ref> anciently *muk-suk (Japanese ''moku-Sukušuku''), subsequently written H" Hf <> with the addition of the classifier No. 140. I recently had occasion to indicate an ancient Tibetan transcription of the Chinese word in the form ''bug-sug'', 4 <ref>T'oung Pao, 1916, p. 500, No. 206. </ref> and this appears to come very near to the Iranian prototype to be restored, which was *buksuk or *buxsux, perhaps *buxsuk. The only sensible explanation ever given of this word, which unfortunately escaped the sinologues, was advanced by W. Tomaschek, 6 <ref>Pamir-Dialekte (''Sitzber. Wiener Akad.'', 1880, p. 792). </ref> who tentatively compared it with Gllakl Gīlakī (a Caspian dialect) buso ''būso'' ("alfalfa"). This would be satisfactory if it could be demonstrated that this buso ''būso'' is evolved from *bux-sox or the like. Further progress in our knowledge of Iranian dialectology

will no doubt supply the correct form of this word. We have to be mindful of the fact that the speech of those East-Iranian tribes, the advance-guard of Iran proper, with whom the Chinese first came in contact, has never been committed to writing, and is practically lost to us. Only secluded dialects may still harbor remnants of that lost treasure. We have to be the more grateful to the Chinese for having rescued for us a few words of that extinct language, and to place *buksuk or *buxsux on record as the ancient Ferganian appellation of ''Medicago sativa''. The first element of this word may survive in Sariqoll Sariqolī (a Pamir dialect) ''wux '' ("grass"). In WaxIWaxī, another Pamir idiom, alfalfa is styled ''wujerk''; and grass, wiiL ''wüš''. "Horse" is ya$ ''yaš'' in WaxIWaxī, and ''vurj '' in SariqollSariqolī.<ref>Cf. R. B. Shaw, On the Ghalchah Languages (''Journal As. Soc. Bengal'', 1876, pp. 221, 231). According to Tomaschek (''op. cit.'', p. 763), this word is evolved from *bharaka, Ossetic ''bairāg'' ("good foal"). </ref>

1 Bretschneider2 Bretschneider<ref>Bot. Sin., pt. III, p. 404.</ref> was content to say that ''mu-su '' is not Chinese, but most probably a foreign name. Watters, in his treatment of foreign words in Chinese, has dodged this term. T. W. Kingsmill8 Kingsmill<ref>''Journal China Branch Roy. As. Soc.'', Vol. XIV, 1879, p. 19.</ref> is responsible for the hypothesis that ''mu-su '' "may have some connection with the M7761K17 PoTavrj ''Μεδικὴ βοτάνη'' of Strabo." This is adopted by the Chinese Dictionary of Giles.* <ref>No. 8081, wrongly printed Μεδική. The word ''βοτάνη'' is not connected with the name of the plant, but in the text of Strabo is separated from ''Μεδικήν'' by eleven words. ''Μεδική'' is to be explained as scil. ''πόα'', "Medic grass or fodder."</ref> This Greek designation had certainly not penetrated to Fergana, nor did the Iranian Ferganians use a Greek name for a plant indigenous to their country. It is also impossible to see what the phonetic coincidence between *muk-suk or *buk-suk and medike mēdikē is supposed to be.

The least acceptable explanation of ''mu-su '' is that recently propounded by Hirth, 6 <ref>''Journal Am. Or. Soc.'', Vol. XXXVII, 1917, p. 145.</ref> who identifies it with a Turkish burfak''burčak'', which is Osmanli, and refers to the pea. 8 <ref>''Kara burčak'' means the "black pea" and denotes the vetch.</ref> Now, it is universally known that a language like Osmanli was not in existence in the second century B.C., but is a comparatively modern form of Turkish speech; and how Can Čaṅ K'ien should have picked up an Osmanli or any other Turkish word for a typically Iranian plant in Fergana, where there were no Turks at that time, is unintelligible. Nor is the alleged identification phonetically correct: Chinese ''mu'', *muk, *buk, cannot represent ''bur'', nor can ''su'',