Slavery in Late Ming China

Abstract

Slavery is not a commonly discussed topic in late imperial Chinese history—indeed, as much as late imperial China is not an area usually discussed in world and global slavery studies. The ubiquity of the “feudal” analytical framework in Chinese historiography and the image of the archetypal “slave” have resulted in the widespread idea that there was no slavery in (early-) modern China (and Eastern Asia). This chapter reevaluates this widely accepted preconception by shedding light on the overlooked and unique succession of “slave revolts” that erupted in China’s economic centers between the 1630s and the 1660s. The focus on this particular moment in Chinese history and on the sources produced in its aftermath serves as an entry point to understanding the life experiences of Chinese nubi, a category of enslaved people that persisted until its formal abolition in the last years of the imperial regime.

Introduction

Was there slavery in late imperial China? Rigorously speaking, the answer to this question should be no. The medieval Latin word “sclavus” (and all its declinations) was not part of the languages used in late imperial China. Furthermore, as a universal category, “slavery” did not enter Chinese societal and legal debates until the end of the nineteenth century. And since the fall of the imperial regime in 1911, historians have hardly ever associated any form of late imperial bondage with “slavery,” whatever its definition.

At the same time, from the sixteenth century onward Western observers frequently reported practices “analog” or “similar” to slavery (in today’s parlance) in China. When the imperial government abolished the “buying and selling of people” in 1910, it presented the reform as a commitment to the abolition of slavery (in line with the international standards of the so-called “civilized” nations).¹ Also, in spite of their reluctance to use the “slave” label, many historians of China still implicitly define late imperial bondspeople with reference to archetypal “slaves”—i.e., as persons subjected to extreme forms of social and legal discrimination, over whom masters exercised rights tantamount to “property” for the purpose of “exploitation.” As a consequence, analyses of human bondage in late imperial China often leave historians with an impossible conundrum: to translate bondage categories as “slaves,” at the risk of distorting the Chinese context considerably; or not to translate, at the risk of perpetuating unintelligibility and the all too widespread (and similarly distorting) perception of China’s cultural alterity.

Despite its narrow focus on labor relations, the approach proposed in 2016 by Marcel van der Linden offers an interesting alternative to “universal” taxonomies of social relations of domination and an opportunity to overcome the conundrum.² This chapter thus attempts to “dissect” nubi bondage in late Ming China (1368–1644) in a more dynamic way. As a social and legal category, nubi shows many similarities with “slaves” in other historical contexts. It existed long before the Ming and persisted long after the fall of the dynasty in the middle of the seventeenth century, as well as after its legal abolition in 1910. Nubi also played the role of a conceptual matrix used to define various relations of subordination in Ming China. And it was present, with substantial variations, in medieval Japan (nuhi) as well as in early modern Korea (nobi).³

For the purpose of contextualization, this chapter focuses on one singular moment in the changing history of nubi bondage. The late Ming crisis and the violent transition from Ming to Qing rule (1644–1911) saw the eruption of many armed revolts across China, including hitherto unknown nubi revolts. Nubi revolts were so unique in Chinese history that contemporaries felt the need to forge a new term to characterize them (nubian, literally “slaves’ catastrophe”). They wrote extensively on the topic of nubi revolts, thereby providing invaluable information about late Ming bondage practices. This chapter offers a brief introduction to nubi revolts and uses the sources produced in the aftermath of this particular moment as an entry point to analyze the three “moments” of nubi bondage.

Enslavement Contracts: The Fiction of Voluntary Bondage

Let us start with one short document selected both for its exemplarity and its uniqueness. By a contract dated 23 December 1645, a man named Jiang Guanda from Qimen district (Huizhou prefecture) was returned to his “former master.” Cataloged as a “self-sale contract” by the Museum of Anhui province, the document reads as follows:

New marriage contract by Jiang Guangda.

Being unable to feed and clothe myself because of scarcity, I formerly called on a middleman to be sold, as a couple with my wife (two persons), and placed in the service of the Hong [family]. As a marital present, I received the full sum of sixteen taels of pure silver [ca. 600g]. After we crossed the gate [of the Hongs’ house] we served without fail. During the yiyou year [1645], as the seditious Wan Biao and others enlisted crowds and built an organization to plunder riches and kill masters, I wrongly joined the crowds. I retrieved the original contract from my master without returning the marital present. [Since then,] I have submitted to my punishment without a word. Now that [the situation] has been settled and clarified, thanks to pawnbroker Siguan, I establish a new marriage contract, by which I am returned to my former master, whom I will serve forever. From now on, would I turn my back again on my master and run away or steal, may I be fully punished according to the law. Fearing that there is no trace, I establish the present [contract] as proof.⁴

Like many Chinese contracts transferring authority over a person to the head of another household (as a wife, concubine, uxorilocal husband, bonded tenant, adopted child, etc.), Jiang’s contract is labeled as a “marriage agreement.” It is nonetheless representative of the few enslavement contracts preserved from the Ming-Qing period.

Using standard formulae that circulated in vernacular almanacs, the contract briefly recounts the circumstances that led Jiang Guanda and his wife to sell themselves and, later, to be returned to the Hong family. As the principal contracting party, Jiang singled out poverty as the main incentive to enter into a bondage agreement. Destitution was a common motivation for selling oneself and one’s children, but it was also a necessary justification. Laws and norms during the Ming period did not prohibit self-sales, but they made enslavement a monopoly of the judiciary and ownership of enslaved people the privilege of a small and poorly defined group of “meritorious officials.” In theory, “honorable people” (i.e., commoners—as opposed to “mean people,” a social and legal group encompassing nubi, entertainers, outcast communities, and unclean professions) could not be privately enslaved. In practice, however, (self-)sales were tolerated as alternatives to starvation. Thus, almost all surviving contracts from early-modern China (like Jiang Guanda’s) open with ready-made references to “poverty” to legally justify enslavement.

Another typical feature of this contract is its emphasis on the “voluntary” nature of the transaction. Whatever the purpose of a sale, the holders of legal authority (usually parents) had to declare their consent in the opening section of an agreement. These statements served to defuse suspicions of coercion and trafficking. Indeed, whereas self-sales and sales of children were tolerated on grounds of poverty, “trafficking” was illegal and fiercely repressed.⁵ As Chinese families were by nature “transactional,” trafficking did not amount to buying and selling people; neither was it limited to trading slaves.⁶ What defined trafficking was the act of usurping the power held by others over their subordinates (children, wives, concubines, or nubi). Consent from the legitimate owners and guardians was thus an absolute precondition for lawful transactions.

As it unfolds through preserved written agreements like Jiang Guanda’s, nubi bondage thus presents the characteristics of a form of contract and of voluntary slavery. Borrowing from Orlando Patterson’s terminology, enslavement was the result of a process of “extrusion”: individuals were extracted from the society of “commoners,” severed from their natal kinship ties, permanently relegated to the socio-legal group of “mean people,” and re-socialized as nubi under the exclusive authority of the father-like figure of a household head.⁷ Combined with the prohibition on trafficking, the transactional nature of enslavement seemingly made it a non-coercive process: the choice was formally made “willingly” by oneself or by one’s father (whose authority was indisputable), and the relationship so created was understood as mutually beneficial. Consequently, “coercion” was not even conceivable in that context.⁸ This helps explain why bondage in late imperial China is often presented as a form of “humane” and “mild slavery,” or sometimes even as a poverty-relief system of sorts. Yet, this picture of a pacified form of bondage should be taken with a grain of salt. Like normative sources, contracts are crucial to appreciating the underpinning rationale for enslavement, but they mainly convey the perspective of the masters’ society, not the actual practices.

Coercion can only be suspected behind the standardized phrasing of late imperial contracts. The stereotyped mention of “scarcity,” for instance, says nothing about the true motivations for pledging oneself and one’s wife to “serve forever.” Even if no pressure was exerted on Jiang, his decision was still the result of a “constrained choice” dictated by economic hazards. Coercion also transpires behind the polished depiction of Jiang’s apparently straightforward submission to punishment and voluntary return after a liminal phase of escape and violent resistance. His participation in a revolt aimed at “plundering riches,” “killing masters,” and retrieving enslavement contracts also suggests resentment against a situation he had allegedly entered “willingly.”

Reference to armed insurrection is probably the most unusual detail in Jiang’s contract. Late Ming nubi revolts were so unprecedented that they drew unusual attention to bondage practices (a topic formerly deemed too trivial to be worth writing about). The sources produced during that particular moment in Chinese history are extremely useful to explore the many dimensions of nubi bondage and to dissect its coercive dimensions. The revolts themselves, a collectively unsuccessful attempt to “exit despite impediments” in van der Linden’s terminology, are little known to historians of slavery, and require a brief presentation.

From Shadows to Light: Nubi Revolts

Nubi revolts erupted amidst the multifaceted late Ming crisis. They lasted for three decades and culminated in the cataclysmic years of 1644–1645—better remembered for the fall of the Ming, the Manchu conquest of China, and the sequels of enduring “peasant revolts.” Starting in the 1630s, nubi took up arms in more than 30 districts and caused “a catastrophe like no other in thousand years.”⁹ Some revolts lasted for years, but most were short-lived and unfolded in a similar pattern. Fueled by the nearby presence of rebel and military troops or by rumors of emancipation allegedly decreed by the Manchus, nubi rose up in “hundreds,” “thousands,” and “tens of thousands” to bring destruction upon their masters. They burnt and pillaged houses, humiliated and killed masters, and systematically demanded the return of their contracts. Soon enough, they faced brutal repression, only to be remembered as renegades by the new dynasty, although it had occasionally used them as local proxies in its conquest.

The sources documenting nubi revolts were mostly written to condemn the insurgents. Very few authors attempted to dig into their root causes and to propose reforms, but all attributed them either to the “ungratefulness” of the enslaved or to the excesses and brutality of their masters. Marxian historians, who have written extensively on the subject, mainly analyzed these revolts as another manifestation of a nascent “class consciousness” in a context of exacerbated “antagonisms” and inequalities. Their conclusions are based on solid evidence, but nubi revolts are more complex and multicausal than it seems.

Nubi revolts had at least one common denominator: opportunity. With very few exceptions, nubi never took arms in isolation but in the context of a surrounding insurrectionary environment (the advance of the Manchu Banners or the struggles of Ming loyalists and the so-called “peasant” armies). In other words, although persistent nubi unrest is attested long before the end of the Ming, its transformation into full-fledged revolts was directly correlated to the collapse of the sociopolitical order that, in ordinary times, contributed to maintaining nubi under control and to empowering masters.

Opportunity, however, does not explain why this form of insurrection erupted at this particular moment in Chinese history. The transformation of bondage practices provides further explanation. Despite the absence of even approximate figures, evidence points at a steady increase of privately-owned nubi from the fifteenth century onward, reaching unprecedented numbers by the end of the Ming. The economic growth of the Ming period, the commercialization of the economy, urbanization, and a growing taste for luxury goods and external signs of wealth all contributed, among other factors, to increasing the demand for enslaved manpower. The concentration of land ownership, rampant inequalities, the burden of taxes, and the effects of a disastrous social and humanitarian crisis simultaneously contributed to fuel the supply side of the market.

The greater presence of nubi in late Ming society not only contributed to raising the levels of violence used to control the enslaved, as evidenced by many sources, but also to increasing their role as agents of violence in an often-brutal society. If the majority of nubi were strictly disciplined and sometimes subjected to unbridled physical violence, a growing minority was no stranger to exercising brutality on their own account or as the “claws and teeth” of their powerful masters. Employed in private militias, used as the strong arms of their masters, and sometimes connected to groups of outlaws, many revolt leaders were ambiguous figures with a shady past who maintained close ties with marginals and who had formerly been entrusted with prominent roles in the local defense against other insurrections. Pan Mao, the leader of the Liyang revolt (Jiangsu, 1645), for instance, was a former “hereditary servant” turned soldier and brigand, who set up a rebel organization after he was handed the defense of the city against the Manchu armies. The unprecedented numbers of the enslaved and their familiarity with the use of violence played a significant role in the ways nubi resentment unfolded in the late Ming insurrectionary context.

The greater place of nubi in society also contributed to changing attitudes toward enslavement. Although (self-)enslavement was permanent and hereditary, in the late Ming period, an ever-growing fraction of the population seems to have been conceiving it more as a convenient lifeline to overcome transitory economic difficulties. The resultant ambiguities are well illustrated by the variety of demands formulated by the insurgents. Nubi revolts are usually depicted as a homogeneous struggle for “personal freedom.” Some of them undoubtedly demanded unconditional emancipation and propagated egalitarian mottoes. In Jintan (Jiangsu, 1644), for instance, nubi took an oath at the city temple and decreed: “Heaven and Earth are tumbling, the honorable and the mean are reversed, why must we remain enslaved forever?”¹⁰ Others, however, rather demonstrated a shared impatience with changes to the rules of bondage, in particular the ways out of it. This is well exemplified by the demands of Yu Boxiang in Taicang (Jiangsu, 1645) that bondage be limited to one generation and no longer hereditary. Another illustration is provided by the third revolt that broke out in Macheng (Anhui) in 1651, sparked by the rumor of an edict allowing nubi to redeem themselves. Before taking arms, their leader, Fang Jihua, attempted to negotiate changes in bondage practices. A “Proposal about enslaved and masters in nine points” was submitted to the local gentry. Its content has not survived, but it seems to have concerned, among other things, the fiscal liability of enslaved people, rather than immediate emancipation.¹¹

Changing attitudes toward enslavement not only surface in the ways in which people engaged with enslavement in the late Ming, but also in numerous texts of moral inspiration reflecting upon the nature and functions of bondage in society. In a period of growing competition and uncertainty (even for literati), marked by increased social and geographic mobilities, intellectual speculation about the human condition flourished. The greater visibility of bondage, its manifest abuses, a growing sense of crisis, and the perceived threats that nubi posed to the social order drew significant attention to the ethical standards of enslavement. Although a few radical thinkers spread subversive messages attacking traditional hierarchies and questioning the moral foundations of enslaving others, the reflection on bondage mostly remained within the framework of Confucian ethics and aimed at correcting excesses and restoring social harmony. In the process, late Ming moralists outlined the contours of a proper and ideal(ized) vision of human bondage. This vision insisted on acknowledging that nubi shared the same human nature and ought to be treated accordingly. The ideal they professed was marked by “benevolence” toward fellow humans on the verge of starvation. They considered bondage as a mutually beneficial relationship framed by “human sentiments” that could not arise from taking advantage of the weakness of others. In general, this vision was no more than a form of paternalism that did not challenge established hierarchies and the existence of human bondage. It nonetheless contributed to set ethic boundaries to human exploitation and to weaken the ideas that nubi were inferior by nature and that bondage was by default permanent. Its effects on practices remained marginal, but they can be sensed in late Ming judgments on contested nubi identities and in the steady development of charitable activities to help the poor and to free the enslaved. This vision of bondage also contributed to influencing a population of enslaved who was not completely immune to popular education and its subversive messages.

The late Ming nubi revolts thus epitomize the tensions that pervaded society in general and bondage relations in particular. They highlight the many ambiguities of bondage and the discrepancies between practices and norms. In Confucian terms, denominations no longer matched realities. Together with other factors, these contradictions help explain the eruption of nubi armed violence.

Entry: Becoming Nubi

How were individuals enslaved in late Ming China? Enslavement was not among the five regular punishments prescribed by Ming law. Although one of them, “penal servitude,” is said to have originated in ancient slavery, it then mainly consisted of temporary penal labor. However, Ming-era normative sources and legal specialists still envisioned punishment as the only lawful way to produce nubi. According to jurist Wang Kentang (1549–1613), enslavement was an ancillary punishment imposed on the families of criminals convicted of crimes against the state (treason, rebellion, and sedition). As he notes, nubi were “men and women convicted by extension of liability” and “seized by the administration.”¹² The founder of the dynasty had sentenced people to enslavement on various occasions, but in its final version (1397), The Great Ming Code included very few provisions for penal enslavement. By the end of the dynasty, punishment by law did not produce significant numbers of enslaved. Neither did capture in war, which was sometimes considered by legal commentators another legitimate source of nubi. Like penal enslavement, war capture is poorly documented and does not seem to have reached levels comparable to those observed later during the Qing conquest of China and Western territories.

The main and unanimously accepted path to enslavement in late Ming China was “sale for a price” (i.e., private contract enslavement mediated through money). In the ideal framework depicted above, bondage was socially acceptable only when mutually beneficial and voluntary. Reciprocity and willingness were deemed necessary to instill in nubi a sense of gratitude and to produce a harmonious relationship in which they would be “content with their lot,” and naturally obey and respectfully submit to the authority of the head of the household. In exchange for surrendering themselves to a master, nubi (or their parents) frequently received monetary compensation upon signing the contract and were always given assurance that their basic needs would be taken care of (including marriage and burial). Said otherwise, only the contract of voluntary (self-)sale—or the illusion of it—guaranteed the proper fabric of a natural bond shaped by “human sentiments.”

Due to their lack of reciprocity, alternative modalities of entry into bondage were either controversial or denied any legitimacy. “Being born in the household” was another common way to become nubi, as illustrated by a provincial proclamation issued in the aftermath of the 1651 revolt in Macheng, which listed only four “legal” ways into bondage: penal enslavement, war capture, sale, and birth.¹³ After the last serious nubi revolt in Guangshan (Henan, 1658), prefect Jin Zhen (1622–1685) nonetheless proposed to limit entry only to “sales for a price.” Jin Zhen did not mention birth, but he might have had in mind the violent rejection of heredity voiced during the revolt. To him, apart from selling oneself and being sold by parents, other modalities of entry were disloyal and unlikely to create a balanced, reciprocal, and solid relationship. Among those, “commendation” was considered particularly harmful.¹⁴

Commendation proceeded from a more genuinely “voluntary” choice, but its legitimacy was strongly disputed. Commendation took various forms. The practice known as toukao (literally to “pledge” oneself and “lean on”) consisted in offering oneself (or money, according to some authors) to an influential master in exchange for subsistence and protection. Another practice, known as touxian (literally to “pledge” oneself and “offer”), consisted in securing similar protection by offering one’s properties (or the properties of one’s former master). In the first years of Qing rule, commendation also consisted in giving oneself to members of the Manchu Banners (touchong), although the practice was outlawed in 1646. Commendation had become so common in the late Ming that various authors recount how cohorts of aspirants rushed to the lists of examination laureates in search of the name of a potential master to whom they could offer themselves. A form of clientelism, commendation was unanimously decried as a fraud motivated only by self-interest (offering not only protection against taxes and prosecutions but also potential access to wealth and power). Late Ming masters who accepted too many commended people were later judged responsible for the loosening of hierarchies and the proliferation of disloyalty among enslaved people, who “inevitably” fought back when the opportunity arose.

Aside from “voluntary” forms of enslavement and birth, becoming nubi was also the result of physical compulsion and deception. Among the causes of the slave revolts, Jin Zhen notes that “seizing by force” was common practice in the late Ming: poor people were frequently lured by false promises or kidnapped, sequestrated, enslaved, and held captive by means of false contracts and threats of prosecution. Such practices, which funneled people into trafficking networks that purveyed various markets in people (sometimes over very long distances), are well documented all across Ming and Qing China. The Ming Veritable Records, for instance, mention the case of a gang of traffickers who, in the 1480s, abducted vagrant and displaced people in Jiangxi province to provide purchasers with “household slaves” and “tenant slaves.”¹⁵

Forcible enslavement was also a potential outcome of indebtedness. Ming officials repeatedly complained about the practice of seizing debtors (and their relatives) in repayment of a debt, which was not only immoral and illegal,¹⁶ but also decreased tax incomes. Although illegal, the practice was common enough for pettifoggers to circulate model complaints against the undue appropriation of children by moneylenders.¹⁷ Yet, children could also be placed as collateral for loans. Models of pawn agreements read like nubi contracts and leave no doubt as to the outcome of a default on repayment.¹⁸ If forcibly seizing debtors was illegal, pawning a child only differed from an outright sale in that it was delayed and conditional.

Finally, two typical forms of compulsion are worth mentioning: the diversion of other social and labor identities; and a phenomenon that could be called contamination. Jin Zhen, again, testifies to the practice of forcing hired workers (gugong) and tenants (dian) into permanent bondage. The legal status of “hired worker” was very similar to nubi, with the difference that it was temporary and that gugong were considered “mean people” only in relation to their employer. The practice that Jin Zhen describes was a mix of hired work contract and uxorilocal marriage, by which men agreed to work for a fixed period of time in exchange for an enslaved woman to marry (one such contract set the term to 22 years).¹⁹ Yet, once the term was over, employers threatened to report gugong as “runaway slaves” and retained them permanently. Tenants, for their part, were “honorable people.” Landowners, however, had made it a habit to call them “tenant-slaves” (dianpu) so as to extract labor and services indefinitely, increase the rent as they pleased, and even sell their children and wives, as they could with nubi. The practice was not limited to Guangshan district. It is evidenced in Huizhou prefecture, in Guangdong province, and all across Southern China. It was also common enough in the late Ming for pettifoggers’ manuals to include model accusations for “confusing tenants with slaves.”²⁰

The late Ming period is thus characterized by a widespread extension of nubi status to other contract relationships (hired work, uxorilocal marriage, tenancy, but also adoption) and a blurring of social and labor identities. This extension sometimes took the form of more subtle contamination. With the importance given to the concrete performance of social roles in assessing nubi identities, acting like an enslaved person could lead to actual enslavement. Such was the fate of a man named Tang Yuan in the early seventeenth century. For reasons that remain unclear, Tang had moved with his aunt and her husband, two nubi in the service of the Han family in Huating district (Jiangsu). Tang was undoubtedly of commoner status, but for sharing the roof of his relatives’ master for years and for accepting money to help him marry, Tang was reported as a runaway slave when he left to settle on his own. Although Tang ultimately committed suicide, the magistrate concluded that he had benefited from the Hans’ benevolence. This fact alone had created a bond similar to signing a bondage contract. In the magistrate’s terms, it had created a “difference between master and slave” (zhupu mingfen).²¹ For failing to acknowledge the porosity of nubi status, Tang Yuan chose a radical exit (suicide) over a life in bondage. The Han family, on the other hand, knew that demonstrating benevolence combined with the passing of time were key factors in establishing nubi status.

In the late Ming, the paths leading to bondage had multiplied and extended far beyond the framework envisioned by the founder of the dynasty. This multiplicity led to a steady growth of the nubi population. It also contributed to changes in the perception of bondage as a permanent condition, and in the meaning of being enslaved.

Nubi Experience as Life in “Service”

Historians disagree about the most adequate way to characterize nubi bondage. Many studies have attempted to evaluate the level of oppression and violence nubi experienced, to elaborate taxonomies based on the tasks they performed, or to measure the discrimination they suffered. Approaches based on material factors (on labor in particular), however, face several limitations. First, because the sources available seldom provide information about the labor that nubi performed. Second, whatever the chosen criteria, nothing was more different from one nubi than another nubi. Third, no criterion alone seems likely to distinguish nubi from other social groups.

Characterizing nubi according to their functions and labor roles thus generally proves unhelpful. The ideal-type male tilling the soil and female doing household work was still prevalent in Ming times, but in terms of gender distribution of labor, this was not different from what was expected in commoners’ households. Aside from household and fieldwork, nubi performed multiple tasks and were present at almost every level of every sector of activity, including, in extremely rare cases, the upper echelons of the prestige ladder, as imperial examinations laureates. Nubi could be land tenants, porters, estate managers, militiamen, traders, artisans, accountants, wet nurses, housekeepers, gardeners, shopkeepers, envoys, etc. They performed many tasks on a single workday and at various stages in their life cycles. Some of them had no productive utility at all and did not perform work. They were nonetheless always useful to their masters, like the ostentatious processions of fine-dressed retainers with whom rich scholars paraded on the streets in the sixteenth and seventeenth centuries.²²

So, rather than searching for one distinctive labor condition, let us look at the variety of nubi conditions through the lens of their autonomy. Taking autonomy as an analytical criterion, Ming-Qing observers as well as modern historians have come to distinguish two extreme archetypes of nubi. Seemingly irreconcilable (and both deemed equally responsible for nubi revolts), these archetypes illustrate the apparent contradictions of nubi as a condition and the range of autonomy that they enjoyed in practice.

On the one hand, we find the great majority of nubi living in extremely harsh conditions and subjected to ongoing violence. In an ideal relationship shaped by “natural feelings,” violence was deemed unnecessary and should have remained exceptional and moderate. From the precepts for nubi management outlined in lineage regulations, however, masters were much more concerned with discipline and control than with demonstrations of benevolence and compassion. The brutality of nubi life is well documented, and the majority of authors testifying to it can hardly be suspected of excessive sympathy. An exhaustive list of the recorded forms of violence would be endless, ranging from deprivation to cruel punishments and sexual abuses. To Zhang Lüxiang (1611–1674), late Ming masters had simply ceased to treat nubi “with humanity” as they were expected to. Many suffered from cold and hunger, worked from dawn to dusk, and performed “hundreds of services” until exhaustion. They were often denied the right to mourn their parents; their wives and daughters were sexually abused; and it was frequent, says Zhang, that masters killed nubi and disposed of their bodies without being prosecuted.²³

The levels of violence unleashed by nubi during the Ming-Qing transition also offer a mirror image of everyday brutality. Retaliation against cruel masters was the most common incentive for nubi to take up arms. Besides killing masters, insurgents performed rituals of role inversions and brought their masters to judgment. They forced them to serve, throwing and beating them on the floor as soon as they failed to comply with orders, as was the case in Ji’an (Jiangxi) in 1644. Zhang Mingbi (1584–1653) extensively recounts how such trials were carried out in Jintan in 1645: while beating their masters, after each blow and after each cry of pain, nubi kept asking “does it hurt?” and why they used physical violence when they knew how hurtful it was.²⁴

On the other hand, we also find nubi enjoying high levels of autonomy. In Huizhou, where land was scarce and properties scattered, nubi were often settled on isolated properties like tenants to cultivate the land and watch over their masters’ remote properties (like graveyards and precious plots of commercial woods). There are also testimonies of nubi leasing their workforce as tenants to others than their owners.²⁵ Evidence of wealthy and powerful nubi also abound in late Ming sources. An anecdote from the Miscellanea about Songjiang recounts the stupefaction of one Mr. Xu when discovering that an old nubi of his owned a bed padded with sable fur and dressed with dragon-embroidered robes.²⁶ However, the archetype of the autonomous nubi is mainly epitomized by the despised figure of the so-called haonu (literally “porcupine slaves,” or “brazen bondservants”).²⁷

The antithesis of the enslaved who were exploited and oppressed, haonu have been depicted as an uncontrolled body of elite slaves who abused their masters’ influence, reputation, and protection for their own benefit. To Shen Li (1531–1615), they accumulated wealth by diverting their masters’ properties, tyrannized the local society, changed masters as they pleased, and ultimately caused the ruin of many families.²⁸ Haonu, however, is neither a social nor a legal category. It is a judgmental label, a criticism of the permissiveness of masters, and a denunciation of the diversion of nubi status for the purpose of personal advancement. As noted by Wang Jiazhen at the beginning of the Qing dynasty, haonu were people who “fed from their owners’ benevolence,” and a serious threat to the local social order. This was illustrated by the numerous inter-familial vendettas led by haonu and by the frequent eruptions of local violence provoked by the exactions they committed for their own sake or in their masters’ name. One such famous example is the episode that opposed the inhabitants of Huating to the renowned painter and statesman Dong Qichang (1555–1636).

By the level of autonomy they demonstrated, the wealth they accumulated, and the power they exercised in local society, haonu and their involvement in the late Ming revolts have been an embarrassing challenge to the narratives depicting nubi as a homogeneously oppressed “class.” Some historians chose to simply ignore this contradictory figure. Others concluded that haonu belonged to the “ruling class.” This apparent contradiction, however, can be resolved by highlighting what best characterizes nubi is that they were people who “served” (usually “forever”). From contracts to lineage regulations and court cases, depictions of nubi experience always come down to this unique feature: “service” (yi)—a notion never clearly defined that cannot be reduced to “domestic service.”

In enslavement contracts, “to serve” is expounded as “obeying orders.” Wang Jiazhen, for instance, explains that once a contract was signed, nubi “would never stand and walk side by side [i.e., like brothers] for the rest of their days. When called to accomplish a task, they would not lose a second.” Wang thus underlines that nubi, as a status and as a condition, was permanent, that to be “in service” produced a strong relational asymmetry, and that it barely amounted to obeying orders.²⁹ An early seventeenth-century model accusation addressing cases of nubi abductions confirms that the essential function of nubi was to “work in place of others.” It argues that taking them away from their master amounted to cutting off the latter’s hands.³⁰ To “serve” thus meant that nubi were men, women, and children of all tasks, and that their functions and their autonomy were only determined by the demands (and the permissiveness) of their masters. In this regard, the despised figure of the powerful and protected haonu no longer appears as an anomaly. It is a manifestation of the polymorphism of nubi, whose autonomy (and lack thereof) depended exclusively on what masters allowed them to do.

Being “in service,” however, was not an exclusive feature of nubi. Hired workers, for instance, had to serve their “employers” in a similarly asymmetric relationship. What distinguished nubi, however, was that they were exclusively and permanently defined by that single role, both within the household and in society. Whereas hired workers were in a position similar to nubi vis-à-vis the household head, they remained “honorable” in relation to outsiders and within their own household. Nubi, on the contrary, remained permanently at the bottom of the social and household hierarchies. Besides, all of their formal social relations were mediated through the person of their master. When committing a crime, their position vis-à-vis the other party was not evaluated directly, but always indirectly according to the hierarchical and/or ritual bonds between their owner and the other party. In other words, the bond between nubi and master was a total and exclusive relationship, which granted masters constant and almost absolute control. If actual coercion was not necessary, its potential was present at all times. The ubiquitous analogy with the father-children relationship invested masters with almost absolute power and impunity. Like children, nubi could neither accuse their master nor disobey orders; like fathers, masters held extended disciplinary powers bestowed by the law and guaranteed by imperial and lineage justice (provided that they did not kill nubi).³¹

This potential for permanent and almost absolute control is what best characterizes nubi bondage. Compulsion could be mobilized (by threat) and realized (through discipline and reporting to lineage and imperial justice) at all times during the relationship, even against haonu. It was all the more crucially felt in the impediments to exiting the relationship.

Exit: The Unraveling Knot

Lawful ways out of bondage were very few in Ming China. Whatever the entry modalities and the levels of autonomy, exit was seldom a choice that nubi could make on their own. The two principal paths out mentioned in Ming law were redemption and emancipation. Both could be granted by masters or imposed by judgment.

Redeeming oneself and redemption by relatives were probably the most common exit scenarios. The prevalence of such lawful ways out of bondage cannot be evaluated, but according to Wang Jiazhen and others, doing so required “lots of money” and the approval of one’s master.³² Lifting the impediments to redeeming was a major demand in the late Ming. The first documented revolt (in Macheng, 1630) was for instance triggered by the rumor of an edict allowing redemption. It unfolded as a movement demanding that masters be compelled to allow redemption and that nubi gain greater control over the formal mechanisms of exit.³³

Emancipation was mainly enacted as a gesture of benevolence and gratitude. Examples of emancipations by testament or by symbolically burning bondage contracts are fairly rare and usually found in edifying stories and model biographies. Zhang Lüxiang recounts the story of a third-generation loyal nubi in Wuxing (Zhejiang), who was emancipated after taking his master out of prison.³⁴ In the context of a rampant social crisis, late Ming sources nonetheless testify to the growing practice of buying bondspeople only to grant them emancipation,³⁵ and court cases provide evidence that emancipation was an actual, albeit exceptional, prospect—as was the case for two slaves that one Mrs. He allowed to “leave the household” in her last wills.³⁶

Emancipation and redemption were sometimes pronounced by magistrates. Since nubi could not appeal to justice against their masters, traces of plights for emancipation are quite rare in Ming China. The ideal of a harmonious and reciprocal bondage relationship could nonetheless work in their favor, provided that the local authorities decided to get involved. When presented with masters who had long forsaken their (limited) obligations and who treated nubi with (unusual) cruelty, magistrates could offer redemption or order emancipation. Such situations occurred when masters were brought before justice for other crimes, or when they attempted to claim rights over former nubi property. As prefectural judge of Songjiang (Jiangsu) in the 1600s, Mao Yilu (?–1629) once emancipated the family of a man violently killed by his master, in full compliance with the Code’s provisions.³⁷ In cases of extreme cruelty, however, Mao not only formally broke the bonds of servitude but also imposed compensations. So he did when called upon to judge a man named Zhang Ying who had been apprehended for other crimes. Upon discovering that Zhang had willingly blinded one nubi named Gu Liang and sold the latter’s wife away because Gu had objected to him having sexual intercourse with her, Mao Yilu sentenced Zhang to paying Gu 10 taels of silver and a piece of land so that he could effectively “leave the household” and make a living of his own.³⁸

The frequency of emancipations and redemptions can hardly be evaluated, but the crucial point is that bondage contracts seldom stipulate the terms of exit—except in rare and ambiguous configurations mixing various identities with nubi (like tenant-slaves and bonded uxorilocal husbands). Nubi could never decide to exit on their own volition. Wealth and influence offered negotiating leverages, but not the power to exit unilaterally.

Other paths out of bondage either increased vulnerability or preceded a phase of re-entry into bondage. None of the alternatives to proper emancipation and redemption did permit to recover commoner status and the relative degrees of autonomy associated with it. “Eviction” from the household was a convenient means for masters to informally break the bonds without formally relinquishing control. An extreme example of this practice is found in early seventeenth-century model complaints against masters who evicted male nubi to secure sexual intercourse with their wives and threatened to report them as runaways.³⁹

Some nubi revolts persisted for years, but all ended in the same manner. Leaders were sentenced to death and cohorts of insurgents ended up begging for forgiveness before being returned to their former masters (like Jiang Guanda).⁴⁰ Records of successful escapes only tell us that no one heard about the runaways anymore, that they joined the ranks of outlaws, or that they were caught and returned to their masters.⁴¹ Thus, although “turning one’s back on a master” was a possible way out, it implied living under the threat of arrest, prosecution, and return into servitude. To those who actually managed to escape, it often meant commending oneself to another, more powerful master. Such was the case of many haonu, like the wealthy Lu Zhaofang who escaped his master’s racketeering around 1615 by submitting to a more influential one⁴²; or like the many nubi who joined the Manchu Banner armies as slaves and soldiers in the 1640s, despite the prohibitions.

The question that remains, though, is whether exit ensured autonomy—as much as autonomy can be construed in a context where obligations and hierarchies shaped the many dimensions of interpersonal relations. In view of the numbers of commoners attracted by the protection offered by enslavement in the late Ming, access to commoner status was not necessarily desirable. With a few significant exceptions, nubi revolts did not take the form of struggles for “freedom,” but rather demanded adaptations of the rules and conditions of bondage. Exit to commoner status may have seemed desirable for many reasons (e.g., as an escape from cruelty), but less so without the prospect of minimal economic independence. Unfortunately, nubi usually cease to exist in the sources when their behavior and their qualities cease to be exemplary (for good or for ill), but it is clear that not all wished to thread into autonomy when offered to do so. Among other examples, the two slaves of Mrs. He mentioned above preferred to stay in their mistress’s household despite being offered emancipation after 40 years in service. When achieved formally and legally, exit did mean status improvement, but it did not necessarily involve better material conditions and the end of subordination.

With regard to exit, gender was a crucial factor. Being a woman offered additional paths out of nubi status, but those paths resulted in very little improvements in terms of autonomy. As a married woman, the wife of Jiang Guanda could not expect to exit servitude by marriage, but she could expect sexual abuses and even being sold away and separated from her husband, as some examples mentioned above illustrate. Despite being very strict with regard to “fornication,” Ming law made no provision to prevent sexual intercourse between male members of the household and enslaved females (early Ming normative sources even placed female nubi within the reproductive and sexual spheres of their masters). To unmarried girls, enslavement certainly increased the risk of sexual abuses, but the promise of “benevolence” also provided some assurance of marriage when coming of age (unless they were denied marriage to avoid contact between household women and enslaved men, as was the practice in some merchant families). Marriage, however, offered no automatic access to commoner status. Besides being married to male nubi (which closed the door to prospects out of bondage), female nubi could be sold as concubines, and sometimes as principal wives to commoners, as in the above-mentioned case of a commoner offered a female nubi in uxorilocal marriage in exchange for 22 years of labor. Concubines being “honorable,” the shift from nubi to concubine was a social improvement. However, concubines were closely assimilated into enslaved women and their actual condition may not have differed much. Although marriage offered an additional road out for enslaved women and could improve their status and even their condition, it was never their choice to make.

Finally, exit did not necessarily end the cycle of bondage and subordination. The only way out envisioned in contracts was death, be it natural (as implied by the formulae stating that nubi would serve “forever” or “all their life”) or accidental (as suggested by the clauses absolving buyers of responsibility in case of “unfortunate” death). Death could also be “chosen” as a last resort (as in the case of Tang Yuan) or could be the result of excessive discipline and abuses. However, death did not end the cycle of enslavement. Historians still disagree as to whether nubi bondage was hereditary. The localized use of labels like “hereditary slaves” in Huizhou and in other places has led to the conclusion that heredity was not the rule. The absence of formal rules on heredity also led late Qing officials in charge of the abolition of slavery in China in 1910 to state that heredity was an exception and a subversion of the fundamental principles of bondage. Yet, examples of hereditary transmission abound, and nubi revolts demonstrate that ending it was a staunch demand among the insurgents. A telling case is found in Mao Yilu’s Brief Accounts of Judgments in Songjiang—also testifying to the many varieties of nubi condition—in which a man adopted by his wealthy nubi uncle was left with the choice between inheriting the wealth and the status of his uncle, or neither of them.⁴³

In addition to the perpetuation of bondage through heredity, exit to commoner status left significant stigmata. The stains of bondage were not explicitly formalized in the Ming as they would be in the Qing period, when people of nubi ancestry were barred from partaking in the imperial examinations for three generations. However, as noted by Wang Jiazhen and others, even after paying the price of emancipation, nubi could “never be on the same level” as commoners.⁴⁴ They continued to demonstrate “the difference between master and slave” by providing services, showing deference, and honoring their former masters and their offspring, sometimes for generations.

Conclusion

The world that nubi lived in during the late Ming was one of excesses, confusion, and ambiguities. It appears all the more ambiguous to us, who tend to evaluate all forms of subjugation according to modern categories that we have learned to regard as universal. To understand this world properly, we need more than the norms that transpire through the elusive content of the Great Ming Code; and more than references to absolute “freedom” and “slavery.” Identifying forms of slavery in contexts where heteronomy is the norm rather than the exception is always a delicate task, for there is hardly one single criterion that distinguishes enslavement from other forms of domination. In Ming China, saleability was not an exclusive feature of captives traded to be enslaved, since children and women were customarily sold to assume a wide variety of roles. Being cast out from the realm of “honorable people” was also not an exclusive feature of nubi, who were nevertheless the largest component among various groups of “mean people.” Being subjected to the almost absolute power of a household head was not a distinguishing feature either, since the power exerted by masters mimicked the authority that fathers held over their children. What distinguished nubi, however, was that enslavement produced a total and unique form of excommunication from the standards of belonging to society and of membership in the household, which reduced people permanently to the sole function of “serving” others and to one single direct social relationship (with their master).

Faced with the effects of an unprecedented growth of the enslaved population, of changing practices, and of the increased use of bondage as a means to overcome economic difficulties, late Ming thinkers paid particular attention to outlining an ethics of bondage that insists on its voluntary, reciprocal, and benevolent nature. Examining the many paths into bondage demonstrates that the ideal of a mutually beneficial contract of slavery was a fiction, since it resulted from a wide array of constrained choices. More important is the fact that coercion was present—effectively or potentially—at all time during the relationship, and even more when it came to exiting it. Although the ways into bondage might have had an effect on the condition of nubi and on the levels of autonomy they could enjoy in practice, those had little effect on the possibility to end the relationship. In other words, despite the illusion of the contract, nubi bondage was a total relation of subjugation safeguarded by the sociopolitical order. For a brief moment, the collapse of that very order made armed revolts and demands for reforms possible. The new dynasty violently suppressed the revolts and sent nubi back to their masters. In the course of the eighteenth century, the Qing would nonetheless pay increased attention to clarifying and formalizing the norms of nubi bondage, without putting an end to it until 1910.

Notes

• 1. Claude Chevaleyre, “Under Pressure and out of Respect for Human Dignity: The 1910 Chinese Abolition,” in Distant Ripples of the British Abolitionist Wave, ed. Myriam Cottias and Marie-Jeanne Rossignol (Trenton: Africa World Press, 2017), 147–98.

• 2. Marcel van der Linden, “Dissecting Coerced Labour,” in On Coerced Labour. Work and Compulsion after Chattel Slavery, ed. Marcel van der Linden and Magaly Rodríguez García (Leiden: Brill, 2016), 291–322.

• 3. On nobi in Chosŏn Korea (1392–1910), see the chapter by Sun Joo Kim in this volume.

• 4. Ming Qing Huizhou shehui jingji shiliao congbian, ed. Anhuisheng bowuguan, vol. 1 (Beijing: Zhongguo shehui kexue chubanshe, 1988), 554–55 (author’s translation).

• 5. Claude Chevaleyre, “Human Trafficking in Late Imperial China,” in Slavery and Bonded Labor in Asia, 1250–1900, ed. Richard Allen (Leiden: Brill, 2021), 168.

• 6. Johanna Ransmeier, Sold People: Traffickers and Family Life in Northern China (Cambridge, MA: Harvard University Press, 2017), 2–13.

• 7. Orlando Patterson, Slavery and Social Death: A Comparative Study (Cambridge, MA: Harvard University Press, 1982), 41–44.

• 8. Claude Chevaleyre, “Acting as Master and Bondservant: Considerations on Status, Identities, and the Nature of Bondservitude in Late Ming China,” in Labour, Coercion and Economic Growth in Eurasia, 17th–20th Centuries, ed. Alessandro Stanziani (Leiden: Brill, 2013), 253–60.

• 9. Fu Yiling, Ming Qing nongcun shehui jingji (Beijing: Sanlian shudian, 1961), 95.

• 10. Gu Cheng, Mingmo nongmin zhanzheng shi (Beijing: Zhongguo shehui kexue chubanshe, 1984), 342.

• 11. William Rowe, Crimson Rain: Seven Centuries of Violence in a Chinese County (Stanford: Stanford University Press, 2007), 155–56.

• 12. Wang Kentang, Da Minglü fuli jianshi (Tōyō Bunka: Oki collection, pref. 1612), 20: 15b.

• 13. Hubei tongzhi (1921), 69: 7b.

• 14. Guangshan xianzhi [1786] (Guangshan: Guangshan xianshu, 1889), 19: 16a–18b.

• 15. Ming Xianzong shilu (Taipei: Zhongyang yanjiuyuan lishi yuyan yanjiusuo, 1966), 4744.

• 16. Jiang Yonglin, The Great Ming Code/Da Ming Lü (Seattle: University of Washington Press, 2005), 105–6.

• 17. Buxiang Zi, Fajia tou danhan [pref. 1618], Legalizing Space in China, online ed. by Sun Jiahong and Gong Rufu, 151, https://lsc.chineselegalculture.org/Asset/Source/lscDocument_ID-241_No-01.pdf.

• 18. Zhang Chuanxi, Zhongguo lidai qiyue huibian kaoshi (Beijing: Beijing daxue chubanshe, 1995), 1069.

• 19. Ibidem, 1063.

• 20. Buxiang Zi, Fajia tou danhan, 97.

• 21. Mao Yilu, Yunjian yanlüe [n.d., Wanli era], in Lidai panli pandu, ed. Yang Yifan and Xu Lizhi, vol. 3 (Beijing: Zhongguo shehui kexue chubanshe, 2005), 574.

• 22. He Liangjun, Siyou zhai congshuo [1569] (Beijing: Zhonghua shuju, 1997), juan 35, 320–21.

• 23. Zhang Lüxiang, Yangyuan xiansheng quanji [1704] (Beijing: Zhonghua shuju, 2002), 575.

• 24. Fu Yiling, “Mingmo nanfang de dianbian, nubian,” Lishi yanjiu, no. 5 (1975): 62–63.

• 25. Zhang Lüxiang, Yangyuan xiansheng quanji, 1024.

• 26. Yunjian zazhi [1615] (Jinan: Qi Lu shushe, 1997), 3: 7b–8a.

• 27. Harriet Zurndorfer, Change and Continuity in Chinese Local History: The Developpement of Hui-Chou Prefecture 800–1800 (Leiden: Brill, 1989), 203.

• 28. Chen Hongmou, Xunsu yigui [1742] (Jinan: Qi Lu shushe, 1997), 2: 20a.

• 29. Wang Jiazhen, Yantang jianwen zaji [Qing], in Taiwan wenxian shiliao congkan, vol. 5 (Taipei: Taiwan datong shuju, 1984), 30.

• 30. Buxiang Zi, Fajia tou danhan, 161.

• 31. Jiang Yonglin, The Great Ming Code, 183–85, 198–200.

• 32. Wang Jiazhen, Yantang jianwen zaji, 30.

• 33. Macheng xianzhi [1904], 17: 11b.

• 34. Zhang Lüxiang, Yangyuan xiansheng quanji, 910.

• 35. Shaoxing fuzhi [1584], 45: 15b.

• 36. Mao Yilu, Yunjian yanlüe, 562.

• 37. Mao Yilu, Yunjian yanlüe, 580; Jiang Yonglin, The Great Ming Code, 184–85.

• 38. Mao Yilu, Yunjian yanlüe, 417–18.

• 39. Buxiang Zi, Fajia tou danhan, 79.

• 40. Qi Zhongmin gong nianpu, [pref. 1837], in Taiwan wenxian shiliao congkan, ed. Wang Siren et al., vol. 107 (Taipei: Taiwan datong shuju, 1987), 150.

• 41. Mao Yilu, Yunjian yanlüe, 441.

• 42. Shen Bingxun, Quanzhai laoren biji, in Wuxing congshu (Beijing: Wenwu chubanshe, 1986), 3: 11b–12a.

• 43. Mao Yilu, Yunjian yanlüe, 512.

• 44. Wang Jiazhen, Yantang jianwen zaji, 30.