1. For examples, see Loewe, Michael, Records of Han Administration (Cambridge: Cambridge University Press, 1967), vol. 2, 101–3Google Scholar. See also document 6 below.

2. For one productive example, see Richter, Matthias, “Textual Identity and the Role of Literacy in the Transmission of Early Chinese Literature,” in Writing and Literacy in Early China: Studies from the Columbia Early China Seminar, ed. Feng, Li and Branner, David Prager (Seattle: University of Washington Press, 2011), 206–36Google Scholar. A binary division between “administrative” and “literary” by no means accounts for all excavated sources: “legal” texts (e.g., those from tombs at Shuihudi 睡虎地 and Zhangjiashan 張家山), and “technical” texts (e.g., the “daybooks” [rishu 日書] and medical texts) do not fit easily into either category.

3. For example, when we compare a registration record written on a wooden board found in a rubbish heap with a copy of the Laozi elegantly transcribed on silk and interred in a tomb, differences in content are by no means more salient than differences in media and archaeological context. See Giele, Enno, “Excavated Manuscripts: Context and Methodology,” in China’s Early Empires: A Re-appraisal, ed. Nylan, Michael and Loewe, Michael (Cambridge: Cambridge University Press, 2010), 114–34Google Scholar. Some silk letters have been recovered from the desert northwest, as noted in Giele, “Private Letter Manuscripts from Early Imperial China,” in A History of Chinese Letters and Epistolary Culture, ed. Antje Richter (Leiden: Brill, 2015), esp. 407–11.

4. For emphasis on this point, see Sanft, Charles, Literate Community in Early Imperial China: The Northwestern Frontier in Han Times (Albany: State University of New York Press, 2019)Google Scholar.

5. Possible links between imperial unification and the emergence of different textual and literary categories remain a matter of debate. Martin Kern has written that during the Han dynasty “edicts, petitions to the throne, and court debates” were “new forms of writing that developed together with the imperial state.” See Kern, “Early Chinese Literature, Beginnings Through Western Han,” in The Cambridge History of Chinese Literature, ed. Kang-i Sun Chang and Stephen Owen, vol. 1, To 1375 (Cambridge: Cambridge University Press, 2010), 109. These sources, however, probably had different developmental trajectories. For obvious reasons, imperial edicts arose with the establishment of the empire; the Shi ji 史記 (comp. c. early first century b.c.e.) explicitly links the two; Shi ji, comp. by Sima Qian 司馬遷 (c. 145–c. 86 b.c.e.) (Beijing: Zhonghua, 1959), 6.236. But royal commands, petitions, and court debates all had pre-imperial precedents. See David Schaberg, “Functionary Speech: On the Work of Shi 使 and Shi 史,” in Facing the Monarch: Modes of Advice in the Early Chinese Court, ed. Garrett P. S. Olberding (Cambridge, MA: Harvard University Asia Center, 2013), 22–31.

6. Editor’s note: that explains why the wenxue 文學 (“document drafters”) have often been cast as “Literary Talents” by those who have not examined their administrative status.

7. My thanks to one of the anonymous reviewers, who suggested situating administrative texts on a continuum.

8. A telling fact: according to the CHANT database, the pre-imperial Lüshi chunqiu 呂氏春秋 (comp. 239 b.c.e.) uses the word shu 45 times and ji just 5 times.

9. In particular, documents called shu 書 seem to have been sealed, while those called xi 檄 were not. For a discussion with relevant citations, see Chen Yunqing 陳韻青, “Qin Han wenshu xingzheng zhong de ‘feng’ yu ‘yin’” 秦漢文書行政中的“封”與“印,” paper presented at conference, Chutu wenxian yu Han Tang fazhi shi yanjiu 出土文獻與漢唐法制史研究, Department of History, Peking University, November 28, 2021 (the paper is under review for publication as of March 2022).

10. Or even “memo,” a possibility first suggested by Harrison Huang, whom I am happy to thank here. At the end of this essay, I speculate that perhaps the meaning of ji as “remember” or “memorize” first achieved prominence during Western Han, a point impossible to prove given the paucity of securely datable materials from the period. Still, claims in Wang Li 王力, chief editor, Gu Hanyu zidian 古漢語字典, (Beijing: Zhonghua, 2000), 1262, that the “base meaning” (ben yi 本義) of ji is “memorize” (ji zhu 記住) are overly confident. See n. 12 below.

11. To take the extant Chun qiu 春秋 (Annals) commentaries as an example, the Gongyang 公羊, Guliang 穀梁, and Zuo zhuan 左傳 typically use shu 書 to describe the act of writing. Verbal uses of ji do not occur in the Guliang and Zuo zhuan (in fact, the word does not occur at all in the former). By contrast, they are more common in the Gongyang, where the word ji is mostly confined to the standard question–answer formula used to categorize a recorded event as a “disaster” (zai 災) or “prodigy” (yi 異). The formulaic question “Why was this written?” (何以書) is answered by the equally formulaic statement “To record a disaster” (記災也) or “To record a prodigy” (記異也).

12. The number increases to just five if we include both 紀 (OCM *kəʔ) and 記 (OCM *kəh), which partially overlap in the Zuo zhuan, not surprising given their cognate relationship. The latter word appears only three times, all in the Guan Zhong speech discussed here. The word ji 紀 usually refers to the name of a realm and secondarily as “guiding line” or “regulation,” with occasional verbal uses as “regulate” or, once, “calculate.” Wolfgang Behr noted that 記 probably derived from 紀, which itself emerged from the root li 理 (OCM: rəʔ; “to divide, regulate, mark”). See Behr, “Language Change in Premodern China: Notes on its Perception and Impact on the Idea of a ‘Constant Way,’” in Historical Truth, Historical Criticism, and Ideology: Chinese Historiography and Historical Culture From a New Comparative Perspective, ed. Helwig Schmidt-Glintzer, Achim Mittag, and Jörn Rüsen (Leiden: Brill, 2005), 17. Schuessler, Axel, ABC Etymological Dictionary of Old Chinese (Honolulu: University of Hawai‘i Press, 2006), 298Google Scholar, presented a similar understanding. Cf. Wang Li, Gu Hanyu da zi dian, 1262 (see n. 10 above).

13. Chun qiu Zuo zhuan zhu 春秋左傳注, annotated by Yang Bojun 楊伯峻 (Beijing: Zhonghua, 1981 [2009]), vol. 1, 318 (Xi 僖 7.3); Zuo Tradition / Zuozhuan / 左傳, translated and introduced by Stephen Durrant, Wai-yee Li, and David Schaberg, 3 vols. (Seattle: University of Washington Press, 2016), vol. 1, 287.

14. See Zuo Tradition, trans. Durrant et al., vol. 2, 1075, with supporting examples.

15. The most famous example is the Chun qiu, an annal associated with the state of Lu 魯. The topic of pre-imperial annals is controversial and the limited evidence requires careful interpretation. According to the Shi ji, the Chancellor Li Si 李斯 famously argued that the “archival office” (shi guan 史官) should burn anything that was not a “Qin record” (Qin ji 秦記) though the passage does not explain the term’s meaning (Shi ji, 6.255). The best essay on the book burning story remains Jens Østergaard Peterson, “Which Books Did the First Emperor of Ch’in Burn? On the Meaning of Pai Chia in Early Chinese Sources,” Monumenta Serica 43 (1995), 1–52. The “Fei ming xia” 非命下 chapter of the Mozi 墨子 speaks of “records of Shang, Zhou, and Xia” (商周虞夏之記), but we cannot automatically understand this statement as a reference to annals in royal courts.For an overview of the chief annals in pre-imperial times, see Yuri Pines, “Chinese History Writing Between the Sacred and the Secular,” in Early Chinese Religion, Part One: Shang Through Han (1250 BC–AD 220), ed. John Lagerwey and Marc Kalinowski, 2 vols. (Leiden: Brill, 2009), vol. 1, 316–23. There is no consistent terminology for “archives” or “libraries,” let alone a clear distinction between the two kinds of institutions, in the early Chinese or ancient Mediterranean sources. See Nylan, Michael, “On Libraries and Manuscript Culture in Western Han Chang’an and Alexandria,” in Ancient Greece and China Compared, ed. Lloyd, G. E. R. and Zhao, Jingyi Jenny (Cambridge: Cambridge University Press, 2018), 373–409CrossRefGoogle Scholar. For shi 史 as “archivist,” see Durrant, Stephen, Li, Wai-Yee, Nylan, Michael, and van Ess, Hans, The Letter to Ren An & Sima Qian’s Legacy (Seattle: University of Washington Press, 2016), esp. 18–21Google Scholar.

16. Lüshi chunqiu zhuzi suoyin 呂氏春秋逐字索引, Institute for Chinese Studies Concordance (Hong Kong: Commercial Press, 1996), 22.6/149/6–9 (“Cha zhuan” 察傳). The translation is mine, but see The Annals of Lü Buwei, trans. John Knoblock and Jeffrey Riegel (Stanford, CA: Stanford University Press, 2000), 584.

17. The story says only that Zixia “asked about it” (wen zhi 問之), but does not describe anybody actually consulting an annal or chronicle stored by the Jin court.

18. See Lüshi chunqiu, 11.2/54/4–5 (“Zhi zhong” 至忠); The Annals of Lü Buwei, 245. The Shuoyuan 説苑 (comp. c. 17), “Li jie” 立節 chapter (juan 4), contains this anecdote, though it has the king shooting a pheasant, not a rhinoceros. Shuoyuan 説苑 (Han Wei congshu 漢魏叢書 ed.), 12.1–2; accessed via Scripta Sinica database, http://hanchi.ihp.sinica.edu.tw.

19. A close parallel occurs in the “Jin teng” 金縢 chapter of the Shang shu 尚書 (Documents). That narrative similarly describes a seemingly transgressive act, which proves to be in fact a display of the highest loyalty after the king and his officials consult a record stored away in a secure location (in this case, not a “storeroom” but a “metal-bound coffer” [金縢之匱]). See Shangshu zhu shu 尚書注疏 (Shisan jing zhushu 十三經注疏 ed., 1815), juan 13, 187-2; accessed via Scripta Sinica database, http://hanchi.ihp.sinica.edu.tw. In the “Jin tang,” the official under suspicion is Zhougong 周公 (the Duke of Zhou).

20. Durrant et al., The Letter to Ren An, 20.

21. Lüshi chunqiu, 13.6/67/23; The Annals of Lü Buwei, 295.

22. Jia Yi Xinshu zhuzi suoyin 賈誼新書逐字索引, Institute for Chinese Studies Concordance (Hong Kong: Commercial Press, 1996), 5.1/33/19–20 (“Fu zhi” 傅志). As the concordance editors noted (see n. 13), the actual text reads gu she 故設, but she here is almost certainly corrupt, since a parallel passage from Guoyu 國語 uses the term gu zhi 故志 (zhi and ji being synonyms). The Kongzi jiayu 孔子家語 contains a version of the story from the Lüshi chunqiu about Zi Xia’s interpretation of “three pigs” (see above) in which the questionable phrase comes from a text called shi zhi 史志 (not shi ji 史記, as in the Lüshi chunqiu).

23. See Guo Qingfan 郭慶藩, Zhuangzi ji shi 莊子集釋 (Beijing: Zhonghua, [1961] 2014), vol. 2, 411 (juan 12, “Tian di” 天地).

24. E.g., in the “Tian dao” (Way of Heaven) chapter, Wheelwright Bian criticizes Huan Gong 桓公 for being too obsessed with old books, the mere “dregs” (zao po 糟魄) of dead sages. Zhuangzi ji shi, 2.493–4.

25. To quote William Hung, “I have not calculated all of the so-called ji because there are so many different kinds that it is difficult to determine their total number” (所謂記無算者，以其種類多而難計其數者). Hung was referring primarily to ji associated with ritual texts, but the statement reflects the larger diversity of ji in Han sources. See Hong Ye 洪業 (William Hung), “Liji yinde, ‘Xu’” 禮記引得序, Shixue nianbao 史學年報 2.3 (1937), 288b.

26. Han shu 漢書, Ban Gu 班固 (32–92 c.e.) et al. (Beijing: Zhonghua, 1962), 30.1765.

27. Han shu, 30.1707.

28. Han shu, 30.1713.

29. Han shu, 30.1709.

30. According to the titles, the first of these “traditions and records” on the wuxing was by Liu Xiang 劉向 (79/78–8 b.c.e.) and the second by the Shang shu expert Xu Shang 許尚. Note that the “Wu xing zhi” (Treatise on the Wuxing) in the Han shu regularly cites zhuan, marked by “a tradition states” (zhuan yue 傳曰), but the phrase “a record states” (ji yue) occurs only once.

31. The Chun qiu category of the “Yi wen zhi” contains two texts with nian ji in the title: (1) Taigu yilai nian ji 太古以來年紀 (Chronology Since High Antiquity); (2) Han da nian ji 漢大年紀 (Great Chronology of the Han). Texts in the Chun qiu category with ji 記 (“records”) in the title include two texts related to the Gongyang commentary to the Chun qiu and one text entitled Han zhu ji 漢著記 (Notable Records of the Han), which Yan Shigu’s commentary likens to a “record of daily life” (qi ju zhu 起居注) from his own era during the Tang (Han shu, 30.1713–14). The Shi ji and Han shu themselves used the word ji 紀 to refer to the pre-unification annals of ruling houses (only in the Shi ji) and to individual emperors (in Han shu as well). A clearer division between 紀 and 記, with the former referring more specifically to texts arranged in an explicitly chronological sequence, seems to mark a change from pre-imperial times. Perhaps the Lüshi Chunqiu played a role in this change, since the titles of the first twelve sections of that text, each referring to a month of the year, end with the word ji 紀. Note, however, that only two other texts in the “Yi wen zhi” contain ji 紀 in their titles, and neither can be unambiguously understood as annals of ruling houses. The two titles are Chen Shou Zhou ji 臣壽周紀 (Zhou Records by Official Shou) (Han shu, 30.1745); and Zi gu wu xing xiu ji 自古五星宿紀 (Annals of the Five Planets and the Lunar Lodges from Ancient Times) (Han shu, 40.1766).

32. See Colin McKenzie, Bibliography and the Sociology of Texts (Cambridge: Cambridge University Press, 1999).

33. For instance, in his annotation to a Shi jing 詩經 poem, Zheng Xuan 鄭玄 (127–200 c.e.) quoted a text that he called Liji 禮記, but the quoted statement is found in the Yili 儀禮 (i.e., not the received Liji). Furthermore, the quote is from the chapter “Shao lao kui shi” 少牢饋食, which, unlike other Yili chapters, does not contain a section called ji. See Ye, Hong, “Liji yinde, ‘Xu,’” 282a; and Xing Wen, “New Light on the Li ji 禮記: The Li ji and the Related Warring States Period Guodian Bamboo Manuscripts,” Early China 37 (2014), 528–29Google Scholar. For a broader discussion of jing and textual authority during Han, see Michael Nylan “Classics Without Canonization: Learning and Authority in Qin and Han,” in Early Chinese Religion: Part One, ed. Lagerwey and Kalinowski, vol. 2, esp. 721–76.

34. For a detailed discussion, see Yamada Toshiaki 山田利明, “Yili zhong ‘ji’ de wenti—guan yu Wuwei Han jian” 儀禮中“記”的問題—關於武威漢簡, trans. Diao Xiaolong 刁小龍, in Zhang Huanjun 張煥君 and Diao Xiaolong, Wuwei Han jian ‘Yili’ zhengli yu yanjiu 武威漢簡「儀禮」整理與研究 (Wuhan: Wuhan daxue, 2009), 332–51.

35. No specific claim is made here about whether this shift necessarily implies an increase in literacy rates, a complicated problem requiring separate treatment. Two studies that address literacy in relation to Han administrative documents from the northwest are Robin Yates, “Soldiers, Scribes, and Women: Literacy Among the Lower Orders in Early China,” in Li and Branner, Writing and Literacy in Early China; and Sanft, Literate Community in Early Imperial China. But the evidence is too scanty and too disputed to posit literacy rates, let alone upward and downward changes in literacy rates.

36. For an introduction, see Yates, Robin, “The Qin Slips and Boards from Well No. 1, Liye, Hunan: A Brief Introduction to the Qin Qianling County Archives,” Early China 35–36 (2012–13), 291–329Google Scholar. See also Korolkov, Maxim, The Imperial Network in Ancient China: The Foundation of Sinitic Empire in Southern East Asia (London: Taylor and Francis, 2021), esp. 1–34 (Introduction)CrossRefGoogle Scholar.

37. The point receives detailed treatment in Takamura Takeyuki 高村武幸, Shin Kan kandoku shiryō kenkyū 秦漢簡牘史料研究 (Tokyo: Kyūko Shoin, 2015), 159–86. As Takamura notes, many kinds of administrative documents known from Han are absent in the published Liye corpus, and Takamura argues that this picture probably will not change significantly even as more of the Liye documents are published. I have confirmed Takemura’s assertion that the word does not appear in the transcriptions and photographs of the boards contained in the two volumes published so far of the Liye corpus. See Hunan sheng wenwu kaogu yanjiu suo 湖南省文物考古研究所, Liye Qin jian (yi) 里耶秦簡（壹） (Beijing: Wenwu, 2012) and Liye Qin jian (er) 里耶秦簡（貳） (Beijing: Wenwu, 2017). Of course, full publication of all legible strips is needed before reaching a definitive conclusion.

38. Several boards refer to people (or the same person?) named Qiuwu 囚吾, serving in different official posts. See 8-681, 8-1610, and 8-1783+8-1852, identified in You Yifei 游逸飛 and Chen Hongyin 陳弘音, “Liye Qin jian bowuguan cang di jiu ceng jiandu shiwen jiaoshi” 里耶秦簡博物館藏第九層簡牘釋文校釋, Jianbo wang 簡帛網, December 22, 2013, www.bsm.org.cn/?qinjian/6146.html, accessed on August 30, 2022.

39. Liye Qin jian (yi), 41 (8-167), 70 (8-472), 137 (8-1011). My reconstruction and translation of the letter is indebted to the transcription and annotations provided in Chen Wei 陳偉, ed., Liye Qin jiandu jiaoshi 里耶秦簡牘校釋 (Wuhan: Wuhan daxue, 2012), 101. The transcription partially follows Chen Wei, but omits 8-194, since that small fragment does not seem to fit physically with the other three, larger fragments and disrupts the letter’s flow.

40. For instance, is the theft connected to the boat or the transported minor official (zushi 卒史)? Or were Kuan and Qiuwu pursuing robbers unrelated to the boat, which then had to be brought back to Qianling County by other means, prompting Commandant Jing to write to Assistant Magistrate Gong? For the latter, see Zhu Shengming 朱聖明, “Liye Qin jian suo jian Qin dai Qianling xian gong chuan xiangguan wenti yanjiu” 里耶秦簡所見秦代遷陵縣公船相關問題研究, Gudai wenming 8.2 (Apr. 2014), 48.

41. Other common phrases in early imperial letters include zuxia 足下 (at [somebody’s] feet) and fudi 伏地 (prostrate). For a useful table of common epistolary terms, see Enno Giele, “Private Letter Manuscripts from Early Imperial China,” in A History of Chinese Letters and Epistolary Culture, ed. Antje Richter (Leiden: Brill, 2015), 425–26. Secondary studies refer to this document as a letter. See, e.g., Zhu Shengming, “Liye Qin jian suo jian,” 48.

42. Giele, “Private Letter Manuscripts,” 403–4, notes problems with the category of “private,” but observes that “personal” is perhaps no better. Individual officials, after all, could exchange letters about government business; indeed, Jing’s letter to Gong seems to be a good example.

43. Commandant Jing’s letter seems to follow a pattern evident in the Liye manuscripts: county magistrates (the highest official in the local bureaucracy) and assistant county magistrates channeled communications from lower officials to other counties or to Dongting commandery headquarters. Jing would not have been able to send such a request directly to the magistrate of a different county of Youyang. See Zhu Shengming 朱聖明, “Qin dai difang guanyuan de wenshu chuandi zhiquan—yi Liye Qin jian yidi tongji wenshu wei zhongxin de kaocha” 秦代地方官員的文書傳遞職權—以里耶秦簡異地同級文書為中心的考察, Nandu xue tan (Renwen shehui kexue xue bao) 38.1 (2018), 31–39.

44. The calligraphy of the letter appears hasty when compared with official documents from Liye (e.g., 8-135A+B, also about a missing boat, but a summary of a formal inquiry), or even in the letter between friends that Giele translated (for images, see “Private Letter Manuscripts,” 465).

45. My discussion of excavated ji has benefited from: Ukai Masao 鵜飼昌勇, “Kan dai no bunsho ni tsuite no yichi kōsatsu—‘ki’ toyiu bunsho no sonzai” 漢代の文書についての一考察−「記」という文書の存在, Shisen 9 (1988), 18–30; Lian Shaoming 連邵名, “Xi yu mu jiang zhong de ji yu xi” 西域木簡中的記與檄, Wenwu chunqiu (1989), (Z1), 21–27, 69; Nakayama Shigeru 仲山茂, “Kan dai ni okeru chōri to zokuri no aida” 漢代における長吏と属吏のあいだ, Nihon Shin Kan shi gakkai gakuhō 3 (2002), 13–42; Sumiya Tsuneko 角谷常子, “Kandoku no keijō ni okeru yimi” 簡牘の形状における意味, in Henkyō shutsudo mokkan no kenkyū 邊境出土木簡の研究, ed. Tomiya Itaru 富谷至 (Kyoto: Hōyu, 2003), esp. 98–104; Takatori Yūji 鷹取祐司, “Kan kan shoken bunsho kō—sho, keki, ki, fu” 漢簡所見文書考--書⋅檄⋅記⋅符, in Tomiya Itaru, Henkyō shutsudo mokkan no kenkyū, 119–60; Fujita Takao 藤田高夫. “Kan ki gōshiki” 官記偶識. Kansai daigaku bungaku ronshū 56.2 (2006), 39–51; and Takamura Takeyuki, Shin Kan kandoku, esp. chaps. 1–2.

46. During Qin and Han, county commandants and assistant county magistrates had the same salary rank (Han shu, 19a.742), so Jing and Gong in theory would have been relative equals. Zhu Shengming, “Qin dai difang guanyuan,” 34, notes that the Liye documents show the assistant county magistrate using gao to address the commandant on one strip (8-69), but on another strip (9-112) using gan gao with the commandant and gao with the county bailiff (sefu 嗇夫).

47. Qianqiu was a squad under the Yumen 玉門 company, which reported to the Yumen commandant of Dunhuang commandery. So Qianqiu was being ordered to go from the squad to the commandant’s bureau. See Bai Junpeng 白軍鵬, Dunhuang Han jian jiaoshi 敦煌漢簡校釋 (Shanghai: Shanghai guji, 2018), 17. Zhang Defang speculates that the strip refers to a storage facility of some kind in Qianqiu. See Dunhuang Maquanwan Han jian jishi 敦煌馬圈灣漢簡集釋, ed. Zhang Defang 張得芳 (Lanzhou: Gansu wenhua, 2013), 624.

48. Dunhuang Maquanwan Han jian jishi, 148 (excavation no. 79.DMT12:31).

49. The latter was located near the Xiaofangpan 小方盤 site. For the two locations, see Wu Rengxiang 吳礽驤, Hexi Han sai diaocha yu yanjiu 河西漢塞調查與研究 (Beijing: Wenwu, 2005), map 12.

50. Dunhuang Maquanwan Han jian jishi, 160 (excavation no. 79.DMT12:108).

51. Jiandu zhengli xiaozu 簡牘整理小組, ed., Juyan Han jian 居延漢簡, vol. 2 (Taipei: Zhongyang yanjiuyuan Lishi yuyan yanjiusuo, 2015), 23.

52. The pattern in these texts is for guan 官 to refer to the “office of the company” (hou guan 候官), while fu refers to the “bureau of the commandant” (wei fu 尉府) (see later examples below).

53. My interpretation here partially depends on observed differences in handwriting, which are subtle but I believe significant. These include the different manner of writing si 四 between recto and verso, while the two zhe 者 on verso are obviously and substantially different.

54. Fujita Takao, “Kanki gōshiki,” 41. The phrase ai lian 哀憐 appears in other documents, including letters. See, e.g., Juyan Han jian, vol. 2, 138, 157.10A+B.

55. Fujita understands the notification from the company on recto and the statement about the runner on verso to have been written at the same time by the company issuing the notification; Fujita, “Kanki gōshiki,” 41. Fujita worked, however, with an older volume of the Juyan strips, which did not fully transcribe all characters on lower left recto and lower right verso. The newer volumes from the Institute of History and Philology with revised transcriptions are more detailed and thus provide information unavailable to Fujita. I have three outstanding questions: First, why would a different company be able to send the Fourth Company runner, unless the Fourth Company runner was somebody in the home company responsible for making deliveries to the Fourth Company? Second, why are there subtle, but evident differences in calligraphy between recto and verso (see n. 53 above)? Third, how did the runner know which document to deliver? Perhaps this request was communicated orally.

56. The dot after the character pi 匹 suggests that the entries for chariots and horses constitute one unit and should be read together. There is no evidence of a dot between chariots and horses.

57. Jianshui jinguan Han jian 肩水金關漢簡, ed. Gansu sheng bowuguan 甘肅省博物館, et al., (Shanghai: Zhongxi, 2012), vol. 2, bk. 2, 163 (black and white photograph).

58. For discussion of a four-sided ji, sent from the commandant headquarters and calling upon subordinate companies to search for men previously sent out to deliver an order, see Takatori Yūji, “Kan kan shoken bunsho kō,” 127.

59. Juyan Han jian, vol. 3, 242 (293.2+293.1).

60. Compare, for instance, the rather sloppy fu 府 in document 3 (recto) with its counterpart in document 5 (recto).

61. Jianshui Jinguan, vol. 4, bk. 2, 280 (black and white photograph). I have not found any studies of this strip.

62. Fujita, “Kanki gōshiki,” argues that the notched ji he analyzed, which feature gaps between the upper and lower portions of the text, were all bound and sealed. All the strips he discusses, however, are notched in the middle, not toward the top, unlike document 7.

63. Juyan Han jian, vol. 2, 210.

64. A point also noted in Fujita, “Kanki gōshiki,” 48.

65. Antje Richter, Letters and Epistolary Culture in Early Medieval China (Seattle: University of Washington Press, 2013), 77–78.

66. Sanft, Literate Community in Early Imperial China, 152, observed that documents using baiji 白記 often (though not exclusively) also indicate the full name and surname of the writer. Richter, Letters and Epistolary Culture, 78, notes that bai alone in letters seems to be “hierarchically neutral,” being used by both superiors and inferiors.

67. Jianshui Jinguan, vol. 5, book 2, 308 (black and white photograph).

68. Jianshui Jinguan, vol. 5, book 2, 164 (black and white photograph).

69. Jianshui Jinguan, vol. 5, book 2, 103 (black and white photograph).

70. Giele, “Private Letter Manuscripts,” provides a brief overview of the Tianchang find. For the excavation report, see Tianchang shi wenwu guanli suo 天長市文物管理所 and Tianchang shi bowuguan 天長市博物館, “Anhui Tianchang Xi Han mu fajue jianbao” 安徽天長西漢墓發掘簡報, Wenwu 2006.11, 4–21. For a detailed discussion of the letter translated here, see Yang Zhenhong 楊振紅, “Jizhuang Han mu ‘Ben Qie’ shu du de shidu ji xiangguan wenti” 紀莊漢墓「賁且」書牘的釋讀及相關問題, Jianbo yanjiu (2009 [2011]), 1–13.

71. In other words, Ben Qie spent just two days in Luoyang.

72. The excavation report transcribed the character after fa 發 as bing 兵 (“army”), but Yang Zhenhong, following another interpretation, rendered it as yu 與.

73. The meaning of zuo you 左右 is unclear. It could be a reference to Ben Qie himself (He Youzu’s interpretation), Ben Qie’s attendants (argued by Yang Zhenhong), Meng (Antje Richter, private communication, April 2014), or just “unnamed parties” (Michael Loewe, private communication, March 3, 2015). Given that Ben Qie immediately mentions members of his family, I suspect that zuo you might refer to family members traveling with him who had their own histories of bad behavior and thus know how to act with care when being grilled.

74. Following the interpretation of He Youzu 何有祖, who glosses gu as shu 屬. See He Youzu, “Anhui Tianchang Xi Han mu suo jian Xi Han mu du guankui” 安徽天長西漢墓所見西漢木牘管窺, Bamboo and Silk Manuscripts (bsm.org.cn), December 12, 2006, www.bsm.org.cn/?hanjian/4703.html, accessed March 18, 2022.

75. For the image, see “Anhui Tianchang Xi Han mu fajue jianbao,” 21, image #26 (verso), #27 (recto). The transcription follows that provided in Yang Zhenhong, “Jizhuang Han mu ‘Ben Qie’ shu du de shidu ji xiangguan wenti.”

76. The phrase is famously used at the very end of the “Letter in Reply to Ren An” (Bao Ren An shu 報任安書). See Han shu, 62.2736. It is perhaps a variant of the famous saying from the “Xi ci zhuan” 繫辭傳 chapter of the Yi jing 易經: “Words do not fully convey intent; writing does not fully convey words” (言不盡意，書不盡言).

77. If “zou ji” were submitted to the ruler (to date we have no instances of this), we would translate as memorialized notes. I hope to discuss this in a future publication.

78. The exception is Gu Yong’s 谷永 (d. 8 b.c.e.) memorial to Wang Feng 王鳳 (d. 22 b.c.e.) (Han shu, 85.3454). At the time, however, Wang Feng was in charge of the government. Submitting a document to Wang, the uncle of Chengdi, might not have been too dissimilar from giving a document to the emperor or empress dowager.

79. See, e.g., Han shu, 51.2338 (jian); 72.3061 (jie); 76.3220 (jian); 85.3454 (xie).

80. Han shu, 78.3284.

81. Shi ji, 123.3162. The phrase could also be rendered as “written notations,” perhaps as a way to refer to the foreign writing script. I thank one of the anonymous reviewers for this suggestion.

82. Han shu, 93.3731. It is unclear from the story whether the shu ji 書記 and shu mentioned in this story are the same thing. My reading here takes shu ji as the actual documents exchanged, with shu referring to Chunyu Zhang’s “writings” or his writing style. A story in the “Wai qi zhuan” chapter, however, mentions people tong shu 通書 (circulating letters) (see Han shu, 97b.3983). Elsewhere, we read that Jia Shan “waded and hunted through writings and notes” (涉獵書記). Because of this “wading,” Ban Gu writes that Jia was a man who who “could not be considered a pure classicist” (不能為醇儒). Perhaps the compound term shu ji helped underscore, in pejorative terms, this varied background of learning. See Han shu, 51.2327.

83. Note, however, that in the Han shu the texts of edicts are prefaced only by the word zhao. The word zhao shu occurs only in quoted statements in which the speaker refers to an edict.

84. For Xie Guang’s memorial, see Han shu, 97b.3990–96. Wilber translated zhao ji as “private edict,” without making all of the distinctions given here. See C. Martin Wilber, Slavery in China During the Former Han Dynasty (Chicago: Field Museum of Natural History, 1943), 424–32.

85. Han shu, 97b.3991.

86. To write a response on the back of an official edict (zhao) seems unimaginable.

87. See, for instance, examples from a presentation by a Smithsonian archivist: Mitch Toda, “The Evolution of the Memo,” November 17, 2011, www.slideshare.net/SIArchives/evolution-of-the-memo, accessed March 17, 2022.

88. Liezi jishi 列子集釋, ed. by Yang Bojun 楊伯峻, (Beijing: Zhonghua, 2016), 3.115–16 (“Zhou Mu wang” 周穆王). As Yang notes, some versions of the received Liezi end the anecdote with ji zhi 記之 and others ji zhi 紀之. Needless to say, the persistent interchangeability of the characters long after the Han is a reminder that any distinctions between the two made in the Han shu “Yiwen zhi” had a limited impact.