国立情報学研究所 - ディジタル・シルクロード・プロジェクト
『東洋文庫所蔵』貴重書デジタルアーカイブ

too far away, and moreover was known as Andargiri to the Portuguese; the form « Draguin » must
be quite modern.

Phillips has even tried to connect the names Dagroian and Hua-mien. As the Chinese map
of the early 15th cent. mentions west of Samudra Ta-Hsiao-Hua-mien, «the Great and Small Tattooed
Faces », Phillips isolated Ta-Hua-mien to read it, in the Amoy dialect, « Dakolien » = « Dragoian,
or Dagoyam ». This was his solution of 1895; ten years earlier, and also in the Amoy dialect, he
had read « Toa-sio-hoe [or Ko]-bin », and that also gave « Dagroian » (cf. Y, II, 297). Such fantastic
solutions must be eschewed.

A more likely explanation is reached if we start from Na-ku-êrh (*Nagur; Yule is mistaken
when thinking it is identical with Ting-k'o-êrh, which is not Sungora as said in Y¹, I, 82, but Treng-
ganu). An alternation n⌣d is not without examples (cf. Ferrand, in JA, 1919, II, 225), and
*Nagur might represent the same original as « Dagroian ». The difficulty lies in the final -ian.
Although we know double forms such as Soli and Soliyân (the Cholas), attributable to Persian
plurals, there are no cases of such plurals in Polo's text, and I hesitate in seeing one in « Dagroian ».
Having written this, I see from Bijdragen, Vol. 77, 8, that Rouffaer had already thought of a possible
equivalence « Dagroian » = Na-ku-êrh; and, before him, it occurred also to Gerini, in the midst of
several delirious etymologies (Researches, 675); the suggestion is also made in Fujita's commentary
of the Tao-i chih-lio, 70 b. To occur independently to different scholars, the equivalence must have
a certain degree of plausibility.

Na-ku-êrh represents *Nagur. Being the name of a Battak kingdom, it can hardly be separated,
as to the name at least, from the ancient Battak kingdom of Nagur. But this ancient kingdom of
Nagur was more to the south-east, in the region extending from the Toba See to the Panè River
(cf. Encycl. van Ned. Indië², IV, 4-5). On the other hand Rouffaer saw « Nagore » in the Na-ku-
êrh of the Chinese. This would be an ancient name of Pedir (or Pedje; but the name of Pedir is
not recent; even the Chinese knew it in 1537; cf. Groeneveldt, Notes², I, 246). Perhaps the
two are reconcilable. The Chinese map mentions together « the Great and Small Tattooed Faces »;
the « Great Tattooed Faces » were perhaps the Battak kingdom of Nagur of the Toba See-Panè
River, and the « Small Tattooed Faces » a Battak kingdom of Nagur (« Nagore ») in the region of
Pedir. I am not certain that the *Nagur can be put as far west as Pedir, since they must be east
of Lidè (cf. TP, 1915, 147), and Lidè itself is east of Pedir in Barros's list of Sumatran states.
I must add nevertheless that the mountain producing sulphur mentioned in connection with Na-
ku-êrh (TP, 1915, 148) may be the one referred to by Beaulieu towards Pedir.

It has been supposed by Schlegel and hypothetically accepted by Hirth and Rockhill
(HR, 66, and Rockhill in TP, 1915, 146) that the name of the Battaks of North-Western Sumatra
appeared perhaps in 1225 under the form 拔㝵 Pa-ta in Chao Ju-kua; the Na-ku-êrh are
expressly said to be the Pa-ta of Chao Ju-kua in the recent edition of the Ying-yai sheng-lan published
by Fêng Ch'êng-chün (p. 27; cf. TP, 1936, 215). Cœdès already observed in 1918 (BEFEO,
XVIII, VI, 11) that Chao Ju-kua's «Pa-ta» must have been on the Malay Peninsula, and Rouffaer
(in Bijdragen, XLVII, 93, 135), while accepting that the Pa-ta may be Battaks, says that they must
have then been continental Battaks. I think this solution must be abandoned. In Chao Ju-kua's
work, written at Ch'üan-chou (Zaitun), the transcriptions are still in agreement with the ancient