1. Karlgren, Bernhard, On the Authenticity and Nature of the Tso Chuan,” Gøteborgs högskolas årsskrift 32.3 (1962), 3–65 Google Scholar; rpt. Taipei: Ch'eng-wen, 1968. See also his “The Early History of the Chou Li and Tso Chuan Texts,” Bulletin of the Museum of Far Eastern Antiquities 3 (1931), 1–59 Google Scholar. Some passing remarks on the Zuo zhuan also appear in his “The Authenticity of Ancient Chinese Texts,” loc. cit. 1 (1929), 165–83.

Early versions of several segments of this discussion were presented at various venues over a number of years, most often at the meetings of the Warring States Working Group organized by E. Bruce and Taeko Brooks under the aegis of the University of Massachusetts, Amherst. A fuller version was presented at the Columbia University Early China Seminar on May 15, 2004. I am indebted to numerous participants in these meetings for their comments and suggestions. What appears here is, of course, my own responsibility. Readers unfamiliar with the Warring States Project website may wish to consult the following url's for the Brooks’ views on the Zuo zhuan (http://www.umass.edu/wsp/wst/a-e/dj/) and the Chun qiu (http://www.umass.edu/wsp/wst/a-e/cc/). On a recent paper by Taeko Brooks, see n.151 below.

2. On the debate, see Schaberg, David, A Patterned Past: Form and Thought in Early Chinese Historiography (Cambridge, MA and London: Harvard University Asia Center, 2001), 316 Google Scholar; Pines, Yuri, Foundations of Confucian Thought: Intellectual Life in the Chunqiu Period, 722–464 B.C.E. (Honolulu: University of Hawai'i Press, 2002), 257n58Google Scholar; Early Chinese Texts: A Bibliographic Guide, ed. Loewe, Michael (Berkeley, CA.: Society for the Study of Early China and Institute of East Asian Studies, 1993), 69–70 Google Scholar. The most recent indictment of the New Text position is Takao, Hirase 平勢隆郎, Saden no shiryō hihan teki kenkyū 左傳の史料批判的研究 (Tokyo: Tokyo University Tōyō bunka kenkyūjo, 1998)Google Scholar.

3. See Pines, , Foundations, 257n58Google Scholar; Schaberg, , Patterned Past, 316–17Google Scholar.

4. For some basic works, see Early Chinese Texts, 67–76 passim. Both of the books introduced below offer very extensive (even if not entirely exhaustive) bibliographies on these subjects

5. On both works, see n.2 above. See also, for Pines, , “Intellectual Change in the Chunqiu Period: The Reliability of the Speeches in the Zuozhuan as Sources of Chunqiu Intellectual History,” Early China 22 (1997), 77–132 CrossRefGoogle Scholar, especially 122–24 and, for Schaberg, , “Remonstrance in Eastern Zhou Historiography,” Early China 22 (1997), 133–79CrossRefGoogle Scholar. Schaberg's book also contains extensive discussion of the Guo yu 國語.

6. Schaberg, , Patterned Past, 438n28Google Scholar.

7. On the conviction that, with a text such as the Zuo zhuan, familiarity breeds insight, this section employs examples based on the Zuo zhuan record concerning the state of Chu 楚, material that has been at the heart of my use of the text for some years now.

8. Hirase Takao's list of components (Saden no shiryō hihan teki kenkyū, 486) makes fewer distinctions. On the basis of their particle usage (a topic discussed below) and role in the Zuo zhuan, I here combine several components hypothesized in my earlier, unpublished analyses.

9. See Schaberg, , Patterned Past, pp. 306, 319–20Google Scholar. It could be that this is the reason, as much as the touting of it as a commentary on the Chun qiu, that the original name given to the text was Zuo shi 左氏 Chun qiu. I cannot here delve into comparing these redactions with the other two surviving chronicles, the Zhushu jinian 竹書紀年 and, undoubtedly in an altered version, the Qin 秦 chronicle as transmitted in the Shi ji 史記. By “canonical Chun qiu,” I here refer to the version to which the Zuo zhuan was eventually attached, which is two years longer than that associated with the Gong yang 公羊 and Gu liang 榖梁 commentaries.

10. In addition to those below, there are differences in diction such as 十有二月 (Chun qiu) and 十二月 (Zuo zhuan Chun qiu) and dating inconsistencies due to use of different calendars.

11. Throughout, I utilize the passage designations given in Combined Concordances to Ch'un-Ch'iu, Kung-yang, Ku-liang and Tso-chuan (rpt. Taibei: Chinese Materials and Research Aids Service Center, Inc., 1966)Google Scholar, with one alteration (see n.14 below). CQ is an abbreviation for Chun qiu and ZZCQ for Zuo zhuan Chun qiu.

12. Interspersed here is the sentence 脩惠公之好也 “In order to restore the good relations [with them established under] Hui Gong” (in my scheme, a gloss; see below).

13. I.e., they treat either Lu events or those relating to neighboring states that typically appear in the extant Chun qiu. Most such entries occur in the passages labeled fu 附 (“supplement” to the Chun qiu material) in the Harvard-Yenching Concordance. (This category includes the entire final segment of the text that extends chronologically beyond the Chun qiu; i.e., from Ai 17, 4th month on.)

14. In order to facilitate location of fu entries, indicated by lower case roman numerals in the Harvard-Yenching concordance, I add the passage number under which they can be found; e.g., in the present case the concordance has Yin 1ii, which here becomes Yin 1.2ii (except for Ai 17 on, which is entirely “supplemental” material, in which case I preserve the diction of the concordance; i.e., Ai 19i).

15. For a summary of allusions in Eastern Zhou texts to historical records of various kinds, see Pines, , Foundations, 253–4n31Google Scholar and Schaberg, , Patterned Past, 319–20Google Scholar. On official compilations, see Pines, , Foundations, 253n29, 253–54n31Google Scholar; Schaberg, , Patterned Past, 7, 319 Google Scholar.

16. See Yin 9.7i: 北戎侵鄭. 鄭伯 禦之 “The Northern Rong invaded Zheng. The Bo of Zheng resisted them,” which is followed by discourse material.

17. E.g., Xiang 3.9: 許靈公事楚, 不會于雞澤 “Gong Ling of Xu, being subservient to Chu, did not attend the convocation at Jize” serves to explain the attack by Jin 晉 on Xu 許 mentioned in both Chun qiu and Zuo zhuan Chun qiu.

18. For instance, Yin 1.5vi. Here the thirty-four graphs between the first four, which I treat as Miscellany, probably added at the time of compilation of the text, and the final six that are commentary, read like reworded Annals material.

19. For analytical purposes I include in this category quoted written material, e.g., messages.

20. A classic example of Miscellany unrelated to narrative is Xiang 7.9, which is composed entirely of miscellaneous material, although some of it could have been reworked from Annals. All but the first five graphs (= Zuo zhuan Chun qiu) of Ai 24iv also is Miscellany. Despite this, since most Miscellany is found in tandem with narratives, I treat it below in the same category.

21. The latter quality also distinguishes Miscellany from Annals material.

22. With some frequency more than one type occurs in connection with the same statement (e.g., Schaberg, , Patterned Past, 178 bottomGoogle Scholar). On commentary, see Pines, , Foundations, 253n30Google Scholar.

23. These have been studied and compared by Henry, Eric in his “‘Junzi Yue’ Versus ‘Zhongni Yue’ in the Zuozhuan ,” Harvard Journal of Asiatic Studies, 59 (1999), 125–61CrossRefGoogle Scholar. See also Yiren, Zhang 張以仁, “Guanyu Zuo zhuan ‘junzi yue’ de yixie wenti 關於左傳 ‘君子曰’ 的一些問題,” KongMeng xuekan 孔孟學刊, 3.3 (Nov. 1964), 28–30 Google Scholar.

24. Those attributed to Confucius (by whatever name) are always introduced by yue 曰; the junzi comments, variously by 曰, 謂, or 是以知.

25. The Chun qiu entry here (辛未, 取郜. 辛巳, 取防 “Xinwei [day], [we = Lu] took Gao; Xinsi [day], we took Fang”) gives the impression that Lu acquired these towns on its own.

26. Not all such statements belong in this category. For instance, in Yin 5.1 非禮也 is an element in a shu yue 書曰 (“the Chun qiu states” comment; on which see below). Most readers probably assume that the value judgments are original to the text. I suspect this is not the case.

27. In cases (such as Yin 5.1) where the Chun qiu entry is repeated in a shu yue gloss and a li ye comment follows, I take the latter as relating to the shu yue statement (not to the Chun qiu entry).

28. This comes after an extended dialogue that would not have been recorded in state annals.

29. Literally, “[the Chun qiu] states.”

30. I.e., this Chu event (otherwise of no particular concern in Lu) is recorded in the Lu chronicle only as an object lesson in unseemly behavior of an official.

31. Schaberg, , Patterned Past, 178–79Google Scholar, follows Zhao Guangxian 趙光賢 in taking comments introduced by fan 凡 (“in general” or “whenever”) as a distinct category. At least in my sample, all fan statements follow either a shu yue gloss (in Yin 7.2, 9.30) or a li ye comment (as here and, for instance, Xiang 12.4—on which see below). Thus, these fan statements may represent a sub-commentary. Chen Pan 陳槃 also discussed these elements of the text in his “Lun Zuo zhuan ‘fan li’ yu Liu Xin zhi guanxi 論左傳凡例與劉歆之關係,” Minzhu pinglun 民主評論, 8.2, 32–37 Google Scholar.

32. The passage can be translated in different ways, depending on one's views on the nature of the kinship structure and its attendant mourning system.

33. See Henry, “‘Junzi Yue’ Versus ‘Zhongni Yue’ in the Zuozhuan,” 148.

34. Given the ambiguity in the use of the designation “Chunqiu” in early times (see Pines 266n151), it is not impossible that the junzi comments derive from one or both of the works by Zoushi 鄒氏 and Jiashi 夾氏 listed in the Han shu 漢書 as commentaries on the Chun qiu (see, for instance, Early Chinese Texts, pp. 67–68). The idea that they are the work of Liu Xin 劉歆 has been safely put to rest (Henry, “’Junzi Yue’ Versus ‘Zhongni Yue’ in the Zuozhuan,” 128–29).

35. Both Pines and Schaberg treat these at various points in their books; See particularly Pines, Foundations, Appendices 2 and 4. For an estimate of the amount of such material in the text, see especially Pines, , Foundations, 22–23, 257n63Google Scholar; Schaberg, , Patterned Past, 98–103 passim Google Scholar, 181–82, 191–98 passim.

36. Foundations, 17–23.

37. Foundations, 25–26.

38. In defense of his position that the discourses were recorded at the time, Pines (wisely in my opinion) does not place much weight on references to one category of scribes recording words and another, events (254n42). For an expression of skepticism on the matter, see Schaberg, , Patterned Past, 7, 328n22Google Scholar.

39. Foundations, 220, Table 4.

40. He also lists a “Chun qiu” but this presumably is the Chun qiu itself, not my Zuo zhuan Chun qiu.

41. Patterned Past, 172–74, 322–23.

42. I am a rather puzzled by a seeming duplication in this scheme: “fragmentary Chun qiu” in the first category and “the Chun qiu chronicle” in the third.

43. These are the subject matter of his chapters 5 and 6.

44. Pines, , Foundations, 217–20Google Scholar.

45. See also his “Intellectual Change in the Chunqiu Period,” esp. 122–24. Groundwork on this shift was done by He Leshi 何樂士 in “Zuo zhuan qian bagong yu hou sigong di yufa chayi 左傳前八公與後四公的語法差異,” Gu Hanyu yanjiu 古漢語研究 1 (1988), 56–65 Google Scholar. Pines also appeals to another shift, from 其 to 豈 as markers of rhetorical questions. I have not yet had a chance to investigate this from the standpoint of my proposed components, but my samples do not show any significant increase of 豈 over time.

46. He does, however, take note of interpolated material; see n.35 above.

47. These are taken from The Digitalized Indexed Zuo zhuan being compiled at El Colegio, Mexico (see http://mezcol.colmex.mx/zuozhuan_2003). Eric Henry has compiled a slightly different computation (see http://www.umass.edu/wsp). I utilize the former because it includes graph counts for individual years.

48. I chose this segment rather arbitrarily, and before encountering graph-counts for individual years in the text. On the latter basis, the middle of the text falls instead near the end of the Duke Xiang section (ca. years 19–24).

49. This ignores the penultimate Dao 悼 Gong 4, certainly a later addition.

50. In my sample, this involves use of 於 in the compound 於是 (“thereupon”) and, less frequently, 於 as a full verb (e.g., Yin 5.1: 皆於農隙講事也). 於 as the fusion of 于 and 之 is absent.

51. There are, however, only 11 such cases in my sample, a mere 3.2% of the 343 total occurrences of the two particles.

52. However, the rate of the replacement in my sample is far less than for Pines. Here, the figure for the last era is about 2.4 times that for the first. In Pines’ analysis, the figure for his late sample (3.6) is over ten times greater than that for his early sample (0.333). That the shift is apparent within the Zuo zhuan indicates that (at least in this case) the Zuo zhuan compiler did not thoroughly standardize the grammar of his material (see Pines, , Foundations, 20, 253n25Google Scholar), perhaps because by his time the rule was no longer operable.

53. The rise in the Discourses here is far less than found by Pines. The incidence of 於 to 于 in his final period (Foundations, 220, Table 4, which computes to 13.7) is about six times higher than that for the early period (2.2), more than four times the comparable increase between my early and late samples (1.38).

54. The instances of 于 in commentary all relate to the differences between the Zuo zhuan Chun qiu and the Chun qiu, and most are of the shu yue variety; the three cases of 於, all in junzi statements (Yin 3.3i [2], Xiang 5.13), relate to other components.

55. For the three chronological samples, the 於: 于 ratio for the combined Discourses and Discourse-related categories are 2.0, 1.6, 1.5. Adding the Miscellany to these results in a pattern of 1.6, 1.4, 1.8.

56. Foundations, 220, Table 4.

57. I.e., both use 于 exclusively.

58. Jens Østergård Petersen, “The Distribution of ‘於’ and ‘于’ in the Zuozhuan 左傳: A Stylistic Approach,” unpublished paper cited by Pines (Foundations, 218).

59. Pines, (Foundations, 17–18, 252n18Google Scholar) sees the Chun qiu as belonging to “the culticritual strand of ancient Chinese historiography,” that it was addressed to the spirits, and that it was written in a “highly formulaic language [that] often concealed actual historical events.” (Foundations, 17)

60. The idea is not new: both Karlgren, , “The Authenticity and Nature of the Tso Chuan,” 32, 41–49 Google Scholar, and Pulleyblank, Edwin G., “The Locative Particles yü 于, yü 於, and hu 乎,” Journal of the American Oriental Society 106.1 (1986), 1–13 CrossRefGoogle Scholar, addressed the matter, but neither delved sufficiently deep into it. The two graphs originally certainly were different—obviously on orthographic, but also on phonological grounds: on the latter, see Karlgren, , “Authenticity”, 42 Google Scholar; Baxter, William H., A Handbook of Old Chinese Phonology (Berlin and New York: Mouton de Gruyter, 1992), 37, 44 CrossRefGoogle Scholar.

61. It may well be, then, that this is the explanation for the single exception in either the Zuo zhuan Chun qiu or the Commentary (Yin 6.2: 始平于齊也).

62. These vary in length from three graphs (Xiang 5.5) to eleven (Xiang 7.2; the final two graphs in the Zuo zhuan entry in this case [旱也] I take to be commentary).

63. Among these are two cases (Xiang 1.4, 2.9) in which there is a Zuo zhuan counterpart to a CQ entry in terms of subject matter, but the wording is substantially different.

64. The season is supplied in Xiang 5.5.

65. The following seven graphs I take to be commentary.

66. Concerning the Chun qiu mention of a meeting of the lords of Zheng 鄭 and Cai 蔡 in 710 (Huan 2), the Zuo zhuan says that the meeting was because they “had begun to fear Chu” 始懼楚也. But this statement is a gloss and, in any case, would not have derived from the Chu chronicle.

67. On the use of the plural in translating zu here, see Blakeley, Barry B., “King, Clan and Courtier in Ancient Chu,” Asia Major, 3rd Ser., 5.2 (1992), 5 and 38 Google Scholar (genealogical chart).

68. On the social and political significance of this event, see Blakeley, “King, Clan and Courtier in Ancient Chu.”

69. This marked the end of the reign of King Wen 文.

70. Two of these Chun qiu entries have no Zuo zhuan counterpart: 1). A notice of the first Chu envoy to Lu 魯 in Zhuang 23 (= 671). Perhaps the Chu chroniclers did not yet consider this important enough to record, even though all other such events are present in the Zuo zhuan and so could have been recorded in the Tao wu. 2). In Xi 喜 19 (= 641) the Chun qiu lists Chu among the participants in a “covenant” meng 盟 but the Zuo zhuan does not. Perhaps this lacuna in the Zuo zhuan is the work of the Zuo zhuan compiler.

71. Three others (in Xi 28, Wen 文 1, Wen 10; = 632, 626, 617, respectively) have parallels in the Chun qiu. All of these record the deaths of prominent Chu figures, so it seems that by the late seventh century major Chu events were considered important enough in Lu to be included in its chronicle.

72. The entries analyzed here are limited to those concerning events of a level of significance or of subject matter typical of the received chronicles. Others could have derived from the Tao wu if it differed somewhat from those chronicles, but it will be assumed here that they originated in some unattested type of record, some type of Annals. Down to 605, there are a number of such entries. Two appear at the appropriate point in the text (Zhuang 19, Zhuang 30), but most are found out of chronological order (under Xi 7, Wen 14, Zhuang 14, Zhuang 19, Zhuang 30, Zhao 7).

73. For instance, the Chu destruction of Xi 息 is mentioned under Zhuang 14 (= 680), but from the fact that the woman King Wen absconded with at the time had by that year given birth to two sons, the elimination of Xi must have taken place in 683: see Shin'ichiro, Takezoe 竹添光鴻, Saden kaisen 左傳會箋 (1907; rpt. Taipei: Feng-huang, 1977), 3.42 Google Scholar.

74. Many of the late fourth-century b.c.e. Baoshan 包山 tomb No. 2 documents are dated on the basis of a major event of the previous year; see Baoshan Chu mu 包山楚墓 (Beijing: Wenwu, 1991), 1: 364–69Google Scholar. Given the precision in dating in the Zuo zhuan, however, the Chu chronicle, at least in Chunqiu times, must have been organized by reign years.

75. In one case, a Jin attack on Chu (Wen 3 = 624, the Chun qiu supplies the day, while the Zuo zhuan offers only the month). The latter could be the work of the compiler of the text.

76. The dates of two of the unconventional entries are supplied in the text under Zhuang 7 (= 688). The dates of two others can be inferred (see Zhuang 14 and Wen 14).

77. The Chun qiu does contain one notice of Chu during these years (the envoy dispatched to Lu in 671; see n.70 above). Either the Chu chroniclers did not find this worthy of recording or the Zuo zhuan compiler considered it of no great significance. The Shi ji, “Chu shijia” 楚世家 describes the events surrounding the assassination of Du Ao and the accession of King Cheng and the early years of his reign. The former event could well have gone unrecorded in Chu but have been notorious enough to be known elsewhere. The account of Cheng's early years, however, is patently fictionalized in the Shi ji; see my “King, Clan and Courtier in Ancient Chu,” 6–7.

78. It is mentioned only by inference and post-facto, with no hint as to the year, under Wen 14. On the other hand, the “Chu shijia” dates the event to the twelfth year of Mu's reign (= 614), so Sima Qian 司馬遷 must have had an alternative source for this event.

79. Even though most are totally absent elsewhere in the Zuo zhuan, several of them may be alluded to obliquely elsewhere in the text. These are: a submission by Sui 隨, perhaps mentioned under Zhuang 4, an attack on Shen under King Wen, probably either one alluded to under Zhuang 18 or in Zhuang 6, the transformation of the states of Shen and Xi 息 into xian 縣, presumably hard on the destruction of Xi, mentioned under Zhuang 14. See n.73 above.

80. The evidence for this is too cumbersome present here. We may mention, however, that if one of these events (an attack on the strategically located state of Shen) took place before the first dated notice of Chu in 706, the Zuo zhuan compiler probably would have considered it of sufficient significance to warrant mentioning.

81. The passage also contains a chronological error (its assertion that Dou Ban 鬭般 became Lingyin 令尹 upon the death of Ziwen 子文 does not accord with other information supplied in the text).

82. E.g., the material concerning the state of Deng 鄧 under Zhuang 18.

83. The tops of bamboo strips, where dates presumably would have been most often located, might easily have broken off.

84. Foundations, 16, 18, 23–26, 35–39, 179, 250n6, 252n18.

85. Patterned Past, especially 315–24, 438–39n37.

86. In some four decades of mining the Zuo zhuan for the study of the Chunqiu period, I have taken an implicit, albeit not openly stated, approach to the problem, namely, to avoid recourse to the discourses if at all possible. Since my research interests have centered on socio-political matters, this is not especially difficult, as specific data on these matters, in contrast to statements about them, are seldom found in the discourses.

87. Egan, Ronald C., “Narrative in the Tso Chuan ,” Harvard Journal of Asiatic Studies 37.3 (1977), 340–45, 349–50CrossRefGoogle Scholar, also points to these.

88. Schaberg, , Patterned Past, pp. 192–207 Google Scholar, utilizes another Chu cycle, one concerning King Ling 靈.

89. Only two kings and a royal consort (who we are about to meet) merit positive treatment.

90. This statement may have derived from the Tao wu.

91. All translations are my own.

92. That a female should offer such sage advice was of sufficient note that it was included in the Lienü zhuan 列女傳 compilation of Han times (Sibu congkan 四部叢刊 [Shanghai: Shangwu, 1922?] Shi bu 史部, 265–67; University of Virginia Library Chinese Text Initiative: http://etext.lib.virginia.edu/chinese/lienu/browse/Lienu.html).

93. There is no Miscellany material in these entries.

94. The Discourse-related materials are so obvious that I do not specifically mark them.

95. As it stands, the following statement is, I think, intended in the spirit of “You miss my point … ”

96. In my view, the following Annals entry has been “borrowed” by both discourse sets for narrative purposes.

97. Chu Set 2a is tentatively tied to Set 1a, in that both involve military matters; but it is more strongly tied to the exchange of opinions about the futures of Sui's Jiliang and Shaoshi.

98. Sui Set 1a, with its indirect discourse, is presumably, at least in its present form, the work of a later compiler or creator of this material.

99. The two positions on tactics (Sui Discourse Set 1c–f [804]) have the flavor of a conflict between a more traditional (Chunqiu-era?) view expressed by the Shaoshi and a more pragmatic (Zhanguo-era?) approach offered by Ji Liang.

100. Having been intrigued by this material for some time, my attention was piqued when I came upon Pines alluding to the passage. He suggests (Foundations, 149) that the zhong 忠 in Ji Liang's speech (Sui Set 2a), which involves a unique understanding of it as a superior's obligations to inferiors, not the opposite, as in the Confucian view, may be an early understanding of the term that later disappeared. It is equally possible, however, that it is an alternative later one. That both 忠and 信 appear in this speech makes the late fourth- to early third-century “Zhongxin zhi dao 忠信之道” text unearthed at Guodian 郭店, Hubei, worth considering in this connection: see bowuguan, Jingmenshi 荊門市博物館, Guodian Chu mu 郭店楚墓 (Beijing, Wenwu, 1998), 163–64Google Scholar.

101. Qiebi, however, is little more than a foil for Bobi, whereas the Shaoshi is a fully developed “character.”

102. It is possible, of course, that something happened in the interim to change the balance of influence in Sui. But if so, we have no way of knowing that.

103. This is, in fact, the basic thrust of Schaberg's entire argument, i.e., that the narratives in general reflect the ideas and didactic objectives of Zhanguo times, not of the Chunqiu-era actors in them; see especially, Patterned Past, 7, 11, 12, 259–64. On the matter, see also Lewis, Mark Edward, Writing and Authority in Early China (Albany, NY: State University of New York Press, 1999), 130–32Google Scholar.

104. Prior to Schaberg, the literary qualities have been analyzed by: Wang, John C.Y., “Early Chinese Narrative: The Tso-chuan as Example,” Chinese Narrative, ed. Plaks, Andrew (Princeton: Princeton University Press, 1977), 3–20 Google Scholar; Egan, , “Narrative in the Tso Chuan,” esp. 343, 346 Google Scholar; and Yiren, Zhang, “ Zuo zhuan meixue di hexie lilun 左傳美學的和諧理論,” Kong Meng xuebao 孔孟學報 47 (April, 1984), 175–91Google Scholar.

105. Thus, in my opinion, Pines' contention that the discourses are reliable, even if to some extent reworked, goes too far.

106. For a list of these, see Schaberg, , Patterned Past, 406n59Google Scholar.

107. The literature on these stories is extensive. They are discussed at a number of points by both Pines (e.g., Foundations, 30–31, 244–46) and Schaberg (e.g., Patterned Past, 191–98 passim).

108. For an example, see Egan, , “Narrative in the Tso Chuan,” 338 Google Scholar.

109. E.g., Yanzi chunqiu 晏子春秋 and a good portion of Zhanguo ce 戰國策. On the idea that such sagas represent “table talk”, see Schaberg, Patterned Past, 320–21.

110. Chunqiu shiyu 春秋事語, discovered in 1973 at Mawangdui, has provided some insight into the evolution of narrative material such as is found in the Zuo zhuan by Han times; see Pines, , Foundations, 45 Google Scholar; Schaberg, , Patterned Past, 4, 328n11, 440n54Google Scholar. Also worth considering is the Zonghengjia shu 縱橫家書; see xiaozu, Mawangdui Han mu boshu zhengli, Zhanguo zonghengjia shu 戰國縱橫家書 (Beijing: Wenwu, 1976)Google Scholar. On a more recent find, see so, Hunan sheng wenwu kaogu yanjiu, “Hunan Cili xian Shibancun Zhanguo mu 湖南慈利縣石板村戰國墓,” Kaogu 考古學報 1995.2, 173–207 Google Scholar (mentioned by Pines, , Foundations, 255n45Google Scholar).

111. He also takes into account material in the Guo yu, although not to the same extent as Schaberg.

112. Pines, , Foundations, 90 Google Scholar.

113. Pines, , Foundations, 92 Google Scholar. I am not content with his translation of weiyi, but this is not the place to pursue the matter.

114. Pines, , Foundations, 94 Google Scholar.

115. Pines, , Foundations, 94–97 Google Scholar.

116. Pines, , Foundations, 97–103 Google Scholar.

117. Schaberg, , Patterned Past, 293 Google Scholar.

118. The allusions to li I cite in this section are limited to ones appearing in narrative discourses (not commentary) in which the word is substantively characterized or defined, excluding ones pertaining to rites or ceremonies (already under that heading in Western Zhou times). A single discourse quite often contains material relevant to more than one topic.

119. Xi 7.4 (= 653).

120. Xi 24.2iii (= 636); Xi 28.23 (= 632); Xi 33.2 (= 627); Wen 15.12 (= 612); Cheng 2.4 and 2.10iii (= 589); Cheng 8.10 (= 583); Cheng 12.3 (= 579).

121. Huan 18.1 (= 694), on male/female roles; Xi 7.1 (= 653), on father/son; Xi 11.1i (= 649), on superior/inferior.

122. Huan 2.11 fu (= 710): 禮以體政 “li is the embodiment of government;” Zhuang 23.3 (= 671): 禮所以整民也 “li is that by which the people are ordered.”

123. Pines, , Foundations, 95 Google Scholar. He also (p. 97) takes the 君子勤禮, 小人勤力 of Cheng 13.2 (= 578), which he translates as, “superior men are diligent in observing ritual (li), while petty men exhaust their [physical] strength,” as evidence of class distinctions, but that this distinction is meant as a value judgment is surely open to question.

124. Of course, it also is required that there be no counter-examples among those not cited.

125. Needless to say, when there is a single case, the point being made is on very shaky grounds.

126. Failure to offer guidance to a representative sample is endemic in his discussions of philosophical concepts. With respect to not citing all available evidence when it is limited in scope, using the discussion of the late Chunqiu era in Chapter 3 as an example: Emphasis on social hierarchy is apparent in two passages not cited by Pines (Zhao 1.8vii = 541 and Zhao 12.8 = 530). Similarly, his point that li had metaphysical significance is considerably reinforced by Zhao 26.7v (= 516). And the idea that li and yi became distinguished from each other is borne out by a passage (Zhao 25.2 = 517) not alluded to by Pines.

127. Pines, , Foundations, 97–98 Google Scholar.

128. 十二月。北宮文子相衛公以如楚 Twelfth month. Beigong Wenzi attended the Gong of Wei on a journey to Chu.

129. 宋之盟故也 This was due to [provisions of] the covenant [recently concluded] in Song.

130. 過鄭，印段迋勞于棐林。如聘禮而勞辭。文子入聘。子羽為行人。馮簡子於子大叔逆客

When passing [the capital of] Zheng, Yin Duan [of Zheng], on account of their difficulties on the journey, went forth [to greet them] at Fei Lin. [In this] he [behaved] as if on a courtesy call but used words [appropriate to] a visit of consolation. Wenzi entered [the city] to pay a [return] courtesy call. Ziyu [of Zheng] served as internuncio [and] Feng Jianzi and Zi Dashu greeted the guest.

131. 事畢而出，言於衛侯曰… When the affair was over and [Wenzi] departed, [upon returning to the entourage] he spoke to the Hou of Wei saying …

132. 其數世之福也，其無大國之討乎。詩曰: 誰能執熱。逝不以濯。禮之於政。如熱之有濯也，濯救熱，何患之有

This [will cause it] prosperity for generations, and [this] will save them from attacks by the greater states. The Shi says: “Whoever can hold something hot [must have first] dipped [his hand] in water.” In governance li is like [this] dipping in water. Having dipped to be saved from [the effects of] heat, what injury can there be?

133. 子產之從政也。擇能而使之。馮簡子能斷大。子大叔美秀而文。公孫揮能知四囯之為，而辨於大夫之族姓班位，貴賤能否。而友善為辭令。神諶能謀。謀於野則獲。謀於邑則否。鄭國將有諸侯之事。子產乃問四國職為於子羽。則使多未辭令。於神諶乘以適野。使謀可否，而告馮簡子使斷之。事成，乃授子大叔使行之。以應對賓客。是以鮮於敗事。

When Zichan was in charge of governance, he chose the able and employed them [in tasks]. Feng Jianzi was able to make judgments in important matters. Zi Dashu was pleasing in appearance and eloquent. Gongsun Hui [= Ziyu] was knowledgeable about the affairs in neighboring states and could distinguish the lineages, clans and positions, ranks and capabilities of [their] great officers. He also not to the implications of li in 542 b.c.e. Turning to Schaberg, his book centers on three aspects of the Zuo zhuan narratives:135 their structure and modes of presentation, the views on and uses of history and their philosophical outlook expressed in them, and what all of this suggests about those responsible for producing them. It is divided into two sections. The subject of the first is the speeches (Discourses, in my terminology). The analysis here focuses on their rhetorical modes (Chapter 1), the qualities of wen 文 (“literary patterns”) they exhibit and employ (Chapter 2), and their intellectual content. The latter is dealt with in two chapters. Chapter 3 addresses views of the extrahuman word (Heaven and Earth, the Five Phases, Yin-yang theory, etc.); also was skilled at composing official documents. Pi Chen was good at strategy, [although while] he was successful when planning in the countryside, [he was] not when doing so in the city. When Zheng was about to be involved in matters with [other] feudal lords, Zichan inquired of Ziyu [=Gongsun Hui] the actions of neighboring states and ordered him to compose [the appropriate] documents. [He then] rode in his carriage with Pi Chen into the countryside and had him consider whether the plans would succeed or not. He [then] told Feng Jianzi of [Pi Chen's plan] and asked him to assess it. When the matter had been decided, he assigned Zi Dashu to carry it out, having him make the replies to [any] visiting guests. [In this way], seldom did matters go awry.

134. In fact, only a quotation from the Shi jing offers any clue to the content of the term li.

135. He also treats or points to nearly five dozen passages from the Guo yu 國語.

136. He does not deal systematically with the rather short and prosaic materials falling into my Annals category, which he treats as integral aspects of the narratives.

137. The book is emphatically not, in the contemporary vernacular, an “easy read.” The prose is often dense and the terminology challenging to the uninitiated. Nevertheless, the rewards are considerable. Of particular note are analyses of the Zuo zhuan narratives from the perspectives of the disciplines of Rhetoric (in connection with the Discourses) and Narratology (with respect to the anecdotes and anecdote series). Included in both are numerous considerations of comparative issues, especially with respect to the Greek tradition. There is much that is instructive and provocative for students of at least three disciplines—Literature, Thought, and Historiography. Subjects for future consideration suggested by the book are numerous. Here I will mention only one. Schaberg basically analyses anecdotes that exhibit most or all of the elements he identifies for the narrative genre, but a comparison of these with more truncated anecdotes might provide some insight into broader narrative and rhetorical tradition, the process of transmission of the material and/or the compiling the Zuo zhuan.

138. Schaberg, , Patterned Past, 6 Google Scholar.

139. He explains (Patterned Past, 8) that his terminology in this connection is “deliberately – and necessarily – vague.” On “author” and “narrator,” see Patterned Past, 328n26. Schaberg uses these terms in the plural (e.g., pp. 178–79), for he presumes (p. 8) that the text is the work of more than one individual.

140. E.g.: “These [shu yue] passages … are probably not part of the original anecdotal material … ” “Marked judgments … lie at the margins of anecdotes without themselves constituting anecdotes.” “The fanli remarks were likely fairly late additions to the Zuozhuan composite.” Schaberg, Patterned Past, 174, 184, 179, respectively.

141. Schaberg, , Patterned Past, 174 Google Scholar.

142. Or, as I have suggested in cases such as this, on the difference in diction between the Chun qiu entry and its Zuo zhuan Chun qiu equivalent, in the present case, a much condensed version of the former.

143. Schaberg, , Patterned Past, 177–79Google Scholar, quotation on p. 178.

144. The two variants of the same rather elaborate discourse in the episode noted by Pines (Appendix 3) could just as well have been from two different oral sources as from two written versions, as posited by Pines.

145. The narratives from the Dou Cycle examined above illustrate the point. Some of the discourses could belong to the first category, but it seems appropriate to assign others, distinguished by a greater ideological thrust and/or rhetorical sophistication, to the second.

146. Unfortunately, we are required by circumstances to make such judgments essentially on the basis of the Zuo zhuan materials themselves. For, except for what can be gleaned from bronze inscriptions and whatever ideas one is comfortable in assigning to Confucius himself, barring the discovery of Chunqiu-era materials of appropriate date, the insurmountable problem is the absence of very little indeed in the way of “control” material.

147. Identity aside, I stand with those who find the stylistic consistency of the text as pointing to a single hand, at the very least for a final editing.

148. Schaberg accepts the consensus position of the mid-fourth century. Pines offers evidence to “cautiously suggest the fifth century b.c.e. as a possible date of the Zuo compilation, and the mid-fourth century B.C.E. as the latest date when a more-or-less fixed version could have appeared.” (Foundations, 29–34; quotation from the latter) The mid-fourth century date he alludes to is 360 b.c.e. (p. 31).

149. Pines, , Foundations, 23 Google Scholar.

150. I here combine scattered elements of Schaberg's discussion of the matter (Patterned Past, 316–17, 22–24), with the hope that in doing so I have not done him an injustice.

151. I do part ways, however, on the commentary material. I am inclined (along with many others) to start with the hypothesis that the personalized glosses, at least the junzi ones, date to later than Schaberg proposes. As for the explanatory glosses, if my hypothesis that the shu yue comments are not specifically aimed at the Chun qiu but rather at differences between its diction and that of Zuo zhuan Chun qiu holds up in the end, these could date only to after the specific tying together of the two texts, whenever that might be determined to be.

In a recent paper (“Heaven, Li, and the Formation of the Zuozhuan,” Oriens Extemus 44 [2003/04], 51–100), A. Taeko Brooks argues that the ideas in the Zuo zhuan are of Zhanguo (4th century) date and on this basis dates the entire text to that time (although seeing it as being generated by accretion). Since, however, the analysis of ideas touches almost exclusively on the narratives (essentially the Discourses), from my perspective her argument is relevant only to that segment of the text.

152. On this basis, the Zuo zhuan as a whole, using the 179,779 graph total (in the El Colegio Mexico computation, see n.47 above) would require nearly 9,000 slips.

153. This, however, is undoubtedly on the low side in terms of narrative material. It also contains 120 graphs of miscellaneous material unconnected to narratives, which do not figure in the following computations.

154. These total something like 1,490 graphs out of 4,935, or 30.2%.

155. We have seen that the Zuo zhuan is somewhat selective for the early Chunqiuperiod history of Chu, but basics of its later history as well as that of the other major states are recorded in about as much detail as one would expect of the originals.

156. In Yin Gong, the Discourse and Discourse-related components plus the Miscellany occurring in the same passages total about 2,090 graphs, or 42.37% of the total. Pro-rated for the entire Zuo zhuan, this works out to ca. 76,175 graphs.

157. See n.105 above.

158. Shanghai bowuguan cang Zhanguo Chu zhushu 上海博物館藏戰國楚竹書, chief ed. Chengyuan, Ma 馬承源 (Shanghai: Shanghai Guji, 2001, 2002, 2003, 2004, 2005)Google Scholar.