1. All subsequent dates in this essay are b.c.e., unless otherwise noted. In a very few cases, b.c.e. has been retained simply to situate the reader.

2. See, e.g., Han shu 漢書 (Beijing: Zhonghua, 1962), 68.2940, in “Huo Guang Jin Midi zhuan” 霍光金日磾傳. Liu He’s father Liu Bo 劉髆 was the first King of Changyi, and a little over two years after Liu He’s exile in 74, Liu He was granted the title of Noble of Haihun (Han shu, 8.257).

3. Early publications include Jiangxi sheng Wenwu kaogu yanjiusuo 江西省文物考古研究所, Faxian Haihunhou 發現海昏侯 (Nanchang: Jiangxi jiaoyu, 2015) and Wuse xuanyao: Nanchang Handai Haihun houguo kaogu chengguo 五色炫曜：南昌漢代海昏侯國考古成果 (Nanchang: Jiangxi renmin, 2016). In 2019, CCTV aired a three-part series titled “Haihunhou,” see https://tv.cctv.com/2019/07/25/VIDA6470jTHonIXYYcKmfQ0u190725.shtml (accessed March 29, 2021).

4. The slip was first reproduced in Jiangxi sheng Wenwu kaogu yanjiusuo, Nanchangshi bowuguan 南昌市博物館, and Nanchangshi Xinjianqu bowuguan 南昌市新建區博物館, “Nanchangshi Xi Han Haihunhou mu” 南昌市西漢海昏侯墓, Kaogu 考古 2016.7, 45–62 (photo on p. 61).

5. Han shu, 30.1716 lists three versions of the Lun yu that were part of Liu Xiang’s bibliographic survey. The auto-commentary to the twenty-two-chapter version says: “There are, additionally, ‘Wen wang’ and ‘Zhi dao’” (多問王知道). A comment by Ru Chun 如淳 (fl. 221–265 c.e.) clarifies: “‘Wenwang’ and ‘Zhidao’ are chapter titles” (問王、知道，皆篇名也).

6. In 2017, Wang Chuning 王楚甯 and Zhang Yuzheng 張予正 noted that the recto of the Haihunhou slip has a close parallel in Gansu slip 7ETJ22.6 in Gansu jiandu baohu yanjiu zhongxin 甘肅簡牘保護研究中心 et al., Jianshui Jinguan Hanjian 肩水金關漢簡 (Shanghai: Zhongxi, 2011), and identify the passage occurring in these geographically distant finds with the Qi Lun yu. See their “Jianshui Jinguan Hanjian Qi Lun yu de zhengli” 肩水金關漢簡《齊論語》的整理, Zhongguo kaogu 中國考古, August 16, 2017, http://chinesearchaeology.net/cn/kaoguyuandi/kaogusuibi/2017/0816/59268.html. While zhi 智 (wisdom) is used in the excavated Haihunhou slip, its homophone zhi 知 (know) appears in the Jianshui Jinguan parallel, which is otherwise nearly identical, albeit broken and missing the Haihunhou exemplar’s final clause. Charles Sanft takes issue with their identification of these texts with the Qi Lun yu in “Questions about the Qi Lunyu,” T’oung Pao 104.1–2 (2018), 189–94. Kyung-Ho Kim’s historical survey of the question is “Popularization of the Analects of Confucius in Western Han and the Discovery of the Qi Lun: With a Focus on the Bamboo Slips Unearthed from the Haihunhou Tomb,” Sungkyun Journal of East Asian Studies 19.2 (2019), 213–32. Another of the Gansu slips (73EJH1:58 E24) strongly suggests that the name of the other additional Qi version chapter, “Wenwang” (Questioning the King), might actually be “Wenyu” 問玉 (Questions about Jade), as it exactly overlaps with part of a Shuowen jiezi 說文解字 entry about jade that ties the virtues of a king to qualities of jade.

7. After an initial set of general articles published in Wenwu 2018.11, Wenwu 2020.6 contained: Zhu Fenghan 朱鳳瀚, “Xi-Han Haihunhou Liu He mu chutu Shi chutan,” 西漢海昏侯劉賀墓出土竹簡《詩》初探, 63–72; Tian Tian 田天, “Xi-Han Haihunhou Liu He mu chutu ‘liyi jian’ shu lüe,” 西漢海昏侯劉賀墓出土“禮儀簡”述略, 73–75; Chen Kanli 陳侃理, “Xi-Han Haihunhou Liu He mu chutu Lun yu ‘Zeng Xi yan zhi’ jian chushi,” 西漢海昏侯劉賀墓出土《論語》“曾晳言志” 簡初釋, 76–79; and Yang Bo 楊博, “Xi-Han Haihunhou Liu He mu chutu ‘fangzhong’ jian chushi,” 西漢海昏侯劉賀墓出土“房中”簡初識, 80–82, 96.

8. Zhu Fenghan, ed., Haihun jiandu chulun (Beijing: Beijing daxue, 2020) has chapters devoted to the following bamboo slip texts (listed with authors and page numbers): Shi 詩 (Zhu Fenghan, 79–110), Bao fu 保傅 (Han Wei 韓巍, 111–25), Yi 儀 (Tian Tian, 126–33), Chunqiu 春秋 (Chen Suzhen 陳蘇鎮, 134–40), Lun yu 論語 (Chen Kanli, 141–63), Xiao jing shuo jie 孝經説解 (He Jin 何晉, 164–203), Diao wang fu 悼亡賦 (Zhao Huacheng 趙化成, 204–13), Liu bo 六博 (Yang Bo, 214–31), Yi zhan 易占 (Li Ling 李零, 232–44), Bu xing 卜姓 and Qu yi 去邑 (Lai Zulong 賴祖龍, 245–54), Cizhu 祠祝 (Tian Tian, 255–67), and Fangzhong 房中 (Yang Bo, 268–76). Some of these pieces overlap with the Wenwu articles. Note the page numbers do not match those appearing in the table of contents.

9. The recto and verso of the three relevant slips (Haihun jiandu chulun, 176) contain a dialogue between Hou Jun 后軍 and Wuma Ziqi 巫馬子期 about the scope of the injunction found in the Mengzi 孟子 and several Han texts that “if you see [an animal] alive, you do not eat its corpse,” (jian qi sheng bu shi qi si 見其生不食其死). Chen Kanli judiciously points out that the inclusion of a second “Zhidao” text does not necessarily mean that the tomb contains the “Qi Lun yu” (161). He further argues that since the slips with Lun yu parallels were mixed with others that contain parallels to sections of texts like the Li ji 禮記, or slips without transmitted parallels, it is not clear if the collection should be seen as a Lun yu ancestor text, or simply a Kongzi-centered transmission (179). The inclusion of a chapter number on the verso appears to mark these three slips as a free-standing quotation from a “chapter 21,” rather than part of a Haihunhou version of the Lun yu.

10. An initial transcription of the text was published in Wang Yile 王意樂, Xu Changqing 徐長青, Yang Jun 楊軍, and Guan Li 管理, “Haihunhou Liu He mu chutu Kongzi yijing” 海昏侯劉賀墓出土孔子衣鏡, Nanfang wenwu 2016.3, 61–70, 50. Wang Yile 王意樂 and Wu Zhenhua’s 吳振華 chapter “Kongzi yijing chudu” 孔子衣鏡初讀 in Zhu Fenghan, Haihun jiandu chulun, 353–91, revises the transcription.

11. On the top panel, there is Red Phoenix (Zhu Que 朱雀) with Xiwang Mu 西王母 to the left and Dongwang Gong 東王公 to the right. Under Xiwang Mu on the left panel is White Tiger (Bai Hu 白虎), and under Dongwang Gong on the right panel is Azure Dragon (Qing Long 青龍). The image on the bottom panel is blurry, but the rhapsody identifies it as Black Crane (Xuan He 玄鶴). See Li Ziliang 劉子亮, Yang Jun 楊軍, and Xu Changqing 徐長青, “Handai Dongwang Gong chuanshuo yu tuxiang xintan: yi Xi-Han Haihunhou Liu He mu chutu ‘Kongi yijing’ wei xiansuo” 漢代東王公傳說與圖像新探——以西漢海昏侯劉賀墓出土“孔子衣鏡”爲線索, Wenwu 2018.11, 81–86.

12. The upper register has Kongzi on the left and Yan Hui on the right facing each other. The middle register contains Zigong on the left facing and Zilu who is facing forward. The lower register has Ziyu on the left facing away and Zixia on the right, reading. Each image is accompanied by between six and twelve columns of text about the lives of Kongzi and the disciple.

13. Both Andrew Hardy, “Imagining the Sage” (M.A. thesis, University of California, Berkeley, 2019), 64–65 and Jue, Guo 郭珏, “The Life and Afterlife of a Western Han ‘Covered Mirror’ from the Tomb of Marquis of Haihun (59 b.c.e.),” Journal of Chinese History 3.2 (2019), 1–30, 26–27Google Scholar) translate the poem. Guo also provides a convincing description of the mirror’s role within the tomb context, with the aim of arguing against the name “Kongzi dressing mirror.”

14. Editor’s note [M. Nylan]: Michael Loewe has remarked to me (personal conversation, December 15, 2021) that this pairing of Xiwang Mu and Kongzi and his disciples is perhaps the most notable feature of the mirror stand, with the goddess standing perhaps for the unseen world and the Kongzi circle for the sociopolitical.

15. Wang Chuning 王楚甯, “Jiangxi Nanchang Xi-Han Haihunhou Liu He mu chutu ‘Kongzi jing ping’ fuyuan yanjiu” 江西南昌西漢海昏侯劉賀墓出土“孔子鏡屏”複原研究 Wenwu 2022.3, 52–63.

16. See Csikszentmihalyi, Mark, “Interlocutor Collections, the Lunyu, and Proto-Lunyu Texts,” in Confucius and The Analects Revisited: New Perspectives on Composition, Dating and Authorship, ed. Hunter, Michael and Kern, Martin (Leiden: Brill, 2018), 218–40Google Scholar.

17. Analects 11.3 contains a similar taxonomy.

18. For the Shi ji reliance on the Analects in these chapters, see Hunter, Michael, Confucius Beyond the “Analects” (Leiden: Brill, 2017), 167ffCrossRefGoogle Scholar.

19. Durrant, Stephen W., The Cloudy Mirror: Tension and Conflict in the Writings of Sima Qian (Albany: State University of New York Press, 1995), 9Google Scholar.

20. Han shu, 30.1717. Huang, Chin-Shing, Confucianism and Sacred Space: The Confucius Temple from Imperial China to Today (New York: Columbia University Press, 2020)Google Scholar translates tu fa as “icons,” while Lai Guolong 來國龍 argues that the text was not visually oriented but rather consisted of tuji 圖籍 and fadian 法典; “Han Jin zhi jian hegui shu de shanbian he guishen hua de yuanliu,” 漢晉之間劾鬼術的嬗變和鬼神畫的源流, in Yishu shi zhong de Han Jin yu Tang Song zhi bian 藝術史中的漢晉與唐宋之變, ed. Yuan Juanying 顏娟英 and Shi Shouqian 石守謙 (Taipei: Shitou, 2014), 63–94, 66n22.

21. Boxes in the Chinese and ellipses in the English indicate missing or illegible graphs in the mirror text, based on the 2020 transcription in Zhu Fenghan, Haihun jiandu chulun. I have also indicated Zhu’s guesses based on parallel passages in the English version. Subscript numbers inserted into the text indicate column numbers on the manuscript, and bullet points are internal section markers in the original text. Column numbers mark the beginning of the column.

22. Wang Yile and Wu Zhenhua, “Kongzi yijing chudu,” 368.

23. 紇與顏氏女野合而生孔子⋯⋯禱於尼丘得孔子. Shi ji (Beijing: Zhonghua, 1959), 47.1905.

24. Sima Zhen’s 司馬貞 (679–732 c.e.) commentary justifies Shuliang He’s marriage to a daughter of the Yan clan by explaining he had already had nine daughters but only one son, and that son had a medical problem with his leg. Zhang Shoujie’s 張守節 (fl. 736 c.e.) subcommentary extends this approach, citing a numerological explanation for why men can and should marry before sixty-four sui, and saying that since Shuliang He was older than that, it was irregular (Shi ji, 47.1905).

25. Yang Jun 楊軍, En Zijian 恩子健, and Xu Changqing 徐長青, “Haihuhou fajue yu lishi wenhua ziliao zhengli yanjiu” 海昏侯發掘與歷史文化資料整理研究, Jiangxi shifan daxue xuebao (zhexue shehui kexueban) 江西師範大學學報（哲學社會科學版） 2018.1, 104–15.

26. The Shi ji reading of “pray” is supported by an early parallel from the Kongzi jiayu (Shandong: Shandong renmin, 1989), 9.10b (“Ben xing jie” 本姓解), that says they “privately prayed to Mount Niqiu for an auspicious outcome” (私禱尼丘山以祈焉) about his gender.

27. Shi ji, 47.1906, compare Lunheng jiaoshi 論衡校釋, ed. Huang Hui 黃暉, 4 vols. (Beijing: Zhonghua, 1990), 26.1089.

28. Wang Yile and Wu Zhenhua, “Kongzi yijing chudu,” 368.

29. Shi ji, 47.1909.

30. Shi ji, 47.1908 makes this connection via Meng Xizi’s 孟釐子 deathbed testimonial regarding Kongzi; as Meng explains to his heir: “I have heard that although the descendants of a sage may not hold office, they will necessarily understand [ritual]. Now, in his youth Kongzi is good at the rites, and this shows he understood ritual” (吾聞聖人之後，雖不當世，必有達者。今孔丘年少好禮，其達者歟？吾即沒，若必師之。).

31. Shi ji, 47.1938: “Kongzi used the Odes, Documents, ritual, and music in his teaching. In all, he reached three thousand people, but seventy-two personally mastered the Six Attainments” (孔子以詩書禮樂教，弟子蓋三千焉，身通六藝者七十有二人). Note that while the transmitted Shi ji has 72, the Tang zhengyi 正義 commentary specifies 77.

32. In turn, the same elements are reused in the geographical treatise of the Han shu (“Dili zhi, xia” 地理志下), 28B.1662.

33. Wang Yile and Wu Zhenhua, “Kongzi yijing chudu,” 368.

34. Here, I have departed from the transcription, presuming that the phrase “ministers kill their rulers, children kill their parents” is a set phrase, and so bi 必 might be a copyist’s error for fu 父.

35. Wang Yile and Wu Zhenhua, “Kongzi yijing chudu,” 368.

36. Wang Yile and Wu Zhenhua, “Kongzi yijing chudu,” 368.

37. Editor’s note [M. Nylan]: as it is absent from the Xunzi’s list of the Classics.

38. Hanmo’s, Zhang Authorship and Text-Making in Early China (Boston: De Gruyter, 2018)Google Scholar argues that Hu Sui’s conversation was with Sima Tan (265), and that the entirety of chapter 130 of the Shi ji was compiled by Sima Qian’s grandson Yang Yun 楊惲 (285). On Yang Yun’s role, see also below. Klein’s, Esther Sunkyung Reading Sima Qian from Han to Song: The Father of History in Pre-Modern China (Leiden: Brill, 2019), p. 32ffGoogle Scholar, offers a careful discussion of the identity of “Taishi Gong.”

39. Shi ji, 120.3285.

40. Huainanzi jiaoshi 淮南子校釋, ed. Zhang Shuangdi 張雙棣 (Beijing: Beijing daxue, 1997), 9.1010 (“Zhu shu” 主術). See The Huainanzi, edited and translated by John Major, Sarah A. Queen, Andrew Seth Meyer, and Harold D. Roth (New York: Columbia University Press, 2010), 335.

41. See Wang Gang, “‘Zhou shi mie’ yu ‘Gongyang xue’ wenti: Nanchang Haihunhou mu ‘Kongi yijing’ wen fawei” “周室烕”與《公羊學》問題：南昌海昏侯墓“孔子衣鏡”文發微, Shehui kexue zhanxian 社會科學戰線 2019.4, 87.

42. Wang Yile and Wu Zhenhua, “Kongzi yijing chudu,” 368.

43. Han shu, 62.2737. Prior to 66, serving as Bureau Head of the Left (Zuo Cao 左曹), Yang Yun was the fourth in a chain of five officials who relayed accusations against the Huo clan (Han shu, 68.2931). Shi ji, 20.1066 (“Jianyuan yilai houzhe nianbiao” 建元以來侯者年表) says that Yang Yun (d. 55) was enfeoffed as Noble of Pingtong in 61, in part as reward for his role in implicating Huo Yu 霍禹 (d. 66), brother of Empress Huo (Huo Chengjun 霍成君, d. 54), in the Huo clan’s treasonous plot. Michael Loewe explains how the opponents of the Huos were rewarded with titles in 66: “The edict granted an amnesty to all those who had been tricked by the Huo family into compliance and whose degree of complicity had not yet been reported; and nobilities were conferred on a number of men who had been concerned in revealing the plot.” Loewe, Crisis and Conflict in Han China, 104 B.C. to A.D. 9 (London: George Allen and Unwin, 1974), 139. Interestingly, the mirror texts, with clear links to numerous chapters of the Sima oeuvre preserved by the anti-Huo hero Yang Yun, were found in the tomb of an emperor deposed by the Huo.

44. Han shu, 68.2940.

45. That this emphasis on Gongyang-style readings of the Spring and Autumn is so central to the Haihunhou portrayal of Kongzi is an important aspect of this find.

46. Wang Yile and Wu Zhenhua, “Kongzi yijing chudu,” 370.

47. The justification for reading this as a title is addressed at length below.

48. Kongzi jiayu and Shi ji both specify that Yan was from the state of Lu, but the Haihunhou treatment omits this information.

49. This passage is parallel to Analects 12.1 (Lun yu jishi 論語集釋, ed. Cheng Shude 程樹德 [Beijing: Zhonghua, 1990], 24.817–24). Note the shift in the Haihunhou version from Yan Hui to Yan Yuan once it moves past this first exchange.

50. Parallel to Analects 9.11 (Lun yu jishi, 17.593–598).

51. This line does not appear in the Analects or other disciple texts.

52. Only the first part of Analects 7.11, is quoted here, and also in the Shi ji. See Lun yu jishi, 13.450–53. In the Analects, the second part of 7.11 is a somewhat unrelated dialogue between Zilu and the Master about military virtue.

53. This passage does not appear in the Analects. Shi ji, 61.2188 prefaces the saying with the words 孔子哭之慟 “Kongzi wept bitterly for him” (a version of the beginning of Analects 11.10, as in Lun yu jishi, 22.758–59: “When Yan Hui died, the Master wept bitterly for him. His followers said: “Our Master is weeping bitterly.” The Master said: “Am I? For whom should I weep bitterly if not for such a man?” (顏淵死，子哭之慟。從者曰：「子慟矣。」曰：「有慟乎？非夫人之為慟而誰為！」). Kongzi jiayu, 9.1a, adds a final assessment of Yan that includes the saying before it ends on a more general evaluation: “When he was twenty-nine, his hair turned white, and he died at the young age of thirty-one. Kongzi said, ‘Since I have had Hui, my followers have grown increasingly close.’ Hui was famous for his virtuous actions, and Kongzi praised his benevolence” (年二十九而髮白，三十一，早死。孔子曰：「自吾有回，門人日益。」回以德行著名，孔子稱其仁焉).

54. In the case of Zigong, the Shi ji narrative contains tropes of misrepresentation and shady business dealings. When a dialogue about wealth and poverty in Analects 1.15 is quoted, the second part, in which Zigong is complimented by Kongzi, is left out (70.2195). By contrast, the capsule biography from Haihunhou says nothing about Zigong’s clever speech or Analects 1.15 dialogue, and only briefly touches on his economic activities in a neutral fashion. Instead, the Haihunhou passage ends on Zigong’s effusive praise of Kongzi also found in Analects 19.25 (Wang Yile and Wu Zhenhua, “Kongzi yijing chudu,” 372). The portrayals of Zigong illustrate how despite formal similarities, the two sets of disciple biographies selectively use information to paint rather different portraits.

55. Shi ji, 67.2203.

56. See Hunter, Confucius, on the “between Chen and Cai” stories. The phrase “Zilu was serving as steward for the Ji clan” appears in the “Liqi” 禮器 chapter of the Li ji, while Analects 13.2, has the same phrase, but describes a different disciple, Zhonggong 仲弓.

57. See the “Gu xiang” 骨相 chapter, see Lunheng jiaoshi 論衡校釋, 3.123. Almost the same phrase appears in the chapter “Zilu chu jian” 子路初見 from the Da Dai Li ji 大戴禮記.

58. Wang Yile and Wu Zhenhua, “Kongzi yijing chudu,” 376.

59. Sui shu (Beijing: Zhonghua, 1973), 32.936. See a fuller discussion in Csikszentmihalyi, “Interlocutor Collections.” Qing reconstructions of Lun yu Kongzi dizi mulu are divided between thirty-eight or thirty-nine disciple versions by those scholars who believed the original text only covered the disciples who appeared in the Analects, and seventy-something disciple versions by those who believed the original text covered all who were mentioned in the Shi ji disciples chapter.

60. Especially interesting in this respect is Zheng Xuan’s citation of the title of the book, which the Sui shu says he wrote, in Pei Yin’s 裴駰 Liu Song period commentary to Shi ji, 67.2189: “Zheng Xuan said: ‘The Catalog of Kongzi’s Disciples in the Analects says he is from Lu’” (鄭玄曰：「孔子弟子目錄云魯人」).

61. Shi ji, 70. 2226. Byung-joon Kim 김병준 pointed me to this passage in connection with the Dizi ji.

62. Loewe, Michael, “Attitudes to Kongzi in Han Times,” Journal of Asian History 55.1 (2021), 1Google Scholar.

63. See Gu Shikao 顧史考 (Scott Cook), Shangbo zhushu Kongzi yulu wenxian yanjiu 上博竹書孔子語錄文獻研究 (Shanghai: Zhongxi, 2021).

64. Alexus McLeod writes that “questioning” here is a technical term for a process that “operates through presenting questions to clarify, asking for clarification of certain points of a view,” which in turn operates in tandem with a method of appraisal based on objections to the point of view. See McLeod, , “A Reappraisal of Wang Chong’s Critical Method through the Wenkong Chapter of the Lunheng,” Journal of Chinese Philosophy 34.4 (2007), 588Google Scholar. Wang Chong’s view of Kongzi is pushing back against a view that Kongzi and his disciples were more talented than people of the current age (夫古人之才，今人之才也). Reading between the lines, Wang argues against views that certain Kongzi texts were inerrant.

65. Wang Yile and Wu Zhenhua, “Kongzi yijing chudu,” 370.

66. Shi ji, 47.1935.