purpose. As he generally introduces them chronologically, at the time of their occurrence, he seems at times merely to increase the mass of indigested matter; but by and by we find what he has thus related to stand in the relation of cause to something subsequently chronicled. But his method with these additions to the text, which are yet connected with it, is very various. As Too Yu says, 'Now he anticipates the text to show the origin of an affair; now he comes after the text [with his narrative] to bring out fully the meaning; now he lies alongside the text to discriminate the principles in it; and now he appears to cross the text to bring together things that differ:-thus various according to what he considered the requirements of the case. '3 What is very surprising is that he does not appear to be conscious of frequent discrepancies between the details of his narratives and the things as stated by Confucius. Now and then, as on VI. xviii. 6, he says that the text conceals the nature of the fact; but generally he seems insensible of the untrustworthiness of the representation in it.

Let it be understood, however, that Tso does not give the details. of every event which the Classic briefly indicates. We must suppose that where he does not do so, his sources of information failed him, and he was obliged to leave the notice of the text as it was. There is the erroneous or defective entry in III. xxiv. 9,-"The duke of Kwoh.' On it Tso says nothing. So on the five paragraphs of Chwang's 26th year he has nothing to say, while he introduces brief narratives of two other things, for the latter of which only we can account as being given with an outlook into the future. Generally speaking, the information given in the Chuen is scanty or abundant in proportion to its distance from or nearness to the era assigned to its compilation. The 18 years of duke Hwan, B.C. 710-693, occupy in the following Work 37 pages; the 15 years of duke Ting, B.C. 508494, 50 pages. The 32 years of Chwang, B.C. 692-661, occupy 59 pages; the 32 of Ch'aou, B.C. 540-509, 173 pages. This certainly gives us for the Work one attribute of verisimilitude.4

3傳先經始事,或後經以終義或依經以辯理,或錯 經以合異隨義而發:

;-see Too's preface.

4 I take the opportunity to advert

here to a question which has produced no end of speculation and discussion among the scholars of China.—Why does the Ch'un Ts'ew begin with duke Yin? Might we not have expected the sage to go back to the first origin of the State of Loo? I believe that the only reasonable answer to these inquiries is this,-that the annals of the State previous to duke Yin's rule had been altogether lost, or were in such a miserable state of dilapidation and disarrangement that nothing could be made of them. We might have expected a sentence or two from the sage to enlighten us on the subject; but his oracle is dumb. Neither does the Chuen say anything about it. How different the practice of writers of history in the West!

The second view of Tso;-to give a general) view of the history of China during the Chun Ts'ëw period.

But while Tso intended his Work to be a commentary on the text of the Ch'un Ts'ëw, I believe that he had in view another and higher object, and wished to give his readers a general view of the history of the country throughout all its States during the Ch'un Ts'ëw period. The account of the Chuen quoted above from Too Yu carries us a considerable way to this conclusion. Tso shows the origin and issue of many events, one phase of which merely is mentioned in the text. The unconnected entries of the classic are thus woven together, and a history is made out of them. But the new matter introduced by him is so very much, and often having no relation to anything stated in the text, yet calculated to bring the whole field of the era before us, and to indicate the progress of events on towards a different state of the kingdom, that we must suppose this to have been a prominent object in the author's mind. This characteristic of the Work has not escaped the notice of native scholars themselves. As early as the Tsin dynasty, Wang Tsëeh preferred to it the commentary of Kungyang on this account. 'Tso's style,' said he,' is so rich, and his aim so extensive, that he is to be regarded as an author by himself, and not having it for his principal object to illustrate the classic.'s Nearly to the same effect is the account of Tso's Chuen given by Wang Cheh of the Sung dynasty. After praising Tso as a skilful reader of the old histories and collector of various narratives, so that he accumulated a very complete account of the events in the Ch'un Ts'ëw, he yet adds:-' But though his book was made as an appendix to the classic, yet, apart from and outside that, it forms a book by itself, the author of which was led away by his fondness for strange stories, and carried his collecting them beyond what was proper. He was remiss in setting forth the fine and minute ideas of the sage, but yet his Work has a beginning and end, being all the compilation of one hand.' Chinese scholars write of Tso under the influence of their admiration and veneration for the sage. I could wish that he had written altogether independently of the Classic, in which case we might have had a history of those times as complete as a man

5王接日左氏辭義贍富自是一家書不主為經發;-sce

the, Bk. 169, p. 3. In Bk. 174, p.3, there is quoted from him his contrary view of Kung

yang:-公羊附經立傳經所不書傳不妄起於文為儉通 經為長

knowing only the heroes and events of his own country could make. It is not too much to call Tso the Froissart of China. The historical novel called 'The History of the various States' shows the use which can be made of his narratives. They lie necessarily in my pages so many disjecta membra, but some one may yet give, mainly from them, an account of the closing centuries of the feudal state of China that shall be found to have an universal interest.

6. Three more points in regard to Tso's Work have yet to be considered: the manner of his composition; how far his narratives are entitled to our belief; and whether there is reason to believe that additions were made to them by writers of the Ts'in and Han dynasties. By the manner of Tso's composition I do not mean the general character of his style. There is but one opinion as to that. It is acknowledged on all hands that he was a master of his l'eculiarity of Tso's composition. art. Condensed, yet vivid, he is eminently pictorial. The foreign student does not for some time find it easy to make out his meaning, but by and by he gets familiar with the style, and it then has a great charm for him. In the words which the foremost of French sinologues once used to me of him, Tsó was un grand ecrivain.1 But the peculiarity which I have in view is the way in which Tso constantly varies the appellations of the actors in his narratives. Very often they are named by their sacrificial or honorary epithets which were not given to them till after their death, so that it is plain he did not copy out the contemporaneous accounts or records which we suppose him to have had before him, and some critics have from this contended that the narratives were entirely constructed by himself, not drawn from historical sources. 2 But such a conclusion is more than the premiss will justify. Tso might very well call his subjects of a former time by the titles which had been accorded to them after their death, and by which

王哲日左氏善覽舊史兼該說得春秋之事亦甚 備其書附經而作然於經外自成一書故有貪惑異 說采极過當至於聖人微信頗亦疎略,而大抵有本末 蓋出於一手之所撰述

1 I select only two Chinese testimonies of the excellence of Tso's style. The first is from

Seun Sung (荀崧) of the Tsin dynasty:一其書善禮多膏腴美辭張本 繼末以發明經意信多奇偉學者好之 The other is from Choo E-tsun of the present dynasty:一匪獨詳事也,文之簡要不可及

E. g., Lew Hwang (劉貺) of the T'ang dynasty says:-左氏年序諸侯列 會,具舉其諡知是後人追修非當世正史也

2

men generally would in his days speak of them. What is really perplexing is that in the same account the same individual is now called by his name, now by his honorary epithet, and now by his designation, or by one or other of his designations if he had more than one, so that the narrative becomes very confused, and it requires considerable research on the part of the reader to make out who is denominated in all this variety of ways. To give only one example:-in the account of the battle of Peih, in the 12th year of duke Seuen, of the leaders on the side of Tsin, we have, 1st, Seun Lin-foo, who by and by is styled Hwan-tsze;3 2d, Sze Hwny, who is variously denominated Woo-tsze of Suy, Suy Ke, and Sze Ke, while elsewhere he is called Woo-tsze of Fan; 3d, Seen Hwoh, also called Che-tsze, and elsewhere Yuen Hwoh, or Hwoh of Yuen; 4th, Seun Show, called also Che Chwang-tsze and Che Ke;6 5th, Han Keuch, by and by Han Heen-tsze; 6th, Lwan Shoo, by and by Lwan Woo-tsze;8 7th, Chaou Soh, by and by Chaou Chwang-tsze;9 and 8th, Keih K'ih, by and by Keih Heen-tsze.10 Similar instances might be quoted in great number. Chaou Yih says that such a method of varying names and appellations was characteristic of the style of that time. If, indeed, it was characteristic of the time, I must think that Tso possessed it in an exaggerated degree. The confusion produced by it in his Work seems to have led to its cure. Sze-ma Ts'ëen and the writers of the Books of Han are careful, at the commencement of their biographies, to give the surname, name, and designation or designations of their subjects, so that the student has none of the perplexity in reading them, which he finds with Tso's Chuen.

The other two points regarding the Work, which I indicated are of more importance, and I will consider them together. Have we Are Tso's narratives reliable? Were reason to receive Tso's narratives as they supplemented or added to. reliable, having been transcribed by him from pre-existent records with merely such modifications of style as suited his taste? Or did he invent some of them himself?

Or were they added to by writers in the Ts'in dynasty and that of

8 荀林父;桓子 *士會;隨武子;隨季士季范武子 5 先穀子;原穀 6荀首;知莊子;知季 7 韓厥: 韓獻子 8 欒書;欒武子。趙朔;趙莊子 10郤克:

郤獻子 11篇中或用名或用字或用謚號蓋當時文 法如此 ;—see Chaou on the Ch'un Tsew, Ch. Z

the Former Han? It is difficult to reply to these questions categorically. What has the greatest weight with me in favour of Tso's general credibility is the difference between his commentary and those of Kung-yang and Kuh-lëang. What of narrative belongs to the latter bears upon it the stamp of tradition, and evidently was not copied from written records but from accounts current in the mouths of men. It is, moreover, of comparatively small compass. Their Works must have been written when the memory of particular events in the past had in a great measure died out. If Tso's sources of information had been available for them, they would, we may be sure, have made use of them. The internal evidence of the three Works leaves no doubt in the mind as to the priority of Tso's. And as they all made their appearance early in the Han dynasty, we are carried back for the composition of Tso's into the period of Chow. As his last entry is about an affair in the 4th year of duke Taou, who died B.C. 430, and he mentions in it the Head of the Chaou family in Tsin by his honorary epithet of Sëangtsze, which could not have been given before 424, we can hardly be wrong in assigning Tso to the fifth century before Christ. This brings him close to the age of Confucius who died in B.C. 478. Tso may then have been a young man ;-he could hardly be a disciple enjoying that intimate association with the sage which Lew Hin, Pan Koo, and other Chinese scholars were fond of asserting.

But to maintain the general credibility of Tso's Chuen as having been taken from authoritative sources and records acknowledged as genuine among the States of China when he wrote, leaves us at freedom to weigh his narratives and form our own opinion on grounds of reason as to the degree of confidence which we ought to repose in them. There are few critics of eminence among the Chinese who do not allow themselves a certain amount of liberty in this respect. Ch'ing E-ch'uen laid down two canons on the subject. "The Chuen of Tso,' he says, 'is not to be entirely believed; but only that portion of it which is in itself credible.'12 To this no objection can be taken; but he opens a very difficult question, when he goes on, 'We should from the Chuen examine the details of the events referred to in the text, and by means of the text discriminate between what

12 程子曰左傳不可全信信其所可信者爾以傳考經 之事迹以經別傳之眞偽;-see the 經義考, Bk. 169, p.5.