266 155. CIN
There are other grounds for doubt. The Greek and Latin s- can represent not only the s-of an Oriental original, but also sr and even e". As a matter of fact, the Syriac word for « silk » is . éràyâ, and a Syriac work of the 3rd cent. A.D. mentions the Séres as Serâye, an ethnical form derived from *Ser (cf. Mo, 23). If we add that a theory has been started that the name «Séres» originally referred to the inhabitants of Chinese Turkestan (HERRMANN, Das Land der Seide, 27), that the ancient name of Kàgyar was perhaps *Sarag (see « Cascar »), that «Sarag» was the ancient name of Lo-yang among people of Central Asia and that Pers. säräh (< Ar. saraq) and Mong. *sirkäg, Manchu sirge bears a disquieting resemblance to o n pcxc v, it will appear that the problem of the names onpcx6v, Séres and o-f must remain in abeyance until it has been studied afresh in great detail.
The earliest mention of the second name, which is the one connected with Polo's « Cin », occurs in the form *Oiv (?) at the end of the 1st cent. A.D. in the Periplus of the Erythraean Sea (cf. HERRMANN, Das Land der Seide, 37; the views expressed there by HERRMANN are different from the ones he maintained in his article «Thinae» in PAULY-WlssowA). This country lies at the extreme north of the (Indian) Sea, beyond Chrysê. There is in the interior a great city called Diva (read *eiv?) whence the various products called o-npcx6v — wool, thread, and linen — are carried by land via Bactria to Barygaza (-= Broach), and also across the Ganges to Limyrikê (coast of Malabar). I feel much hesitation in giving *eiv as the form intended by the author of the Periplus; it does not appear in the index of the last editor, H. FRISK (Göteborg, 1927, p. 126), where we find only (-)11..x. All mentions of the name occur in §§ 64 and 65. The first mention, at the beginning of § 64, ... sic <Ocv,s> Tcva T67rov, is due to a correction of K. MÜLLER; HERRMANN, Das Land der Seide, 37, would favour a bolder emendation Eis Oiv Tnv xwpay. Although MÜLLER's text has often been accepted (it is the one quoted in YULE, HobsonJobson2, 197), FRISK retains the original reading El's Tcva T67rov, without comment; I shall do the same, since nothing can be said of the original form of the name on the basis of the correction. The other mentions are : 1. (§ 64) : 7r6Àcs fiso-6yecos ~eyio rj, AeçDo..évn Oiva, â ' r3.5 ...; 2. Eis
~E T))v Ocva TaÛTnv ...; 3. (§ 65) . É7r) T))v O'vvop âv TŸ)S 01v6S ...; 4. ... TES auvoplas aû'ri5v
xai ToiV Tics Ocv6s.... COEDEs (Textes grecs et latins relatifs à l'Extrême-Orient, 24) followed
MÜLLER and corrected the first Oiva to Oïvac (occurring in Ptolemy, VII, 3, 6, etc., and Martianus of Heraclia, I, 16) ; here again, FRISK retains the original reading. Half a century ago, VON GUTSCHMID, relying on the three other passages in which he saw the genitive and accusative of a name eiv, had corrected in the first case Oiva, a¢' ijs to ()iv, cap' 45 (Kleine Schriften, III, 604), which seemed at first sight reasonable, since a nominative Ohm cannot become Ocv65 in the genitive. But, according to FRISK, Oi:•a is an accusative employed as a nominative; in view of the genitive Ocv6s, the name is an -v word, «the nominative of which, *eis, has been replaced by the accusative ». FRISK adds that there was in Egypt a town of the same name, only mentioned too in the oblique cases (except once Ot's in Stephen of Byzantium; also called 001v/S, eIvcs, it was the well-known head-city of the Thinite name; cf. «Thinis » in PAULY-WISSOwA). He believes that, for the Chinese city as well as for the Egyptian one, the name (read « the form of the name »?) is due to popular etymology. Quite independently, my colleague BENVENISTE has suggested to me that the declension gen. Ocv65, acc. Olva, almost too regular for a foreign name which had
IIIF Color HighRes
