Christoph Hess

In 19th-century England, afternoon tea would not have been complete without a serving of Keemun – one of the Victorians’ most prized black teas. This was not for a lack of choice. A secret operation headed by the Scottish botanist Robert Fortune, who donned Chinese-style garments to travel incognito to China’s tea regions, had seized the secrets of tea production and successfully established plantations in British India from the 1850s on. But many British consumers kept a taste for Chinese teas, especially the malty and slightly smoky Keemun, which, when scented with bergamot, made an excellent Earl Grey. What they may not have known was that their love for a cup of Keemun connected their warm salons to a group of Chinese serfs who picked the tea leaves from the steep mountain slopes of Qimen (old: Keemun) County in East China. 

Louis William Wain, Afternoon at home (early 20th century). Victoria & Albert Museum Department of Prints and Drawings and Department of Paintings.

By the late 19th century, the serfs of Qimen County were one of the last vestiges of serfdom in China, which had been widespread across the densely settled region at the lower reaches of the Yangzi River until the early 18th century. However, despite the fact that China had probably the largest population of coerced labourers in the early modern world, we know very little about the spread of serfdom, and even less about the reasons for its sudden decline in the late 17th century.  

Studying the development of serfdom in China matters not just for tea drinkers. It matters, above all, for the sake of historical completeness. The sheer size of China’s past serf population makes it indefensible to leave it ignored. But Chinese serfdom also raises important questions about the nature of labour coercion. Some economic historians believe, for instance, that slavery and serfdom flourish where labour is scarce – think of the antebellum US South, or early modern Russia. But labour was abundant in China. This was especially true for the Lower Yangzi Region, where serfdom thrived.  

The landscape around Qimen County in 2024. Author’s photo.

Others have argued that the political institutions of different countries spurred the decline of labour coercion in some places over others. But serfdom declined only in some parts of 17th-century China, while other parts of the same polity, such as Qimen, clung to it. However you look at it, the case of Chinese serfdom seems like a strange historical anomaly. 

Given that so little is known about these people, I had to start my research with three very fundamental questions: how were serfs different from ordinary people in late imperial China? What kinds of tasks did they do for their masters? And last, why did serfdom decline in some places, but not in others?  

What made a serf a serf? 

Serfs belonged to a group of people that the legal codes of the Ming and Qing dynasties referred to as jianmin 賤民, that is, “base people”. Some people became jianmin because their occupations were seen as spiritually unclean, such as prostitutes or beggars, but this category also included actors and musicians. In some ways, Chinese serfs were luckier than these groups. Since they were not viewed as “untouchable”, they could hope for some basic participation in their local societies. In other ways, serfs were worse off than even the untouchables. While the son of a musician was not tarnished by the stigma of his father, the status of serfs was hereditary. This meant, above all, that the imperial examinations would forever be denied to them and their children. 

What else do official statutes tell us about the legal status of serfs? Not much. To maintain its leverage over landlords, the imperial court granted the privilege of owning serfs to government officials only. Thus although common merchants and landlords still owned indentured labourers in droves, they disguised them as in-laws or adoptive sons to avoid government sanctions. One example is a marriage contract from 1580, which disguised the sale of an eleven-year-old serf as an exchange of betrothal gifts: 

Wang Tao… sets out this marriage contract. Now that he is short of money, he sells his yinan義男 [i.e. adoptive son, a euphemism for “serf”] by the name of Tianbao, born in the fourth month of the Yin-Earth-Snake Year [i.e. April/May 1569], to the Cheng family as their yinan. In return, Wang will receive 2 silver taels and 800 copper coin in betrothal gifts. The silver is to be received immediately, and Tianbao will be at the disposal of his new master without any further dispute… [1]

Sales of children like Tianbao were common. In a sample of 40 contracts of serf sales that I have collected so far, 18 serfs were below the age of fifteen when sold or resold. Some were as young as five. People could of course only be sold if they had already assumed servile status, which happened either through inheritance or, most commonly, through “commendation”: destitute smallholder peasants “commended” themselves and their land to a landlord in exchange for basic food, shelter, and protection from taxes. 

Female serfs and slaves had characteristically unbound feet. This photo was probably taken in Fujian Province around 1900 and shows a bare footed slave, a girl with bound feet, and a Christian with unbound feet. University of Bristol, Maxwell Family Collection.

Systems of coercion varied across eastern China and over time, but shared four key traits throughout: 

1. Labour coercion: serfs had to provide labour services for their masters.
2. Demographic controls: masters decided when, and whom, their serfs could marry. Cases of masters ordering infanticide of serf children are known, but we do not know how common this was.
3. Mobility controls: serfs were not allowed to leave their assigned plots.
4. Heritability: children of serfs became serfs as well. Serf status could be passed on by both mothers and fathers. 

A typical serf “cultivated the master’s land, will be buried on a mountain belonging to the master, and lives in a shelter provided by the master”, a recurrent formula in many serf contracts that gives perhaps the most succinct answer to the question “what made a serf a serf?”. The formula makes clear that Chinese serfs submitted to their masters in life and death. In exchange, they could expect land to till, a place to live in, and spiritual shelter – being buried on a mountain with good fengshui was a ticket to a cosy afterlife. 

The economy of serfdom

Serf labour was integral to the economy of highland regions like Qimen County, where it endured well into the 20th century. Working on the slopes of rugged mountains was a demanding and often dangerous task, for which wage labour was hard to come by. The best tea leaves grew on the highest and steepest slopes, which made them especially hard to pick. 

Even more often than for harvesting tea, highland serfs were used for cutting wood, especially the heavy pine trees that flourished on the granite peaks around Qimen. The exquisite wood carvings (see image below) and palace-like temples that the region is still known for would not have been possible without heavy inputs of serf labour.

Serfs were used in the cultivation, cutting, and transportation of pine wood for heavy wooden structures and fine woodwork. Author’s photo.

Serfs were not only used to cut timber but also played an important role in the cultivation of trees. Pines need around two decades to grow, during which they need to be pruned and protected. The region’s landlords solved this commitment problem through heavy inputs of bonded labour. 

Unlike serfs in the lowland regions of the Yangzi Delta, who lived on or around the landlord estate, highland serfs mostly lived in satellite settlements around the village where the dominant local kin group lived (the “lineage seat”). I have reconstructed this type of structure around the village of Chawan in Qimen County (map 1 below).  

A first line of settlements immediately surrounded the central village. Serfs living in these places could be called into the village at the strike of a gong, often to perform ritual dances, music, or domestic work. Five serf settlements to the north of Chawan managed the lineage’s grain storages. Fire-bearing bondservants secured a safe passage on the road down to these grain storages. Last, specialised defence serfs secured the main access road to the county seat of Qimen and formed a line of defence against politically unstable areas to the south of the village. It is likely that these defence serfs also protected the lineage tombs and surrounding forests, by which they also maintained spiritual protection: tombs and their surroundings were the most important elements for creating an auspicious fengshui around Chawan Village. 

Map 1. Serf settlements in Qimen county in the early 20th century. “Bondservant” is synonymous with “serf”. Author’s illustration, based on Ye Xian’en 葉顯恩 “guanyu Zhongguo de dianpu zhi” 關於中國的佃僕制. Zhongguo shehui kexue, 1 (1981), p.187. Base map from AMap.

Serfdom in the lowland regions along the Yangzi River looked more like what was common across Eastern Europe in the early modern period. Serfs on the plains did not live in dedicated settlements, but in barns or small houses scattered across their master’s manor. Both in the mountains and in the Delta region, most serfs were put to work in farming. Lowland serfs were, however, rarely seen in ceremonial contexts. Their role was defined in more strictly economic terms, and could include more skilled tasks like weaving and spinning.  

Many landlords in the Delta region entrusted the management of their estate to serf-managers, who collected rents and oversaw the labour services of lower-ranking serfs. Until the early 17th century, this division of labour had largely worked in the landlords’ favour because it kept their servile labourers divided. Some serf-managers became reviled as “brazen serfs” (haonu) who, as one local source put it, “were gutting the common people like a dead animal or fish” [2]. But when the political order of the Ming state crumbled in the 17th century, serf-managers turned that power against their masters. 

The global 17th-century crisis and the decline of serfdom 

Historians of Europe have coined the notion of a “General Crisis” to refer to the succession of wars, famines, and dynastic transitions that ravaged the continent in the 17th century. As the Thirty Years War entered its final decade in central Europe, almost a decade of unusually cold and dry years put the nail in the coffin of China’s Ming dynasty. In the volatile political climate, serfs across the east and south of China felt that the time had come to end their servile status. 

The official histories are silent on the serf revolts that broke out in the two decades between 1640 and 1660. Local sources, by contrast, leave no doubt about their intensity. A contemporary observer reports: 

1645: Riot… with one cry there are a thousand echoes. Each [serf] goes to his master’s house and demands he annul the contract. If there is even a small delay in producing the contract, they will set fire to the house, and there have been reports of serfs tying up their masters. [3]

The revolt was not driven by economic discontent alone. Serfs strove to end their legal subordination for good; a fact clearly reflected in the demand to have their contracts handed to them and that “serfdom shall not be extended to their sons and grandsons”. Similar revolts broke out throughout eastern China (map 2 below). 

Source: Andreas W. Mixius, “Nu-pien” und die “nu-p’u” von Kiangnan : Aufstände, Abhängiger und Unfreier in Südchina 1644-45, in Hamburg: Gesellschaft für Natur- und Völkerkunde Ostasiens (1980). Population density in 1580 from Cao, Shuji, The Population History of China (1368-1953) (Brill, 2024), Appendix 1. Shape files from China Historical GIS. Base map from ESRI.

If the number of serf uprisings is any indication for how widespread serfdom was regionally, it is striking that most were recorded in the most densely settled parts of eastern China, where free labour should have been too cheap to justify coercion. So why did indentured labour exist? 

It is difficult to answer this kind of question for a serf population that officially was not recognised to exist – even though de facto it made up around 20-30 percent of the total population in the middle and lower Yangzi Valley. How to write a quantitative history of something that was never officially registered? 

Systems of coercion and the economy of skill 

My answer was to build the story of Chinese serfdom from the bottom up, parsing through gazetteers, contracts, diaries, and other local sources for any mention of serf labour that I could find. Because these sources have survived very unevenly, the total number of observations in different places and over time does not tell us much about how widespread serfdom was. But the 1,500 observations I have collected so far for the late Ming and early Qing dynasty reveal important differences in the structure of serf labour. These differences help explain why it declined in some places following the 17th-century crisis, while it persisted in others. The data also suggest reasons for the rise of serfdom in places with astonishingly high population densities. 

Document describing a court case between a serf and owners in the early 19th century. From a private collection, kindly provided by Prof Wang Zhenzhong.

So what encouraged the rise of serfdom in the Ming? Hereditary serfs existed alongside free tenants from at least the 14th century onwards, and qualitative sources suggest that both systems operated in tandem. Serfdom expanded in times of crisis, when smallholder peasants commended themselves to large landlords. Its scope decreased when the economy recovered. Serfs in the lowland regions were therefore almost entirely agricultural and assigned to productive rice paddies. Temporary changes like the decay of local tax collection structures in the late Ming could make it attractive for free tenants to reduce their economic burden by entering serfdom. The structure of their work, though, never fundamentally differed from that of free smallholder peasants.  

Peasants who transitioned between the two systems did therefore not need to acquire new skills. They could not always decide freely, and landlords often fought to keep their serfs in place. But the costs of transitioning between serfdom and free tenancies were low for peasants on the river plains. 

The serfs who picked tea leaves for Victorian afternoon teas were in many ways different from ordinary tenant farmers. Not only did they live in their own clearly demarcated satellite settlements, but they and their peers were assigned demanding tasks to cut tea leaves, wood, or bamboo in the mountains. Farmland given to serfs mostly consisted of low-quality mountain plots, while productive fields for rice and wheat were tilled by free tenants with different tools and techniques.  

Highland serfs were thus highly specialised in tasks for which few unindentured occupations existed in the region. And while lowland slaves were almost never considered spiritually unclean, highland society emphasised that serfs were inferior in all domains. Spiritual submission was enforced by having serfs perform polluting music and dances for rituals at their masters’ ancestral hall. 

Advertisement for Tung Chuen Chang & Co, who shipped Keemun tea to the UK, continental Europe, and the US. From a private collection, kindly provided by Prof Wang Zhenzhong.

Keemun tea is still enjoyed widely across China and the world. The serfs who lifted it to fame more than a century ago are now widely forgotten, even in Qimen County itself. But their story is an important one. It shows the power of specialisation for creating a dependent workforce with few outside options. It is also a story of how religious beliefs like ritual and fengshui could be harnessed to maintain coercive institutions over half a millennium. 

References 

[1]  Anhui sheng bowuguan 安徽省博物館 (ed.). Ming-Qing Huizhou shehui jingji ziliao congbian 明清徽州社會經濟資料叢編. Beijing: Zhongguo shehui kexue chubanshe, 1988, p.552. 

[2]  Kun-Xin liang xian zhi 崑新兩縣志 (Daoguang period), vol.18, p.19b. 

[3] Wang Jiazhen 王家楨. Yantang jianwen zaji 研堂見聞雜記, 1671, p.54. Full text on ctext.org. 

Further reading 

• Brook, T., The Confusions of Pleasure: Commerce and Culture in Ming China (University of California Press, 1999). 
• Brook, T., The Price of Collapse: The Little Ice Age and the Fall of Ming China (Princeton University Press, 2023). 
• Brown, T. G., Laws of the Land: Fengshui and the State in Qing Dynasty China (Princeton University Press, 2023). 
• Domar, E. D., ‘The Causes of Slavery or Serfdom: A Hypothesis’, in The Journal of Economic History 30:1 (1970), 18-32. 
• Grove, L., and Daniels, C., State and Society in China: Japanese Perspectives on Ming-Qing Social Economic History (University of Tokyo Press, 1984). 
• Hansson, A.. Chinese Outcasts: Discrimination and Emancipation in Late Imperial China (Brill, 1996). 
• McDermott, J. P. ‘Bondservants in the T’ai-Hu Basin During the Late Ming: A Case of Mistaken Identities’, The Journal of Asian Studies 40:4 (1981), 675–701. doi:10.2307/2055676. 
• Postan, M., ‘The Chronology of Labour Services’, in Transactions of the Royal Historical Society 20 (1937), 169-193. 
• Wiens, M. C., ‘Kinship Extended: the Tenant-Servants of Hui-chou’, in Kwang-Ching Liu (ed.), Orthodoxy in Late Imperial China (University of California Press, 1986), pp. 231-254.