National Institute of Informatics - Digital Silk Road Project
Digital Archive of Toyo Bunko Rare Books

432   183. COTTON

xx' [1887], 230) to a passage in which the Fukienese author of the Tung-hsi yang k'ao (Hsi-yinhsiian ts'ung-shu ed., 4, 7 b; cf. also 6, 18 a; 7, 12 b), who wrote in 1617-1618, speaks of the tou-

lo-mien of the north-western part of Sumatra « which is now called   if   to-lo-lien ». HIRTH
read the last character nien, a pronunciation for which I can find no authority in Mandarin; yet he was certainly right when he identified this to-lo-lien with the later to-to-ni. Let it be remarked by the way that this older form to-lo-lien does not support the connection made by the Tz'û yüan

between the ni of to-lo-ni and the ni used alone in Huang T'ing-chien's poem. As to   to-lo,

it has been customary, since WATTERS' day, to say that it was also written ÿ   to-lo and that

the latter to-lo was sometimes used as a designation of the cotton tree (cf. WATTERS, loc. cit. 439; GILES, Chin.-Engl. Dictionary, No. 11302; TARANZANO, Vocabulaire des Sciences, ti, 743). But in

Chinese mediaeval texts ÿ   to-lo (*td-/c1), as a botanical term (in other contexts it has been a
designation of a cosmetic box; cf. T'ai-p'ing yü-lan, 717, 5 b), always transcribes Skr. tala, the palmyra, Borassus flabelliformis (cf. Hobson-Jobson, s. v. « talee » and « talipot ». In Essays relating to Indo China, 2nd series, I, 193, GROENEVELDT, who adopts a dialectical pronunciation « to-lo » of tou-lo, refers the reader to D'HERVEY DE SAINT-DENYS, Ethnogr. des peuples étrangers, Méridionaux, 474, n. 33; but this is a palpable error since, in that note, D'HERVEY DE SAINT-DENYS speaks of to-lo, tala, not of tou-lo. Moreover, although TARANZANO gives both forms, in all the Chinese examples of to-lo-ni I know, the characters ;rj:', P to-lo are used (and these we find also in the earlier to-lo-lien of the Tung-hsi yang-k'ao), not ÿ 6 to-lo, and I have met the latter form only in the to-lo jung of the Chêng tzû t'ung, a woollen velvet. It looks as though WATTERS'S assignation of the secondary meaning « cotton » to ÿ{ to-lo were a mistaken inference drawn from the term to-to-ni. That the to-lo of to-lo jung and to-lo-lien or to-lo-ni is due to a confusion with tou-lo is probable enough (as was already suspected in 1868 by MAYERS, in Notes and Queries, II, 95; cf. also HIRTH, China and the Roman Orient, 249), but the confusion is of late popular origin (later than the phonetic evolution from *td-lei, to to-lo), and it has nothing to do with the old term to-lo which, as a botanical name, only represents tala, never tiild. My provisional conclusion is that to-lo-ni is an adaptation in Mandarin of a local term used in Amoy and Chang-chou. In the Amoy and Chang-chou dialect, this name is to-lô-ni" (cf. C. DOUGLAS, Chin.-Engl. Dict. of ... Amoy, 335-336), and it is a matter of very frequent occurrence in that dialect that an initial 1-, so pronounced in literary usage, becomes n- in popular speech. Thus to-Id-n£" exactly represents the to-lo-lien of the author of the Tung-hsi yang k'ao, himself a native of Chang-chou. The to-lo-ni of modern Mandarin, in its turn, may transcribe the popular Amoy pronunciation to-16-nî in which the final -n is practically mute. If so, lien would be the true earlier form of n£" -- ni, and its connection with the ni used by Huang T'ing-chien must probably be discarded. The only remaining difficulty would be to explain lien, since there are in the Amoy dialect interchanges between 1- and m-, but

not in the case of   mien, which is there pronounced biên in literary usage, and m£" in popular
speech. On the other hand, the other solution would remain possible, if we should suppose that, as is often the case in the Amoy dialect, the final -n of lien and ni" is itself of secondary appearance, and that the 1- of lien instead of n- is a mistaken purism on the part of the author of the Tung-hsi yang k'ao; but I hold this to be less probable. To-lo jung may, in its turn, have been created on the model of to-to-ni. In my opinion, to-lo-lien and to-to-ni, as well as the later terms formed