We see all governments as obscure and invisible.
Francis Bacon, The Advancement of Learning (1605)
Few early Chinese sources matched Francis Bacon's extremism in advocating for a ruler's absolute right to secrecy in deliberations and decisions,Footnote 1 but most recognized some level of secrecy as essential for governing.Footnote 2 The concern is broadly evident in discussions of government and rulership, whether to ensure effective operations or create a sense of majestic authority.Footnote 3 Scholars have not ignored this fact, offering studies of a) the importance of secrecy in military strategy and intelligence gathering;Footnote 4 b) bureaucratic procedures from pre-imperial and early imperial times, particularly those designed to ensure documents remained confidential;Footnote 5 and, especially, c) an ideal model of rulership in which the sovereign remained obscured, outside the system of laws and regulations that he directed.Footnote 6 Beyond the question of criminal sanctions backed by a supposed obsession with state secrecy and surveillance,Footnote 7 however, students of early China have mostly ignored the problem of breakdowns in confidentiality. How were failures in maintaining secrecy framed and discussed?
This article answers this question by exploring the use in early sources of several terms, including xie 泄 (or 洩) and lou 漏 (both meaning “to leak”), as well as zhou 周 and mi 密 (“confidential,” among other meanings). The leaking metaphor in particular passes with surprising smoothness into contemporary English and our endless debates about information leaks.Footnote 8 We are accustomed to understanding the leak, almost an inevitable by-product of confidential military and central government activities, as a transgression between secret and open spheres performed for a variety of reasons.Footnote 9 Such concerns are by no means absent from Chinese discussions of the leak: for instance, as this article will show, early sources reflect an expanding scope of secret information at the imperial court as well as critiques of that expansion. The terms given above, however, referred as much to states of perfect behavior and physical cultivation as they did to the status of different kinds of information. In this vein, note the following passage from the “Xici zhuan” 繫辭傳 (Commentary on the appended sayings) chapter of the Yijing 易經 (Classic of changes), which as discussed in greater detail below was cited with some regularity in the final decades of Western Han (202 BCE–9 CE):
君不密則失臣,臣不密則失身,幾事不密則害成。 是以君子慎密而不出也。
If the ruler is not discreet then he will lose his minister. If the minister is not discreet then he will lose his life. If pivotal matters are not kept discreet then disasters will occur. This is why the ruler is careful about discretion and is not forthcoming.Footnote 10
The passage offers two interlinked claims. First, rulers, officials, and secret matters alike all had to remain discreet, and second, the ruler should therefore be “not forthcoming” (bu chu 不出), a proscription against noticeable actions that could include the ruler's physical location, expressions, or verbal statements. Rather than drawing a line between “secret” and “open” matters, then, the passage emphasizes a comprehensive mode of circumspect behavior. The world it evokes is one in which inadvertent leaks, from mistaken movements to slips of the tongue, are just as important and receive as much critical attention as those that are purposeful.
One of the anchoring observations of this article, however, is that the earliest extant citations of this passage, made by court officials in the final decades of the Western Han, completely omitted the final statement about the ruler. Instead of a meditation on how to maintain the ruler's power or the power of his government, then, by late Western Han the passage was understood to articulate a shared, idealized state of circumspect behavior that allowed all parties to evade disaster.Footnote 11 How did such a vision emerge? This article attempts to answer this question by tracing, in three sections, debates and representations of leaking. The first section starts with a detailed discussion of the terms given above (xie, lou, zhou, and mi), which in early texts sometimes evoke secret information but just as frequently refer to idealized states of physical cultivation. Relevant chapters from the Guanzi 管子 are enlightening in this regard, for they draw upon medical theories of qi 氣 circulation to construct a vision of rulership in which vital essences and information are not leaked from the sovereign's body. This model, also evident in other texts, was rejected in the Xunzi 荀子, which advocated an open and highly visible sovereign whose power comes from visibility, not secrecy. The second section starts by noting that the question of the ruler's visibility is central to stories from the Shiji 史記 (Records of the Senior Archivist, comp. ca. 86 BCE) about the First Emperor of Qin (Qin Shihuang di 秦始皇帝; r. 221–210 BCE). The First Emperor, of course, is famous for his obsession with secrecy, but less recognized is that the Shiji suggests this obsession created institutional and spatial divisions within the capital designed to prevent the spread of confidential information. In this sense, the Shiji story reflects anecdotes from the Han Feizi 韓非子 that explore the spaces and audiences at court for leaked information, and the constraining power they exerted over rulers and officials alike.
The third and final section delves into further institutional changes during Western Han that formalized such constraints, allowing officials to celebrate adherence to the highest codes of professional responsibility, highlight their membership in the most elite circles, and remind everybody (including the emperor) of conventions that required circumspection of all parties at court. The case of leaking thus provides an enlightening way to write a cultural history of officialdom, during a period when many of the formal regulations and informal norms that governed the behavior of officials were still being worked out. If the earliest discussions of leaking, discussed in the first section, were informed by medical theories applied primarily to the perfected body of an idealized sage ruler, the late-Western Han proscriptions against leaking discussed in the final section are far from a reflection of the ruler's power. The desire to properly manage the ruler's body, both internally and externally, as well as the information that supported his authority, led to regulations that officials and others at court used in extra-legal ways, expressing their status and advancing a vision of elite political action bound by shared norms.
Leaking Bodies, Obscured (and Visible) Rulers
Starting from around the fourth century BCE, as Romain Graziani has noted, a burgeoning interest in a “holistic discourse grounded in qi” prompted a variety of investigations that created a “continuum between medicine and philosophy.”Footnote 12 Discussions of leaking rested on this same continuum, for medical texts, on the one hand, and discussions of rulership that drew upon ideas of self-cultivation or “nurturing life” (yang sheng 養生), on the other, all equally evinced concern about qi circulation and the possibility of vital essences oozing out of the body. As we will see via close readings of essays contained in the Guanzi, however, the further texts on this continuum moved away from discussions of medical concepts and toward questions of rulership and governance, the more emphasis was placed on the dangers of leaking and the importance of maintaining a perfectly cultivated and hardened body that expelled (and revealed) nothing. The Xunzi will help us clarify this model, for it rejected its assumptions, dismissing concerns about leaking and emphasizing instead a highly open and visible form of rulership.
Detailed study of the medical side of the continuum, entailing as it would complicated technical questions in the history of early Chinese medicine, must remain outside the bounds of this article.Footnote 13 The important point for our purposes is that medical practitioners were highly concerned with movement of qi and other essences across boundaries. Whether to retain or expel depended on the substance, as Nathan Sivin has noted:
To sustain life, the body can be neither completely open nor completely closed. Food and qi must enter without admitting agents of disease; wastes must be excreted without allowing the body's vitalities to leak out.Footnote 14
In early medical texts xie and lou, “leaks” into and out of the body and within the body, could be good or bad. Thus the text *Shi wen 十問 (Ten questions) from Mawangdui 馬王堆 tomb 247 (sealed in 168 BCE) articulates problems caused by “Yin essence leaking out” (陰精漏泄),Footnote 15 while the “Si qi tiao shen da lun” 四氣調神大論 (Comprehensive discourse on regulating the spirit via the four seasonal qi) chapter of Huangdi neijing su wen 黃帝內經素問 (Inner classic of the Yellow Emperor—basic questions) mentions the salutary importance of “allowing qi to leak away” (使氣得泄) during the summer.Footnote 16 Notwithstanding Sivin's use of the verb “leak” as a deleterious action, usage in early medical texts suggests xie and lou were neutral actions: whether or not such flows were harmful or helpful depended on the type of substance, amount, and timing, among other factors.
Texts that articulated cultivation practices and “nurturing life” (yang sheng 養生) ideas for the development of ideal bodies, especially the ideal bodies of sagely beings and rulers, drew upon a similar language of inner and outer, “bad” and “good” leaking. Both are evident in “Nei ye” 內業 (Inner Training), one of the most famous chapters from the Guanzi and an important early source for practices of physical self-cultivation.Footnote 17 The problem of “bad” leaking is evident in the text's opening discussion of “essence” (jing 精) and a following treatment of the Dao 道. The two key concepts mostly receive similar treatment,Footnote 18 but when it comes to descriptions of their movement, contrasting language forms a model of the body that implies the dangers of leaking. According to “Nei ye,” essence “flows” (liu 流) and can be “stored within the chest” (cang yu xiong zhong 藏於胸中),Footnote 19 forming a “wellspring” (quan yuan 泉原) and “pool of qi” (qi yuan 氣淵). If this wellspring does not “dry out” (he 涸), then the “nine apertures” (jiu qiao 九竅) become “penetrating” (tong 通) and “reach to the ends of Heaven and Earth” (窮天地).Footnote 20 In contrast, the Dao does not “flow” but “goes” (wang 往) or “comes” (lai 來) and “fills the form” (chong xing 充形).Footnote 21 Since, as the chapter states, “in all cases, the Dao is certain to be bounded (zhou) and close (mi), broad and expansive, hard and secure” (凡道必周必密, 必寬必舒, 必堅必固),Footnote 22 presumably Dao-filled bodily forms (xing) would attain the same properties. How a body could simultaneously be both “bounded and close” and “broad and expansive” is not entirely clear.Footnote 23 Nonetheless, the language of “Nei ye” suggests a contrast between the behavior of “essence” (jing) and the Dao, outlining a cultivated body animated by internal flows and stores of essence, contained within a compacted, strengthened form that protects the “pool of qi” inside.
At the same time, even while the “Nei ye” thus emphasizes the dangers of leaks from this pool, the statement that the bodily apertures become “penetrating” (tong) hints at passages allowing interchange between interior and exterior. Moreover, at the very end we find mention of “leaking” (xie) in a positive context:
得道之人,理丞而屯泄,匈中無敗。節欲之道,萬物不害。
For the person who has attained the Dao, lines in the skin excrete and hair follicles leak (xie), while within the chest nothing is corrupt. With this Way (dao) of moderating desires, the myriad things cause no harm.Footnote 24
Commentators have struggled over the phrase “lines in the skin excrete and hair follicles leak,”Footnote 25 but typically cite a parallel passage in the “Tai zu” 泰族 chapter of Huainanzi 淮南子 (comp. 139 BCE) to strengthen the interpretation reflected in this translation.Footnote 26 Both texts thus evoke the medical understanding of the body described by Sivin, “neither completely closed nor completely open,” but applied to self-cultivation practice: the refined essence remains stored within the chest, while regular and modulated leaks (xie) out of the protective skin create a perfectly regulated, sagely body. The “Nei ye” thus explicitly recognizes the sage's homeostatic stability, of leaks outward that accompany the preservation of essence inward, even if the opening passages focus primarily on hardening the body's exterior and preventing leaks.Footnote 27
When we move to other Guanzi chapters and further down the continuum, away from purely medical understandings and toward expressly political formulations of ideal rulership, we see familiar terminology, but even the limited recognition of positive leaking seen in “Nei ye” is absent. The “Shu yan” 樞言 (Pivotal sayings) chapter, for instance, uses language reminiscent of “Nei ye,” in particular the word zhou, in order to argue that the ruler's body should reveal nothing:
周者,不出于口,不見于色,一龍一蛇,一日五化之謂周。故先王不以一過二,先王不獨舉,不擅功。
“Bounded off” (zhou) refers to nothing emerging from the mouth or appearing in the expression, even while now a dragon, now a serpent, five transformations occur in one day.Footnote 28 Prior kings thus did not go from the [unified] one to the [divided] two. They did not act on their own or arrogate achievements [for themselves].Footnote 29
While some details are elusive, the passage's overall picture is clear: the “bounded off” (zhou) body of an ideal ruler becomes clothed in a kind of exterior armor,Footnote 30 never emitting speech or revealing facial expressions in irregular or uncontrolled fashion. Indeed, such irregularities are impossible, since the physical person of the ruler is entirely unified and free of internal divisions.Footnote 31 Readers familiar with Warring States philosophical texts will no doubt recognize this sort of imagery from several chapters of the Laozi 老子, which describe the ideal sage ruler as entirely “still” (jing 靜) and “not engaging in purposeful action” (wu wei 無為), while his subjects “transform themselves” (zi hua 自化) and “bring themselves into alignment” (zi zheng 自正).Footnote 32 The Han Feizi, of course, famously developed such ideas even further, with the “Zhu dao” 主道 (Way of the ruler) chapter describing a perfectly still, even “invisible” ruler who does not reveal his intentions or plans until the moment of action.Footnote 33
Through its reference to a perfectly unitary and still body, the “Shu yan” chapter bridges the gap between discussions of self-cultivation and calls for secrecy in rule. Other Guanzi chapters, however, much more explicitly call for discretion, using the same words zhou and also mi, in the sense of “secret” or “confidential.” For instance, the “Fa fa” 法法 (Model laws) chapter warns that, “when secret plans are not kept discreet (mi) it is dangerous” (幾而不密殆). It then goes on to argue that the ruler who is not zhoumi, here best translated as “confidential,” will be unable to attract truthful and righteous advisors to his court. As a result, factions and cliques will form and eventually cause the ruler's downfall.Footnote 34 The Han Feizi contains an almost identical passage, as well as appeals for the ruler to remain “confidential.”Footnote 35 The Guanzi chapters thus allow us to move from a medically informed image in “Nei ye” of a body “bounded and close” (zhou mi), capable of storing vital essences (though not impermeable), to a ruler “bounded off” from his advisors and revealing no intentions. In terms of qi and leaking, then, movement along the continuum from medicine to philosophical models of ideal rulership entailed a closing off, with the homeostatic stability of a regulated physical body transforming into a bounded, obscured, and confidential political body.
The Xunzi, however, took the circularity implied by the word zhou and developed it in a different direction, using it to mean “all-encompassing” or “comprehensive.” This sense of the word occurs frequently in the text and helps clarify the issues at stake, for it emphasizes the importance of being “all-encompassing and close” (zhou mi) when it comes to cultivating the best kind of knowledge, which the text emphasizes is much more important than cultivating a perfect physical body. A long passage in the “Ru xiao” 儒效 (Achievements of the classicists) chapter, for instance, describes a hierarchy of self-cultivation practices that ends with the highest program possible: that of the sage (sheng ren 聖人). Sagely practices, we read, cannot be compared to mere physical cultivation, since “to take nurturing life as one's own supreme Dao is the virtue of a commoner” (以養生為己至道,是民德也). The passage, however, retains zhou mi as characteristic of the sage's “deliberations” (zhi lü 知慮).Footnote 36 In other words, the ruler's knowledge must be complete in a way that cannot be achieved by mere physical cultivation. The Xunzi thus borrowed language associated with “nurturing life” practices (e.g. those in the “Nei ye” chapter of the Guanzi), but recast zhou as a description of the sage ruler's all-encompassing, comprehensive knowledge, doing so in a way that downplayed the connection drawn in the Guanzi between physical self-cultivation and the Dao.Footnote 37
Indeed, the Xunzi explicitly condemns the connotation of zhou as closed off or “bounded.” The “Zheng lun” 正論 (Aligned judgments) chapter, for example, opens by noting that “vulgar purveyors of doctrine” (世俗之為說者) claim, “the way of the ruler benefits from being closed off” (主道利周). The Xunzi immediately disputes the statement, arguing that when “standards are hidden then subordinates will not move” (儀隱則下無動也).Footnote 38 At the very end of the “Jie bi” 解蔽 (Undoing fixations) chapter, meanwhile, we read that, “achieving success by being bound off and failing due to leaking (xie)—this the enlightened lord will not have” (周而成,泄而敗,明君無之有也).Footnote 39 Somewhat astonishingly, then, the Xunzi refutes the idea that the ruler should close himself off, be secretive, and prevent leaking. Indeed, advocating a model of shining openness, the text implies that leaking simply cannot happen, because nothing is hidden. With the ruler emerging as a standard for all to follow, even his deliberations must be based on comprehensive, all-encompassing knowledge.Footnote 40 The Xunzi thus helps us clarify the issues at stake, for, as we saw above, many other texts made the opposing claim that the ruler's secrecy and guardedness were central to effective rule. Whether or not such debates about sagely cultivation, the ruler's body, secrecy, and leaking influenced legal regulations during the Warring States period cannot be fully answered, due to limitations in pre-Qin sources.Footnote 41 The salient issue for our purposes is that such regulations do not receive particular emphasis in our extant sources. As we will see in the next section, the picture does begin to shift in Qin and early Western Han, though only in an oblique way: stories were circulating that linked the ruler's demands for secrecy not so much to explicit legal regulation, but rather to spatial and institutional divisions at court, creating a plethora of considerations that constrained both rulers and officials.
Secret Spaces and Leaking Officials
The texts analyzed above, while focused almost entirely on the ruler, still evince some recognition that he was surrounded by officials. Note, for instance, the reference in the “Fa fa” chapter of the Guanzi that only if the ruler is circumspect will he be able to prevent the formation of factions and cliques at court. Such references, however, primarily depicted officials as an undifferentiated audience, reacting to the ruler in a generic space. Strikingly, then, stories in the Shiji about the First Emperor of Qin foreground questions of space and audience, explicitly claiming that the ruler's need for secrecy—rooted in his desire not just for safety but also to commune with the gods—resulted in spatial and institutional divisions at the capital. Anecdotes in the Han Feizi, meanwhile, extended this discussion of space and the circulation of privileged information from rulers to officials.
The First Emperor of Qin was famously obsessed with secrecy, demonstrated most spectacularly by his decision to build covered walkways to hide his location. Surely security considerations were one factor driving this decision, as some scholars have claimed.Footnote 42 However, a closer look at the narrative in the “Basic Annals of the First Emperor of Qin” of the Shiji, still our most important source for the First Emperor's life and activities in the Qin capital of Xianyang 咸陽, shows that motivations less strategic and more divine were at play: specifically, concerns about secrecy and leaking were driven by the First Emperor's obsession with meeting the gods. According to the narrative in the “Basic Annals,” toward the end of his life the emperor became increasingly obsessed with immortal deities on elusive ocean islands and mysterious, life-prolonging drugs. One of the men advising him on such matters, Lu Sheng 盧生, eventually persuaded the First Emperor to hide himself in his palaces, even from his own advisors, since “if vassal ministers know where the ruler is residing, this will cause disfavor with the gods” (人主所居而人臣知之,則害於神).Footnote 43 Lu concluded by advising the emperor that drugs of immortality could “probably be obtained” (殆可得也) if he made his place of residence secret. The emperor agreed and immediately ordered all walkways between his palaces covered up and forbade anybody, on pain of death, from revealing his location. Later, however, the emperor was angered to see that his Chancellor had nonetheless followed him on a journey to Mt. Liang. When he noticed that the Chancellor immediately pulled back and reduced his entourage, he furiously exclaimed: “This means that a person in my palace has leaked my words” (此中人泄吾語).Footnote 44 When no one admitted to divulging his supposedly secret location, the emperor executed all in attendance. Subsequently, we read, nobody knew the emperor's location, while “all audiences and the conferral of decisions to ministers took place in the palace at Xianyang [the capital]” (聽事,群臣受決事,悉於咸陽宮).
As we saw in the previous section's discussion of essays in the Guanzi and Xunzi, pre-imperial texts had already addressed the ruler's closing off of body, facial expression, and words. The story of the First Emperor cloistering himself away shows a demand for confidentiality about both what the emperor said (“my words,” wu yu 吾語) and his physical person, with both related to specific places. Concerns about location and space surely existed prior to the First Emperor, but while the Guanzi and other texts emphasized the idea that the patterns of the Dao allowed the ruler to maintain a “bounded and close” body, the Shiji narrative of the First Emperor highlights the importance of space as a confidential location in which the ruler communed with spirits. Moreover, while the First Emperor's obsession with secrecy supposedly sprang from his search for the spirits of immortality, the “Basic Annals” suggests that his demand for confidentiality led to a spatial division of administrative work. The palace at Xianyang theoretically became the sole location for making official decisions, and officials would organize their work accordingly.Footnote 45 Such normative claims, of course, did not reflect actual practice. After all, when the First Emperor died while traveling in his “covered chariot” (wen liang 轀輬), a small group of advisors concealed the emperor's body (and its stench) while continuing to issue decisions from the chariot in his name, secretly destroying sealed letters, and forging new documents that established a different heir to the throne.Footnote 46 Shiji stories of the First Emperor's seclusion and death, then, evince a sensitive understanding of attempts by both rulers and officials to establish and navigate boundaries, boundaries that isolated different audiences and activities as well as conferred authority upon central government directives.
Such concerns find direct reflection in stories and anecdotes found in the Han Feizi, most of the chapters of which can be associated with the man Han Fei—who actually served at the court of the future First Emperor of Qin—and would have been known and in circulation at the Western Han court. While parts of the Han Feizi famously assert the ruler's dominance over ministers who appear as inherently suspect figures, liable to commit acts of betrayal,Footnote 47 other parts of the text examine political dynamics from the perspective of officials. Importantly, they also explore the different spaces and audiences through which leaks travelled, in a manner reminiscent of the Shiji stories about the First Emperor. In a story from one of the “Chu shuo” 儲說 (Stockpiled persuasions) chapters, for instance, one Tangxi Gong 堂谿公 describes to Lord Zhao 昭侯 two different chalices: an expensive but bottomless jade chalice, and a humble ceramic one that “did not leak” (bu lou). Even though the jade chalice was “supremely valuable” (zhi gui 至貴), it could not hold any water, so “who indeed would pour liquid into it?” (人孰注漿哉). Tangxi Gong goes on to equate the jade chalice with the ruler who leaks information to his ministers: “Even if he has sagely intelligence, because of his leaks he does not fully exhaust his techniques [of rule]” (雖有聖智,莫盡其術,為其漏也). The story concludes with a disturbed Lord Zhao, who goes so far as to change his sleeping patterns in order to prevent inadvertent disclosures of information:
昭侯聞堂谿公之言,自此之後,欲發天下之大事,未嘗不獨寢,恐夢言而使人知其謀也。
Lord Zhao heard Tangxi Gong's statements, and from then on, whenever he wanted to engage in a major effort in All Under Heaven, he never failed to sleep by himself, since he feared that while dreaming he would talk and thus allow others to know his plan.Footnote 48
An alternate version of this story immediately follows, as is common in the “Chu shuo” chapters. It concludes by identifying more explicitly these “others”: the ruler's wife and consorts (qi qie 妻妾).Footnote 49 While the story emphasizes the importance of the ruler remaining discreet, recalling the essays analyzed in the previous section, it does not call attention to the ruler's perfect cultivation; indeed, the bottomless jade chalice hints that the ruler's magnificent appearance did not guarantee discretion, while the final passage about sleeping separately suggests Lord Zhao did not trust that his own body would maintain the required level of confidentiality. As we saw above, the First Emperor of Qin similarly cloistered himself away to ensure secrecy.
The Han Feizi makes clear, however, that such concern about spaces and audience were not only the province of the ruler. In another anecdote from the “Chu shuo” chapter, we read first that the ruler is like a chariot hub who collectively transports all archers (cf. Laozi, chapter 11), before going on to note that that this hub-like function of the ruler can only be accomplished if he hides his desires from view. Similarly, “if statements and words circulate then the vassal will find it difficult to speak, while the ruler will not be divine” (辭言通則臣難言,而主不神).Footnote 50 The chapter cites stories about leaking, including one Gongsun Yan 公孫衍, also called Xi Shou 犀首 (“Rhinoceros Head”?), to illustrate the point:
甘茂相秦惠王,惠王愛公孫衍,與之間有所言,曰:「寡人將相子。」甘茂之吏道穴聞之,以告甘茂。
甘茂入見王,曰:「王得賢相,臣敢再拜賀。」
王曰:「寡人託國於子,安更得賢相?」
對曰:「將相犀首。」
王曰:「子安聞之?」
對曰:「犀首告臣。」
王怒犀首之泄,乃逐之。
Gan Mao served as minister to King Hui of Qin. King Hui favored Gongsun Yan, and secretly had a word with him, saying: “I plan to make you minister.” One of Gan Mao's subordinate officials heard what he said through a carved hole and then reported it back.
Gan Mao then went in for an audience with the king, saying: “My king has obtained a worthy minister, and your servant dares to repeatedly bow in congratulations.”
King Hui said: “I have entrusted the realm to you, so why would I then seek a worthy minister?”
Gan Mao responded: “You plan to appoint Xi Shou [Gongsun Yan] minister.”
King Hui said: “Where did you hear this?”
Gan Mao responded: “Xi Shou told me, your servant.”
The King was furious about Xi Shou's leaking, and had him exiled.Footnote 51
Though King Kang took Xi Shou to a separate room in order to speak with him privately, one of Gan Mao's underlings was able to listen to the conversation through a hole in the wall. That Gan Mao had spies at his disposal is hardly surprising, and in this version of the story Xi Shou is a rather hapless figure, easily framed and dispatched by Gan Mao's strategic disclosure of information to the king.
The variant version of the story that immediately follows describes Gongsun Yan as a famous general from Liang who flees to Qin after being convicted of a crime. The King of Qin favored him, much to the worry of the Qin general Chu Liji 樗里疾 who, fearful of being replaced, carved a hole in the “place where the king often spoke in secret” (王之所常隱語者). During one of these confidential conversations, Gongsun Yan gave advice on a military campaign, and in response the king said he would appoint him prime minister. Having overhead the exchange, General Chu quickly related the details to the “gentlemen of the palace” (lang zhong 郎中). By the end of the day, all of them knew about the plans for the attack and Gongsun Yan's appointment; by month's end, “within the borders of the realm everybody knew about it” (境內盡知之). General Chu accused Gongsun Yan of divulging the plans, but the nervous king expressed confusion, stating falsely (perhaps to protect himself) that he had never spoken with Gongsun. In response, General Chu characterized his rival as an itinerant guest, one whose loyalty was suspect since he had fled Liang. As General Chu put it, “in his heart he is alone, so he made these statements to wed himself to the populace” (其心孤,是言自嫁於眾).Footnote 52 The first version of the story is a relatively simple tale of an official spying on a king's conversation and using it to dispatch a potential rival. The more complicated second version, however, shows the king and his officials deceiving each other and jockeying for political gain via different spaces and audiences, from the court and its courtiers to the entire realm and its inhabitants.Footnote 53
Such concerns are the central topic of “Shuo nan” 說難 (Difficulties of persuasion), an essay by Han Fei that attained great fame during Western Han for its eloquent articulation of the predicaments faced by officials trying to convince rulers of a course of action.Footnote 54 The dangers of leaking were one of these predicaments, and the chapter exhorts would-be persuaders that, “affairs are successful because of discretion, while words fail because of leaking” (事以密成, 語以泄敗). The line evinces a broad understanding of “leaking” that extends beyond a premeditated and direct disclosure of privileged information.Footnote 55 For instance, if a persuader, “when speaking happens upon a matter that has been concealed” (語及所匿之事), he will be in danger, presumably because the ruler will assume the speaker obtained the information through secret channels, whether or not the persuader understands the matter to be confidential.Footnote 56 Meanwhile, the persuader must also fear his rivals at court, for even if he managed to convince the ruler of a given course of action, “intelligent people might surmise the plan from the outside and obtain it, so that it is leaked to the outer world” (知者揣之外而得之,事泄於外). In such cases, the ruler will “necessarily assume that you [the speaker] leaked it, putting you in danger” (必以為己也,如此者身危).Footnote 57 Officials must beware not only of how their statements will be perceived by the ruler, but also by others at court, in the “outer world.”
The Han Feizi and stories from the “Basic Annals of the First Emperor” thus show an overriding concern with audiences for confidential information (whether or not everybody understands it to be secret) within confidential space. That space was conceived as simultaneously housing and protecting the ruler's sacred body, demarcating areas for administrative work, and serving as an arena in which officials jockeyed for political advantage. This third and final concern, as we shall see in the next section, received the bulk of attention starting in the last century of Western Han, when institutional patterns and regulations emerged that both attempted to control leaking and allowed officials to assert authority and status.
Leaking in the Late Western Han: Institutions, Laws, and Status
The division of administrative space that the Shiji claims emerged after the First Emperor concealed his location, even if it was at best honored in the breach, remained an ongoing pattern and problem into the Western Han. The sources examined in greater detail below suggest that such concerns only grew in intensity, with late Western Han providing evidence for regulations that proscribed leaking privileged information from specified spaces. Such proscriptions emerged alongside a recognition of confidentiality as a shared norm that bound all at court and, especially, served as a characteristic of exemplary officials.
A full history of the spatial and institutional growth of the Western Han capital and its palaces, as well as the rules and regulations governing them, cannot be attempted here. It is sufficient for our purposes to note that the palaces of Chang'an 長安 came to be tightly secured by three different guard corps, each under the direction of separate government ministries, which patrolled nested areas of the imperial palaces.Footnote 58 Access to these different palace spaces was controlled by a registry, which guards would check before allowing entry into a given space. The most closely monitored spaces were the so-called “inspection zones” (xing zhong 省中), sometimes called the “forbidden zones” (jin zhong 禁中). These areas were the emperor's most private chambers, only accessible with special permission or by possession of a supernumerary title (e.g. ji shi zhong 給事中) that allowed free entry. We do not know much about the zones, which were not necessarily fixed in one location (or even limited to one palace).
The same could be said about the location of secret archives and document repositories, which were sprinkled through different palaces in Chang'an. The number and scope of such archives likely increased, especially after the reign of Wudi 武帝 (r. 141–87 BCE) and up to the reign of Chengdi 成帝 (r. 33–7 BCE), when that emperor famously launched an effort to collect and collate manuscripts from around the empire.Footnote 59 Even if the details remain unclear, the growth of imperial archives must have entailed a concomitant expansion in security regulations designed to restrict access to the archives and the activities that took place therein. Parallel to these developments was the creation and rise in stature of the Imperial Secretariat (Shangshu 尚書), the director of which was in charge of receiving and forwarding official documents and imperial edicts (see the poem by Yang Xiong discussed below). By the end of the Western Han, the Secretariat had become a larger and more organized office,Footnote 60 and the director of the office was one of the most important officials in the empire, with several incumbents going on to serve as Chancellor.Footnote 61 Though admittedly sketchy, our overall picture is one of rising influence as well as increasing institutional growth and differentiation.Footnote 62
As a result, even if we accept the Shiji account that the First Emperor established spatial divisions and institutional patterns in order to ensure his secrecy, by the end of the Western Han our sources suggest that the object of secrecy measures and regulations was no longer so focused on the ruler. Incidents and stories dating to the last century of Western Han demonstrate that while legal regulation and punishment of leaking were ultimately connected to the emperor, the action had shifted to officials accusing others and being accused themselves of different leaking crimes. Almost all of these accusations come relatively late, after the reign of Wudi. Several of them receive mention in the “Table of the Many Offices and Ministerial Posts” (Bai guan gong qing biao 百官公卿表) included in the Hanshu 漢書 (History of the Han, comp. ca. 100 CE). The earliest such accusation in the “Table,” in an entry dating to 77 BCE, reads as follows:
蒲侯蘇昌為太常,十一年坐籍霍山書泄秘書免。
Su Chang, the noble of Pu, was appointed Superintendent of Ceremonial. Eleven years later (66 BCE), he was charged with passing documents to Huo Shan and dismissed for leaking secret palace writings.Footnote 63
This incident was part of a larger purge that eliminated the Huo 霍 family from the imperial court at the start of Xuandi's reign.Footnote 64 The salient point for our discussion is that “leaking” involved not the spoken disclosure of information, which was so central to the Han Feizi stories as well as the paranoia of the First Emperor in the Shiji, but rather the unauthorized release of what amounted to classified documents (if of an unspecified nature).Footnote 65 The story bolsters the narrative of institutional growth and regulation of imperial archives detailed above: while details remain unclear, the growth of imperial archives would have entailed a concomitant expansion in security regulations designed to restrict access to the archives and the activities taking place therein.
Concern about oral leaks of information, however, hardly went away, and the connection between the regulation of space and the regulation of information was not limited to imperial archives and confidential documents.Footnote 66 After the reign of Wudi, we begin to see regular charges of “leaking words from the inspection zone” (泄省中語).Footnote 67 Most of the accusations mentioned in the Hanshu came from Shi Xian 石顯, the powerful court insider during the reign of Yuandi 元帝 (48–33 BCE), who included “leaking words from the inspection zone” among charges lodged against several officials he believed threatened his position.Footnote 68 We need not conclude that all such accusations were trumped up pretexts to be used in politically based attacks.Footnote 69 At the same time, accusations of leaking necessarily invoked court regulations even while they simultaneously indicated social status, access to privileged spaces, and proper comportment as a good official.Footnote 70 This latter point receives confirmation in a story about Kong Guang 孔光 (d. 5 CE), the director of the Secretariat, who was famous for being so circumspect that he refused to mention even the type of trees growing within the inspection zone.Footnote 71
It is not necessarily the case, however, that not leaking indicated fidelity to the law and the good of the empire. Depending on the context, circumspection could actually subvert imperial institutions. The Hanshu records a scene, for instance, in which Zhu Bo 朱博, while serving in 15 BCE as Metropolitan Superintendent of the Left (Zuo Pingyi 左馮翊), recruited a man from his district to be a personal client and spy. After discovering the man had paid bribes to avoid a mutilating punishment, Zhu dismissed all attendants, called in the man for a personal meeting, and gained his confession. Rather than penalizing him, however, Zhu offered total amnesty in exchange for becoming an ally. Naturally, the man agreed to serve his new patron to the death. With a warning to “not leak my words” (毋得洩語), Zhu asked the man to put down in writing anything he heard that might be to Zhu's advantage. The story concludes by stating that the man subsequently became Zhu's intimate confidant, his “eyes and ears” (耳目).Footnote 72 Such recruitment and employment of spies by officials was no doubt common,Footnote 73 but the specific reference to leaking vividly shows the bond Zhu forged with his new underling. In the context of the Hanshu narrative, the warning against leaking immediately underscored the privileged nature of their relationship, one that existed outside of the bureaucratic structure. Circumspection and “leaking,” then, were contingent behaviors whose moral value (not to mention legality) depended on the person or institution receiving the protection of confidentiality. Who or what deserved circumspection and why? The correct answers to such questions were not necessarily obvious, but the fact that by the late Western Han such dilemmas were being framed separately from the ruler himself (e.g. leaks were from the “inspection zone,” not the ruler) shows how far the problem had evolved from Warring States times.
Everybody at court, then, including the emperor, faced pressure from norms and regulations that proscribed certain kinds of leaking. It was in this context that the “Xici zhuan” passage translated earlier in this article, though only the first portion predicting disaster if rulers, officials, or secret affairs are indiscreet (bu mi 不密), emerged as a commonplace quoted in statements from late Western Han officials. Explicitly characterizing it as a line from the Changes (Yi 易), Liu Xiang 劉向 (79/78–8 BCE), cited the phrase toward the end of a memorial that decried the influence of the imperial distaff (waiqi 外戚) Wang 王 family and urged Chengdi to distance himself from its powerful members.Footnote 74 For Liu, then, the holistic, tripartite understanding of circumspection articulated in the line supported his case that the Wang family rightly remained outside the regular order of the imperial court. It is somewhat ironic, then, that when he ascended the throne of his newly established Xin 新 dynasty in 9 CE, Wang Mang 王莽, scion of the Wang family, also cited the phrase, though only its third and final clause—“when pivotal matters are not kept discreet then disasters will occur” (機事不密則害成)—to explain the danger of leaking from the inspection zones and the Secretariat.Footnote 75
Meanwhile, just a few years earlier, during the reign of Aidi 哀帝 (r. 7–1 BCE), the high official Shi Dan 師丹 (d. 3 CE) was accused of letting memorials leak (lou xie) so that others “transmitted copies and spread hearsay throughout the empire” (傳寫流聞四方). Arguing for an investigation, the accusers invoked the second clause of the “Xici” statement, that “officials who are not discreet will lose their life” (臣不密則失身). Though Shi Dan was accused of leaking written memorials, in a careful statement after the investigation Aidi equated his high officials as “Our vital organs” (朕之腹心) and criticized Shi for a more general lack of circumspection, mentioning his “deliberations that were not discreet” (lü bu zhoumi 慮不周密).Footnote 76 This final move by Aidi, from a specific act of leaking to an assessment of Shi Dan's overall behavior and character, shows that the emergence of regulatory proscriptions against leaking oral and written statements from secure palace zones did not replace a holistic understanding of circumspection as the mark of a perfectly refined person. Indeed, the truly cultivated official would ideally embody regulatory norms in such a way that the laws themselves would be unnecessary.
In his “Shangshu zhen” 尚書箴 (Admonition on the Secretariat), the late Western Han exegete Yang Xiong 揚雄 (53 BCE-18 CE) advanced precisely such a vision of exemplary service, in the process suggesting that the best kind of official followed the same norms as the ruler and could in fact more capably accord with them. Writing in the voice of an idealized incumbent, Yang Xiong described the behavior of a director of the Secretariat perfectly fulfilling the duties of the office. Yang celebrated the Secretariat as the one “Issuing and reporting on the royal commands / As the throat and tongue of the king” (出入王命 / 王之喉舌).Footnote 77 He then celebrates the perfect cultivation of the ideal Secretariat:
Midway through, however, the poem turns toward rulers, writing “The Annals criticizes leaked words / The Changes says that without confidentiality ministers are lost” (春秋譏漏言 / 易稱不密則失臣).Footnote 79 Yang thus paired our now-familiar Changes statement about rulers with a reference to a passage from the Annals (Chunqiu 春秋), which states that the pre-imperial state of Jin 晉 put to death a high official, after which another man fled to a different realm. According to the Gongyang 公羊 and Guliang 穀梁 commentaries to the Annals, the passage refers to an incident in which an official successfully convinced the king of Jin not to send a particular general on a military expedition. The king then demoted the general, explaining to him the advice he had received from the official. In retaliation, the demoted general killed the official and fled. Both commentaries blame the ruler for this turn of events, condemning him for “leaking” to the general.Footnote 80 By applying stories about leaking rulers to illustrate the ideal behavior of a model Secretariat, the admonition both confirms a shared norm of confidentiality that applied equally to all figures at court and suggests that a Secretariat director who conformed to the duties of confidentiality relevant to his office had the potential to act in a more responsible and upright manner than an actual ruler.
Coda: Leaking and Laws, Politics and Position
Histories of spy-craft tell us that ancient and modern states alike have developed sophisticated measures for gathering and deploying intelligence for strategic purposes in war or political struggle. States, concerned as they inevitably must be with governing populations, enacting policies, and ensuring security, are necessarily interested in managing and controlling information. Such a concern seems characteristic of almost all states, whether or not they seek monopoly control over the accumulation and dissemination of this information. While more detailed study of laws, regulations, and practices related to secrecy and leaking in early China might yield helpful information about innovative solutions, it would not upend our understanding of the underlying problem that confronted the contending realms of the Warring States period and the Qin and Han empires that succeeded them. The evidence assembled in this essay thus suggests that the significance of regulating confidentiality and leaking can and should be understood in terms broader than the strategic requirements of states or the legal and bureaucratic procedures designed to ensure confidentiality. Rather, secrecy and the fear of leaking helps illustrate the convergence of laws, institutions, political considerations, and status, all of which informed rhetorical representations of court actors. The subject allows us to trace a different kind of legal and administrative history through an investigation of sources that can by no means be understood strictly as “legal” or even “administrative.”
Indeed, this article explored a history of leaking that traces back not to the strategic demands of states and militaries, but to visions of sagely rulership that drew upon medical concepts, particularly qi circulation, and self-cultivation practices. Without repeating all of the interpretations advanced above, we might note that across the medical-philosophy continuum, from a medical body that leaked essences to achieve homeostatic balance to the political body of a ruler that was “bound off” (zhou) and confidential, the boundaries that separated inner from outer were an unchanging but implicit concern. Such concerns were made entirely explicit in the Shiji stories of the First Emperor, tracing as they did the emergence of spatial and institutional boundaries at court to the emperor's concerns about maintaining secrecy and preventing leaks that denied him access to the spirits. A prominent story during Western Han, then, was that the divine status of the ruler's person, not just security or strategic concerns, drove secrecy regulations at the imperial court. The anecdotes in the Han Feizi explored the implications of such a spatially and institutionally segmented court, bound by confidentiality, that both rulers and officials were forced to navigate.
By the late Western Han, this trend toward shared norms and confidentiality regulations had accelerated, along with a growth in secret spaces and institutions bound by quasi-legal proscriptions (e.g. to not leak secret documents or conversations from the “inspection zone”) that seem to have almost taken on a momentum and logic of their own, or at least a logic removed from the direct concerns and actions of the ruler. Leaking or not leaking thus presented ethical problems, since circumspection could solidify extra-official factions as much as the stability of the entire court itself. Nonetheless, the popularity of the “Xici zhuan” line and Yang Xiong's poem on the Secretariat shows that court officials did not hesitate to elevate regulatory demands for secrecy to an almost sanctified norm equally expected of rulers and officials. Concerns about leaking, as expressed through regulations and ethical standards, rooted in practices of self-cultivation and idealized patterns of behavior, provided a language for officials to partake of a divine power that had settled across the entire court, including the ruler but by no means limited to him.
Even if writers in the Eastern Han and post-Han period continued to elaborate on the problem and dynamics of leaking in new ways,Footnote 81 such a reflexive understanding of this status of the court, and its implications for leaking secret information, endured. An illustrative example, which relates to the changed understanding of leaking charted in this article, can be seen in commentary by Du Yu 杜預 (222–284 CE) to a passage in the Zuozhuan 左傳 (comp. ca. late fourth century BCE?). For such a massive text, full of stories about political conflict, anecdotes about leaking are surprisingly few. One, however, describes King Kang 康王 (d. 545 BCE) of Chu 楚 issuing an order to execute his chief minister. Before actually carrying out the execution, the king met three times with the minister's son, who happened to serve in the king's royal guard. After eventually mustering the courage to tell him that his father was going to die, King Kang asked the son and guard if he would remain in royal service. He responded by noting that the king would undoubtedly be unable to keep him in the guard post after his father was executed. At the same time, he stated, “leaking your order (ming) will result in a heavy punishment, so I also will not do that” (洩命重刑,臣亦不為).Footnote 82
The story, then, associates leaking a ruler's “command” (ming 命) with punishment, but interpretive difficulties make it impossible to clarify the precise connection between the two: we can, after all, understand the punishment not as a specific reaction to leaking a command, but rather a potential result of a series of events that would be set in motion by the leak. In King Kang's plan to execute the prime minister, the prime minister's son might have feared that if he betrayed the king's plans, his father would launch a pre-emptive rebellion that would lead to his own execution.Footnote 83 Taken on its own terms, then, the Zuozhuan anecdote highlights the risks of elite politics, but does not go much further. Du Yu did go further, however, papering over ambiguities within the story and arguing that the actor simply wished to avoid punishment, since “leaking a ruler's command explains the heaviness of the crime” (漏泄君命,罪之重).Footnote 84 Du Yu, writing decades after the collapse of Eastern Han, thus cast leaking a ruler's command as a formally proscribed act that merited punishment, suggesting a legalistic, cause-and-effect relationship between infraction and response so reflexive as to be mechanistic. Based on the evidence assembled in this article, however, we should not solely cast Du Yu's interpretation as the product of an ever-more elaborate legal code designed to privilege the ruler's power. It can also be understood as a coda to the story told above, an accretive process by which different understandings about leaking and secrecy at the imperial court layered on top of each other. Writing in this context, Du Yu transformed a story of back-stabbing political intrigue by assuming a conventionalized understanding of rules and norms that constrained the statements and speech of court actors, including rulers and officials alike.