¹ From Max Weber: Essays in Sociology, trans. Gerth, H. H. and Mills, C. Wright (New York, 1946), p. 209Google Scholar (referred to hereafter as: Weber, Sociology). Davis, Arthur K., “Bureaucratic Patterns in the Navy Officer Corps,” Social Forces XXVII (1948), 145.Google Scholar

² Gibbon wrote that the Roman Empire “was supposed to contain above sixteen hundred thousand square miles”; see Gibbon, Edward, The History of the Decline and Fall of the Roman Empire, ed. Bury, J. B., (New York: Heritage Press, 1946), I. 21Google Scholar. The area known as “China Proper” comprises some fifteen hundred thousand square miles, and the Chinese Empire was far more extensive than this as early as 100 B.C. In population, however, the Roman Empire appears to have greatly exceeded the contemporary Chinese Empire.

³ Any careful comparison of the governmental systems of the Chinese and Roman empires, in the centuries just before and after the beginning of the Christian Era, makes it evident that in terms of centralized administration China was far more developed. No doubt the Roman territories were under more firmly centralized control, guaranteed by the formidable and highly mobile Roman armies. But most of the business of local administration was left by the Romans (no doubt wisely in terms of their objectives) to the cities that dotted the Empire. See Rostovtzeff, M., The Social and Economic History of the Roman Empire, 2nd ed. (Oxford, 1957)Google Scholar (referred to hereafter as: Rostovtzeff, Roman Empire), I. 187, 192–194, 252; Belot, Émile, Histoire des Chevaliers Romains (Paris, 1866, 1873), II. 77Google Scholar; Boak, Arthur E. R., A History of Rome to 565 A.D. (New York, 1923)Google Scholar (referred to hereafter as: Boak, Rome), pp. 198–199. Max Weber has commented on “the lack of a bureaucratic apparatus” under the Roman Republic; see Weber, , Sociology, p. 211Google Scholar. Even in the Roman Empire, no bureaucratic structure comparable in scope and complexity to that of contemporary China was developed. Certainly the Byzantine Empire had a large and complicated bureaucracy. But that empire was, in the first place, only a fraction of the size of the Chinese. And Byzantium inherited from Rome the emphasis that gave military officiais precedence over civilians. This fact, together with the necessities of defence, gave rise to the organization into districts called themes, each controlled by a military governor who administered it somewhat in the manner of a “vice-emperor”; see Cambridge Medieval History (referred to hereafter as: Camb. Med. Hist.) IV (New York and Cambridge, 1927) 731–734Google Scholar; Byzantium, edited by Baynes, Norman H. and Moss, H. St. L. B. (Oxford, 1949; reprint of 1953), p. 290Google Scholar. While these military governors were usually held in check by the emperor, this organization did not resemble that of the modern centralized bureaucratic state to the same degree as did that of China, which was normally controlled by civilian bureaucrats. Concerning other factors that tended to operate against complete centralized control of the Byzantine Empire, see Kantorowicz, Ernst H., “‘Feudalism’ in the Byzantine Empire,” in Feudalism in History, ed. Coulborn, Rushton (Princeton, 1956)Google Scholar (referred to hereafter as: Coulborn, Feudalism), pp. 151–166.

⁴ See Hsien-ch'ien, Wang, Han-shu Pu-chü (1900)Google Scholar (referred to hereafter as: Han-shu) 19A.31, and Yü-ch'üan, Wang, “An Outline of the Central Government of the Former Han Dynasty,” in HJAS XII (1949)Google Scholar (referred to hereafter as: Wang, “Han Government”) 136–137.

⁵ See Creel, H. G., “The Fa-chia: ‘Legalists’ or ‘Administrators’?” in Studies Presented to Tung Tsopin on His Sixty-fifth Birthday (Taipei, 1961)Google Scholar (referred to hereafter as: Creel, “Fa-chia”), pp. 631–634.

⁶ It has commonly been supposed that examination did not become important in connection with the bureaucracy until the seventh century A.D. It is true that the pattern of examination that continued without essential change to the twentieth century was established only at that time, but a variety of examinations played an increasingly important role in connection with the Han bureaucracy. The examinations for men recommended as “Filial and Incorrupt” are commonly said not to have been “open” because the candidates had to be recommended, and it is assumed that the officials would have recommended only their relatives and members of powerful families. But a crucial fact is often overlooked: those recommended had to pass the examinations. If a candidate proved conspicuously incapable, both the candidate and the recommender might be punished very severely; see Han-shu 17.4a, 19B.17a, 88.5a. The result was that, since officials were compelled to make recommendations, they had to find men of ability regardless of birth or wealth, and there is concrete evidence that numerous men of humble origin in fact did enter the bureaucracy. For an excellent discussion of this situation see Ch'ien Mu, Chung-kuo Li-tai Cheng-chih Te-shih^a, 2nd ed. (Hongkong, 1954)Google Scholar (referred to hereafter as: Ch'ien, Te-shih), pp. 10–14.

⁷ Han-shu 88.4a–6a. Ch'ien, Te-shih, pp. 10–14.

⁸ The impressive bureaucracy of ancient Egypt, concerning which voluminous records have survived, does not seem to have included any apparatus for the objective weighing of merit. See Erman, Adolf, Aegypten und Aegyptisches Leben im Altertum, revised by Hermann Ranke (Tübingen, 1923), especially pp. 55–145Google Scholar. Wilson, John A., The Culture of Ancient Egypt (Chicago, Phoenix Books edition, 1957), p. 49Google Scholar, writes that the pharaoh “as a god, was the state,” and that the evidence indicates that the officials were “appointed by him, responsible to him alone, and holding office subject to his divine pleasure.” See also “The Question of Feudal Institutions in Ancient Egypt,” by William F. Edgerton, in Coulborn, Feudalism, pp. 120–132.

⁹ In the course of a fairly extensive study of Roman administration, I have been able to find no evidence of the systematic use of objective criteria for the control of official careers, either under the Roman Republic or the Empire. It appears that the emperor made all appointments, promotions, and dismissals at the imperial pleasure. See Boak, , Rome, p. 268Google Scholar; The Development of the Civil Service, by various authors (London, 1922), p. 16Google Scholar; Rostovtzeff, , Roman Empire, I. 47, 185–186.Google Scholar

¹⁰ To what extent anything resembling what we know as “civil service examination” may ever have been practiced in the Byzantine Empire is a difficult problem, which must be reserved for discussion in my as yet uncompleted history of the institution of examination. In any case there does not seem to have been much use of objective criteria in the actual management of the bureaucracy; Charles Diehl writes (Camb. Med. Hist. IV. 726–727Google Scholar) that the Byzantine Emperor “kept a close supervision over administrative affairs, appointing and dismissing officials at his pleasure, and advancing them in the complicated hierarchy of dignities according to his caprice.” And the manual on statecraft that has come down to us from the hand of a Byzantine emperor shows very little concern with impersonal or objective criteria; see Constantine Porphyrogenitus, De Administrando Imperio, Greek text edited by Gy. Moravcsik, English translation by R. J. H. Jenkins (Budapest, 1949).Google Scholar

¹¹ See Ch'ien, , Te-shih, pp. 10–14Google Scholar. Wang, , “Han Government,”Google Scholar passim.

¹² See Wang, , “Han Government,” pp. 145–146Google Scholar and passim. Ch'ien, , Te-shih, pp. 2–10.Google Scholar

¹³ Han-shu 24A.3b, 9b–10a, 20a; 24B.8a; 84.9a. Wang, , “Han Government,” pp. 145–146Google Scholar. In contrast, Rostovtzeff (Roman Empire, 1. 515Google Scholar) says: “The Roman state never had a regular budget, and when it was faced with financial difficulties, it had no fixed and stable reserve to draw upon … the usual way of getting money, according to the principles of the city-state, was to demand it from the population either by means of extraordinary taxation or by means of requisitions and confiscations.” These latter devices were not unknown in Han China, but they were unusual abuses rather than normal policy. Concerning the Roman lack of interest in the principles of public finance, see Stevenson, G. H., Roman Provincial Administration (Oxford, 1939; 2nd impression, 1949), p. 133.Google Scholar

¹⁴ Han-shu 24A, 24BGoogle Scholar. Swann, Nancy Lee, Food and Money in Ancient China (Princeton, 1950).Google Scholar

¹⁵ Weber, Max, Gesammelte Aufsätze zur Religionssoziologie, I (Tübingen, 1920), 278Google Scholar. Weber, Max, The Religion of China, trans. Gerth, Hans H. (Glencoe, Illinois, 1951), p. 252.Google Scholar

¹⁶ In connection with Weber's interpretations of passages in the Analects, it should be noted that he explicitly (The Religion of China, p. 250Google Scholar) cites the translation of James Legge in the Chinese Classics as his source. He writes (ibid., p. 143): “On the one hand, when viewed from the perspective of Confucian reasons of stste, religion had to be ‘upheld for die people.’ The order of the world, according to a word of the Master, could not be maintained without belief. Therefore, the retention of religious belief was politically even more important than was the concern for food.” This is obviously a reference to Analects 12.7. But the text makes it perfectly clear that the “belief” in question is not religious belief, but confidence in the government or the ruler; this is the interpretation of Legge and, I believe, of all other translators and commentators.

Again Weber writes (ibid., p. 163): “‘Where we are three I find my master,’ Confucius has allegedly said, which meant, I bow to the majority.” This must refer to Analects 7.21, which Legge translates, “The Master said, ‘When I walk along with two others, they may serve me as my teachers. I will select their good qualities and follow them, and their bad qualities and avoid them.’” Other scholars vary slightly in their interpretations, but none brings in Weber's idea of “bowing to the majority,” which is inconceivable on the basis of the text.

These and other distortions of the sense of Chinese materials by Weber cannot be put down to simple ignorance. The fact is that the Analects, either in the Chinese original or in Legge's or another translation, does not yield a picture of Confucius that will square with Weber's conception of him. Weber forced it to do so.

¹⁷ Parsons, Talcott, Structure and Process in Modern Societies (Glencoe, Illinois, 1960), p. 20Google Scholar, points out that “a minimal description of an organization will have to include an outline of the system of values which defines its functions …” It is precisely the Chinese system of values of which Weber seems to have had very little understanding or appreciation. Thus (The Religion of China, p. 232Google Scholar) he tells us that in the Chinese character there is “the absence of an inward core, of a unified way of life flowing from some central and autonomous value position.” Again Weber tells us (ibid., p. 235) that the Confucian “way of life could not allow a man an inward aspiration toward a ‘unified personality.’” But surely few ways of life, anywhere, have been so characterized as the Confucian by “an inward striving toward a ‘unified personality.’” One wonders where Weber could have acquired such ideas. He prefaces these remarks by stating (ibid., p. 231) that for information concerning the Chinese character “the sociologist essentially depends upon the literature of the missionaries.” It would be a little hard on “the missionaries,” however, to give them all the credit. It was pointed out in the preceding note that in interpreting some passages in the Analects Weber seems to have paid singularly little attention to the judgments of James Legge, a very great missionary scholar. Weber appears to have been unfortunate in selecting which missionaries to believe.

The defects of Weber's treatment of Chinese bureaucracy are those that we should expect when a master of brilliant generalization, possessing inadequate and inaccurate information, attempts to reduce to tidy formulae a range of phenomena that resists comprehension even by the most informed specialists. An example of his uncontrolled generalization is the manner in which he uses the term “literati.” He tells us (The Religion of China, p. 42Google Scholar) that the literati were “ritualist advisers' of the princess” and (ibid., p. 108) that “the ‘literati’ of the feudal period … were first of all proficient in ritualism.” He also informs us (ibid., p. 41) that Shang Yang was “a representative of the literati,” whereas Shang Yang was very far indeed from being a “ritualist.” Again he says (ibid., p. 111): “If one may trust the Annals, the literati, being adherents of the bureaucratic organization of the state as a compulsory institution, were opponents of feudalism from the very beginning.” This is confused and confusing. No matter how we define “literati,” whether we limit it to “Confucians” or make it more inclusive, the men it must denote include individuals and groups whose attitudes on these matters were various and conflicting. This variety and conflict are the stuff of which institutional history is made. Such use of nullifying generalizations reduces history to a meaningless tableau.

¹⁸ See Gouldner, Alvin W., “Discussion of Industrial Sociology,” in American Sociological Review XIII (1948), 397–398.Google Scholar

¹⁹ Weber, , Sociology, p. 196Google Scholar, writes: “Modern officialdom functions in the following specific manner: I. There is the principal of fixed and official jurisdictional areas, which are generally ordered by rules, that is, by laws or administrative regulations, 1. The regular activities required for the purposes of the bureaucratically governed structure are distributed in a fixed way as official duties. 2. The authority to give the commands required for the discharge of these duties is distributed in a stable way and is strictly delimited by rules concerning the coercive means, physical, sacerdotal, or otherwise, which may be placed at the disposal of officials. 3. Methodical provision is made for the regular and continuous fulfilment of these duties and for the execution of the corresponding rights; only persons who have the generally regulated qualifications to serve are employed.

“In public and lawful government these three elements constitute ‘bureaucratic authority.’… Bureaucracy, thus understood, is fully developed in political and ecclesiastical communities only in the modern state.…”

Weber can hardly be blamed for not realizing that these characteristics were fulfilled as early as Han times; he was no Sinologist, and even Sinologists in his day hardly appreciated the degree to which bureaucratic techniques and regulations were elaborated in early China. Today, when information about them is gradually becoming more available, the facts are sometimes difficult to credit. Bünger, Karl, “Die Rechtsidee in der chinesischen Geschichte,” in Saeculum III (1952), 211Google Scholar, writes that whereas in the Roman Empire administrative law was little developed, China had already at an early date a body of legal regulations for the administration of the state of “astonishing completeness.” See also Hulsewé, A. F. P., Remnants of Han Law I (Leiden, 1955), 5–9Google Scholar. For other criteria of bureaucracy that have been called exclusively “modern,” but can be found in China at an early date, see note 51 below.

Weber characterized the Chinese bureaucracy, as distinguished from that of the modern Occident, as “patrimonial” and “prebendal” (Weber, , Sociology, pp. 207, 243Google Scholar). He explains “patrimonial” recruitment of the administrative staff (Weber, Max, The Theory of Social and Economic Organization, trans. Henderson, A. M. and Parsons, Talcott [New York, 1947]Google Scholar [referred to hereafter as: Weber, Organization], p. 342) as recruitment: “From persons who are already related to the chief by ties of personal loyalty.… Such persons may be kinsmen, slaves, dependents who are officers of me household, clients, coloni, or freedmen.” But long before Han times bureaucratic recruitment in China had come to embrace strangers and even “foreigners,” and from the beginning of Han on the bureaucracy was recruited from very diverse sources.

Weber's characterization of the Chinese bureaucracy as “prebendal” is even more curious. He writes (Sociology, p. 207Google Scholar): “We wish to speak of ‘prebends’ and of a ‘prebendal’ organization of office, whereever the lord assigns to the official rent payments for life, payments which are somehow fixed to objects or which are essentially economic usufruct from lands or other sources.” But the Chinese bureaucracy, from at least as early as Hare times, has for the most part been a regularly salaried bureaucracy, paid from the treasury of the state. Indeed, Weber gives another very complicated definition of “prebends” (Organization, p. 351Google Scholar) which would include as such the payment of salary in money to an official by the ruler. Weber even says (Sociology, p. 108Google Scholar) that in the United States of America in the nineteenth century “as the price of victory, the true booty object of the office-prebend was held out precisely at the presidential election.” In Weber's lexicon “prebend” appears to have been a term of uncommon flexibility.

It appears impossible to agree that the Chinese bureaucracy was either “patrimonial” or “prebendal” in any meaningful sense.

²⁰ Creel, “The Fa-Ma.” See also Creel, H. G., “The Meaning of Hsing-ming,” in Studia Serica Bernhard Karlgren Dedicata (Copenhagen, 1959), pp. 199–211Google Scholar. Shen Pu-hai was chancellor of the state of Han^c, which of course should not be confused with the Han^d dynasty.

²¹ In my opinion most but not all of the quoted materials attributed to this book actually did stem from Shen Pu-hai, though not necessarily in the form in which we have them. I rather doubt that he wrote the book; it was probably put together from sayings attributed to him, somewhat in the manner of the production of the Confucian Analects.

²² Liu Pei, the founder of the Shu-Han dynasty, who reigned 221–223 A.D., is quoted as saying in instructions to his son that “the books of Shen [Pu-hai] and Han [Fci-tzu] increase one's sagacity; they should be read and recited.” See Shen, Yang, Tan-ch'ien Tsung-lu^eGoogle Scholar (1588), 11.7b. His chancellor, the famous Chu-Ke Liang, is reported to have written out with his own hand several works, including that of Shen Pu-hai, for the use of his ruler's heir; see Er-shih-ssu Shih (Shanghai, , T'ung-wen Shu-chü ed.; 1884Google Scholar), Shu-chih 2.20a. The Sui emperor Wen (reigned 590–604 A.D.) is said to have withdrawn his favor from Confucians and given it to “the group advocating hsing-ming^t and authoritarian government” (Er-shih-ssu Shih, Sui-shu 75.2b); hsing-ming is the method of control of the bureaucracy associated with the name of Shen Pu-hai. Discussions of Shen's doctrines by scholars of Sung and even Ming dynasty date still seem to indicate some familiarity with their actual nature; see Yang, Tan-ch'ien Tsung-lu, 11.7a–8a.

²³ On the role of Chang Chü-cheng see Hucker, Charles O., “The Tung-lin Movement of the Late Ming Period,” in Chinese Thought and Institutions, ed. Fairbank, John K. (Chicago, 1957), pp. 133–134, 139–140.Google Scholar

²⁴ I am indebted to Professor Robert B. Crawford for bringing to my attention this information concerning Chang Chü-cheng. The doctrines of Shen Pu-hai are twice referred to in the biography of Chang Chu-cheng in Fu Wei-lin, Ming-shu, in Chi-fu Ts'ung-shu^g (1879)Google Scholar, 150.13b 38b. Fu (who died in 1667) wrote: “[Chang] Chü-cheng was by nature severe and illiberal. He was fond of the methods of Shen [Pu-hai] and Han [Fei-tzu], and used clever techniques to control his subordinates.” In his appraisal at the end of the biography Fu said, “[Chang] Chü-cheng took as his model the surviving practices of Shen [Pu-hai] and Shang [Yang].” It is interesting that the name of Shen Pu-hai appears in both of these statements. It is rather usual to attribute Fa-chia ideas to every strict official, but the specific reference to techniques of administrative control may indicate that even as late as the seventeenth century there was still some genuine knowledge of the nature of the doctrines of Shen Pu-hai.

²⁵ Creel, , “The Fa-chia” pp. 610–611.Google Scholar

²⁶ In ibid., p. 612, note 22, I have listed the works in which these quotations occur. It should be noted, however, that they are very cryptic and must be collated and compared with other materials before they are illuminating. I plan to publish my collated texts, with notes, translation, and introduction, in the near future.

²⁷ I place “rational” in quotation marks because I am dubious of the scientific usefulness of this term. It is not easy certainly to distinguish “rational” from “non-rational” conduct within one's own cultural system; to apply the concept unerringly so as to evaluate actions performed in other cultural contexts is difficult indeed.

²⁸ Shang Yang died in 338 B.C., one year before Shen Pu-hai. He was chancellor of Ch'in^h, which bordered Shen's state of Han^c on the west. Although the Shang-chün Shu was not written by him, we can get at least an approximate conception of his ideas from it and other works. Han-fei-tzu, who died in 233 B.C., was a scion of the ruling family of Shen's own state of Han. The work called the Han-fei-tzu is very composite, but undoubtedly contains some chapters written by him. Han-fei-tzu tried to combine the doctrines of Shang Yang and Shen Pu-hai, but in the process missed much of what was most important in the philosophy of the latter. It is interesting that when, as very commonly, the name of Shen Pu-hai is mentioned along with that of Shang Yang or of Han-fei-tzu, Shen's seems always to be mentioned first, although the other two figure far more prominently in Chinese literature. Concerning the roles of these three men, see Creel, “The Fa-chia.”

²⁹ Ch'ün-shu Chih-yao¹, ed. Wei Cheng, (Ssu-pu Ts'ung-k'an ed.), 36.266b–27a. Creel, , “The Fachia,” p. 613.Google Scholar

³⁰ Ch'ün-shu Chih-yao 36.26a–27a. Lü-shih Ch'un-ch'iu (Ssu-pu Pei-yao ed.) 17.7a. Frühling und Herbst des Lü Pu We, trans. Wilhelm, Richard (Jens, 1928), p. 270.Google Scholar

³¹ Ch'ün-shu Chih-yao 36.25b. T'ai-p'ing Yü-lan (edition of 1879)Google Scholar 638.4b. Han-fei-tzu (Ssu-pu Pei-yao ed.) 16.6b.

³² Ch'ün-shu Chih-yao 36.27a. T'ai-p'ing Yü-lan 638.4b. Han-fei-tzu 11.11b–12a. Creel, , “The Fa-chia,” pp. 613–620.Google Scholar

³³ “The Vizier of Egypt,” translation by Wilson, John A., in Ancient Near Eastern Texts Relating to the Old Testament, ed. Pritchard, James B. (Princeton, 1950), pp. 212–214Google Scholar. “Egyptian Instructions,” trans. John A. Wilson, in ibid., pp. 412–424.

³⁴ “Advice to a Prince,” in Babylonian Wisdom Literature, by Lambert, W. G. (Oxford, 1960), pp. 110–115.Google Scholar

³⁵ “The Words of Ahiqar,” trans. H. L. Ginsberg, in Ancient Near Eastern Texts Relating to the Old Testament, pp. 427–430.Google Scholar

³⁶ Kautilya's Arthaśāstra, trans. Shamasastry, R., 5th ed. (Mysore, 1956)Google Scholar; Das altindische Buch vom Welt- und Staatsleben, das Arthçāstra des Kautilya, trans. Meyer, Johann Jakob (Leipzig, 1926).Google Scholar

³⁷ Plato, , “The Republic”Google Scholar and “Laws,” in The Dialogues of Plato, trans. Jowett, B., House, Random ed. (New York, 1937), I. 591–879, II. 407–703Google Scholar. Aristotle, “Politics,” trans. Jowett, Benjamin, in The Basic Works of Aristotle, ed. McKeon, Richard (New York, 1941), pp. 1127–1316.Google Scholar

³⁸ Cicero, , De Re Publica, De Legibus, with an English translation by Clinton Walker Keyes (Cambridge and London, 1928; reprinted 1943).Google Scholar

³⁹ Constantine Porphyrogenitus, De Administrando Imperio.

⁴⁰ al-Mulk, Nizām, The Book, of Government or Rules for Kings, trans. Darke, Hubert (London, 1960).Google Scholar

⁴¹ Machiavelli, Niccolò, The Prince and The Discourses, Modern Library edition (New York, 1940).Google Scholar

⁴² A psrt of the rc3son for this is undoubtedly the fact that the area envisaged for direct rule was in general smaller than that contemplated by Shen Pu-hsi. This may seem odd, since the Roman dominion of Cicero's day was far larger than Shen's state of Han. But Rome, like most of the Occident until a late date, rermined wedded to the city-state conception. Chinese thinkers, however, even when China was split into small units, were aiming at a government that would unite and administer all of China—and for them “China” meant all of the “civilized” world.

⁴⁸ The works I have read for purposes of this comparison include, among others: From Max Weber: Essays in Sociology, trans. Gerth and Mills. Max Weber, The Theory of Social and Economic Organization, trans. Talcott Parsons. Weber, Max, The Religion of China, trans, by Gerth, Hans H.. Barnard, Chester I., The Functions of the Executive (Cambridge, Mass.; 1951)Google Scholar. Simon, Herbert A., Administrative Behavior, and ed. (New York, 1960)Google Scholar. Parsons, Talcott, The Structure of Social Action (New York and London, 1937)Google Scholar. Parsons, Talcott, Structure and Process in Modern SocietiesGoogle Scholar. Blau, Peter M., The Dynamics of Bureaucracy (Chicago, 1955)Google Scholar. Blau, Peter M., Bureaucracy in Modern Society (New York, 1956; 7th printing, 1961)Google Scholar. Reader in Bureaucracy, ed. Merton, Robert K., Gray, Ailsa P., Hockey, Barbara, and Selvin, Hanan G. (Glencoe, Illinois, 1952Google Scholar; second printing, 1960). Essays on the Scientific Study of Politics, ed. Storing, Herbert J. (New York, 1962)Google Scholar. I have also consulted, but not read in toto, Weber, Max, Gesammelte Aufsätze zur Religionssoziologie (Tübingen, 1920–1921).Google Scholar

⁴⁴ These similarities will be described in some detail in my study of Shen Pu-hai. A single example may be cited here. The definition of “authority” given by Chester I. Barnard has been die subject of some discussion in social science literature; see Essays on the Scientific Study of Politics, ed. Storing, pp. 132–142Google Scholar. Barnard himself wrote that it was “so contrary to tie view widely held by informed persons of many ranks and professions, and so contradictory to legalistic conceptions, and will seem to many so opposed to common experience,” that he found it necessary to justify it with some pages of discussion; see Barnard, , The Functions of the Executive, pp. 163–175Google Scholar. Yet it is quite clear that Barnard's theory of the nature of authority is entirely in consonance with that of Shen Pu-hai; see Pei-t'ang Shu-ch'ao, comp. Yü Shih-nan (ed. of 1888)Google Scholar, 45.9b.

⁴⁵ Bryson, Lyman, “Notes on a Theory of Advice,” in Political Science Quarterly LXVI (1951), 322Google Scholar, speaks of “the Chinese maxim of political administration which states that the good administrator does nothing; he does nothing, in order to give his executives the opportunity to do their best.” Shen Pu-hai repeatedly says that the skillful ruler “does nothing,” and various contexts show that the purpose is that his ministers shall not be mere “yes men” but shall rather develop their own ideas and exert themselves. See: Ch'ün-shu Chih-yao 36.26b–27a. Han-fei-tzu 13.6a. See also Creel, , “The Fa-chia,” pp. 613–614.Google Scholar

⁴⁶ In a classic paper, Professor Teng examined this whole question and showed conclusively the Chinese a influence, in particular, upon the introduction of civil service examinations in Britain in the nineteenth century; see Teng, Ssu-yü, “Chinese Influence on the Western Examination System,” in HJAS VII (1943), 267–312Google Scholar. My colleague Professor Donald F. Lach has for many years been contributing through his publications to our understanding of Chinese influence on Europe, and studying the question of Chinese influence on the Prussian civil service examinations. He has done a grest deal of work on it, but has not as yet published his results. He has most kindly, however, given me access to his unpublished data. He has established that a written examination for legal officials was inaugurated in Berlin in 1693; this appears to be the earliest written examination known in Europe (the statement of the article on “Examination” in the current Encyclopaedia Britannica to the contrary notwithstanding!). Lach has only gone so far as to say that the Prussian institution of examination “was possibly influenced by the example of China”; see Lach, Donald F., The Preface to Leibniz' Novissima Sinica (Honolulu, 1957), p. 37Google Scholar. But on the basis of his evidence, and of other materials that bear on the problem, I believe that the probability of such influence is very great indeed.

⁴⁷ Camb. Med. Hist. VI (New York and Cambridge, 1929; reprint of 1936), 131.Google Scholar

⁴⁸ Camb. Med. Hist. V (New York and Cambridge, 1926; reprint of 1929), 205–206, VI. 148Google Scholar. Haskins, Charles H., “England and Sicily in the Twelfth Century,” in The English Historical Review XXVI (1911), 433–447. 641–665.CrossRefGoogle Scholar

⁴⁹ Naturally I am fully aware of such matters as the Norman penchant for organization and centralization, which operated independently of any Chinese influence. But I think it was precisely because European states of the twelfth and thirteenth centuries were developing along lines parallel to those taken earlier in China, that they were able to appreciate and utilize certain Chinese techniques. I suspect that diffusion, as distinguished from acculturation, is most likely to occur between cultures that have independently developed culture patterns having a good deal in common.

⁵⁰ The brevity and simplicity of this definition of “feudalism” will undoubtedly make it unacceptable, if not shocking, to many scholars. I would like to point out that if I am guilty of error it springs from stupidity rather than from ignorance. I have been studying feudalism as an institution for more than thirty years, and participated in the conference at Princeton which resulted in the volume Feudalism in History, edited by Coulborn.

An excellent resumé of the literature debating the nature of feudalism is given in Hall, John W., “Feudalism in Japan—a Reassessment,” in Comparative Studies in Society and History V (1962), 15–51.CrossRefGoogle Scholar

In my opinion there is widespread confusion between the definition of feudalism as an institution, and the description of institutional complexes designated as feudal that have existed at various times and places. Many scholars give detailed descriptions of the political, social, and economic conditions associated with feudalism in one or more particular situations, and declare that “true” feudalism must include all of these details. It is undoubtedly true that the feudal pattern of government has strong tendencies to produce certain concomitant phenomena; a military aristocracy, a code of “chivalric” honor, a hierarchically organized society, heriditary privilege, etc., etc. But these occur in various guises, they may occur independent of feudalism, and any particular feudal system may at some time lack one or more of such stigmata. A definition, however, should embody only the sine qua non.

“Feudalism” is not, after all, a thing; it is a term, and a term that is perhaps no older than the seventeenth century. It means what it is defined to mean, and in a sense anyone may define it as he wishes if he does so clearly and uses it consistently. My functional definition is designed to make “feudalism” manageable as a concept for the analysis of institutional history. It is intentionally minimal and does not include some factors that are perhaps inseparable from feudalism, but I think that it does imply them.

It may be objected that I am presuming to promulgate a “private definition” of feudalism. In fact, the shoe is on the other foot. I worked out this definition, insofar as I am aware, independently. But having done so, I find that it seems to correspond closely with the sense intended by those who first used the term. Bloch, Marc, Feudal Society, trans. Manyon, L. A. (Chicago, 1961; 2nd impression, 1962), pp. xvii–xviiiGoogle Scholar, writes that it appears to have been first employed “to designate a state of society” by the Comte de Boulainvilliers, and that Montesquieu first gave wide currency to the conception. “To Boulainvilliers and Montesquieu, living in an age of absolute monarchy, the most striking characteristic of the Middle Ages was the parcelling out of sovereignty among a host of petty princes, or even lords of villages. It was this characteristic that they meant to denote by the term feudalism …”

If others wish—and clearly a great many do—to list an array of political, social, and economic phenomena that have occurred in particular situations, and assert that where these phenomena are duplicated, and only there, “feudalism” exists, that is their privilege. But the earlier, simpler, more functional sense of the term seems to me to have much greater usefulness as a conceptual tool. The burden of proof rests upon those who would replace it with more complex and particularized formulae.

There are, no doubt, those who would disagree with the proposition that feudalism should be under-stood as “a system of government,” but this is clearly the position taken by many of the most serious students of the institution. See: Weber, Organization, pp. 351, 373–381. Stephenson, Carl, Medieval Feudalism (Ithaca, New York; 1942), pp. 14, 32, 94Google Scholar. Coulborn, , Feudalism, pp. 4, 9, 16–18Google Scholar. Bloch, , Feudal Society, pp. 187, 401, 441.Google Scholar

Concerning the delegation of authority—of what I have called “limited sovereignty”—two authorities in addition to Boulainvilliers and Montesquieu, who were cited above, may be quoted. Stephenson (Medieval Feudalism, p. 32Google Scholar) writes that “seignorial government originated as a delegation of power by the monarchy …” Joseph R. Strayer, in “Feudalism in Western Europe” (in Coulborn, , Feudalism, pp. 17–18), says: “Effective feudal government is local, and at the local level public authority has become a private possession. Yet … kingship survives … and by the thirteenth century most lawyers insist that all governmental authority is delegated by the king …”Google Scholar

⁵¹ This definition is of course much simpler than some scholars would consider necessary, but again I think it includes or implies the essentials. I do not propose such a simple definition because this is necessary in order to represent early Chinese government as “bureaucratic.” On the contrary, I believe that one can find in the bureaucracy of the Han dynasty nearly all of the many factors that Weber includes in his complex descriptions of modern bureaucracy. See: Weber, , Organization, pp. 329–341.Google Scholar

Weber, , Sociology, pp. 196–204Google Scholar. All three of the characteristics that Weber lists in the latter work, p. 196, as found fully developed “only in the modern state,” were in fact fully present in Han times (see Note 19 above). Of the factors listed in his fuller description of modern bureaucracy, diose most likely to be called in question are specialized technical competence and pensions.

While it is true that the Chinese civil service (like that of the British Empire) has in theory preferred the “generalist” rather than the “specialist,” there was such emphasis on career experience, and such definite sequence of office, that almost any man who attained a high post had survived a rigorous process of practical training and selection. For this process in Han rimes see Wang, , “Han Government,” pp. 178–182Google Scholar; for the Sung period see Edward A. Kracke, Jr., Civil Service in Early Sung China—960–1067 (Cambridge, Mass.; 1953) (referred to hereafter as: Kracke, Civil Service), pp. 87–90, 118–125; for the Ming see Hucker, Charles O., “Governmental Organization of the Ming Dynasty,” in HJAS XXI (1958), 30Google Scholar; for the Ch'ing see Yang, C. K., “Some Characteristics of Chinese Bureaucratic Behavior,” in Confucianism in Action, ed. Nivison, David S. and Wright, Arthur F. (Stanford, Calif.; 1959)Google Scholar (hereafter referred to as: Yang, “Bureaucratic Behavior”), pp. 136–146. The Yen-t'ieh Lun gives us what appears to be a reasonably faithful report of a debate at the court of a Han emperor early in the first century B.C. between the high officials—the “ins”—and the scholars criticizing them—the “outs.” It is noteworthy that the “ins” lay great emphasis on administrative expertise, while the “outs” deprecate it from the Confucian point of view; see Discourses on Salt and Iron, trans. Esson M. Gale (Leyden, 1931), pp. 59–61, 86–91, 106–111. It appears probable that technical competence has always played a greater role in Chinese government than Confucian theory would allow. But most of the accounts upon which our knowledge of Chinese government is based were written from the Confucian standpoint. Pensions for higher members of the bureaucracy existed as early as the Former Han period. At first there were individual grants, but in 1 A.D. the principle was established that officials above a certain grade received one-third of the pay of the office from which they retired, for life. See Hsü T'ien-lin, Hsi-Han Hui-yao ¹ (Kuo-hsüeh Chi-pen Ts'ung-shu edition), pp. 434–435. Under the Sung dynasty retirement benefits were apparently even more general and liberal; see Kracke, , Civil Service, pp. 82–83Google Scholar. It is interesting to note that this situation apparently deteriorated later. C. K. Yang, writing primarily of the Ch'ing dynasty situation, says that “there was, of course, no provision for pensions” (“Bureaucratic Behavior,” p. 158Google Scholar). Western scholars have some tendency to take recent practice, for which information is more readily available, as representative of the whole history of bureaucracy in China; this illustrates the fact that such procedure is often misleading.

A number of social scientists have found Weber's complex description of bureaucracy as an “ideal type” to be, however stimulating, insufficiently flexible to be used as a practical norm; instead they propose less rigid criteria. See: Friedrich, Carl J., “Some Observations on Weber's Analysis of Bureaucracy,” in Reader in Bureaucracy, pp. 27–33Google Scholar. Gouldner, Alvin W., “Discussion of Industrial Sociology,” in American Sociological Review XIII (1948), 396–400CrossRefGoogle Scholar. Blau, , The Dynamics of Bureaucracy, pp. 1–3, 202–203Google Scholar. Blau, , Bureaucracy in Modern Society, pp. 14, 19, 60–66Google Scholar. Yang, , “Bureaucratic Behavior,” pp. 134–137, 350 (note 4).Google Scholar

⁵² The tendency of bureaucracy to inhibit initiative has often been noted. See Merton, Robert K., “Bureaucratic Structure and Personality,” in Social Forces XVIII (1940), 561–565.Google Scholar

⁵³ A number of Chinese scholars believe that there was clearly a feudal system before Chou, in the Shang dynasty. They may be correct, but in my opinion the evidence is not conclusive.

⁵⁴ See Ssu-ho, Ch'i, “A Comparison between Chinese and European Feudal Institutions,” in Yenching Journal of Social Studies IV (Peiping, 1948), 1–13.Google Scholar

⁵⁵ Weber, , The Religion of China, p. 86Google Scholar, says, “The sib … in the occidental Middle Ages was practically extinct.…” See also Bloch, , Feudal Society, pp. 137–142, 181Google Scholar. A feature of the European system that was lacking in China was the division of loyalty whereby an individual owed loyalty equally to his paternal and maternal relatives, so that in the same lineage the obligations of kinship differed in each generation; see ibid., pp. 137–139.

⁵⁶ This does not mean that it did not have some importance, for of course it did. But while insisting on this, Bloch also repeatedly points out that in principle blood ties were “foreign to” European feudalism; see Bloch, Feudal Society, pp. 123–142, 181, 443. In the Chinese context such an idea would seem very strange.

⁵⁷ On the political importance of the family in' recent times see: Yang, , “Bureaucratic Behavior,” pp. 157–159Google Scholar. Hsien Chin Hu, The Common Descent Group in China and Its Functions (New York, 1948), pp. 53–63, 95–100. Maurice Freedman, Lineage Organization in Southeastern China (London, 1958), pp. 73–76, 114–125. During the twentieth century the family has played a diminishing role. Its present status under the Chinese Communists is certainly greatly reduced, but appraisal of the current situation1 is difficult. At the other end of history we know very little in any detail about the family in the Shang period, but what we do know permits the hypothesis that it was already very important.

⁵⁸ The expression kuo-chia ^m is supposed to derive from the Mencius; see The Chinese Classics, trans. James Legge, II, Mencius, 2nd edition (Oxford, 1895), p. 295. If so, the usual sense of the expression is not that in which it was originally employed, but this is a familiar semantic phenomenon.

⁵⁹ Weber, , The Religion of China, p. 86Google Scholar. On the following page Weber says that “the cohesion of the sib undoubtedly rested wholly upon the ancestor cult.” This is surely an exaggeration. The great functional utility of the kinship organization kept the ancestor cult alive, even among agnostic Chinese intellectuals; the two reinforced each other.

⁶⁰ Durkheim, Émile, Le Suicide (Paris, 1897), pp. 264–288, 428–451Google Scholar. Parsons, , The Structure of Social Action, p. 377.Google Scholar

⁶¹ Durkheim, , Le Suicide, pp. 432–434.Google Scholar

⁶² The early history of Chinese culture is as yet little known. It is evident, however, that as late as the beginning of the first millennium B.C. “Chinese” culture was completely dominant only over a relatively limited area. Its great extension was the result, in my opinion, primarily of a process of acculturation promoted by the attractive power of the culture (in which kinship solidarity was a major factor), rather than by military conquest or colonization, though both of these latter processes have certainly occurred. It goes without saying that as Chinese culture spread it was modified, and took over elements from other cultures; an instance of this process, involving the culture of Ch'u, will be dealt with in this paper. But the dominant pattern, however modified, remained “Chinese.”

⁶³ Shih were also started in other ways, but they need not concern us. The whole subject of the shih is one on which, considering its importance, we have remarkably little clear information. The difficulty of understanding it may be judged by the fact that Ssu-Ma Ch'ien, writing around 100 B.C., did not distinguish between hsing and shih, and appears to have confused them. See Kametaro, Takigawa, Shih-chi Hui-chu K'ao-cheng” (Tokyo, 1932–1934)Google Scholar (referred to hereafter as: Shih-chi) 8.2–3.

⁶⁴ In modern usage both hsing and shih mean “surname,” but to transfer this to antiquity is to compound confusion. In a sense both were surnames, but almost every aristocrat had both a hsing and a shih. But the hsing was hardly ever used, anciently, except in connection with women (it is demonstrably for this reason, and not as is sometimes alleged as a relic of matriliny, that many hsing had the element meaning “woman” as a part of the character).

⁶⁵ Max Weber (Gesammelte Aufsätze zur Religionssoziologie I, 314–316, 373–379Google Scholar) did not distinguish hsing and shih, but called both Sippe. Haloun, Gustav, “Contributions to the History of Clan Settlement in Ancient China I,” in Asia Major I (1924), 76–111Google Scholar, used “clan’ as a general term for kinship groupings, and referred to the hsing and shih by the Chinese characters. Granet, Marcel, La Polygynie Sororale et le Sororat dans lit Chine Féodale (Paris, 1920), p. 50Google Scholar, spoke of the “familles seigneuriales,” some of which shared the same hsing, and apparently distinguished the shih (although he did not cite the term) as “branches d'un même tronc.” In another work, Granet, La Religion des Chinois (Paris, 1922), p. 40Google Scholar, he referred to the designation of the shih as “un cognomen et non pas un nom de famille”; the groups thus designated he called “les rameaux d'une même famille.” See also Granet, , La Féodalité Chinoise (Oslo, 1952), pp. 99, 125, 173–174Google Scholar. Maspero, Henri, La Chine Antique, revised edition (Paris, 1955), pp. 100, 103Google Scholar, translated hsing as “clan” and shih as “famille.” Eberhard, Wolfram, A History of China, trans. Dickes, E. W. (London, 1950), pp. 26, 50–51Google Scholar, uses “family” to denote both hsing and shih. Hu, The Common Descent Group in China and Its Functions, pp. 11–12Google Scholar, briefly describes this historical phenomenon but employs the romanizations hsing and shih.

⁶⁶ There is some variation in the terminology used by anthropologists to denote such kinship groups, in this and other countries.

⁶⁷ The Ch'un Ts'ew, with the Tso Chuen, trans. Legge, James, in The Chinese Classics V (London, 1872), pp. 343Google Scholar (Duke Ch'eng, year 2), 714 (Duke Chao, year 26) (hereafter this work will be referred to as “Legge, Tso-chuan”; in every case the reference will not be to Legge's translation, but to the Chinese text, and the page number will be followed by the name of the reigning duke of Lu and the number of the year of his reign).

The characters used to refer to vassals belonging to other hsing are sheng chiu ^q. These are sometimes translated as “sororal nephews and maternal uncles,” but in fact each character has a number of meanings, more than one of which may apply here. For a discussion of these rather complex phenomena, see Fêng, Han-yi, The Chinese Kinship System (Cambridge, Mass., 1948), pp. 45–50.Google Scholar

Not only the king but also lesser rulers spoke of vassals belonging to other hsing in this way. As used by the king this may have been in part a fiction, since the Chou house may not have intermarried with all of its vassals. Yet there was certainly some such intermarriage. The third century A.D. commentator Tu Yü said that the rulers of the state of Ch'i intermarried with the Chou house for generations; see Ch'un-ch'iu Tso-chtian Chu-su, in Shih-san-ching Chu-su (Nanchang, 1815), 25.25a.Google Scholar

⁶⁸ Kuo-yü (Ssu-pu Pei-yao edition) 2.6a says that a subordinate cannot appeal over the head of his overlord, for if that were permitted “there would be litigation between' father and son, and all hierarchy would be subverted.” On the importance of kinship in political affairs, see Legge, , Tso-chuan, pp. 143Google Scholar (Hsi 5), 354 (Ch'eng 4), 545 (Hsiang 29).

⁶⁹ Weber, , Organization, p. 368.Google Scholar

⁷⁰ Legge, , Tso-chuan, pp. 31Google Scholar (Yin 11), 38 (Huan 2), 81 (Chuang 8), 127 (Min 2), 292 (Hsüan 3), 438 (Hsiang 9), 454 (Hsiang 12), 510 (Hsiang 25), 666 (Chao 17), 858 (Ai 26).

⁷¹ This code was, of course, li ^r. This is commonly translated as “ritual,” but this is most inadequate. The concept was redefined by Confucius, but as we find it in the Tso-chuan it often epitomizes the whole Chinese tradition, and civilized as distinguished from “barbarian” usage. It also prescribed the obligations, religious, social, and political, of everyone from plebeians to rulers; see Legge, , Tso-chuan p. 715Google Scholar (Chao 26) (the attribution of this speech is no doubt apocryphal, but this does not nullify its importance). Li included ritual, but a great deal more.

⁷² Tuan-lin, Ma, Wen-hsien T'ung-k'ao (Ming Chia-ching edition), Preface, p. 9Google Scholar. de Bary, W. T., “Chinese Despotism and the Confucian Ideal: A Seventeenth Century View,” in Chinese Thought and Institutions, ed. Fairbank, John K. (Chicago, 1957), pp. 169, 196–197.Google Scholar

⁷³ Maspero, , La Chine Antique, p. 103Google Scholar, distinguished them but in my opinion did not characterize them well. He wrote that the hsing “n'avait qu'une importance religieuse.” Of the shih he wrote that “elle était essentiellement civile et même administrative.” Both descriptions are inadequate. Granet, , La Polygynie Sororale et le Sororat dans la Chine Féodale, pp. 48–51Google Scholar, describes the aristocratic kinship organization in a highly theoretical manner. Thus he writes that “les seigneuries de même nom [i.e. hsing] ne doivent point se faire la guerre.” No doubt they ought not, but they did it frequently. Again he says that “le lien qui attache le vassal au suzerain étant absolu, les familles ne pouvaient se lier entre elles par des liens d'interdépendance compléte, pas plus qu'un fils de famille, dès qu'existe une autorité domestique, n'est laissée libre de contracter des amitiés qui l'engagent jusqu'à la mort.” But in this case how can we explain the fact that in state after state the authority of suzerains was undermined by strong alliances between their vassals? These conflicts between the theory and historic fact tend to become clarified if we understand the difference between the organization of the hsing and that of the shih.

⁷⁴ Weber, , Organization, pp. 145–146Google Scholar. Weber's term was Verband; “corporate group” is the felicitous translation of Talcott Parsons. Weber wrote that “it is sufficient for there to be a person or persons in authority—the head of a family [etc.] … whose action is concerned with carrying into effect the order governing the corporate group…. Whether or not a corporate group exists is entirely a matter of the presence of a person in authority, with or without an administrative staff.” The shih also fulfilled the requirements of some other definitions of a “corporate group,” since it held property in common and persisted for generations.

⁷⁵ Our knowledge of Western Chou government is so slight as to be almost non-existent. From the fragmentary evidence it appears quite possible that there was, in its early portion, a more effective and centralized administration than critical scholars have usually been willing to suppose. This was first called to my attention by my wife, Dr. Lorraine Creel, on die basis of her study of Western Chou bronze inscriptions. More recently Professor Hsü Cho-yün has pointed out to me further evidence of this kind. That there were functionaries, who may be called “proto-bureaucrats,” is clear; but it is difficult to determine to what extent they may have functioned in the domains of vassals as well as of the king. For a few of many pertinent references, see: Mo-jo, Kuo, Liang-Chou Chin-wen-tz'u Ta-hsi K'ao-shih ^a (Tokyo, 1935)Google Scholar (hereafter referred to as: Kuo, Chin-wen K'ao-shih), pp. 50–103, 35a–39b, 133a–139aGoogle Scholar. The Chinese Classics, trans. James Legge, Vol. III, The Shoo-king, pp. 381, 399, 410–411, 414.Google Scholar

In this connection there always arises the question of the relation of the Chou-li to actual Western Chou government. It is restudied in a recent paper: Broman, Sven, “Studies on the Chou Li,” in Bulletin of the Museum of Far Eastern Antiquities, Stockholm, XXXIII (1961), 1–89Google Scholar. Broman concludes: “Thus the Chou Li depicts a governing system which, in all its essentials, prevailed in middle and late feudal Chou in the various states and has its roots irt the system pertaining to late Yin and early Chou” (ibid., p. 73). For my own part, I should agree that the “roots” of the Chou-li system undoubtedly do lie in Western Chou, to a degree that scholars have often been unwilling to credit. But that the system itself, in anything like the fullness and complexity that we find in the Chou-li, existed at any time in the Chou period does not seem to me to be proved even by the voluminous and interesting evidence that Broman has assembled. My own hypothesis is that the Chou-li is probably a late Chan-kuo elaborstion upon governmental institutions that did exist, and thus is no mere figment of imagination. But to determine what the actual substratum was will require much future research, and may be impossible. For such research, Broman's paper provides material of great value.

⁷⁶ Ssu-ho, Ch'i, Chou-tai Hsi-ming Li K'ao ^t, in YCHP XXXII (1947), 221–223Google Scholar. Stephenson, , Medieval Feudalism, pp. 23–24Google Scholar. Bloch, , Feudal Society, pp. 190–210Google Scholar. There is a widespresd misconception that the hereditary transmission of fiefs is an essential property of feudalism. Thus Chang Yin-lin, Chou-tai ti Feng-chien She-hui ^u, in CHHP X (1935), 803Google Scholar, makes it a part of his definition of feudalism. But hereditary transmission nullifies one of the principal objectives of feudalism: the appointment of vassals of known loyalty and capacity. Circumstances usually bring about hereditary transmission of fiefs, but where it is firmly established feudalism as an effective method of government is usually moribund if not in fact dead.

⁷⁷ Legge, , Tso-chuan, p. 654Google Scholar (Chao 14).

⁷⁸ Pareto, Vilfredo, Traité de Sociologie Générale, French translation by Pierre Boven, revised by the author (Paris, 1917–1919), pp. 1112–1113.Google Scholar

⁷⁹ Confucius and his disciples, as reported in the Analects, treat kinship as the archetype of the social and political order. But it is significant that the only kinship ties mentioned in the Analects are those between parents and children, and those between brothers; see Analects 1.2, 1.6–7, 1.11, 2.5–8, 2.20–21, 4.18–19, 4.21, 8.2, 9.15, 11.4, 11.21, 11.23, 12–5. 12.11, 13.7, 13.18, 13.20, 13.28, 17.9, 17.21, 19.18. Kinship ties beyond the nuclear family, and the question of the duty of loyalty to hsing or shih, simply go unmentioned in the Analects. Certainly Confucius enjoins family solidarity in Analects 13.18, but again this relates only to the nuclear family. Questions of getting and keeping office and wealth, and of loyalty to one's superiors beyond the nuclear family, were deeply involved, according to the traditional morality, with the interests of the shih, yet Confucius argues that they should be determined on the ground of moral imperatives aiming at the good of the whole society; see Analects 4.5, 4.16, 7.15, 8.13.3, 11.16, 11.23, 12.9, 14.1, 14.13, 14.17–18, 15.6.2, 15.8, 15.31.

⁸⁰ Ch'ün-shu Chih-yao 36.25b–27a. Lü-shih Ch'un-ch'iu 17.7a. Frühling und Herbit des Lü Pu We, trans. Wilhelm, , p. 270Google Scholar. T'ai-p'ing Yü-lan 638.4b Han-fei-tzu 11.11b–12a, 16.6b. Merton, , “Bureaucratic Structure and Personality,” pp. 561, 565–566.Google Scholar

⁸¹ Bloch, , Feudal Society, pp. 424–425Google Scholar. The failure of heirs was a principal cause of the reversion of European fiefs to the crown, but in China a failure of male offspring was rare because of aristocratic polygyny. Furthermore, not only sons but also brothers could inherit the leadership of a shih, and therefore could inherit the fief. See also: Fesler, James W., “French Field Administration: The Beginnings,” in Comparative Studies in History and Society V (1962), 76–111.CrossRefGoogle Scholar

⁸² Tsien, Se-Ma, Les Mémoires Historiques, trans. Chavannes, Édouard (Paris, 1895–1905)Google Scholar (referred to hereafter as Mem. Hist.), II.132, 530–532.

⁸³ Chao I, Kai-yü Ts'ung-k'ao, in Ou-pei Ch'üan-chi (1877)Google Scholar, 16.8b–10a. Yao Nai, Hsi-pao Hsüan Wen-chi^w (Ssu-pu Pei-yao ed.) 21a. Ch'i Ssu-ho, Chan-kuo Chih-tu K'ao ^x, in YCHP XXIV (1938), 214Google Scholar, note 369. Bodde, Derk, China's First Unifier (Leiden, 1938), pp. 135–139, 238–243.Google Scholar

⁸⁴ Mem.Hist. II.65–66.

⁸⁵ Fu-pao, Ting, Shuo-wen Chieh-tzu Ku-lin (1928), pp. 3970a–3971bGoogle Scholar. Legge, , Tso-chuan, pp. 181Google Scholar (Hsi 22), 443 (Hsiang 10). Forms of this character appearing in pre-Ch'in bronze inscriptions also contain the element mu meaning “tree” (so-called “radical” .no. 75), probably representing the post from which heads were sometimes suspended. But if we may judge from the Shuo-wen, this component of the character had been dropped by Han times.

⁸⁶ Kuo, , Chin-wen K'ao-shih, p. 2033.Google Scholar

⁸⁷ Yin-lin, Chang, Chou-tai ti Feng-chien She-hui, p. 826Google Scholar, and Chieh-kang, Ku, Ch'un-ch'iu Shih-tai ti Hsien ^zGoogle Scholar, in Yü-kung ^aa VII, Numbers 6–7 (referred to hereafter as: Ku, Ch'un-ch'iu Hsien), p. 179, seem to consider this to be the original sense in which the character was used to denote territory. In my opinion, however, our evidence is not adequate to warrent a firm conclusion on this point.

⁸⁸ A bronze vessel ascribed to the reign of Duke Ling of Ch'i (581–554) records that the duke gave to a retainer a city and two towns and “their three hundred hsien” (Kuo, , Chin-wen K'ao-shih, pp. 202b–205bGoogle Scholar). Bodde, , China's First Unifier, pp. 238–241Google Scholar, says that “it is impossible, however, to suppose that such an enormous number as three hundred hsien could have been given to anyone in Ch'i as this time, when a little later, as we know from the Tso-chuan, the total number of hsien, in the equally large state of Chin, amounted to only forty-nine.” Bodde appears to question the authenticity of the inscription, on this ground alone. But surely it is clear that the term hsien, as used here, does not refer to administrative districts but only to small “suburban” areas associated with towns.

Various cases of the use of the term hsien where it denotes territory, but not the directly controlled administrative district, are cited in Ku, Ch'un-ch'iu Hsien, pp. 180–184.Google Scholar

⁸⁹ See Ku, Ch'un-ch'iu Hsien, pp. 169–170, 180–181, 184–187.Google Scholar

⁹⁰ Shih-chi, 5.16–17. Even if Ssu-Ma Ch'ien, writing this passage around 100 B.C., intended it to mean that Ch'in made these areas into districts that were called hsien, this would not necessarily prove that districts so called existed in Ch'in in the seventh century B.C. The Huai-nan-tzu, written by contemporaries of Ssu-Ma Ch'ien, at one point says that the whole world was possessed by two kings who are traditional exemplars of wickedness, Chieh (whose reign is doubtfully dated 1818–1766 B.C.) and Chou (1154–1123 B.C.). The text says that every region “was [included in their] chün ^ac and hsien”; see Huainan-tzu (Ssu-pu Pei-yao ed.) 13.9b. In fact, however, the territorial division known as the chün appears in the literature even later than the hsien; see Bodde, , China's First Unifier, pp. 139–140, 243–246Google Scholar. And there is no evidence to indicate that chün and hsien existed at the early date to which the passage in the Huai-nan-tzu refers. Furthermore, it is improbable that its author had any special intention to indicate that they did; he was merely using terminology current in his day. The same may well be true of this Shih-chi passage; the expression hsien chih ^ab may simply mean that Ch'in incorporated these areas in its territory. The Shih-chi is a valuable history, but on such points of detail it does not have as much specifically evidential force as the Tso-chuan.

⁹¹ A later passage in the same chapter (Shih-chi 5.36) tells us that in 623 B.C. Ch'in “attacked the king of the Jung and added [to its territory] twelve states, [thus] opening up territory of a thousand li [in extent].” Yet in connection with this major annexation the character hsien is not used at all; this would seem to cast grave doubt on the interpretation' of Shih-chi 5.16–17 as proof that as early as 688 B.C. Ch'in used the hsien as a political institution.

⁹² The Tso-chuan makes no reference to hsien in Ch'in. A passage in the Kuo-yü (Ssu-pu Pei-yao ed.), 8.10b, records a conversation implying that Ch'in had hsien in 651 B.C. But Bodde, China's First Unifier, pp. 243–244Google Scholar, states the evidence which shows that this mention of hsien is probably a later addition to the text.

⁹³ While the Tso-chuan has little information on Ch'in, it appears that Ssu-Ma Ch'ien had a history of the state and used it in writing his chapter on Ch'in in the Shih-chi; see Shih-chi 5.2 (commentary, columns 1–4) and Ch'i, Chan-kuo Chih-tu K'ao, p. 214Google Scholar note 369. This chapter gives no evidence that Ch'in had any special type of organization to assure control by the central government. Compare its late establishment of taxation, in 348 B.C. (Shih-chi 5.54), with the development in Ch'u two centuries earlier, described in Legge, Tso-chuan, p. 512Google Scholar (Hsiang 25). The great advances in government in Ch'in are ascribed to two men, Pai Li Hsi^ad and Shang Yang; neidier was a native of Ch'in and both presumably took administrative techniques with them to Ch'in when they went there.

⁹⁴ Legge, , Tso-chuan, p. 223Google Scholar (Hsi 33).

⁹⁵ See Legge, , Tso-chuan, p. 552Google Scholar (Hsiang 30), where the chancellor asks the old man who is his hsien tai-fu ^af, evidently assuming that no matter where he comes from in the state of Chin he must have one. The idea that a whole state might, in the Ch'un-ch'iu period, have been divided into hsien seems to be contrary to that of most scholars, who view the hsien as comprehending only peripheral territories at that time. On this and various other grounds Ku Chieh-kang, Ch'un-ch'iu Hsien, pp. 190–193Google Scholar, denounces this passage in the Tso-chuan as a forgery inserted by Liu Hsin (died 23 A.D.). However correct that theory may be, some of his criticisms undoubtedly tell against the passage, and may possibly invalidate it as evidence that all of Chin was divided into hsien. But there is other evidence of this. Qnly six years later, in 537 B.C., we are told that Chin had forty-nine hsien which could furnish four thousand nine hundred chariots of war (Legge, , Tso-chuan, p. 602Google Scholar [Choa 5]). By implication, this must mean that these hsien included virtually the whole state, for this is a very large number of chariots; it may be the largest attributed to any state in the Ch'un-ch'iu period. The Tso-chuan (ibid., p. 645 [Chao 13]) reports that eight years later, in 529, when Chin held a review of its military might to overawe the other states, it assembled only four thousand chariots.

⁹⁸ From studying everything that is said about hsien in Chin in the Tso-chuan, I am inclined to think that hereditary transmission was normal. The evidence on specific cases is sometimes complex. In Legge, Tso-chuan, p. 586Google Scholar (Chao 3), Chao Wen-tzu is quoted as saying, “Wen is my hsien.” This claim is apparently based on heredity. Ibid., p. 574 (Chao 1), implies that the ancestral temple of the Chao shih was at Wen. Chao Wen-tzu died in 541, but as late as 493 we find that the tai-fu of Wen is named Chao Lo; see ibid., p. 797 (Ai 2). It would appear therefore that the Wen hsien was hereditary in the Chao shih. Ibid., p. 602 (Chao 5), says that “the Han [shih] draws its revenues from seven towns which [with their attached territories] all amount to hsien”; this would clearly seem to indicate hereditary tenure. Two years later the Han holdings were enlarged with yet another hsien; see ibid., p. 613 (Chao 7).

⁹⁷ References to hsien in states other than Ch'in, Chin, and Ch'u are either of dubious authenticity, or do not appear to refer to the administrative institution. In Legge, , Tso-chuan, p. 592Google Scholar (Chao 4), a speech attributed to an officer of Lu in the year 538 B.C. uses the character hsien, but the speech is fanciful and dubious and in any case does not indicate that the hsien was a political institution in Lu. A conversation recorded in Kuo-yü 2.8b–9b, supposed to have taken place in the period 606–586 B.C. and relating to the state of Ch'en, twice uses the character hsien, but again the conversation is of dubious authenticity and there is no indication that it refers to a political institution. Note 88, above, discusses the bronze inscription of the state of Ch'i which has the character hsien, but clearly not as a governmental institution. Legge, , Tso-chuan, p. 679Google Scholar (Chao 20), gives the only occurrence of hsien in that work as relating to Ch'i, but again not as a governmental institution. It is so mentioned as occurring in Ch'i in the Ch'un-ch'iu period in: Kuo-yü 6.1a–8a. Kuan-tzu (Ssu-pu Pei-yao ed.) 22.8b–9a. Yen-tzu Ch'un-ch'iu (Ssu-pu Pei-yao ed.) 7.11a. Shuo-yüan (Ssu-pu Ts'ung-k'an ed.) 2.15a. For reasons that would require a great deal of space to detail, each of these references is highly questionable. Bodde, China's First Unifier, pp. 240–243Google Scholar, has summarized evidence against them; I would cite evidence additional to his, and I agree with his conclusion that it is doubtful that the hsien existed as an administrative institution in Ch'i, insofar as the Ch'un-ch'iu period is concerned. In Bodde's compilation of data on hsien (ibid., pp. 238–240) he notes references to hsien during the Ch'un-ch'iu period only for Ch'i (which he questions), Chin, Ch'in, Ch'u, and Wu. Ku Chieh-kang, Ch'un-ch'iu Hsien, pp. 170–179Google Scholar, also cites only the same five states. The single reference for Wu, for the year 545 B.C., occurs in Shih-chi 10–11 (Mem.Hist. IV.7). But this same incident is related in Legge, , Tso-chuan, p. 538Google Scholar (Hsiang 28), where the same place is not called a hsien; this raises the possibility that the author of the Shih-chi simply added the term hsien as an ordinary term for such a place. Together with the fact that neither the Tso-chuan nor, insofar as I am aware, any other work says that the hsien existed as a governmental institution in Wu during the Ch'un-ch'iu period, this would seem to make it impossible to give much weight to this reference in the Shih-chi.

⁹⁸ A principal reason is certainly the rather cryptic nature of the reference in the Tso-chuan to the earliest recorded establishments of hsien in Ch'u; see note 115, below. This early date has not been entirely overlooked, however. Both Ku Chieh-kang, Ch'un-ch'iu Hsien, pp. 170–173Google Scholar, and T'ung Shu-yeh, Ch'un-ch'iu Shih ^ah (1946), p. 84Google Scholar, take note of it and say that hsien appeared in Ch'in and Ch'u at around the same time. Chang Yin-lin, Chou-tai ti Feng-Men She-hui, p. 826Google Scholar, appears to have considered hsien to be earlier in Ch'in, but he says that the administration through hsien in Ch'u constituted “the beginnings of the institution of government through hsien and chün.”

In the course of making final revisions of this paper for publication I encountered an essay by Hung Liang-chi (1746–1809) in which he states that “in the Ch'un-ch'iu period the tendency toward the conversion of fiefs into chün and hsien was already under way; it was initiated in Ch'u and continued in Ch'in and Chin”: Hung Liang-chi, Ch'un-chu'iu Shih I Ta-i Wei Hsien Shih yü Ch'u Lun, in Keng-Sheng-chai Wen Chia Chi ^al (1802) 2.1–2. This little essay, less than seven hundred characters in length, naturally does not marshal all the evidence or deal with all the problems, but Hung's incisive mind saw through to the point which most others have missed.

⁹⁹ The evidence for the traditional view that the ruler of Ch'u was enfeoffed in recognition of loyal service by the Chou King Ch'eng (ruled 1115–1079 B.C.), as stated in Shih-chi 40.5–6, seems very inadequate. The Shih-chi (40.5–8) gives the impression that the rulers of Ch'u were loyal vassals until the reigm of Hsiung Ch'ü (887–878 B.C.). But a bronze inscription dated to the reign of King Ch'eng himself speaks of a military expedition by the king against the ruler of Ch'u; see Kuo, Chin-wen K'ao-shih, p. 3Google Scholar. And the original text of the Bamboo Books, as reconstituted by Wang Kuo-wei, says that in 1037 B.C. the Chou king Chao made an expedition against Ch'u; see Wang, Ku-pen Chu-shu Chi-nien Chi-chiao, in Wang Chung-ch'io Kung I-shu ^aj III (1928) 7a. In this work the state of Ch'u is named in fourteen passages, of which ten concern military actions and not one indicates that Ch'u was a vassal state of the Chou house, or subordinate to it (ibid., pp. 73, 9b, 11b, 12b, 15a, 16a, 17b–19a). Such evidence as we have seems to accord best with the hypothesis that Ch'u was an independent state from the earliest time we know until it was conquered by Ch'in in 223 B.C.

¹⁰⁰ The Ch'u ruler is called wang as early as an entry for 706 B.C.: Legge, Tso-chuan, p. 46Google Scholar (Huan 6). I have not made a statistical study on this point, but it is my impression that while rulers of other states arc sometimes called by this title in the Tso-chuan, it is applied earlier and much more commonly to the ruler of Ch'u.

¹⁰¹ Ch'i, Chan-kuo Chih-tu K'ao, p. 213Google Scholar. Hu Hou-hsüan, Ch'u Min-tsu Yüan yü Tung-fang K'ao, in Shih-hsüeh Lun-ts'ung I (1934), 2–4Google Scholar. Legge, , Tso-chuan, p. 185Google Scholar (Hsi 23). Kuo-yü 18.8a.

¹⁰² Kuo-yü 18.3b. Legge, , Tso-chuan, p. 292Google Scholar (Hsüan 3). Fu Ssu-nien, Hsin-huo Pu-tz'u Hsieh-pen Hou-chi Pa, in An-yang Fa-chüeh Pao-kao II (Peiping, 1930), 351–352.Google Scholar

¹⁰³ The origins of the people of Ch'u are difficult to determine with certainty. Some Chinese scholars have considered them to derive from the eastern people known as I. See: Hu, Ch'u Min-tsu Yüan yü Tung-fang K'ao. Kuo Mo-jo, Yin-Chou Ching-fung-ch'i Ming-wen Yen-chiu (Shanghai, 1931)Google Scholar I.51–52. Yang K'uan, Chung-kuo Shang-liu-shih Tao-lun, in Ku-shih Pien 7A (Shanghai, 1941), pp. 93, 119Google Scholar (note 4). A southwestern origin, on the other hand, is posited by Wolfram Eberhard, who writes: “Der Staat Ch'u entstand im Gebiet der Pa-Kultur und verlagerte sich im Lauf der Chou-zeit immer weiter nach Osten”; Eberhard, , Lokal-kulturen im alten China, II (Peking, 1942), 371Google Scholar. See also: Eberhard, , Kultur und Siedlung der Randvölker Chinas (Leiden, 1942), pp. 331–332Google Scholar. Eberhard, , Lokalkulturen im alten China I (Leiden, 1942), 321, 357–359, 363.Google Scholar

Considerable geographical barriers separated the people of Ch'u from those of the north; these alone might have been expected to produce some cultural differences. In the Tso-chuan the many passages telling of events in Ch'u give the impression of a people whose culture, on the whole, differs only in certain respects from that of the Chinese, and they are never stigmatized as utterly different, or non-human, as “barbarian” peoples often are in that work. Waley, Arthur, An introduction to the Study of Chinese Painting (New York, 1923), p. 21Google Scholar, wrote that the people of Ch'u “were not wholly different from the Chinese either in speech or race; the relation may be compared with that of Rome to the Italic tribes.”

Various scholars have pointed to similarities or connections between the people of Ch'u and those of the Shang dynasty. See: Kuo, Chin-wen K'ao-shih, Preface, p. 4bGoogle Scholar. Hu, Ch'u Min-tsu Yüan yü Tung-fang K'ao, pp. 32–38Google Scholar. Fu, Hsin-huo Pu-tz'u Hsieh-pen Hou-chi Pa, pp. 349–370Google Scholar. Eberhard, Wolfram, “Early Chinese Cultures and Their Development: A New Working-Hypothesis,” in Annual Report of the Board of Regents of the Smithsonian Institution, 1937 (Washington, 1938), p. 524Google Scholar. The discovery of bronzes of Shang type in Hupei may lend some corroboration to the theory of such connection; see Te-k'un, Cheng, Archeology in China, Vol. II (Toronto, 1960), 160.Google Scholar

¹⁰⁴ Whatever their connections with Shang culture may have been, the people of Ch'u clearly admired and took over much of that of Chou China. The Shih-ching and the Shu-ching were quoted by nobles and even by rulers of Ch'u; see Legge, , Tso-chuan, pp. 255Google Scholar (Wen 10), 315 (Hsüan 12), 341 (Ch'eng 2), 342 (Ch'eng 2), 751 (Ting 4), and Kuo-yü 17.7, 17.10a, 17.10b, 18.1a. Kuo-yü 17.1a–2a gives a curriculum proposed by a Ch'u officer for the education of the heir of King Chuang of Ch'u (reigned 613–591 B.C.). It includes various historical works, poetry, ceremonial, and music, and is indistinguishable from the curriculum that might have been used in a northern state (unless by the emphasis that is also given to statecraft, which may reflect the development of that art in Ch'u). An important avenue of acculturation is indicated by the anecdotes concerning the wife of King Wu of Ch'u (reigned 740–690), a woman of a smsll state to the north of Ch'u. She is quoted as moralizing to the king in a highly “Chinese” manner, while he treats her words with great respect; see Legge, , Tso-chuan, pp. 60Google Scholar (Huan 13), 76 (Chuang 5). Many other indications give the same impression, so that Ch'i Ssu-ho goes so far as to assert that by the Ch'un-ch'iu period Ch'u had already been entirely assimilated to Chou culture; see Ch'i, Feng-chien Chih-tu yü Ju-chia Ssu-hsiang; in YCHP XXII (1937), 189.Google Scholar

¹⁰⁵ It seems impossible to say whether architecture was more highly developed in Ch'u than in the northern states, but certainly there was great interest in it. Duke Hsiang of Lu (reigned 572–542 B.C.) visited Ch'u in 545 B.C. and remained for half a year. Duke Hsiang was then in the twenty-eighth year of his reign over what was considered the most “Chinese” of the feudal states, and through his travels must have been very familiar with the architecture of the northern region. But he was apparently quite captivated with that of Ch'u, and on his return to Lu erected what was called a “Ch'u palace.” No doubt this provoked criticism in Lu. One of his officers is recorded as saying, “The ruler desires [to live in the manner of] Ch'u, and therefore has built this palace; if he does not return to Ch'u, he will die in this palace.” And in fact, he did See: Legge, , Tso-chuan, pp. 539Google Scholar (Hsiang 28), 544 (Hsiang 29), 559 (Hsiang 31).

King Ling of Ch'u (reigned 540–529 B.C.) built a tower of which he was so proud that he took extraordinary measures to persuade other rulers to attend the celebration of its completion (Legge, , Tso-chuan, pp. 612, 613Google Scholar [Chao 7]). Its size and elaborate ornament are described in Kuo-yü 17.6a. King P'ing of Ch'u (reigned 528–516 B.C.) was criticized for toiling the people by building palaces “without measure”; Legge, , Tso-chuan, p. 674Google Scholar (Chao 19).

The well-known discoveries in the region of Changsha have shown that Ch'u had an art of extraordinary delicacy and vivacity. Lawrence Sickman writes: “The recent finds of magnificently decorated lacquer at Ch'ang-sha are added evidence to support the suggestion of Arthur Waley that painting in the Han Dynasty may have been derived from the State of Ch'u”; Sickman, Lawrence and Soper, Alexander, The Art and Architecture of China (Penguin Books, 1956), p. 31Google Scholar. See also Waley, , An Introduction to the Study of Chinese Painting, pp. 21–23Google Scholar.

On the literary culture of the state of Ch'u, see Liang-chi, Hung, Keng-sheng-chai Wen Chia Chi (1802)Google Scholar, 2.11–12b. Hung considered it to have been the most cultivated state, as regards literature.

¹⁰⁶ In fact references to the men of Ch'u as “barbarians,” by those of other states, are quite rare; for such an instance (by a claimant to the Chou throne who has fled to safety in Ch'u), see Legge, , Tso-chuan, p. 714Google Scholar (Chao 26). Those of Ch'u sometimes speak of themselves as “barbarians,” ironically, as in Kuo-yü 18.8b. There is a good deal of praise of Ch'u rulers and officials in the Tso-chuan, in one case by Confucius; sec for instance Legge, Tso-chuan, pp. 808Google Scholar (Ai 6), 851 (Ai i8).

¹⁰⁷ Legge, , Tso-chuan, pp. 309Google Scholar (Hsüan n), 312–313 (Hsüan 12), 468 (Hsiang 15).

¹⁰⁸ This is, I think, the conclusion that must be drawn from a careful study of the materials, but it is difficult to prove by citation of evidence. Ku Tung-kao (1679–1759) made what may well be the most thorough study ever undertaken of the Ch'un-ch'iu period. He discusses the troubles that beset the administration of the various states and says, “But although Ch'u was a barbarian state, from the Ch'un-ch'iu through the Chan-kuo period, for four or five centuries, its power remained strong among the feudal lords. To the end it was free from the tendency for superiors to tyrannize or for subordinates to usurp their places. It attained the highest excellence of administrative institutions.” Ku Tung-kao, Ch'un-ch'iu Ta-shih Piao, in Huang-ch'ing Ching-chieh Hsü-pien (1886–1888), 102.15b–16a. Ch'i Ssu-ho, Feng-chien Chih-tu yü Ju-chia Ssu-hsiang, p. 192Google Scholar, points out that the southern states of Ch'u, Wu, and Yüeh were more centralized than those of the north, but that Wu and Yüeh were destroyed by war. He also characterizes Ch'in as maintaining centralized control. This is undoubtedly correct. Nevertheless, the territory of Ch'in was smaller than that of Ch'u during most of the period in question. And while we have little knowledge of the government of Ch'in, what we have would seem to show that its administration was much less complex and developed than that of Ch'u; see note 93, above.

¹⁰⁹ Legge, , Tso-chuan, pp. 223Google Scholar (Hsi 33), 312 (Hsüan 12), 425 (Hsiang 5), 493–494 (Hsiang 22), 654 (Chao 14), 848–849 (Ai 17).

¹¹⁰ The only office in Ch'u that seems clearly to have been hereditary is that of Mo-ao, which seems wherever it is mentioned in the Tso-chuan to be occupied by a member of the Ch'ü family (to which the poet Ch'u Yüan, 343–277 B.C., belonged). The history of this office is quite interesting. At the beginning of the Ch'un-ch'iu period, in records of events in 701, 700, and 699 B.C., the Mo-ao appears to have been the principal minister of Ch'u; Legge, , Tso-chuan, pp. 55Google Scholar (Huan 11), 58 (Huan 12), 60 (Huan 13). In an entry for 690 B.C. the Mo-ao is named immediately after the Ling-yin, which may mean that the Ling-yin, who was appointed by the ruler without primary regard to heredity, had already become the chief minister as he was to remain throughout Ch'u history; Legge, Tso-chuan, p. 76 (Chuang 5). Significandy the office of Mo-ao is not even mentioned again for 132 years, and while the office is occasionally named thereafter there is nothing to show that any great importance attaches to it; Legge, Tso-chuan, pp. 468 (Hsiang 15), 494 (Hsiang 22), 497 (Hsiang 23), 512 (Hsiang 25), 601 (Chao 5). One would seem justified in inferring that while the office of Mo-ao continued to be transmitted in the same family, it became a more or less empty honor.

¹¹¹ Mem appear to have been appointed to the higher offices in Ch'u on the basis of proven merit, which would seem to require that they demonstrate their worth in a lower office first. It can be shown that in some cases at least they did move from lower to higher office, but a full demonstration of this would require detailed study which has not yet, insofar as I know, been made. Sometimes they were denoted on the basis of demonstrated incompetence, as in the case of Tzu-hsi who was twice moved from a higher to a lower office, and so resented these successive demotions that he plotted rebellion and was killed; Legge, Tso-chuan, pp. 205 (Hsi 28), 255 (Wen 10).

Clearly there was no definite progression from one office to another in Ch'u, but there are indications that the concept of an official career, in which one rose by achievement from one office to another, existed in Ch'u from an early date. After the great battle of Ch'eng-p'u in which Chin defeated Ch'u, in 632 B.C., the ruler of Ch'u reproached his chief minister who thereupon committed suicide. Duke Wen of Chin exulted at the news, asserting that a certain Wei Lü-ch'en would now be made chief minister and that he would be less devoted to the public welfare; Legge, , Tso-chuan, p. 205Google Scholar (Hsi 28). There is no evidence that in fact this Wei Lü-ch'en did become chief minister. But five years earlier, when the chief minister who later committed suicide was appointed, Wei Lü-ch'en (here called Shu-po) was sufficiently important to be able to question whether the appointment was for the good of the state; Legge, , Tso-chuan, p. 184Google Scholar (Hsi 23). At around the same time as this event the future Duke Wen of Chin, then an exile, spent some time in Ch'u; see Legge, , Tso-chuan, p. 185Google Scholar (Hsi 23), and Shih-chi 39.43–45. Apparently he formed a high opinion of the qualifications of Wei Lü-ch'en, and also believed that such qualifications would assure succession to the office of chief minister.

Somewhat similar evidence is furnished by a statement recorded as having been made by the occupant of the office of chief minister of Ch'u in 479 B.C. He was warned against his protégé, named Sheng, and he replied, “After my death, [in accordance with] the order [of succession] of the state of Ch'u, if Sheng does not become Ling-yin [chief minister] or Minister of War, who will?”; Legge, , Tso-chuan, p. 844Google Scholar (Ai 16). The character tz'u, which in Ch'un-ch'iu and Chan-kuo literature commonly denotes the order of succession to office, occurs as part of the name of certain books said to have been essential to the functioning of the government of Ch'u; see Chan-kuo Ts'e 14.10b.

¹¹² Ssu-p'ing, Mei, Ch'un-ch'iu Shih-tai ti Cheng-chih he K'ung-tzu ti Cheng-chih Ssu-hsiang, in Ku-shih Pien II (Peking, 1930), pp. 165–168Google Scholar, says that in Ch'u, from the beginning of the Ch'un-ch'iu period, feudalism was of a different son from that of the other states. In Ch'u, he says, officials were given fiefs but they merely received the revenues from them, while other officials were sent by the ruler to administer them. According to my definition of “feudalism,” this is not feudalism at all. But I am not sure that in fact we have enough evidence to say that this was definitely the situation in Ch'u. Certainly however the men who held titles that arc translated as those of “administrators” of certain territories were used by the central government to perform functions outside of those territories, in many cases.

In a “pure” feudal situation the ruler who delegates limited sovereignty is able to take it back, and replace one feudatory with another. In some feudal situations this is done. But the delegation of sovereignty over a piece of territory, even though there is some limitation of the sovereignty, commonly permits the vassal and his family to entrench themselves strongly, so that it becomes increasingly difficult for the overlord to dislodge him, or even to prevent his heirs from enjoying the fief. Thus the condition in which sons succeed their fathers as vassals, while not at all an essential property of feudalism as such, is one that by the logic of circumstances tends to develop in feudal states that persist for very long. In Ch'u, while there does seem to have been some inheritance of lands, we find the ruler taking lands away from aristocrats and moving them around with a freedom that seems much greater than that enjoyed by rulers of other states. Clearly the Ch'u aristocrats were on the whole less firmly entrenched than those of other states, which may indicate that they did not enjoy limited sovereignty, i.e., that they were not feudatories. In 529 B.C. King Ling of Ch'u lost his life as a result of a revolt by a number of families which, it is said, had lost their lands and their offices; Legge, , Tso-chuan, pp. 642–643Google Scholar (Ai 13). But Ling was a particularly oppressive king who gained the throne by murdering his predecessor, and it is not clear whether it was because his deprivations were considered contrary to the royal prerogative, or because they were so extensive, that they were so violently resented

In my opinion it is impossible to say, at least without a more careful study than any that has been made to my knowledge, whether or not Ch'u actually had feudalism. But if there was feudalism, the vassals were normally held in such tight control by the central authority that it did not limit the power of the ruler as feudal institutions commonly do.

¹¹³ Ch'i Ssu-ho, Chan-kuo Chih-tu K'ao, p. 198Google Scholar, says that in the Ch'un-ch'iu period ministers were paid only by being given the revenues from estates, while the payment of salaries in grain began in the Chan-kuo period. For the “Chinese” states this seems to be correct, but the chief minister of Ch'u is said to have received a daily allowance of grain and meat from the ruler beginning around the middle of the seventh century B.C.; see Kuo-yü 18.5b. This anecdote might be dismissed as fanciful, but it is supported by another passage in the same chapter of the Kuo-yü (18.5a) which, presumably speaking of the ruler of Ch'u, says that “the king takes constant revenues with which to feed the myriad officers.” Further corroboration is the elaborate survey of lands of the state of Ch'u, for purposes of taxation, made in 548 B.C.; Legge, , Tso-chuan, p. 512Google Scholar (Hsiang 25). In contrast, Kuo-yü 10.14a says that in the state of Chin under Duke Wen (reigned 635–628 B.C.) the duke's officers lived upon the produce of their estates.

¹¹⁴ Legge, , Tso-chuan, pp. 312Google Scholar (Hsüan 12), 342 (Cheng 2), 512 (Hsiang 25). Chan-kuo Ts'e 14.10b.

¹¹⁵ The evidence for the first appearance of the Asien in Ch'u is radier concealed in the Tso-chuan, which is undoubtedly the reason that most scholars have overlooked it. Thus Bodde, China's First Unifier, p. 239Google Scholar, lists 598 B.C. as the date of the earliest mention. But the Tso-chuan entry for the year 478, Legge, , Tso-chuan, p. 848Google Scholar (Ai 17), records the statement by a Ch'u official that under the Ch'u king Wen (reigned 689–675) the states of Shen and Hsi were made hsien by Ch'u. Both Legge, Tso-chuan, p. 78Google Scholar (Chuang 6), and the Shih-chi (Mem.Hist. IV.345) say that King Wen attacked Shen in 688. Shen does not figure again in the Tso-chuan as an independent state. In 664 we encounter the name of a Shew Kung; Legge, , Tso-chuan, p. 117Google Scholar (Chuang 30). This evidendy means Shen had been made a Asien, since kung was one of the two tides used for Asien administrators in Ch'u. Further corroboration is the fact that in 635 we find a Shen army fighting for Ch'u, under command of a Ch'u officer; Legge, , Tso-chuan, p. 194Google Scholar (Hsi 25). Chavannes thought that this indicated that Ch'u annexed Shen in 688: see Mem.Hist. 11.50 note 3, and IV.345 note 2. The extinction (implying annexation) of Hsi by Ch'u is reported in the Tso-chuan under the year 680, but this is a retrospective reference and the event must have happened some time previously; see Legge, , Tso-chuan, p. 91Google Scholar (Chuang 14).

Furthermore, the Tso-chuan tells of two men who were caused to yin districts under King Wu of Ch'u, who reigned 740–690; Legge, , Tso-chuan, p. 97Google Scholar (Chuang 18). While the term hsien is not used here, yin as a title of hsien administrators was regularly used in Ch'u; this looks very much as if the institution of the hsien may have existed in Ch'u even before 690 B.C. In the light of all diis evidence, it appears that die hsien can be documented as having existed in Ch'u approximately as early as, and perhaps much earlier than, 688 B.C., which is die date for which the Shih-chi first uses the term hsien in connection with Ch'in.

¹¹⁶ Legge, , Tso-chuan, pp. 194Google Scholar (Hsi 25), 255 (Wen 10), 359 (Ch'eng 6), 362 (Ch'eng 7), 477 (Hsiang 18), 520 (Hsiang 26), 594 (Chao 4), 719 (Chao 27), 721 (Chao 27), 751 (Ting 4), 844 (Ai 16). Kuo-yü 17.8b–9b, 18.11a.

¹¹⁷ Legge, , Tso-chuan, pp. 844–845Google Scholar (Ai 16). The title “Duke” as used here has caused much unnecessary confusion. Ch'u lay outside the political system of the Chinese states. Its ruler called himself wang, “King,” and called certain of his administrators of hsien, kung, “dukes.” The Chinese commentators, and some Western scholars after them, have called these “usurped” titles.

¹¹⁸ In the northern states we constantly find references to these groups. They are called by the surname used by their members, plus the character shih. The Chi-shin of Lu is a well-known example. The Ch'un-ch'iu Ching-chuan Yin-le (Peiping, 1937), PP 1176–1177Google Scholar, lists no less than fifty-six occurrences of the expression Chi-shih. Other surnames followed by shih are numerous in other northern states. Ku Tung-kao, Ch'un-ch'iu Ta-shih Piao, 89.25a–31b, lists nine surnames used by important kinship groups that are mentioned in the Tso-chttan as holding office in Ch'u. Sun Yao, Ch'un-ch'iu Shih-tai chih Shih-tsu (Shanghai, 1936), pp. 185–188Google Scholar, adds eighteen more surnames of Ch'u individuals mentioned in the Tso-chuan, making a total of twenty-seven. I have checked all twenty-seven with the Ch'un-ch'iu Ching-chuan Yin-te. They occur in the Tso-chuan, followed by the character shih, a total of only fifteen times, as follows (the citations are preceded by letters to facilitate later reference; all are to Legge, Tso-chuan): A, p. 531 (Hsiang 27); B, p. 632 (Chao 11); C, 632 (2nd time, Chao 11); D, p. 642 (Chao 13); E, p. 654 (Chao 14); F, p. 654 (2nd time, Chao 14); G, p. 720 (Chao 27); H, p. 720 (2nd time, (Chao 27); I, p. 720 (3rd time, Chao 27); J, p. 720 (4th time, Chao 27); K, p. 720 (5th time, Chao 27); L, p. 720 (6th time, Chao 27); M, p. 751 (Ting 4); N, p. 758 (Ting 5); O, p. 845 (Ai 16). Since Ch'u sometimes used different terminology from that of the northern states, it might be supposed that in this sense shih was simply replaced in Ch'u by another term denoting a group equivalent to the shih, but in fact this does not seem to be the case. To check this I have examined all occurrences of the nine important surnames listed by Ku Tung-kao; with them I have found only one case of use of such a character. This is tsu which is used together with shih in the passage lettered D.

It is clear, then, that according to the testimony of the Tso-chuan (much our best source for the period) the concept of the shih did not play a great role in Ch'u. Even more significant is the fact that of the nine surnames listed by Ku Tung-kao as those of the kinship groups holding important offices in Ch'u, only a single one is found among this list linked with shih. Thus the political role of the shih in Ch'u was apparently negligible. Furthermore, the earliest of the instances listed above dates from 546 B.C., so that every one of them falls in the final third of the Ch'un-ch'iu period. This certainly permits the hypothesis that the concept and the institution of the shih were taken over by Ch'u from the northern states.

The term shih does occur at an earlier date in connection with the Ch'u kinship group known as the Jo-Ao. But this group was very different from the shih of the northern states. “Jo-Ao” was the name of a ruler of Ch'u who reigned from 790 to 764 B.C. From his sons (aside from his heir who of course continued the ruling line) there descended two kinship groups which used, respectively, the surnames Toubh and Ch'eng. Although they used two surnames, these descendants of Jo-Ao maintained a strong sense of group unity. They composed what was probably the strongest kinship group in Ch'u, until 605 B.C. when they rebelled and were exterminated (except for one man in Ch'u and such fugitives as may have fled the state). Clearly, however, the Jo-Ao group was not a shih in the sense in which this term was used in the northern states. Not only did its members not have a single common surname, but the group apparently had no recognized head (see note 119, below). Significantly the Tso-chuan does not refer to either a Tow shih or a Ch'eng shih. Yet in two passages, under the years 618 and 605 B.C., the Tso-chuan refers to the “Jo-Ao shih”; Legge, , Tso-chuan, pp. 253Google Scholar (Wen 9), 295 (Hsüan 4). Apparently the author of the Tso-chuan used the term shih here analogically, because while this group was not a shih in the northern sense, the northern vocabulary had no term that would precisely fit this Ch'u phenomenon.

¹¹⁹ There are, I think, clear indications that the aristocratic family in Ch'u was, at an early date, organized differently from that of the northern states. We cannot expect a great deal of evidence for this however, for the reason that Ch'u was undergoing rapid acculturation from the north during the Ch'un-ch'iu period. Our information about Ch'u becomes more abundant as communication with the northern states becomes more common, but this same communication tended, pari possu, to wipe out differences. Thus the record inevitably minimizes differences, and magnifies similarities, between Ch'u and the north.

Nevertheless there are significant indications of difference. Legge, , Tso-chaan, p. 228Google Scholar (Wen 1), reports a conversation of 626 B.C. in which the chief minister of Ch'u, discussing the succession to the throne, says: “The constant rule concerning the succession in Ch'u is that a younger son is chosen.” This contrasts sharply with the northern norm according to which the eldest son of the principal wife was supposed to have the chief claim (though it was not always honored). Legge, , Tso-chuan, p. 644Google Scholar (Chao 13), reports a conversation of 529 B.C., a century later, between two officers of the state of Chin in which one of them says, “When there is trouble in the Mi hsing [i.e., in the ruling house of Ch'u] they always establish the youngest scion as ruler; this is the constant rule of Ch'u.” This qualification may merely be the northerner's explanation for this “queer” custom; alternatively it may indicate that in Ch'u, because of acculturation, the practice was beginning to be modified.

Another small indication that the kinship organization was different in Ch'u appears in Kuo-yü 18.2b. King P'ing of Ch'u (reigned 528–516 B.C.) was sacrificed to by one of his sons who was not his heir. In the northern system only the heir (who was also ipso facto the head of the hsing or shih) was supposed to sacrifice to the deceased father, so that this would have been a usurpation of the ruler's function. Yet although the ruler of Ch'u knew of this sacrifice, the text seems clearly to indicate that he did not resent it.

The crucial question regarding the organization of the kinship group in Ch'u is whether it had a head. In the Tso-chuan we find it constantly Indicated, in unmistakable terms, that a certain individual is the head of a certain shih in one of the northern states, but there are few cases of any such clear indication that any individual in Ch'u is the recognized head of a kinship group. And if such leadership over the group did exist in Ch'u it does not appear to have descended, as it normally did in the north, from eldest son to eldest son.

This appears, for instance, from the tale of the complicated viscissitudes of the Jo-Ao kinship group (see note 118, above), which was perhaps the most important in Ch'u. One of its members, Tzu-wen, held the office of chief minister from 664 to 637 B.C. In a northern state, a man holding such an office would normally have been the head of his shih. The younger brother of Tzu-wen, named Tzu-liang, held the office of second importance in the state, that of Minister of War; in a northern state it would have been almost impossible for two brothers to have held the two highest offices unless they had also been brothers of the ruler of the state. After the death of Tzu-wen his son Tzu-yang became chief minister, and Tzu-yüch (son of Tzu-liang) became Minister of War. A plot was formed by another minister, as a result of which Tzu-yang was killed and Tzu-yüch succeeded him as chief minister. Tzu-yüeh thereupon got the members of the Jo-Ao group to rise and kill the minister responsible for the death of Tzu-yang, and they prepared to attack the ruler of Ch'u. In a battle the ruler defeated the Jo-Ao and wiped out the group. While these events were occurring there was also living a son of Tzu-yang, who would seem as a descendant of the line of elder sons to have had (according to the shih system of the north) best claim to be the head of the Jo-Ao group. But he was away on a mission to the state of Ch'i, and apparently had nothing whatever to do with the rebellion. When he heard of it he returned to Ch'u and proclaimed his loyalty to the ruler. He was not put to death with the other members of his group. See: Legge, , Tso-chuan, pp. 117Google Scholar (Chuang 30), 184 (Hsi 23), 295–296 (Hsüan 4). Ch'un-ch'iu Tso-chuan Chu-su 21.23a. Ku, Ch'un-ch'iu Ta-shih Piao, 89.25. All this is rather complicated, but it would seem to show with some clarity that the kinship group in Ch'u did not have the same kind of organization and definite leadership that was the norm for the shih of the northern states.

There is further evidence of this. In the “corporate” shih of the northern states, the head of the shih was regarded as its virtual embodiment. Thus the surname plus shih was not only used to designate the shih but also to designate its head. It is often impossible to determine, for instance, whether Chi-shih means “the Chi shth” or “the head of the Chi shth.” For some cases in which it is fairly clear that a surname plus shih denotes the head of the shih as an individual, see: Legge, , Tso-chuan, pp. 497Google Scholar (Hsiang 23). 498 (Hsiang 23), 707 (Chao 25), 783 (Ting 13), 806 (Ai 6), 827 (Ai 12). In the fifteen passages listed in note 118 above, in which a Ch'u surname is followed by shih, it seems clear that an individual is not designated in those lettered B, C, I, K, and N. An individual may be designated, but this is not certain, in passages D, F, J, M, and O. An individual probably is designated in A, E, G, H, and L. But of all these latter passages, only A concerns a member of one of the nine principal kinship groups of Ch'u listed by Ku Tung-kao. Thus in the entire Tso-chuan the manner of reference often used of shih heads in the northen states occurs only once as designating a member of a Ch'u kinship group of great consequence. This is additional evidence, negative but not negligible, against the existence of the shih as a “corporation” in Ch'u.

¹²⁰ Legge, , Tso-chuan, p. 105Google Scholar (Chuang 23), 107 (Chuang 24), 109 (Chuang 25), 289 (Hsüan 2). Kuo-yü 8.4a.

¹²¹ Legge, , Tso-chuan, pp. 185Google Scholar (Hsi 23), 204–205 (Hsi 28). For knowledge of the Ch'u system of official careers, by Duke Wen of Chin, see note in, above.

¹²² Legge, , Tso-chuan, p. 205 (Hsi 28).Google Scholar

¹²³ Legge, , Tso-chuan, pp. 331Google Scholar (Hsüan 17), 342 (Ch'eng 2), 362 (Ch'eng 7), 365 (Ch'eng 8), 391 (Ch'eng 16), 521–522 (Hsiang 26). Kuo-yü 17.3b–5b Legge, , Tso-chuan, p. 521Google Scholar (Hsiang 26), quotes a statement to the effect that many of the ablest officials of Chin were fugitives from Ch'u.

¹²⁴ It was pointed out in note 115, above, that there is evidence that the hsien as an administrative district may have existed in Ch'u even in the period 740–690 B.C., and that the first clearly datable establishment of a hsien by Ch'u probably took place in 688 B.C. In 598 B.C. we find the ruler of Ch'u saying, “The feudal lords and hsien-kung have all congratulated me”; Legge, , Tso-chuan, p. 309Google Scholar (Hsüan 11). Hsien-kung is the tide of one of the two classes of hsien administrators used by Ch'u. Since the ruler here speaks of “all” of them, this appears to imply that they compose a numerous group. And in 597 B.C. we find reference to the “nine hsien” of Ch'u; Legge, , Tso-chuan, p. 311Google Scholar (Hsüan 12). It is possible, however, that “nine” is used here in the sense of “many,” and that Ch'u had more than nine hsien at this time; see Ku, Ch'un-ch'iu Hsien, p. 172Google Scholar. These and other indications make it quite clear that Ch'u had a system of government in which the hsien played an important role, which was thoroughly established well before the beginning of the sixth century B.C.

For Chin, on the other hand, the earliest reference of any kind to a hsien is from 627 B.C.; Legge, , Tso-chuan, p. 223Google Scholar (Hsi 33). But this merely says that Duke Hsiang gave a hsien as a reward to one of his officers. There is nothing to show that the hsien was at that time an administrative unit in Chin, and the context makes this at least uncertain. References in 594 and 578 B.C. are similarly equivocal; ibid., pp. 327 (Hsüan 15), 380 (Ch'eng 13). But in a conversation of 547 B.C. it is said that Chin plans to give a hsien to a refugee from Ch'u, with the evident meaning that Chin intends to make him one of its officials; ibid., 522 (Hsiang 26). Here, then, in 547 B.C., we have rather definite evidence that the hsien as an administrative unit existed in Chin.

¹²⁵ There was an established hierarchy of office. Probably no official passed through all the stages, but the normal practice was to move from lower to higher office. It was very rare for a man to be appointed chief minister without having served an apprenticeship in at least one lower office. All of the higher offices in Chin were, in title, those of commander or lieutenant commander of one of the various armies, but in fact their functions included civil as well as military administration. This also is a parallel to the Ch'u administration, where the two highest officials performed both civil and military functions.

It is possible to reconstruct individual careers on the basis of the data given in Ku Tung-kao, Ch'un-ch'iu Ta-shih Piao 86Google Scholar, though this sometimes needs to be supplemented. Two examples of careers may be cited. Han Chüeh is mentioned as holding the post of Ssu-ma (at this time an unimportant office in Chin), in 597 and again in 589 B.C. In 588 he is said to have been promoted to be commander of the hsin-chung-chün. In 578 he was commander of the hsia-chün, and by 573 he had been promoted to be chief minister of Chin. For this career see: Legge, , Tso-chuan, pp. 312Google Scholar (Hsüan 12), 339 (Ch'eng 2), 381 (Ch'eng 13). Ch'un-ch'iu Tso-chuan Chu-su 26.5a. Fan Yang appears in 557 as a mere kung-tsu-tai-fu. By 537 he was apparently commander of the hsia-chün. He is said to have become lieutenant commander of the Chung-chün, but the date at which he assumed this office is uncertain. When the commander of the chung-chiin (i.e., the chief minister of Chin) died in 509, Fan Yang apparently succeeded to this office. The difficulty of determining these careers is well illustrated by the case of Fan Yang, for whom the evidence seems reasonably solid but is for the most part inferential. See: Legge, , Tso-chuan, pp. 471Google Scholar (Hsiang 16), 602 (Chao 5), 742 (Ting 1). Ch'un-ch'iu Tso-chuan Chu-su 54.3a, 4a. I am deeply indebted to Professor C. Y. Hsü for his invaluable assistance in working out the careers of these and other Chin officials.

The official career, with an orderly progression from lower to higher office based upon experience snd achievement, was not merely a fact in Chin; it was a recognized concept. See: Legge, , Tso-chuan, pp 350Google Scholar (Ch'eng 3), 437 (Hsiang 11). Kuo-yü 10.17a–18a, 13.3b. The Kuo-yü, 2.14, records 3 conversation in which a Chin officer boasts to an officer of the Chou court that his merit is such that he will surely become chief minister of Chin. The Chou officer replies, “You are certainly worthy, but the stste of Chin in promoting [officers] does not violate the order [of seniority]; I am afraid that [command of] the government will not come to you.” Another Chou officer comments that no less than seven Chin officials have prior seniority over the man in question.

¹²⁶ Shen Pu-hai was a native of Cheng, which adjoined Ch'u. During his lifetime Cheng was conquered by Han, one of the three states into which Chin was divided, of which he became chancellor.

¹²⁷ See for instance de Bary, “Chinese Despotism and the Confucian Ideal: A Seventeenth-century View,” pp. 178–179, 195–197Google Scholar

¹²⁸ Yang, , “Bureaucratic Behavior,” pp. 156–164.Google Scholar

¹²⁹ Barnard, , The Functions of the Executive, p. 120.Google Scholar

¹³⁰ Blau, , Bureaucracy in Modern Society, p. 56.Google Scholar