1. Zuozhuan references in this article are to Bojun, Yang 楊伯峻, Chunqiu Zuozhuan zhu 春秋左傳注, rev. ed. (Beijing: Zhonghua, 1990)Google Scholar, in the form Zuozhuan, Xiang 14.3 (Yang, 1009), which indicates the third section of the Zuozhuan's entry for the fourteenth year of Duke Xiang's 襄公 rule, the relevant passage appearing on p. 1009 of Yang's work. Guoyu references are numbered as in the Shanghai guji edition (Shanghai: Shanghai guji, 1978), which includes Wei Zhao's 韋日召 commentary: Guoyu, Chu 1.7 (550), is the seventh entry in the first juan卷 of the three juan devoted to Chu 楚, and the passage in question appears on p. 550 of that edition.

2. For a demonstration that the Chunqiu and corresponding passages in the Zuozhuan contain accurate records of eclipses, see Hung, William 洪業, “Introduction,” Combined Concordances to Ch'un-ch'iu, Kung-yang, Ku-liang and Tso-chuan, Harvard-Yenching Index no. 11 (1937; reprint, Taipei: Chinese Materials and Research Aids Service Center, 1966), vol. 1, i–vGoogle Scholar.

3. Edward Shaughnessy has advanced convincing arguments for the accuracy and authenticity of the received text of the Zhushu jinian. See his “On the Authenticity of the Bamboo Annals” Harvard Journal of Asiatic Studies 46 (1986), 149–80CrossRefGoogle Scholar.

4. See A Concordance to the Liji, ed. Lau, D.C. and Ching, Chen Fong (Hong Kong: Commercial Press, 1992), “Yuzao” 玉藻, 80Google Scholar; and Hanshu 漢書(Beijing: Zhonghua, 1970), 30.1715 (“Yiwenzhi” 藝文志).Youshi is not attested in either the Zuozhuan or the Guoyu. For references to zuoshi, see Zuozhuan, Xiang 14.3 (Yang, 1009), Zhao 12.11 (Yang, 1340), Ai 17.4 (Yang, 1709); Guoyu, Chu 1.7 (550). Other types of scribe include the Zhou 內史, Guoyu, Zhou 1.12–14 (29–44); and the heroic taishi 大史 of Jin 晉 and Qi 齊 at Zuozhuan, Xuan 2.3 (Yang, 662) and Xiang 25.2 (Yang, 1099) respectively.

5. Karlgren, Bernhard, “On the Authenticity and Nature of the Tso chuan,” Gøteborgs högskolas årsskrift 32, no. 3 (1926), 3–65Google Scholar. William Boltz argues forcefully for derivation, at least of certain sections of the texts, from a shared third written source; see Boltz, , “Notes on the Textual Relation between the Kuo yü and the Tso chuan,” Bulletin of the School of Oriental and African Studies 53 (1990), 491–502CrossRefGoogle Scholar. Comparison of the two texts has suggested to me a fluid, not entirely literary transmission of the two pro to-works in a single milieu before separate transcription and composition; see Schaberg, , “Foundations of Chinese Historiography,” 262–77Google Scholar.

6. For a strong and erudite argument for early dating of the Zuozhuan speeches, see the work of Yuri Pines, including his article in this issue of Early China.

7. The works which introduced this view of historical writing are by now canonical. They include, among others: White, Hayden, Metahistory: The Historical Imagination in Nineteenth-Century Europe (Baltimore: The Johns Hopkins University Press, 1973)Google Scholar; White, Hayden, Tropics of Discourse: Essays in Cultural Criticism (Baltimore: The Johns Hopkins University Press, 1978)Google Scholar, especially “The Historical Text as Literary Artifact”; and Ricoeur, Paul, Narrative and Time, vol. 3 (Chicago: University of Chicago Press, 1988Google Scholar).

8. See the comments of the Siku quanshu 四庫全書 editors in their entry on the Chunqiu Zuozhuan zhengyi 春禾火左傳正義： “Predictions of disaster and blessings recorded in the Zuozhuan are without exception fulfilled; likely they could not avoid post facto joining and matching” (Yunwu, Wang 王雲五, chief ed., Heyin siku quanshu zongmu tiyao ji siku weishou shumu jinhui shumu 合印四庫全書總目題要及四庫未收書目禁燬書目 [Taipei: Shangwu, 1971], 516)Google Scholar. See also the lively remarks of Zhongshu, Qian 錢鍾書 in Guanzhuibian 管錐編（ Beijing: Zhonghua, 1979), 164–66Google Scholar.

9. See, for example, Zuozhuan, Xuan 2.3 (Yang, 662–63) and Xiang 25.2 (Yang, 1099), In the first passage, Dong Hu 董狐 (praised as a “good scribe” by Kongzi 孑匕子) writes the core of one Chunqiu entry for that year, 趙盾拭其君“Zhao Dun killed his ruler.” In the second, unnamed scribes of Qi die in their insistence on recording another assassination, 崔打拭其君 “Cui Zhu killed his ruler”; again the words become the core of the Chunqiu entry on the matter. Note too that mentions of historical writing as a commemoration of misdeeds imply no more than a simple, Chunqiu-style record of names and events (see, for example, Zuozhuan, Wen 15.2 [Yang, 609], and Guoyu, Lu 1.2 [153]).

10. See Guoyu, Lu 1.9 (170), and Lunyu 論語 (Zhuzi jicheng 諸子集成 ed.), 335 (Lunyu 15.6); in both cases, the decision to transcribe a speech is presented as extraordinary.

11. Tadashi, Kamata 鎌田正, Saden no seiritsu to sono tenkai 開 (Tokyo: Taishūkan, 1963), 305–27Google Scholar.

12. Shiji 史記 (Beijing: Zhonghua, 1959), 14.509–10Google Scholar.

13. For some examples of letters, see Zuozhuan, Wen 17.4 (Yang, 625—27), Cheng 7.5 (Yang, 834), Cheng 13.3 (Yang, 861—65), Xiang 24.2 (Yang, 1089–90), Zhao 6.3 (Yang, 1274–77), and Guoyu, Jin 7.3 (438); for a treaty, see Zuozhuan, Xiang 11.3 (Yang, 98990); Zhao 26.9 (Yang, 1475–79) includes a proclamation sent to a number of rulers.

14. For the purposes of this investigation I group the two works together as examples of a single mode of historiography distinct in style and philosophical bent from canonical collections (e.g. Shangshu 尙書), chronicles (Chunqiu and Zhushu jinian), other bodies of anecdotes (e.g. Zhanguoce), and ordered philosophical treatises which include anecdotes (e.g. Lüshi chunqiu 呂氏春秋).

15. I use “Traditionalist” instead of “Confucian” not because the latter would be anachronistic—indeed I take it for granted that this historiography came into being partly among thinkers who considered themselves followers of Kong Qiu 孔丘— but because I wish to emphasize exactly the continuity with the past that the Sage himself prized. “Traditionalist” is also a better rendering of Ru 儒 (see Zuozhuan, Ai 21.2 [Yang, 1718], where the term has a distinctly pejorative connotation).

16. For an example of the epigrammatic answer, see Zuozhuan, Xi 23.6 (Yang, 405), where the Jin 晉 prince Chonger 重耳 (later Duke Wen of Jin 晉文公)says the following to his Di 狄 wife before leaving his Di exile for Qi 齊： “Wait for me for twenty-five years. If I don't come back, then remarry.” She replies: “I am twenty-five now; if after twenty-five more years I remarry, I'll be in my coffin. I'll wait for you.” At Guoyu, Jin 4.23 (386), Duke Wen of Jin asks Guo Yan 郭偃: “In the beginning I thought it was easy to rule the state. Now it's difficult.” He replies: “Difficulty came because you thought it was easy. Now that you think it difficult, it will become easy.” See also Guoyu, Jin 4.3 (344–45), 4.22 (386).

17. E.g. Zuozhuan, Zhao 1.12 (Yang, 1217–23), Zhao 7.7 (Yang, 1289–90); in both passages the great Zheng 鄭 minister Zichan 子產 is consulted about an illness that the Jin duke is suffering from and gives long explanations.

18. E.g. Zuozhuan, Xiang 14.4 (Yang, 1013–14), where the Lu 魯 duke sends an envoy to express his sympathy with Wei 衛 after the Wei ruler has been driven into exile; a Wei minister responds to the formal speech with an equally formal speech.

19. E.g. Zuozhuan Yin 5.1 (Yang, 39), Zhuang 20.1 (Yang, 215–16), and Zhao 6.3 (where Zichan responds to a letter; Yang, 1274–77).

20. E.g. Zuozhuan Min 1.5 (Yang, 257), Wen 1.7 (Yang, 514–15), Xiang 9.1 (Yang, 963–64), Zhao 3.3 (Yang, 1234–37), Zhao 20.6 (Yang, 1415–18), Zhao 20.8 (Yang, 141921), Zhao 26.11 (Yang, 1480–81; the last four all involve Yan Ying 晏嬰 and Duke Jing of Qi 齊景公), Zhao 12.11 (Yang, 1338–41). See also Guoyu, Zhou 1.12 (29–34), Jin 4.8 (352–55), Jin 8.17 (473–76), Chu 1.4 (53441), Chu 2.6 (578–79), Chu 2.9 (583–90). Additional examples may be found in the Qi and Zheng sections as well as the second Yue section of the Guoyu.

21. E.g. Zuozhuan, Xiang 29.13 (Yang, 1161–67), where Gongzi Zha 公子札 of Wu 吳 comments on a performance of traditional dance and music, including what may have been the entire Shijing 詩經 corpus.

22. E.g. Zuozhuan, Wen 13.2 (Yang, 594–95), Xiang 27.4 (Yang, 1131–32), Zhao 1.1 (Yang, 1202–4); Guoyu, Jin 3.8 (333–35), Jin 5.4 (397–99), Jin 6.1 (409–13), Jin 8.5 (455–58).

23. E.g. Zuozhuan, Xi 23.6 (Yang, 410), Xiang 4.3 (Yang, 932), Xiang 8.8 (Yang, 959–60), Xiang 27.5 (Yang, 1134–35), Zhao 16.3 (Yang, 1380–81); Guoyu, Jin 4.10 (360–61).

24. The most impressive example is Zuozhuan, Xuan 12.2 (Yang, 721–28), where Jin ministers confer in a series of long speeches before the battle of Bi 郝.

25. At the end of Zuozhuan, Xuan 12.2 (Yang, 744–47), just after his victory over Jin at Bi, King Zhuang of Chu 楚莊王 delivers the most eloquent and canonically informed speech attributed to any state ruler in the Zuozhuan or Guoyu. In an example of specious speech, Li Ji 驪姬, the treacherous wife of Duke Xian of Jin 晉獻公, has her henchmen convince the duke to lodge his elder sons in border cities (Zuozhuan, Zhuang 28.2 [Yang, 239–41], Guoyu, Jin 1.6 [270–71]).

26. Relatively little work has been published on literary aspects of the Zuozhuan and Guoyu, and almost nothing on the rhetoric of speeches. See, however, Liu Zhiji's 劉矢口幾 Shitong 史通, especially his remarks on speeches in the essay “Yanyu” 言語 (Shitong tongshi 史通通釋, ed. Qilong, Pu 浦起龍 [Shanghai: Shanghai shudian, 1988], 2.1–4Google Scholar); Heidbüchel, Ursula, “Rhetorik im Antiken China: Eine Untersuchung der Ausdrucksformen Höfischer Rede im Zuo zhuan, Herzog Zhao” (Ph.D. diss., Wilhems-Universität Westfalia, 1993Google Scholar; I thank Christoph Harbsmeier for this reference); Egan, Ronald C., “Narratives in Tso chuan” Harvard journal of Asiatic Studies 37 (1977), 323–52CrossRefGoogle Scholar; Egan, Ronald C., “Selections from Tso-chuan: Translation and Analysis” (Ph.D. diss., Harvard University, 1976)Google Scholar; Wang, John C.Y., “Early Chinese Narrative: The Tso-chuan as Example,” in Chinese Narrative: Critical and Theoretical Essays, ed. Plaks, Andrew (Princeton: Princeton University Press, 1977), 3–20Google Scholar; Gaoping, Zhang 張高評, Zuozhuan zhi wenxue jiazhi 左傳之文學價値 (Taibei: Wenshizhe, 1982)Google Scholar; Gaoping, Zhang, Zuozhuan wenzhang yifa tanwei 左傳文章義撢微 (Taibei: Wenshizhe, 1982)Google Scholar. For rhetoric in the Zhanguoce, see Crump, J.I. Jr., Intrigues: Studies of the Chan-kuo Ts'e (Ann Arbor: The University of Michigan Press, 1964), 47–75Google Scholar.

27. Graham, A.C., Disputers of the Tao: Philosophical Argument in Ancient China (La Salle, Illinois: Open Court, 1989), 24Google Scholar, acknowledges the importance of these chains of inference or “sorites” in the rhetoric and reasoning of pre-Qin philosophers.

28. For one extraordinary example of the rhetoric of weaving, see Guoyu, Zhou 3.2 (94–101), analyzed at length in Schaberg, , “Foundations of Chinese Historiography,” 284–311Google Scholar; there is a dear connection there between the patterns of rhetorical structure and the connotations of wen 文, “cultured patterning” and “literary activity.”

29. This rendering of jian 諫, used by both Legge and Couvreur (remontrance) in their translations of the Zuozhuan, has been standard at least since the Reverend Morrison compiled A Dictionary of the Chinese Language (1815; reprint, Shanghai: London Mission Press, 1865)Google Scholar.

30. Applied to early Chinese texts, the traditional tripartite division of Greco-Roman oratory collapses, largely because occasions of speech in the Chinese works are almost exclusively deliberative. The forensic, to the extent that it exists, takes place in the same institutional setting as the deliberative, while those who deliver the display pieces of epideictic oratory almost always do so on the pretext of some deliberative occasion. One brilliant exception is Cai Mo's 蔡墨 discourse on dragons at Zuozhuan, Zhao 29.4 (Yang, 1500–1504), where the only occasion for the speech is the appearance of a dragon; compare Guoyu, Lu 1.9 (1657־]), where the appearance of a strange seabird brings a policy decision which prompts the learned and critical speech.

31. It should be noted here that the legacy of these idealized visions also survives outside of historiography, and in some way reflects Warring States practices. The Mozi 墨子, for instance, refers to the activities of the travelling philosopher/persuader as “reciting the way of the former kings” 誦先王之道 (Mozi xiangu 墨子閒話[Zhuzi jicheng ed.]f 13.287). The word song 誦 “recite” will be prominent in all four of the passages translated below. See below for Han echoes of the language circulation schemes.

32. Jiaxi, Yu 余嘉錫, “Xiaoshuojia chuyu baiguan shuo” 小說家出於稗官說, in ]Yu jiaxi lunxue zazhu 余嘉錫論學雜著 (Beijing: Zhonghua, 1963), 265–79Google Scholar, discusses this passage and several related texts as evidence for the origins of fiction in China.

33. The distinctions among helpers here are somewhat unclear; see Yang Bojun's note on Zuozhuan, Huan 2.8 (Yang, 94), a passage comparable to this one. As for the difference between higher ministers' and lower ministers' helpers, schematization sometimes necessitates divisions which do not necessarily correspond to actual distinctions; we will see other cases below.

34. Lothar von Falkenhausen, noting the reappearance of gong 工 later in the passage, has suggested that the text here might be emended to 公誦蔵諫, gong 公 being the higher ministers in the royal court (personal communication, September 23, 1996). Karlgren, Bernhard, Grammata Serica Recensa (Stockholm: Museum of Far Eastern Antiquities, 1972), 1172a, 1173aGoogle Scholar, gives identical archaic reconstructions (*kung) for the two characters. Both readings are supported by later text parallels: Da Dai liji, 工誦箴諫 (A Concordance to the Dadai liji, ed. Lau, D.C. and Ching, Chen Fong [Hong Kong: Commercial Press, 1992], 20Google Scholar; Huainanzi, 天子聽朝, 〔使〕公卿正諫 (A Concordance to the Huainanzi, ed. Lau, D.C. and Ching, Chen Fong [Hong Kong: Commercial Press, 1992], 80)Google Scholar; and Lüshi chunqiu, 天子聽政, 使公卿列士正諫 (A Concordance to the Lü shi chunqiu, ed. Lau, D.C. and Ching, Chen Fong [Hong Kong: Commercial Press, 1994], 133Google Scholar).

35. Yang, 1017, argues that the verb for the clause should be supplied from the preceding text.

36. The quoted material is now found in the “Yinzheng” 撒征 section of the Shangshu (Jiegang, Gu 顧額岡1), Shangshu tongjian 尙書通檢 [Beijing: Shumu wenxian, 1982], 5Google Scholar); but it would appear that the late forger's source for these lines was this Zuozhuan passage.

Zhang Binglin 章炳麟 relates duo 釋 “bell” or “clapper” to the word yi 譯 “transmit” and adduces other clues about the supposed early practice of language-collection. See Binglin, Zhang, Zhang Taiyan quanji 章太炎全集, ed. Youwei, Wang 王有爲 (Shanghai: Shanghai renmin, 1986), 497–98Google Scholar.

37. Shi Kuang concludes by remarking that Heaven cherishes the people too much to protect the disgraced ruler of Wei. This passage is clearly related to, if not the source of, the ] ater legends concerning poetry-collecting. See Liji, “Wangzhi” 王制 (A Concordance to the Liji, 32); Hanshu, 24.1123 (“Shihuozhi” 食貨志), 30.1708 (“Yiwenzhi”). See also Guoyu, Jin 6.1, discussed below. The first of the two Hanshu passages is the earliest explicit mention of caishi 采詩 in received texts; the author paraphrases a part of Shi Kuang's Xiashu citation and adds the words yi caishi 以采詩, “in order to collect poems.” The passage from the “Yiwenzhi,” though based on echoes of the Liji, “Wangzhi,” adds a claim that the ancient government included an officer in charge of poetry-collection (caishi zhi guan 采詩之官).

38. See Zuozhuan, Xuan 2.3 (Yang, 662–63) and Xiang 25.2 (Yang, 1099), cited in n. 9 above. Later commonplaces about the critical function of literary composition occur in Mengzi 孟子 (Zhuzi jicheng ed.), 6.271 (Mengzi 3B.9; Confucius's work on the Chun-qiu), in all three commentaries on the Chunqiuf and in the Mao 毛 tradition of Shijing interpretation, with its emphasis on historically situated occasions of ci 刺 “criticism.”

39. The normal term for the new composition of a text (such as a poem) is zuo 作 rather than wei.

40. Song 誦 is the recitation either of known poems (Zuozhuan, Xiang 14.4 [Yang, 1011], Xiang 28.9 [Yang, 1149]) or of new compositions (Zuozhuan, Xi 28.3 [Yang, 458], Xiang 30.13 [Yang, 1182]) with the intention of praising or blaming the listener.

41. See Shaughnessy, Edward, Sources of Western Zhou History (Berkeley: University of California Press, 1991) 272–86Google Scholar, for the issues surrounding the dating of King Li's reign.

42. Many speeches and narratives demonstrate the danger of suppressing words. See, for example, Guoyu, Jin 1.5 (266–67), where a ruler who lacks remonstrating ministers is marked for ruin, and Chu 2.6 (578–79), where it is said that the doomed King Fuchai of Wu 吳王夫差“gives free rein to transgressions and suppresses remonstrances” 縱過而醫諫.Since the free flow of critical words to and from the center is often related to the economic arrangements by which property circulates and tribute reaches the court, one might consider Yu's 禹 sagely efforts at flood control a management both of watercourses and of other means of communication.

43. In these idealized passages there are several distinct types of vision impairment among the sage's officials. The blind musicians above (gu 瞽)are characterized by their activity. The “blind” (sou 瞍)and the “sightless” (meng 矇) are distinguished, according to Wei Zhao, by the cause of their impairment: the former lack pupils entirely, while the latter have eyes but cannot see.

44. My figure of speech is supported by the idiom which the Guoyu here puts to clever use: 三年, 乃流王於慮 “After three years, they expelled the king to Zhi.”

45. The word dou 兜 does not yield sense here. Wang Yinzhi 王弓∣之 understood it as a mistake for gu “to hide, to cover over”; Zhang Yiren 張以仁 supports this view (Guoyu jiaozheng 國語酙證 [Taibei: Taiwan Shangwu, 1969], 254–55)Google Scholar. Chen Zhuan 陳綠, writing in the first half of the nineteenth century, suggested that dou is a loan for dou 兜, defined in Shuowen jiezi 說文解字 as “speaking many words'’ (duoyan 多言)； see Guoyu yijie 國語翼解 in Guangya congshu 廣雅叢書 (ed. Shaoqi, Xu 徐紹綮 [Guangzhou: Guangya, 1920], vol. 406, 5.3b–4aGoogle Scholar. Dou 兜 does not appear in Schuessler's, AxelA Dictionary of Early Zhou Chinese (Honolulu: University of Hawaii Press, 1987)Google Scholar, and though it is listed among the characters of Grammata Serica Recensa (117a), Karlgren unaccountably gives no gloss for it.

46. Another peculiar phrase. The word feng 風 resonates with the whole legend of poetry-gathering, which is without question related to the issue we are discussing here; but in glossing the word as cai 采 “to gather,” Wei Zhao stretches the possibilities of grammar and usage. Chen Zhuan, Guoyu yijie, 5.4a, argues that 臁 is a loan for 旅, meaning shanglü 商旅 “travelling merchants” (see Karglren, , Grammata Serica Recensa, 69r, 77aGoogle Scholar).

47. Yixiang is a zuoshi 左史 “scribe of the left” (see n. 4 above), and thus is perhaps uniquely qualified to represent both the past and the practices of representing the past.

48. For the translation of the term zhi 志 as “normative content” rather than the customary “ambition” or “intent,” see Schaberg, , “Foundations of Chinese Historiography,” 184–87Google Scholar. It is important to recall that zhi, in addition to denoting the central message or lesson of a text, is a name for certain kinds of historical texts, which seem (on the basis of passages cited from them) to have been written or unwritten collections of aphorisms well adapted for citation. See the following passages for citations: 志, Zuozhuan, Xiang 4.4 (Yang, 935), 25.10 (Yang, 1106), Zhao 1.12 (Yang, 1220), Zhao 3.8 (Yang, 1242), Ai 18.2 (Yang, 1713);軍志, Zuozhuan, Xi 28.3 (Yang, 456), Xuan 12.2 (Yang, 739); 前志, Zuozhuan, Wen 6.8 (Yang, 552–53), Cheng 15.1 (Yang, 873);周志, Zuozhuan, Wen 2.1 (Yang, 520);仲虺之志, Zuozhuan, Xiang 30.10 (Yang, 1175; Zhong Hui 仲脑 was a minister to Cheng Tang 成湯, founder of the Shang–his words [yan 言] are cited at Zuozhuan, Xuan 12.2 [Yang, 725] and Xiang 14.9 [Yang, 1019]); 禮志, Guoyu, Jin 4.9 (358). For the illuminating use of zhi at Guoyu, Wu 1.3, see the discussion below.

49. Wei Zhao identifies this “Yi” 懿 admonition with Shijing, Mao 256 (“Yi” 抑), which the Mao preface attributes to Duke Wu of Wei. Reconstructions of Archaic Chinese offer Wei Zhao some support; 抑 is , while 穀 is (Karlgren, , Grammata Serica Recensa, 915a, 395d)Google Scholar.

50. The origins of xun and the mimetic associations of these teachings are exemplified in the “Gu ming” 顧命 chapter of the Shangshu: “We succeeded to and preserve the great teachings of Kings Wen and Wu” 嗣守文武大訓 and “Extol in response the luminous teachings of Kings Wen and Wu” 用答揚文武之光訓 (Jiegang, Gu, Shangshu tongjian, 21–22Google Scholar). The latter phrase of course recalls the mimetic formulas of the bronze inscriptions.

51. For examples of songs as expressions of popular sentiment, see Zuozhuan, Xi 5.8 (Yang, 310), Xi 28.3 (Yang, 458), Xuan 2.1 (Yang, 654), Xiang 4.8 (Yang, 940), Xiang 17.6 (Yang, 1032), Xiang 30.13 (Yang, 1182), Zhao 25.4 (Yang, 1460), Ding 14.8 (Yang, 1597), Ai 5.3 (Yang, 1631).

52. See the Liji and Hanshu passages cited above, n. 37.

53. See Zuozhuan, Xiang 14.6 (Yang, 1017).

54. See Guoyu, Zhou 1.3 (10), Chu 1.7 (551). The latter passage also contains the compound 箴儆 “remonstrate and admonish.”

55. Zuozhuan, Zhao 26.11 (Yang, 1480). Yanzi 晏子, speaking here of the relation between filial piety and criticism, foreshadows the Han interest in remonstrance within the family (see discussion below).

56. Zuozhuan, Xuan 12.2 (Yang, 731), where Luan Shu 欒書 of Jin describes how King Zhuang of Chu 楚莊王 has instructed his people in the lessons of their humble past and enjoined them (蔵之), saying that the people's livelihood depends on their willingness to work hard.

57. See Zuozhuan, Xuan 4.3 (Yang, 683), Ding 4.3 (Yang, 1545), Ai 16.5 (Yang, 1704). In Zuozhuan, Ding 4.3 (Yang, 1545), the character is 緘 rather than 鐵, but the named individual is the same as in Ai 16.5, and there is no doubt that the same office is intended. See also Zuozhuan, Zhao 4.7 (Yang, 1255), where different editions have 蔵尹 or 咸尹 for 箴尹.

58. Zuozhuan, Xiang 4.7 (Yang, 936–39).

59. See discussion of de 德 below.

60. A rare instance in Chinese of interrupted speech, and apparently not the result of textual mutilation. See Zhongshu, Qian, Guanzhuibian, 211–13Google Scholar. Here one function of the interruption is to change Wei Jiang's intended citation into a paraphrase; from the Lessons of Xia we get no more than those four characters.

61. Yang, 938, cites the Shiji ji jie 集解 commentary on the “Zhou benji” 周本紀, where Pei Yin 裴駟 quotes a passage from Liu Xiang's 劉向 Bielu 別錄 identifying Xin Jia as a minister of the last Yin king Zhou 糸寸. He remonstrated seventy-five times with his ruler before switching his loyalties to King Wen of Zhou 周文王.

62. Here I follow the interpretation of Takezoe Shin'ichir¯ 竹添進一郎, Saden kaisen 左傳會箋 (1911; reprint, Taibei: Mingda, 1986), 1010.

63. Note that in the game warden's text, wu 武 must refer to the hunting expeditions which were conducted with the instruments of, and as exercises for, war.

64. The significance attributed by the Zuozhuan authors to this speech and the resulting policy is apparent from Zuozhuan, Xiang 11.5 (Yang, 993), where Duke Dao credits Jin's lasting international success to the Rong treaty here advocated by Wei Jiang.

65. According to traditional attributions, “Changdi” was composed either by the Duke of Zhou 周公 (Guoyu, Zhou 2.1 [45]) or by Duke Mu of Shao 召穆公 when he saw that the Zhou house was going into decline (Zuozhuan, Xi 24.2 [Yang, 423]). The Mao preface does not specify an author, but dates the poem to the time of the early Zhou rebellions, supporting the Guoyu attribution; see Shijing (zhushu, Shisanjing 卜三經注疏, ed. Yuan, Ruan 阮元 [1815; reprint, Taibei: Dahua, 1982]), 407Google Scholar. Like all but a few Shijing poems, including the famous “Chixiao” 鶴號 (Shijing, Mao 155) attributed to the Duke of Zhou in Shangshu 尙書,“Jinteng” 金縢, this piece shows no clear signs of individual authorship.

66. For examples see Zuozhuan, Xi 24.2 (Yang, 423–24), Xuan 2.3 (Yang, 657); Guoyu, Zhou 1.1 (1), Zhou 2.1 (45, 51), Zhou 3.3 (109–10) Jin 4.7 (349–50), Jin 9.20 (502–3), Chu 1.8 (556).

67. For examples see Zuozhuan, Yin 3.7 (Yang, 31), Zhuang 24.1 (Yang, 229), Xi 23.6 (Yang, 408), Xi 24.2 (Yang, 420), Wen 7.7 (Yang, 563), Xiang 23.4 (Yang, 1077); Guoyu, Jin 1.9 (279), Jin 4.7 (349), Yue 1.1 (633), Yue 2.6 (652).

68. For examples see Zuozhuan, Huan 2.2 (Yang, 89), Min 2.7 (Yang, 272), Xi 24.2 (Yang, 420–24), Xuan 12.5 (Yang, 748), Zhao 30.2 (Yang, 1508); Guoyu, Zhou 1.1 (2–3), Zhou 2.1 (48), Zhou 3.3 (103–9), Jin 4.7 (350), Jin 9.20 (502–3), Chu 1.8 (554–56), Wu 1.3 (597–99), Yue 2.7 (656).

69. Many remonstrances appeal to the example of the past without recounting historical particulars; see Zuozhuan, Yin 5.1 (Yang, 43), Min 2.7 (Yang, 268; both refer to “the system of times past” 古之制); Guoyu, Zhou 1.6 (15–22), Zhou 1.9 (24–26), Zhou 3.3 (101–3), Lu 1.2 (153), Yue 1.1 (634). References to precedent {gu 故） involve a simitar use of the past, as do discussions of li 禮 “ritual,” a concept of immense importance for the construction of historiographical speech and narrative (Zuozhuan, Zhuang 23.1 [Yang, 225–26] shows how li can be used in remonstrance without other sorts of citation). For an example of gu in remonstrance, see Guoyu, Jin 1.9 (279); and see Guoyu, Lu 1.4 (156), where a duke claims falsely that it is his prerogative to make precedent. Note that gu 故 “precedent” is probably related at its origins to gu 古 “the past.”

70. Zuozhuan, Xi 5.8 (Yang, 307–10).

71. Jin first attacked Guo across Yu in Zuozhuan, Xi 2.2 (Yang, 281–83). There Duke Xian of Jin was at first unwilling to ask Yu for permission, but one of his ministers argued that this action would make Yu a dependent and that Gong zhi Qi, though wise, would not be able to remonstrate forcefully enough to save Yu.

72. The zhi 之 in Gong zhi Qi's name functions as it does in such names as Jie zhi Tui 介之推 (Zuozhuan, 24.1 [Yang, 417–18]) and Meng zhi Fan 孟之反 (Lunyu, 123 [Lunyu 6.15]). See comments of Yang Bojun and Liu Baonan 劉寶楠 on the passages cited.

73. The adage on lips and teeth is also found at Zuozhuan, Ai 8.2 (Yang, 1647–48) and in a number of other pre-Qin works.

74. Yu Zhong founded Yu. The better-known version of the Tai Bo legend includes no failure of compliance; he is said to have yielded to his younger brother Wang Ji and retreated to the barbarian territory which would later become the state of Wu 吳 (see Zuozhuan, Ai 7.3 [Yang, 1641]; Shiji, 31.1445).

75. Guo Zhong and Guo Shu were founders, respectively, of eastern and western Guo; see Yang, 308. To summarize this family tree: The Grand King (Tai Wang) begat Tai Bo (who went to Wu), Yu Zhong (who founded Yu), and Wang Ji, who succeeded his father. Wang Ji in turn begat King Wen (who succeeded him) and Guo Zhong and Guo Shu (founders of Guo). Note that Guo Zhong and Guo Shu are thus nephews to Yu Zhong and brothers of King Wen. Jin was founded by Tang Shu 唐叔, a son of King Wu 周武王.

76. This “treasury of covenants” (meng fu 盟府) seems to have been a place for the storage of certain writings. Records of services rendered to the throne would have been found in the investiture texts which court officials sometimes present to casters in the bronze inscriptions. See Yang, 308; Takezoe, , Saden, 379Google Scholar; and Lewis, Mark Edward, Sanctioned Violence in Early China (Albany: State University of New York Press, 1990), 43–50Google Scholar.

77. Huan Shu was father of Zhuang Bo; both ruled in the Jin city of Quwo 曲沃. Their line ruled Jin starting in Zuozhuan, Yin 5.2 (Yang, 44; 718 B.C.E.); Duke Xian, himself a member of this line, exterminated the other branches in Zuozhuan, Zhuang 23–25 (Yang, 226–33; 671–669 B.C.E.). See Yang, 309.

78. For these difficult lines I follow Yang, 309–10, but see also Karlgren, Bernhard, “Glosses on the Tso chuan,” Bulletin of the Museum of Far Eastern Antiquities 41 (1969), 26Google Scholar: “The people are not careless about the wu quality (in their gifts), but only the virtue––that it (Heaven) (considers as quality:) estimates.”

79. The idea is implicit in Zuozhuan, Zhao 5.4 (Yang, 1267–69; a Chu minister's argument against the mutilation of ]in envoys) and Ai 1.2 (Yang, 1605–6; Wu Zixu's 子胥 speech on the necessity of destroying Yue 越).

80. For an account of some of the cosmic, political, moral, and rhetorical theories which speakers in historiography are made to advance, see Schaberg, , “Foundations of Chinese Historiography,” 327–465Google Scholar.

81. None of the three quotations appear in that portion of our Shangshu which is called the Zhoushu (or in the Yi Zhoushu 逸周書). As noted above, citation is no guarantee that either speakers or historiographers knew these texts through reading, and it is entirely possible that the the only written versions which ever existed are the fragments found in historiographical speeches.

82. For more on the concept of de, see Boodberg, Peter A., “The Semasiology of Some Primary Confucian Concepts,” in Selected Works of Peter A. Boodberg, comp. Cohen, Alvin P. (Berkeley: University of California Press, 1979), 26–40Google Scholar; and Kryukov, Vassili, “Symbols of Power and Communication in Pre-Confucian China (on the Anthropology of De): Preliminary Assumptions,” Bulletin of the School of Oriental and African Studies 58 (1995), 314–33CrossRefGoogle Scholar.

83. See Zuozhuan, Xiang 29.11 (Yang, 1160), for more on the rise of large states at the expense of small states.

84. Prince Chonger of Jin is the best example of this process, which underlies the construction of narratives of all sorts. During the period of exile which precedes Chonger's triumphant return to Jin as duke (posthumously named Wen), and then again during the early years of his rule, he accumulates a de which makes Jin militarily invincible (see the Chu king's speech and Junzhi citation in Zuozhuan, Xi 28.3 [Yang, 456–57]). Working backward from known outcomes, the narrators construct stories that show the power of virtue.

85. Zuozhuan, Yin 6.4 (Yang, 50), Cheng 2.3 (Yang, 799; see below), Cheng 6.11 (Yang, 830), Zhao 21.6 (Yang, 1430), Zhao 22.1 (Yang, 1432), Ding 9.3 (Yang, 1573). Remonstrances which do not cite inherited speech and which have nothing to do with military matters are found at Zuozhuan, Huan 18.3 (Yang, 154), Xuan 9.6 (Yang, 702), Zhao 28.4 (Yang, 1496–97), Ding 10.5 (Yang, 1580), Ai 20.2 (Yang, 1715). All of these are discussed below except Xuan 9.6, which is a borderline case: its appeal to public appearances is couched in language reminiscent of Shijing, Mao 235 (“Wen Wang” 文王).

The Guoyu includes seven remonstrances which do not cite inherited speech. Only two of them (Guoyu, Jin 1.7 [271–72], Jin 3.6 [327]) bear on military matters; in others one may trace principles either related to venerable ancient practices (Guoyu, Zhou 1.7 [22], Lu 1.16 [183]) or indirectly related to military action (Guoyu, Wu 1.2 [595]). Guoyu, Yue 2.1 (641) and Jin 9.6 (488–89; corresponds to Zuozhuan, Zhao 28.4), are discussed below.

86. Zuozhuan, Cheng 2.3 (Yang, 799).

87. According to Zuozhuan, Xuan 17.1 (Yang, 771–72), when Xi Ke once visited Qi on a diplomatic mission, the Qi ruler had his mother hide behind a curtain in court so that she could observe Xi Ke's limp. Xi Ke heard her laughing, was infuriated, and vowed to undertake the attack which has here resulted in Qi's defeat.

88. Zuozhuan, Cheng 6.11 (Yang, 830).

89. Zuozhuan, Zhao 22.1 (Yang, 1432).

90. Zuozhuan, Zhao 23.4 (Yang, 1444–45).

91. Guoyu, Jin 3.6 (327).

92. According to Wei Zhao (327), the Jin ruler's self-destruction is inevitable, and Duke Mu can simply await it.

93. Guoyu, Jin 1.7 (271–72). The remonstrance does not appear in the Zuozhuan, but see Shi Wei's speech at Zuozhuan, Min 1.6 (Yang, 258–59).

94. For a similar remonstrance, with a similar appeal to military and administrative tradition (fei gu ye 非故也, “It is not according to precedent,” says the remonstrator), see Guoyu, Jin 1.9 (279). As it happens, Shensheng is victorious in all of his campaigns against Jin's barbarian neighbors. But critics of the duke's policies are right in predicting doom for Shensheng, who finally commits suicide (Zuozhuan, Xi 4.6 [Yang, 399]; Guoyu, Jin 2.1 [292]).

95. Zuozhuan, Ding 9.3 (Yang, 1573). It is a commonplace that de in a state or ruler translates into military power. See, for instance, the line King Cheng of Chu 楚成 cites from the Junzhi 軍志 as he justifies his reluctance to attack Jin under Duke Wen: “The one who has de cannot be rivaled” 有德者不可敵 (Zuozhuan, Xi 28.3 [Yang, 456]).

96. Zuozhuan, Min 2.7 (Yang, 272). Compare the account in Guoyu, Jin 1.7 (271–72), mentioned above, where the state of Huo, not a Di tribe, is the object of Shensheng's campaign.

97. It will be better for Shensheng to avoid battle, both so that he can show his fiiiality by saving himself from possible injury and so that he can keep the soldiery safe. By fighting he will put himself in danger and aggravate his father's and stepmother's animus. Hu Tu does not mention the problem of disobeying a father's orders, presumably because it was unjust for the duke to make his son a general in the field in the first place.

98. See Chunqiu Zuozhuan zhengyi 舂秋左傳正義 (zhushu, Shisanjing ed.), 1789Google Scholar.

99. Zuozhuan, Huan 18.3 (Yang, 154).

100. Ziyi was Wangzi Ke 王子克, younger brother of King Zhuang 莊王, successor to their father King Huan 桓王.

101. Historiography never records the same material twice in the same way, and the discrepancies in the account of Xin Bo's speech are typical. There is always some variation in the account of an incident or a speech. For instance, Guoyu, Jin 1.9 (281), records Hu Tu's remonstrance, including several identical phrases and similar ideas, but omits the citation of Xin Bo's words. More generally, Zuozhuan and Guoyu taken together include scores of speeches remembered with slight differences. One could add the numerous Warring States and Han anecdote collections, which include more variants of some of the same speeches.

102. Guoyu, Wu 1.3 (598). The rise and fall of King Ling are narrated in exquisite detail in Zuozhuan, Xiang 26-Zhao 13 (Yang, 1114–353) culminating in his suicide at Zhao 13.2 (Yang, 1347); cf. Guoyu, Chu 1.5–1.8 (541–57). See discussion in Schaberg, , “Foundations of Chinese Historiography,” 471–521Google Scholar.

103. Here I follow Wei Zhao's interpretation of these lines. The Zuozhuan anecdotes on theZhanghua Pavilion (Zuozhuan, Zhao 7.2 [Yang, 1283–85], Zhao 7.3 [Yang, 1285–87], Zhao 7.6 [Yang, 1289]) do not include these details.

104. Chu annexes Chen in Zuozhuan, Zhao 8.6 (Yang, 1304–5), and Cai in Zuozhuan, Zhao 11.8 (Yang, 1327).

105. See Zuozhuan, Zhao 11.10 (Yang, 1327–29), in which this neglect is dramatized.

106. Wei Zhao refers the reader to Zuozhuan, Zhao 6.9 (Yang, 1279–80), but see also Zhao 5.8 (Yang, 1270–72).

107. Zuozhuan, Zhao 13.2 (Yang, 1346).

108. See the discussion above, n. 48, of the term zhi and its connection with aphorism collections.

109. See, for example, the comment attributed to Confucius at the very end of Zuozhuan, Zhao 13.2 (Yang, 1341).

110. See Zuozhuan, Zhao 3.3 (Yang, 1234–37), Zhao 20.6 (Yang, 1415–18), Zhao 20.8 (Yang, 1419–21), Zhao 26.11 (Yang, 1480–81).

111. 殷鑒不遠, 在夏后之世 “The mirror for Yin was not far; it was in the age of Xia”; Shijing, Mao 255 (zhushu, Shisanjing ed.), 554Google Scholar.

112. See especially the “Waichushuo” 外儲說 sections; Qiyou, Chen 陳奇猷, Han Feizi jishi 韓非子集釋 (Beijing: Zhonghua, 1974), 611–790Google Scholar.

113. Note that the language of Xin Bo's remonstrance (in Hu Tu's version), if not the narrative frame, is closely echoed by Han Fei; see Chen Qiyou, Han Feizijishi, 932, cited by Yang, 154, who also notes a more remote parallel in the Guanzi 管.

114. Zuozhuan, Ding 10.5 (Yang, 1580).

115. Zuozhuan, Ai 20.2 (Yang, 1715).

116. Duke Ling of Jin implies as much when he tries in vain to forestall an imminent remonstrance by saying, “I know how I have erred, and I will change it” 過矣, 將改之 (Zuozhuan, Xuan 2.3 [Yang, 757]). He is lying about his intention to change, and is dead within the year.

117. See Shuoyuan jiaozheng 說苑校證, ed. Zonglu, Xiang 向宗魯 (Beijing: Zhonghua, 1987), 206Google Scholar; Bohutong shuzheng 白虎通疏證, ed. Zeyu, Wu 吳則虞 (Beijing: Zhonghua, 1994), 236Google Scholar; and A Concordance to the Kongzi Jiayu, ed. Lau, D.C. and Ching, Chen Fong (Hong Kong: Commercial Press, 1992), 25Google Scholar.

118. Zuozhuan, Zhao 28.4 (Yang, 1496–97). See also Guoyu, Jin 9.6 (488–89), which is the same anecdote told (as is often the case) less economically, and with several variants in wording.

119. Probably the son of Wei Shu. In Zuozhuan, Zhao 28.3 (Yang, 1494), he was put in charge of the city of Gengyang in Jin; failing to rule on a lawsuit there, he appealed to a higher court, i.e. the head minister himself.

120. The two are apparently household retainers of the Wei family.

121. It is unclear why drinking precluded eating a complete dinner. Takezoe, , Saden, 1764Google Scholar, suggests that they were (or are claiming to have been) too drunk to eat.

122. Wei Shu is the head of Jin's center army.

123. If one does not insist on the literal truth of all Zuozhuan anecdotes, one may even believe that the adage prompts the sighing rather than vice-versa: explaining Wei Zi's change of heart, the narrators designed an anecdote in which he noticed a divergence from the norm of eating (as expressed in the adage) and was thus taught a lesson.

124. One set of narratives in which the Traditionalist values seem to fail is the grand tale of late Spring and Autumn period usurpation and ducal decline. I have argued (“Foundations of Chinese Historiography,” 622–33, 730–61) that the narrators here shift their work of narrative vindication to a new level, justifying the rise of “Confucianism” or Traditionalism as a school of thought rather than a body of governing practices.

125. For some examples of Gong zhi Qi's remonstrance, see Zhanguoce (A Concordance to the Zhanguoce, ed. Lau, D.C. and Ching, Chen Fong [Hong Kong: Commercial Press, 1992], 162Google Scholar), Xinxu (A Concordance to the Xinxu, ed. Lau, D.C. and Ching, Chen Fong [Hong Kong: Commercial Press, 1992], 47)Google Scholar, Huainanzi (A Concordance to the Huainanzi, 189), Lüshichunqiu (A Concordance to the Lüshichunqiu, 82), and Chunqiufanlu (A Concordance to the Chunqiufanlu, ed. Lau, D.C. and Ching, Chen Fong [Hong Kong: Commercial Press, 1994], 18Google Scholar).

126. See Hucker, Charles O., A Dictionary of Official Titles in Imperial China (Stanford: Stanford University Press, 1985), nos. 787, 831, 836, 865, 5582, etcGoogle Scholar.

127. Most interesting is the one found in Lüshichunqiu (A Concordance to the Lüshichunqiu, 133), in which the tale of the remonstrance before King Li of Zhou (Guoyu, Zhou 1.3 [9–12]; discussed above) is retold (note the thorough rewriting of the Guoyu's depiction of language circulation; and compare Huainanzi [A Concordance to the Huainanzi, 80]). See also the schema which includes a drum that the minister beats when he will presume to remonstrate 傲諫之鼓) in Da Dai liji (A Concordance to the Dadai liji, 20), in Lüshichunqiu (A Concordance to the Lüshichunqiu, 156), and in Jiayi Xinshu (A Concordance to the JiayiXinshu, ed. Lau, D.C. and Ching, Chen Fong [Hong Kong: Commercial Press, 1994], 36)Google Scholar. The Zhouli refers to office of the Remonstrator (sijian 司諫; A Concordance to the Zhouli, ed. Lau, D.C. and Ching, Chen Fong [Hong Kong: Commercial Press, 1993], 16, 25Google Scholar), but perhaps significantly leaves the work of singing and poetry recitation (feng song shi 諷誦詩)to the blind musicians (A Concordance to the Zhouli’ 42); the division of labor is quite different from that envisioned in the Zuozhuan and Guoyu.

128. See references in n. 117 above. The only term the various versions have in common is Confucius's favored type, feng jian 諷諫.

129. See Xiaojing 孝經 (zhushu, Shism jing ed.), 2558Google Scholar.

130. Xunzi jijie 荀子集解 (jicheng, Zhuzi ed Zhuzi jicheng), 167Google Scholar (“Chen dao”).

131. Confucius is made to say 稱詩以諫, 順哉 “How smoothly it works when one refers to a poem in the course of remonstrating” (A Concordance to the Kongzi Jiayu’ ed. Lau, D.C. and Ching, Chen Fong [Hong Kong: Commercial Press, 1992], 76Google Scholar); the remark does not occur in the Zuozhuan, Zhao 12.11 (Yang, 1341), version of Confucius's comment on King Ling's fall.

132. Hence the several accounts of Confucius's esteem for feng jian 諷諫 “indirect remonstrance.”

133. Evidence that fengjian resembles the theatrical remonstrance of Zuozhuan, Zhao 28.4, is to be found in the Wu Yue chunqiu 吳越春秋 (A Concordance to the Wuyuechunqiu, ed. Lau, D.C. and Ching, Chen Fong [Hong Kong: Commercial Press, 1993], 24Google Scholar), where the Wu heir performs a remonstrance for his father, King Fuchai.

134. As, for instance, in the famous case of Zhao Dun 趙盾 and his repeated remonstrances (zou jian 驟諫)with Duke Ling of Jin (Zuozhuan, Xuan 2.3 [Yang, 655–59]).