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

the Kuṣaṇa. I have devoted to the question a special paper, La théorie des Quatre Fils du Ciel,
in TP, 1923, 97-125, and made some additions to it in TP, 1932, 174; in the meantime, FERRAND
had published fresh information from Mussulman sources in BSOS, vi [1931], 329-339. On the
four Sons of Heaven, my main addition of 1932 was Chih-mêng's biography in which it is said that
the pilgrim heard from an arhat in Cashmire explanations on the four Sons of Heaven, a statement
which proves that the Chinese believed in the Indian origin of this system. But the four Sons of
Heaven left out the Mussulmans, Arabs as well as Persians. A fresh development occurred in
Mussulman countries, according to which there were five kings, that of 'Irāq, the « king of kings »,
in the centre, then the « king of men » in China, the « king of wild beasts » among the Turks, the
« king of elephants » (or of « wisdom ») in India, and the « king of fine men » in Rūm (Byzantium).
The Arabian traveller Ibn Wahab (9th cent.) puts this list of the five kings in the mouth of the Chinese
Emperor, who of course can have said nothing of the kind. The five kings are of Mussulman origin,
and FERRAND cites collateral evidence on them. He has overlooked, as I had done myself, a passage
of Ya'qūt which it would be interesting to trace in earlier works : at Kirmānšāh was the famous
terrace where Chosroes Parwēz had assembled around him the fayfūr, or king of China, the ḥaqān
(= qayan) of the Turks, the daher (? = mahārāja), or king of India, and the Qaiṣār, or king of
Rūm (cf. BARBIER DE MEYNARD, Dict. géogr. de la Perse, 438). This is an exact counterpart to
Ibn Wahab's story. Although Mussulman texts speak only of kings, the connection of the list of
five kings with the earlier four Sons of Heaven is not doubtful.
I find it difficult to decide whether the Iranian title of fayfūr was or was not ever used in refe-
rence to sovereigns other than the Emperor of China. HORN (Grundriss der neupers. Etymologie,
71) quotes it from VON HAMMER's, Geschichte der schönen Redekünste Persiens, 94, as being referred
to the Byzantine Emperor; but this would require confirmation, in view of VON HAMMER's care-
lessness, and of a possible confusion with takfūr. Another case is more puzzling. In Armenian,
the title of the Chinese Emperor occurs as Čenbakur < Činfayfūr, « fayfūr of Čin » (cf. HÜBSCH-
MANN, Armen. Grammatik, I, 49), and the Orbelian, professing to be descended from Chinese princes,
adopted the clan-name of Čenbakurian. This form bakur is confirmed by a portion of a Syriac
catena attributed to Theodore of Mopsuestia, extracts from which have been recently translated
by BIDEZ and CUMONT, Les mages hellénisés, II, 117 [cf. also p. 114]; but the passage in which
the term occurs, to judge from its content, is only of the 7th cent. at the earliest. In a list of heathen
countries, the catena speaks of « the Chinese with Bagūr »; there is not any doubt that we have
here, in Syriac, the same type of transcription of baypūr which is represented by bakur in Armenian.
But this may entail wider consequences. LOKOTSCH, s. v. « fayfūr », gives as a fact that the name
occurring in Horace (Odes, 3, 6, 9) and Martial (9, 36) as « Pacorus » is no other than a form of
fayfūr. But there are many Πάκορος, Παχόρης, Pacorus, Bakūr, Afqūrśāh, etc., in Iranian history,
whose names are given in many classical authors besides Horace and Martial; JUSTI, Iranisches
Namenbuch, 238-239, gives 24 of them, and one may be the prototype of the Mañkura, in Chinese
transcription Wang-chün, of the Milindapañha (cf. JA, 1914, II, 405; TARN, The Greeks in Bactria
and in India, 422); but JUSTI carefully distinguishes this Iranian name from Armenian bakur
< baypūr. If LOKOTSCH were right, all of them would be Bakūr < Baypūr, and the form bakūr
of baypūr would be traceable to a date even prior to the Christian era. Although I know of no