Of Wounded Bodies and the Old Manchu Archive: Documenting Personnel Management in the Early Manchu State

This article examines the only bilingual component in the Old Manchu Archive— the record of 228 officers and soldiers’ battle wounds, as well as Nurhaci’s decision to reward them after he became khan in 1616. Departing from previous scholarship that highlights the historiographical impulse that shaped these preconquest documents, I argue that the compilation of archival materials (dangse) under the nascent Latter Jin state was also motivated by the need to recognize the meritorious deeds of military personnel and manage their ranks and privilege in writing. The enigmatic presence of a Manchuonly copy of the battle wound documents, which bears marginal notes concurrent with its active use as a ledger, indicates the rise of a more standardized system of military administration and the ascent of Manchu script as the dominant language for documentation.

This article examines the only bilingual component in the Old Manchu Archive-the record of 228 officers and soldiers' battle wounds, as well as Nurhaci's decision to reward them after he became khan in 1616. Departing from previous scholarship that highlights the historiographical impulse that shaped these pre-conquest documents, I argue that the compilation of archival materials (dangse) under the nascent Latter Jin state was also motivated by the need to recognize the meritorious deeds of military personnel and manage their ranks and privilege in writing. The enigmatic presence of a Manchu-only copy of the battle wound documents, which bears marginal notes concurrent with its active use as a ledger, indicates the rise of a more standardized system of military administration and the ascent of Manchu script as the dominant language for documentation.
負傷的身體與 《舊滿洲檔》： 早期滿洲國家形成中的人事紀錄 邊和 普林斯頓大學歷史系及東亞研究系 摘要 本文考察《舊滿洲檔》中僅見的一批滿漢雙語文書及其滿文複本("寒字 檔"1-42葉)。雙語文書中記載了二百二十八名軍官及士兵在作戰中負傷的類 型及數目，以及努爾哈赤給予他們的賞賜。以往研究中多強調後金及清代修 訂國史對檔案形態及內容的影響，而本文意在指出後金檔冊的形成亦出於軍 事化人事管理之需要。此外，雙語文書的滿文複本存在大量對數目字紀錄的 塗改，顯示該文書曾被作為軍功簿頻繁使用，亦說明後金國家的人事管理逐 漸趨於標準化、滿文的應用趨於普遍化。 conquest Qing history. 5 In recent years, new research on the Old Manchu Archive was greatly enhanced by the publication of a new, much clearer facsimile edition by the National Palace Museum in 2005, under the slightly different title Manwen yuandang (MWYD), as well as the First Historical Archive's publication of the Qianlong-era compilers' annotated version that reflected their interpretations of the old script (Neige cangben Manwen laodang, hereafter NGMW for Neige Manwen). 6 This study primarily cites the 1969 JMZD due to its wider availability but uses images from the 2005 MWYD, with references to MWRT and NGMW, when discussing the rendering of pre-conquest language into standard Manchu.  The Manchuness of the Old Manchu Archive, however, was not taken for granted upon the discovery of those documents in 1931. In one of the earliest published images of the records, Xie Guozhen 謝國楨 (1901Guozhen 謝國楨 ( -1982, then a young scholar working at the National Library in Beijing, noted the existence of "parallel Manchu and Chinese texts" 滿漢文並行 in the set (see Figure 1). Examining the originals together with renowned historian Chen Yinke 陳寅恪  and Feng-kuan 奉寛 (b. 1876), a former Mongol bannerman and scholar who conducted extensive research on the Grand Secretariat collection, Xie highlighted the unconventional Chinese rendering of important tribe names such as Yehe and Ula (葉赫作拽黑、吳喇作兀剌). 7 Chen Chieh-hsien 陳捷先, in his editorial introduction to the 1969 facsimile edition of JMZD, confirmed that the only volume containing bilingual documents was number 16 of the whole set, known as the Hanzidang 寒字檔 (indexed with the Chinese character han by the Qianlong-reign editors according to the order of the sequence of characters in Qianziwen 千字文 [Thousand character essay], a common bibliographical ordering tool) and dated to Tianming 9 (1624), since they were bound together with other documents composed that year. Chen also noted that many Chinese characters in the Hanzidang appeared quite "vulgar" (cusu 粗俗), even erroneous (biezi 別字). 8 The Chinese parts of the Hanzidang were eliminated in the Qianlong standard Manchu edition, as the eighteenth-century editors focused primarily on increasing the legibility of the Manchu text. In this article, I use the bilingual documents of Hanzidang to highlight the transition from a polyglot culture of documentation to standardized Manchu under Nurhaci's reign. 9 As discussed elsewhere by Pamela Crossley and others, the idea of a Manchu state did not coalesce until the 1630s, and the processes by which a majority population of different ancestries and linguistic capabilities subscribed to the language of a minority elite are worth further investigation. 10 The bilingual records of the Hanzidang are particularly important in this regard.
Unlike the majority of entries in the Old Manchu Archive, Nurhaci's edicts (ejehe) recorded in folios 1-28 of the Hanzidang are not chronicles that documented the khan's day-to-day business. Instead, they are undated, recording battle wounds of officers and soldiers as well as the reward and privilege they received in return. Furthermore, the 7. Xie Guozhen 謝國楨, Qing kaiguo shiliao kao 清開國史料攷 (Beiping: Guoli Beiping tushuguan, 1931), Supplement, 1a-1b. Xie also considered the rapid growth of Jianzhou Jurchens from a small tribe to a dominant regional power as an event of world historical significance, on the same level as the astonishing rise of Meiji Japan in the late nineteenth century 8. Chen Chieh-hsien 陳捷先, "Jiu Manzhou dang shulüe" 舊滿洲檔述略, in JMZD, vol.  11 Building on Matsumura's work, I compare Parts A and B in greater detail and offer some tentative explanations as to why such a non-identical duplicate was made and preserved. My primary argument is that the peculiar features of Hanzidang's bilingual documents reflect the importance of managing rank-and-file officers during the formative period of the Latter Jin state. Historians have used the Old Manchu Archive to elucidate the intricate power game among Jurchen leadership, revealing bitter internecine fighting among imperial relatives, and the rituals designed the manage them. 12 Folios 1-41 of Hanzidang shed light on the similarly pressing need to use documentation to effectively administer mid-and lower-ranking military personnel. Here I borrow insights from Katō Naoto's 加藤直人 recent work on the pre-Conquest "escapee archives" (taoren dang 逃人檔), and his discussion of the specific historical conditions that created the archival documents (dangse) and negotiated their functions. 13 The acts of "writing" (Ma. arambi, meaning "to write," also "to make") and recording (Ma. ejembi, related to the act of rulers [ejen]) archival documents constituted a vital part of what Thomas Conlan, in reference to state-building in Japan, has called the "organizational technology" of a fledgling regime. Comparisons of the Latter Jin state with other militarized societies in early modern Northeast Asia should be further explored. 14 From the perspective of the rank-and-file members of the Jurchen state, the existence of bilingual documents in the Hanzidang should not be read as the coexistence of two clear-cut identities (Manchu vs. Chinese), but as a product of a polyglot vernac-ular culture intelligible to the "cultural chameleons" who flocked, voluntarily or not, to join the Manchu cause in the 1610s and 1620s. 15 In the broader Liaodong region, the militarizing communities only gradually took on a more uniform outlook bearing many signature institutions of Manchu administration. A second argument of this paper uses the privilege and ranks granted in the Hanzidang documents to highlight the indebtedness of early Manchu military administration to existing Ming ranks and rules of granting exemption of service to elite members of the society. Tellingly, the edicts do not mention banners (gūsa) as units of organization, but companies (niru) only in Part B, not in Part A. Contrary to how Qing dynastic history depicted the pre-conquest period, the Hanzidang documents as well as their duplicate copy reveal how Manchu military administration took shape as a malleable system situated at the confluence of Ming and Mongol military cultures. 16 The final sections of this paper look beyond the Hanzidang to discuss the overall significance of archive-making (dangse arambi) in the peculiar context of pre-Conquest Manchu state, building on Devin Fitzgerald's discussion of the neologism dang'an in this era. 17 Rising imperial ambitions under Hong Taiji's rule in the late 1620s and 1630s created what Giovanni Stary called the "autochthonous historiography" of the Jurchens. 18 This was the period when earlier documents were systematically scrutinized and, in many cases, heavily edited and redacted, by the newly founded Historiographical Office (Guoshi guan 國史館) in Mukden, motivated by the anticipated compilation of dynastic history. It is thus quite remarkable how the battle wound registers in the Hanzidang survived this round of scrutiny to remind us of the administrative functions of archive-making and the persistent power of documentation to shape lives beyond the upper strata of government.

Feye / Beye: Wounds and Bodies
The twenty-two edicts (ejehe) recorded on the twenty-eight pages of Part A of Hanzidang document the type, number, and geographical location of battle wounds for a total of 228 individuals. Each edict starts with "the Khan says" (Han hendume), line elevated, and proceeds to discuss battle wounds and corresponding rewards for two to seventeen individuals (Table 1). Each group begins with a higher-ranking official who received the highest reward, albeit without calling the person the "captain of the company" (nirui ejen), and the terms gūsa or niru never appear in the edicts. With two exceptions (edicts 6 and 12), each edict starts on a new piece of paper, even though it meant leaving considerable blank space at the end of the previous edict. It appears to be the case that the khan (Nurhaci) was going through the company-like groups one-byone, reading from an existing register compiled for the occasion. 19 Parallel Chinese translations only appear for the description of battle wounds and the reward, but not the khan's personal remarks or additional notes in which the rules of reward are explained (see the next section). In many cases, the Chinese translation of the first entry in an edict appears at the very end of the previous edict, indicating that the Chinese translations were added after Manchu edicts were copied and compiled into a booklet. There are also clear traces of editing and deletion affecting both Manchu and Chinese text, which suggests that both languages were read and edited in a parallel manner.
At first glance, the choice of Chinese characters in rendering personal and place 19. Folio numbers of the Hanzidang here follow MWYD,vol. 4,. In JMZD, the facsimile of Hanzidang corresponds to vol. 4, 1695-1837. Note that folio 99 is reproduced in MWYD but not JMZD. names is frequently quite vulgar, as if the scribe or translator did not care at all whether names looked good or decorous in written Chinese (e.g. Ma. Otonggo written as Ch. 惡通果, using the character e 惡, "bad, offensive, etc."). Homonyms (for example lu 錄 and lu 祿, zhen 陣 and zhen 鎮, shang 傷 and shang 上) were used interchangeably, suggesting that the scribe knew Chinese well as a spoken language but sometimes was uncertain or indifferent with regards to the use of specific characters. After writing out a place name such as Ula, the translator would often add a side note indicating the term as a "place name" (diming 地名), which suggests a certain lack of familiarity with these Chinese terms. At the same time, however, the translator or scribe also demonstrated considerable literacy in a popular sense, even to the extent of overdramatizing what was appropriate to say. For example, the Manchu term han (khan) was always rendered in Chinese as Huangdi 皇帝 (emperor), even though technically Nurhaci had not yet assumed any imperial persona yet. In Manchu, the Khan simply "says" (hendume), whereas the Chinese version used the honorific form of "imperial edict" (chiyu 敕諭). 20 The translator also went out of his way to translate Nurhaci's "son-in-law" (efu) as "imperial son-in-law in the Eastern Palace" (Donggong fuma 東宮付[sic]馬). 21 The Chinese translation evoked the well-known metonymic move of referring to the imperial heir by his cosmic residence in the "Eastern Palace," but conflated the term with "imperial sonin-law" who normally would not aspire to become an heir in the Chinese context. This indicated that the scribe or translator was interested in creating documents in an "imperial" manner that he probably had learned in a performative setting such as popular drama or fiction, but not Ming official culture. The use of both Chinese characters and Manchu script in writing out these documents was not considered a problem. Rather, Part A of the Hanzidang reveals a vernacular sphere of communication that might be characterized as polyglot and fluid. The battles mentioned in Hanzidang had been fought early on during Nurhaci's conquest of rival Jurchen tribes at Hejige (1593), Hada (1599), Akiran (1604), Hoifa (1607), Ihan alin (1608), Hūlha (1609), Yaran (1610), Jakūta (1611), and finally the defeat of Ula (1613). 22 Hostility toward the Yehe was clearly building up, and some officers are recorded in the Hanzidang as having been sent to the Khalkha Mongols, likely in search of an alliance. A man named Hahana, for instance, was praised for "having killed a Yehe man as he went in defense of the Mongol road." 23 All evidence suggests that the edicts are likely to have been made between 1616-1619, in all probability as a gesture to reward loyal soldiers after Nurhaci became Khan in 1616, but prior to the 1619 Battle of Sarhū, where Nurhaci achieved a decisive victory over the Ming-Yehe- Korean alliance. Tellingly, battle wounds caused by firearms, which had been heavily deployed by the enemy at Sarhū, were not mentioned at all. 24 Instead, the world of war and violence in Hanzidang was one dominated by arrows and other cold weapons. The typology of battle wounds taken into account reveals a certain way of reading the bodies of enlisted men in the context of a militarized (and militarizing) society. A typical entry concerns a man named Ilden: Ilden ulade emgeri gidalabuha emgeri langtulabuha emgeri gabtabuha seme . . .

亦兒鄧兀喇(地名)陣上刺一鎗又打一杠子又中一箭 25
At the battle of Ula, Ilden was stabbed with a spear (gidalabuha) one time, hit by a heavy stick (langtulabuha) one time, and shot by an arrow (gabtabuha) one time.

兀吉在兀喇刺二鎗砍上一處 26
Weji, at Ula, had two wounds by being stabbed and one wound by being cut.
Stabbing by spears, hitting by sticks, cutting by (presumably) swords, and arrow-shots constituted four basic types of wounds in Hanzidang. Among those, arrow wounds were so common that the verb gabtabumbi (to be shot by an arrow) was frequently omitted: Looja hoifade emu feye ulade emu feye baha . . .
Without the parallel Chinese translation that specified "arrow" (jian), we would not have been able to tell from Manchu alone what kind of wound (feye) Looja got (baha). The danger of arrow shots has been rendered vividly in another soldier's case:

扎扶你亦哈阿嶺陣上中二十三箭只六箭至肉裏 28
Jafuni, at Ihan Alin [the Ox mountain], obtained twenty-three wounds, only six of them reached the flesh.
Jafuni must have been clad in a set of durable armor. Arrow shots that did not penetrate the flesh (yali) were seen as somehow less significant. Meanwhile, battle injuries suffered by one's horse (morin) were also mentioned, likely owing to the heavy emphasis on cavalry in Manchu military culture. For instance, Bantasi, the leading officer mentioned in the second edict, only had his horse "stabbed once" but still received promotions (Bantasi ulade morin emu feye gidalabuha seme). 29 Elsewhere in the documents, more debilitating wounds such as the loss of eyesight (yasa ehe ohobi), as well as hand or finger injury (gala sacibuha / šumuhun sacibuha), also appear a few times. 30 The selective focus on eyes and hands again may have been due to the primary importance of archery in early Manchu warfare. The examples above give an idea of the general tone of Hanzidang, Part A (ff. 1-28). The Manchu word feye, a word with cognates in other Tungusic languages, means "wounds" but also "key points (a strategic stronghold)." 31 The examination of wounds as a way of visualizing a soldier's valor appears to have been an established practice at least throughout Nurhaci's reign. In 1619, Nurhaci ordered the captains of various companies (niru) to supervise their soldiers. For those who "stood far away at the end of a row, and do not fight with all their might," even if they "bear wounds, it does not count as merit." 32 Although the documents in Hanzidang did reward someone who merely "fought a battle" (Ma. dain de afaha; Ch. shang zhen le 上陣了), as well as people who specialized in various occupations, including a "much-deployed" (takūraha ba labdu 因多差用使之處) shaman and many scholars (baksi), most individuals received recognition according to their battle wounds. 33 The documents do not offer any hint that actual physical examinations of battle wounds were conducted to make these documents. Many of the battles had been fought decades earlier, and the wounds inflicted there would have become quite old scars by 1616. Many men documented here had already been killed (bucehe) at those battles. Instead, the documentation of battle wounds is likely to have been conjured by collectively recognized memories, or even preexisting written records. The Hanzidang documents are therefore not pristine records gained firsthand from the reading of bodies, but the commemoration of wounded bodies retrieved from earlier records and living memory.

Caliyan, Weile (Uile): Stipend and Work
What did the wounded soldiers get in return for their demonstrated loyalty and bravery? Let us consider again the example of Ilden, this time with the complete sentence: Ilden ulade emgeri gidalabuha emgeri langtulabuha emgeri gabtabuha seme jakūn niyalma guwehe Since at the battle of Ula, Ilden was stabbed with a spear one time, hit by a heavy stick one time, and shot by an arrow one time. He got an eight-person exemption.
What does it mean to be "exempt by eight persons?" The bilingual rendition here was informed by context that was obvious to the translator-scribe, but opaque to us, and not done on a word-for-word basis. The Manchu verb guwembi ("to escape; to exempt") indicates the shirking of responsibilities, whereas the corresponding verb in the Chinese version, zhan ("to take possession of ") connotes personal gain on the part of Ilden. Moreover, the Chinese text specified that Ilden's reward involved eight "adult laborers" (dingfu), whereas the Manchu version simply used the generic word for "people" (niyalma). Whereas the Manchu syntax demonstrated considerable flexibility, the Chinese translation was relatively fixed. The recurrence of the Chinese term in various cases helps us ascertain the meaning of niyalma guwehe: The examples above all have different Manchu expressions but identical Chinese expression of zhan dingfu X ming. We might conclude that the reward bestowed to Ilden meant that he could enjoy the service of eight adult men, whose obligations toward the state were now "exempted" (guwehe) and contributed instead to Ilden's "money and grain" (qianliang). From the example of the scholar Hasitu geopi, we can see that those who enjoyed the service of others were by default exempt from service themselves, and that the status as a baksi conveyed the minimal benefit for himself (beyei tele), although for some unknown reason another baksi received much more (four men's labor).
The decision over service exemption as a reward was not made in a haphazard way. As the case of Weihede indicated, "six persons' exemption" constituted the equivalent of the stipend of a "third-rank captain." The document, along with other records in the HZD, indicates a fairly stable "pay-chart" of military ranks adopted from the Ming (Table 2). Some people might receive a higher title while still getting a lower stipend: In addition, an "explanatory" passage repeatedly showed up in virtually all twentytwo edicts in Manchu (but not translated to Chinese): To the one lieutenant (cansun 千總), four persons; to the three sub-lieutenants (basun 把總), three persons each; to the three castle guards (šopu 守堡), two persons each; the armored people that stayed in the Khan's city, the detectives, the gatekeepers, and the artisans, two persons each; the blacksmith and the tilemaker, three persons each; relay station [employees], four persons; swineherds, six persons.
In their annotations, the Qianlong-era editors explained that ulgiyan de ninggun niyalma meant that six people were assigned the pivotal task of raising livestock. 38 We might rephrase it to say that the swineherd's job is "worth" six persons, whose labor by definition was exempted from the state. The early Manchu state took over the Ming military ranks and reassessed their meaning together with other specific needs for expertise, including special guards, artisans and pig keepers.

Saksaha
Vol. 16 The majority of soldiers documented in Hanzidang, however, did not receive any rank that came with a fixed amount of stipend for their battle wounds. Take the example of Weji: Weji ulade juwe feye gidalabuha emu feye sacibuha seme juwan ilan yan i uile waliyambi 兀吉在兀喇刺二鎗砍上一處十三匁之錄 39 Weji, at Ula, had two wounds by being stabbed and one wound by being cut. The work of thirteen taels of silver is being waived.
The keyword here is uile (Standard Ma. weile), which originated from Mongolian üyile meaning "work; act; employment, service," with a secondary sense of "crime" that probably arose later. 40 In his study of Hanzidang, Matsumura Jun translated the sentence above as Weji being "exempt from thirteen taels' offense," a decision with which the Chinese translators of MBRT concurred. Here I suggest that it would be more appropriate to interpret uile here in the former sense (e.g. weilembi "to work"; weilen "work"; weile-i baita "constructions" [Ch. gongcheng 工程]), as it was likely service but not crime that became monetized here. Furthermore, the translation of uile as punishment is difficult to square with the fact that 14 of the 228 individuals discussed in Part A of the Hanzidang had in fact been killed long ago at battles, but the "waiving" (waliyambi) of their uile persisted after death: Here the compound uilere alban clearly indicates uilere to be a particular kind of service (alban) collectively owed to the state by a group or family, not punishable wrongdoings. The passage also reveals that the Khan reserved the power to put an entire company under uilere alban or suspend it for any arbitrary duration of time. It is this service that was measured by silver taels, not crime.
Again, an untranslated explanatory sentence in the Hanzidang sought to clarify these rules: niyalma guwehe niyalmade menggun i uile nakaha. niyalma guwehe akū niyalmade menggun i uile kemuni waliyambi. 43 To those who were exempt personally, their monetized work is canceled. To those who were not exempt personally, their monetized work is still waived.
On a note pasted to the margin, the eighteenth-century editors of the Old Manchu Archive also understood this passage to be about some kind of "service tax" (hahai alban). 44 The important point here is that in this militarized society, men with military service were further divided into two status groups: one with privilege of personal exemption (a total of 50 out of 228), and their monetized work owed to the state (menggun-i uile) was simply canceled (nakaha); the other was without exemption, and their silver duty was merely being "waived" (waliyambi) and could be reinstated anytime. That is to say, the Weji's three battle wounds earned him a lessening of monetized service he owed to the state by thirteen taels. Most entries in Hanzidang involved the lessening of small amounts of silver, rarely exceeding twenty taels. No consistent rules seemed to be in place to convert the number and nature of wounds to the amount of silver. Some suffered more wounds and received less; some only got one arrow shot and received more.
The 228 men covered in the twenty-two edicts were too few to constitute a complete account of all active members of the Latter Jin army. Tumei's company also surely consisted of more than two men who suffered any battle wounds discussed in the edict. In other words, the Hanzidang documents concerned the rank-and-file members of the whole document to 1617. Here I suggest that the perfective oho manggi indicates completion of the action in the future, not necessarily time lapsed since the past. 43. HZD, 2a, and passim. 44. NGMW, 17.352 (Tianming 9). gingguleme kimcici. . . . sere gisun. ainci hahai alban guwebuhe niyalma de menggun guwebure ejehe šangnakū. hahai alban guwebuhekū niyalma de kemuni menggun guwebure ejehe šangnambi sere gūnin dere. "As we respectfully investigate . . . (this sentence) means that for those who have been exempt from the 'labor tax,' they do not get awarded the decree of silver exemption. For those who have not been exempt from the 'labor tax,' they still get the decree of silver exemption." Vol. 16 Jurchen military below the very top aristocratic elite, but still a worthy minority whose merits were recognized by the state in those edicts. They were likely to have been landowning men, heads of their household put under the uilere alban ("service") measured by silver (menggun). The most favored members of their class received "personal exemptions" that got their uilere alban canceled once and for all; the rest used their exceptional military service, measured by battle wounds, to receive individual deductions to their monetized service that may have derived from agricultural surpluses from their land. It is perhaps not surprising that this is reminiscent of Ming rural society in Liaodong, where the Jurchens had recently become the new overlords in the late 1610s but perhaps could not completely reinvent the social order all at once. The distinction between the "personally exempt" (niyalma guwehe) elite and the rest only resembles that between gentry and commoners in a superficial way, as the reciprocal relationship between state and local elite was forged through military service, not the civil service examinations used in Ming China. We shall see that even the temporary privilege granted to elite officers underwent rapid change in the early 1620s.

Between Parts A and B: Duplication and Adaptation
Having examined the bilingual documents in Part A (ff. 1-28) of Hanzidang, we now compare them with Part B (ff. 29-41), which is Manchu-only but written on the same type of paper (researchers have identified it as Korean paper, Gaoli zhi 高麗紙). 45 There are only nineteen edicts recorded in Part B (three fewer than Part A), and a preliminary comparison shows that despite the initial resemblance, Part B in fact rearranged Part A's content substantially (Table 3): Although the total number of individuals discussed in Part B gets close to Part A (228), there were many names that were new to Part B. The edict appears more compact, without line breaks in between individuals. The khan's comments on certain extraordinary individuals also differ from those of Part A, suggesting that the overlap between Parts A and B was not just a matter of scribal duplication, but fully directed and supervised by Nurhaci. If, as I suggested, the edicts of Part A were made between 1616-1619, then what new conditions might have prompted Nurhaci to revisit them and issue a new set of edicts? Again, consider the example of Ilden. In Part B, Ilden's battle wounds were repeated with slight modifications. More importantly, he received a totally different kind of reward: The Khan says.: Ilden of Cergei's Company, got stabbed by a spear once, shot by arrows once, and hit by a stick once at Ula. Thus having made merit, he is sixth rank, with twenty-third taels' worth of work waived. Having canceled one tael and five mace (1.5 taels), there is still twenty-one taels and five mace (21.5 taels) left.

亦兒鄧兀喇(地名)陣上刺一鎗又打一杠子又中一箭占丁夫八名 47
Since at the battle of Ula, Ilden was stabbed with a spear one time, hit by a heavy stick one time, and shot by an arrow one time, he receives an exemption of eight men. 46. HZD, 29b. 47. HZD, 4a. The differences between the two edicts are quite striking. While in Part A, Ilden's name is mentioned in the same edict, following Cergei, without further explanation, Part B spells out that he belonged to the company (niru) of Cergei. Furthermore, whereas in Part A Ilden received the equivalent of eight men's service, in Part B he first got a "sixth rank" (ningguci jergi) on account of the merit he accomplished (gung arabi), and had "twenty-three liang [of silver]" exempted from his monetized work, in the same fashion as the majority of soldiers got in Part A. It did not just happen to Ilden alone. In Part B, everybody got a numerical rank and was given a small amount of work relief, no matter how great their merit. The two-tiered system of Part A was collapsed into one tier in Part B, and now nobody was "personally exempt" (niyalma guwehe) from the numerical reckoning with the state. All of the Chinese military titles were also completely erased, replaced by the numerical ranking from one to thirteen. Upon closer examination, it appears that the last clause of the edict concerning Ilden (underlined in my transcription above) was added as an afterthought, and the word waliyambi "to waive" changed to waliyambihe "was being waived," a change in verb form that allows it to connect to the added clause (see Figure 2). It shows that Part B was an active ledger, not a fixed text, and that Ilden at some point deducted 1 liang and 5 qian of his merit for some unknown reason, which the scribe duly noted in the books. It is not difficult to tell that Part A was probably created earlier than Part B. The handwriting of both the Manchu and the Chinese therein is rough and unadorned, containing many grammatical errors that were partly corrected later. Furthermore, documents in Part A reiterated rules concerning pay and different awards, indicating that the rules were first added to individual edicts and then compiled together. In Part B, by contrast, the scribes knew exactly what they were doing, and their Manchu handwriting is much more uniform in appearance. They made corrections only to the amount of monetized exemption held by each soldier, not to misspellings. Moreover, the monetized merit earned by each soldier invariably decreased, which suggested that the ledger had been in active use. For example, the soldier named Looja initially got eleven liang of exemption, which was first crossed out and changed to nine, and then eight. 48 Weji, who received thirteen liang of silver exemption in Part A on account of his three wounds, appears in Part B as follows: Weji ulade juwe feye gidalabuha emu feye sacibuha seme juwan ilan juwe yan i uile waliyambi juwan nemuci jergi.. 49 Weji, at Ula, had two wounds by being stabbed and one wound by being cut. The duty of thirteen (twelve) taels of silver is being waived.
Wherever there was space at the end of a line, as in the case of Ilden, a sentence was added indicating his "balance." In most cases, the scribe simply made in-line changes to the original number. The need to constantly revisit the record may have been one reason why this set of non-chronicle documents were kept intact. The first set of documents in Part A may have been preserved as references for the initial merit (gung) recognized by the state.
What might have been the institutional context for the creation of Part B alongside Part A? Part A's two-tiered system exploited the familiarity of Ming military titles to residents of the Liaodong region, whereas Part B clearly articulated the company (niru) 48. HZD, 29a. 49. HZD, 29b.

Saksaha
Vol. 16 as a socio-military unit, maintaining an egalitarian spirit in disciplining persons of all ranks. In addition, Part B also specified which four officers below the leader (nirui ejen) could use the title of "adjutant" (janggin 章京, occasionally called "the four adjutants" without listing their names). The creation of the niru as well as the four adjutants as assistants to the leader had been documented very early on in the Old Manchu Archive as one of Nurhaci's foundational designs for the Jurchen state. 50 Yet the absence of institutional terms in Part A, and their reappearance in Part B, suggests that the implementation of Jurchen social structures was by no means so clear-cut. Similarly, Yang Bin 楊賓 (1650-1720) noted during his life as the son of an exile in Ningguta:

邊外文字多書於木。往來傳遞者曰牌子，以削木片若牌故也。存貯年久者曰 檔案。曰檔子。以積累多貫皮條掛壁若檔故也。
Beyond the palisade, texts are mostly written on wood. Those that are passed back and forth are called paizi [plaques], as the wooden pieces are pared thin like plaques. Those that are stored for a long time are called dang'an, or dangzi, for the wooden tablets accumulate to a large number, and are stringed together with hide stripes and hung on the wall, like a crosspiece. 53 He goes on to ponder on the historicity of writing practices, comparing the Manchu use of dangse with ancient bamboo tablets in the "Central Kingdom":

然今文字之书于纸者亦呼为牌子档子。犹之中土文字。汉以前载在竹简。故 曰简。以韦编贯。故曰编。今之人既书于纸。为卷为部。而犹呼之为编为简 也。
However, nowadays texts written on paper are also called paizi and dangzi. This resembles how texts of the central states [of China] had been recorded on bamboo slips prior to the Han dynasty, and so are called jian [bamboo slips]; the slips were stringed together with reeds, called bian [files]. Whereas people today write on paper, making fascicles (juan) and volumes (bu), but still use the terms of bian and jian. 54 Both Wang and Yang noted the rich tradition of creating, storing, and referencing written documents among residents "beyond the palisade." They also noted that the neologism dangse originated in a multi-faceted documentary culture that frequently employed media beyond paper, such wooden tablets that were carefully arranged into a dangse. 55 Going beyond my analysis of the Hanzidang, I now turn to the question of how the idea of dangse, defined as collected files stored in one place, gained traction early on in the expanding regime under Nurhaci. We shall see how the need to manage personnel provided the key context in the numerous occurrences of the word dangse in other parts of the Old Manchu Archive. Yuanji 張元濟 et al. (Shanghai: Shangwu yinshuguan, 1919) On the eighth day, in order to award people to officials who died, the Khan looked at the archival documents. Thereupon, . . .

Back in 1619 [
Tianming 4], he ordered every ranked officer to make a written oath, declaring allegiance and obedience to the Khan. The written oath was to be reviewed by the Khan, "written in the dangse and received" (han tuwafi dangse de arame gaifi). Future offenses would be checked against their oaths and punished accordingly. 58 In the early Jurchen regime, promotion and punishment were not matters to be taken lightly. The Hanzidang testifies to how Nurhaci took great pains in individually assessing his officers' military merit (measured primarily by battle wounds) and reviewed their record frequently. The creation and maintenance of dangse was instrumental in how he managed this kind of administrative matters. We see how the silver exemption balance was crossed out and recalculated, and how entire sentences (specific people) were obliterated with a brush stroke (faitaha). For Hūsitun, who showed up for his battle wounds in Part A of Hanzidang, 59 was later punished on account of dubious activity and erased from the archive: 56. NGMW,17.130 (Tianming 6/5/26). 57. NGMW,17.322 (Tianming 8/7/8). 58. NGMW, Hūsitun, on account of his having delivered the younger child with a young horse in the direction of Zhenjiang (Mao Wenlong's base), establish twenty taels' worth of service and, following that, cancel his merit recorded in the edict archive.
The "ejehe dangse" in which Hūsitun's merit was recorded probably referred to the edicts in Hanzidang. However, in Part B of Hanzidang, Hūsitun seemed to have received an order of rehabilitation: Han hendume.. Hūsitun juwan aniya oho manggi geren i jergide tekini. 61 The Khan says, let Hūsitun sit with others after ten years [have passed].
Again the sequence of Hūsitun's punishment and rehabilitation suggests that Part A of Hanzidang was composed before Tianming 6 (1621) and Part B after that. Overall, the documents give the impression that Nurhaci played with the give-and-take of titles so frequently that it must have become a heavy burden on his everyday schedule as khan.
Later during the Tianming reign, Nurhaci began to lay the foundation for a princely council to oversee state affairs, delegating some of the minute management of personnel to this body of collective governance. In 1623, the princely council considered the promotion of Elcibu to the rank of captain (beiyu), which was then "written into the dangse" (dangse de araha). 62 A few days later, the council deliberated and decided to strip several officers of their ranks. Upon hearing the proposal, Nurhaci asked: This key passage reveals a critical moment, the duplication of the one and only set of archival documents into eight copies. Just as matters of promotion and demotion were now to be deliberated by a group of councilors, the written record of such actions was to be simultaneously established and updated in each office. The Khan's final remark is especially intriguing. He spoke directly to the "scholars" (baksi sa), who were in front of him recording this very document, and said that the duplication of archives was "a good thing" for them as well. If more hands were needed to copy all these extra documents, it would be a boost to the scholars' status (and remuneration). The writing of dangse under the early Jurchen regime was not limited to confidential matters overseen by the Khan and a small group of leaders. In 1621, Nurhaci ordered that every company (niru) was to appoint ten people to "write dangse" (dangse araha). 64 Those documents were then to become the basis for future investigations: in 1625, Nurhaci ordered a comprehensive inspection (baica) of the documents, to see which company had committed the most offences by theft, and which company had none. 65 The company's dangse also carried the function of registering adult males fit for military service, and subject to requests for the most up-to-date information at any time. In 1623, an unsuccessful accusation against Wang Zideng, a high-ranking general in charge of iron melting, shed some light on the practice of archival documentation. A letter of treasonous nature, it was reported, was placed "along with all sorts of dangse documents in a big cabinet" (bithe be amba horho de geren dangse bithei suwaliyame bihe). When a messenger came to fetch "documents of the number of male laborers" (hahai ton i dangse bithe), the letter was also discovered to be treasonous. 66 Upon hearing that, Nurhaci considered the accusation to be absurd. If it were truly a treasonous document, he reasoned, then Wang would have "hidden it elsewhere, in a secret place; how can he hide it by putting it alongside all the dangse documents?" (narhūn bade encu somime asarambi dere. Geren dangse bithei emgi ainu asarambi). Notably, these official documents were considered to be the opposite of "secret" (narhūn), stored in an official's cabinet so that it could be consulted if necessary. The battle wound 64. NGMW,17.140 (Tianming 6/7/19 Let us briefly turn to the latter part of Hanzidang that I have not discussed so far. The 57 folios were written on previously used Ming official documents, sometimes written on the back, sometimes on the margins. It belongs to a type of "court diary" typical in the Old Manchu Archive. Every day, two banners took the responsibility of recording the Khan's whereabouts, including the hour at which he prayed at the shrine; diplomatic letters received and sent; and intricate dances of politics between aristocratic men and women. 67 The tattered look of those documents notwithstanding, their content carried paramount political significance starting from the very early stage of official historiography sponsored by the Manchu state. Officials in the late 1630s examined each piece of paper, made decisions of whether to copy (Erebe ara "Write this!") or not (Erebe ūme "Do not [write] this!"). The process eventually resulted in the sanctioned and sanitized editions of pre-Conquest veritable records (shilu) for Nurhaci's and Hong Taiji's reign periods. In contrast, ff. 1-41 of the Hanzidang, which had been bound together with those later court diaries for unknown reasons, look impressively pristine, unmarked by the later ambition to write and rewrite dynastic history. They were created and preserved to serve an administrative purpose, mediating each individual's privilege and obligations to the state. At the same time, however, this pristine look is perhaps another illusion. We have already seen how Parts A and B of the Hanzidang were likely to have been copies of even earlier documents that had long been lost, and how they presented quite distinct ways of conceiving and distributing merit, even if the number and type of battle wounds inflicted on one's body remained the same. There is something stubborn in these personnel files that resisted the writing and un-writing of dynastic history, to the extent that it might justify the translation of dangse into "archive," prior to the word dang'an becoming the equivalent of "archive" in the twentieth century.
When the National Palace Museum in Taipei first sent out the pre-conquest Manchu documents for facsimile reproduction in the 1960s, the original of the last page of Hanzidang was reportedly lost. 68 The loss of the so-called "image of a female corpse" (nüshitu) in the Old Manchu Archive had sparked media attention since then. 69 If we examine the Ming official document on which the Jurchen scribes wrote, it is clear that the original document was a Ming coroner's report produced in the 28 th year of Jiajing reign (1549). Although most of the Chinese text is now illegible, one could tell that the body image represented the corpse being examined, and the report consists of a list of body parts: "two ankles: no cause [of death]; ten fingers: no cause . . ." In a strange way, the last part of Hanzidang, which had been written on Ming official documents in the local administration that dealt with dead bodies, came to be bound with the first 41 pages that documented wounds resulting from battles fought for the Jurchen state. I came to study the Hanzidang due to my interest in this image. It turns out that seventy years after it was made, the Ming record of a corpse was lifted out of the files and, by chance, bound together with the documentation of over two hundred officers fighting for the insurgent Jurchen state.
The first and last parts of the Hanzidang also found uncanny resonance in their bilingualism. On the last day of the first month in the year of jiazi (1624), the scholar (baksi) in charge finished off his work and decided to have some fun by writing something else on the empty space left on that piece of Ming official document paper. Marking time with the specific year and month (甲子年正月記), the scholar copied a ditty from Poetry of Scholar Xie (Xie xueshi shi 解學士詩), a popular late Ming anthology attributed to the child prodigy Xie Jin 解縉 (1369-1415):

一直往南去 轉來朝向東 門前有大字 便是董郎中 70
Go straight toward the south, Make a turn, then facing east; In front of the door are large characters And that is Doctor Dong's home.
The ditty confronts us with the likely scenario that the person writing the court chronicles for Nurhaci during the first month of Tianming 9 (1624) might have been a native speaker of Chinese, or at least familiar with late Ming vernacular literature. The unintended (and undisciplined) bilingualism at the very end of Hanzidang resonates in an uncanny manner with the bilingual battle wound records, showing how a vernacular, polyglot culture ran deep among rank-and-file members of the early Manchu state that had not yet become the Great Qing. In the 1620s, the scholars who created and managed dangse documents for the khan were not primarily concerned with the moral imperatives of official historiography. Nor was Nurhaci, who toward the end of his life 70. HZD, 98b.