586 195. ÇAITON
had been partly rebuilt, they were called « Carp Walls » ~ ~_ J Li-yü-ch'êng). I may add that in the 10th cent. the city had been surrounded with three concentric walls, all of which were already in a state of complete dilapidation in 976. The poem quoted under the catch-word Tz'û-t'ung-ch'êng in the Yü-ti chi-shêng of 1221 is a lament on the ruin of the walls. As for the modern walls, i.e. those in « carp shape » (cf. ECKE and DEMIÉVILLE, loc. cit. 24), they are the ones which were enlarged in 1352 by a high official « Uighur » of the great Hsi family (Ta-Ch'ing i-t'ung chih, 328, 5 a; he is the « marquess Hsi » of TP, 1912, 451). The result of this inquiry is that the city walls of Ch'üan-chou have had several different names. In Chinese, the same word ch'êng designates the walls as well as the city enclosed within them. Such names do not have of course anything to do with the official appellations of administrative centres, but they may very well have been in popular use. One may perhaps feel somewhat disturbed when noticing that the works of the Sung period speak more often of T'ung-ch'êng than of Tz'ü-t'ungch'êng. Moreover, even starting from Tz'û-t'ung-ch'êng, it is somewhat surprising that the word ch'êng should have been dropped by foreigners; we should feel safer if we could imagine that, popularly, the Chinese used the term Tz'û-t'ung alone. But these difficulties are not in themselves of enough weight to make us reject the equivalence « Zâitûn » -- Tz'û-t'ung proposed by KLAPROTH, if it be phonetically sound (PHILLIPS'S attempt to explain Zâitûn as a transcription of « Guet-kong », the Chang-chou pronunciation of if f Yüeh-chiang [JA%CB, 21, 42; TP, 1890, 229-231] need not be discussed; moreover, cf. ARNAIZ. in TP, 1911, 690-691).
We now come to the second question : Is « Zâitûn » a normal transcription of Tz'û-t'ung? I leave out of account CORDIER'S opinion (Odoric de Pordenone, 271) according to which the Arabs had given to Ch'üan-chou the name of Zâitûn, which in Arabic means « olive » and « olive-tree », because there were at Ch'üan-chou t'ung-trees which the Arabs considered to be a sort of olive-tree. The tz'û-t'ung tree has nothing in common with any sort of olive, and Zâitûn is either a transcription of tz'û-t'ung or has no relation, semantic or other, to it. But is it a good transcription, and, if so, can it be dated?
FERRAND has already tackled the problem, and starting from tz'ü-t'ung, he declared «
tûn » to be an erroneous form for « Zitûn » which he adopted in his Relations de voyages (II, 427, 455), and which he regarded as absolutely correct from the point of view of comparative phonetics. I am not so certain that the change is a happy one. If the tradition about Liu Ts'ung-hsiao be true, the name of T'ung-ch'êng or Tz'ü-t'ung-ch'êng could not have existed for Ch'üan-chou before the 10th cent. ; moreover we have no mention of « Zâitûn » before the 13 th. But then all sources, Mussulman as well as European, know only Zâitûn, « Zaiton », never Zitûn. A text of Abû-l-Fidâ, as quoted by KLAPROTH (JA, y [July 1824], 43) and PAUTHIER (Pa, 528), says that the name of the place, according to travellers, is to be pronounced exactly like the Arabic word for « olive », zaitûn (this is not a quotation from the Geography translated by REINAUD and GUYARD). Andrea da Perugia says that the place is called « Zaiton » « in Persian» (wy, 374). It may or may not be that the vocalism of the name had been influenced by the Arabic word for « olive », but the fact remains that the name was pronounced Zâitûn or « Zaiton » at the time it was known, and that is therefore the form we must retain.
But even apart from the -a-, the transcription is not so satisfactory. Beginning with KLA-
IIIF Color HighRes
