Turning bandits into ‘good citizens’: Coastal violence on the south coast of the Ming Empire in the fifteenth century

Banditry was a recurring problem on the mountainous, island-dotted south coast of the Ming Empire (1368–1644). Scholars mainly discuss banditry on the south coast in the context of growing international trade and focus on the sixteenth century while largely ignoring endemic forms of banditry. This article studies the interaction of bandit groups, coastal communities and Ming officials during the fifteenth century to show that banditry was an integral part of life on the south coast of the Ming Empire. Furthermore, it argues that the inability of Ming officials to understand how banditry was integrated in coastal society limited their capacity to secure long-term solutions.

1. Kwan-wai So, Japanese Piracy in Ming China in the 16 th Century (East Lansing, MI, 1975 This article examines how bandits emerged time and again, sometimes in the immediate aftermath of state intervention. Scholars usually study banditry on the south coast of the Ming Empire in the context of growing maritime trade. Kwan-wai So, one of the first to produce a detailed study of coastal violence in the Ming period, points to the accelerating momentum of maritime trade with Japan as one of the reasons for increased coastal violence in the sixteenth century. 1 Ming historian John Dardess similarly attributes this peak of coastal violence to 'maritime entrepreneurs' engaged in 'smuggling and piracy'. 2 Robert Antony, a scholar of piracy during the late imperial period, dismisses earlier forms of coastal violence as 'petty, localized piracy'. 3 In focusing on exceptional 'waves' of bandit activity, these authors leave endemic forms of coastal violence understudied. This article purposefully draws on cases of banditry from the fifteenth century to illustrate the recurrent nature of banditry on the south coast of the Ming Empire, not in the context of growing maritime trade but as a consequence of the interaction between Ming officials and coastal society.
Ming officials served terms of two or three years in unfamiliar territory with limited resources at their disposal. Efficient governance depended on their ability to cooperate with influential families in their jurisdictions, as well as their capacity to understand local conditions and implement suitable solutions. Ming officials along the south coast often displayed a limited understanding of the reasons why people turned to banditry and their policies did not secure long-term solutions. This article argues that the approach of Ming officials to the issue of banditry in their jurisdictions did not address the social context that gave rise to banditry, which helped to create an environment where banditry repeatedly emerged.
The social context in which bandit groups emerged was historically contingent and shaped by local conditions. Some scholars take a broad approach and argue that banditry emerged in the absence of strong state presence. James Tong, for example, contends that most instances of violence during the Ming period took place in the coastal provinces of Fujian and Guangdong, far away from the political capital in Beijing. 4 Dardess points out that this conclusion is understated because Tong omitted less reliable data from the 'violence-prone provinces' on the southwestern border. 5 However, David Robinson shows that banditry did persist in the capital region precisely because of strong state presence. The complex political landscape around Beijing as well as the nuanced relations between political patrons and 'the less reputable members of local society' created opportunities for bandits to escape punishment. 6 Robinson successfully challenges the broad approach 7. Joseph Dennis, Writing, Publishing, and Reading Local Gazetteers in Imperial China, 1100-1700(Cambridge, MA, 2015. of Tong by studying banditry in its local context. This article similarly treats each case within its local context to show that banditry was an integral part of life on the south coast of the Ming Empire. Those involved in banditry did not write about or record their activities. This article therefore relies on evidence from local gazetteers to study cases of banditry in Fujian and Guangdong. Gazetteers contained the geographic, demographic and administrative information that officials used to familiarise themselves with their jurisdiction, as well as biographies of prominent residents and officials, and descriptions of cultural practices and historical events such as bandit attacks. However, according to Joseph Dennis, the officials and prominent residents involved in the compilation of a gazetteer used the project to showcase their own achievements and promote their own worldview. 7 In other words, the impact of banditry on Ming society left traces in gazetteers, but only careful reading against the grain enables the marginalised voice of those involved in banditry to challenge the dominant narrative produced by Ming officials and prominent residents.
Critical appraisal of descriptions of banditry in local gazetteers provides insight into how the limits of state intervention perpetuated banditry on the coasts of Fujian and Guangdong. The gazetteers from two prefectures proved particularly productive: Zhangzhou and Chaozhou were neighbouring prefectures, the former being the westernmost prefecture in Fujian province, while Chaozhou was the easternmost prefecture in Guangdong province. The mountainous and island-dotted south coast already offered hideouts for bandit groups, and the administrative division between two provinces further hindered Ming officials in their pursuit of bandits. Coordinated pursuit across provincial borders generally required permission from senior officials in Beijing. According to the Guangdong Comprehensive Gazetteer of 1602, 'Chaozhou and Zhangzhou are close to each other. Thieves escape to one or the other and hide, which makes managing affairs difficult.' 8 This administrative obstacle was temporarily resolved when Zhou Xuan, a native from Zhangzhou, was appointed prefect of Chaozhou in 1458. He combined personal connections in Zhangzhou with professional connections in Chaozhou to effectively extend his reach as prefect. In 1459, a group of bandits under the leadership of Huang Yuyi attacked Jieyang, a subordinate county of Chaozhou. Prefect Zhou supervised the soldiers who captured and killed bandit leader Huang. He then offered amnesty to all the remaining bandits on the condition that they cease their violent behaviour. The residents of Jieyang resumed their occupations, and everything seemed peaceful, but as soon as Zhou was transferred, bandits returned. In 1462, the bandits Wei Chonghui and Xu Wanqi besieged the county capital for more than a month. They also robbed the residents of Jieyang and pierced the dykes to flood their houses. In other words, Zhou achieved temporary success, but failed to root out banditry in his jurisdiction. Zhou did not investigate why Huang attacked Jieyang, and he did not consolidate his victory in a manner that addressed the social conditions that gave rise to banditry in the first place.
9. The capital of a county (xian, also translated as 'district') was named after the county itself, hence the capital of Jieyang county was known as Jieyang city. This article argues that Zhou's inability to understand why bandits attacked Jieyang limited the effectiveness of his intervention. Furthermore, his willingness to offer amnesty perpetuated banditry in Jieyang. Zhou did not see how Huang and his followers were integrated in coastal society and he lacked knowledge about the networks that allowed them to regroup and rise again once he left.

Huang Yuyi attacks Jieyang City
Jieyang city 9 was located on the banks of the Rong river where the Northern Rong and the Southern Rong met. One contemporary observer praised the county as a grand estate with fertile land and numerous people who were inclined towards scholarship. The surrounding areas were not as admirable: Now then, outside Xialing is the cold ocean and inland from Huangzhai one passes through [the territory of indigenous] Yao and Zhuang who often remain obstinate and do not respond to leadership. . . . In the early years there were the robbers Huang Shoushan and Zeng Bichang who gathered in the mountains and at sea. . . . Then these stupid, rebellious people such as Wei Chonghui and Xu Wanqi continued their unsuccessful endeavour. 10 The mountains and the sea feature prominently in the writings of many contemporary observers, and their choice of words suggests unease with this unfamiliar terrain. This particular description of Jieyang comes from the hand of Li Hui, a Ming official who was born in Haiyang county, Chaozhou. In this description, Li emphasises familiar elements of Ming society such as agriculture and scholarship. He even claims that Jieyang was like a seaside version of the birthplace of Confucius. Li then contrasts these familiar elements with the unfamiliar mountains and sea, which his imagination populates with stupid, unruly people. His description of Jieyang portrays the county as a wealthy beacon of civilisation surrounded by uncivilised bandits in the mountains and at sea. The contemporary observation of Li is steeped in prejudice against people who did not fit within the agrarian society of the Ming Empire, but his views are also based on real contrasts between different communities on the south coast. Agricultural land was relatively scarce in this mountainous area, with the floodplains of rivers forming an exception. The rivers on the south coast of China run fast when they leave the mountains, but slow down before they empty into the sea. Where the rivers slow down they deposit fertile soil on the adjacent land. In the Ming period, agrarian communities developed along the river banks, between the mountains and the sea. Farmers therefore accumulated agricultural resources in relative isolation among a population that largely made a living from the mountains and the sea. Jieyang city was located on the fertile banks of the Rong river, but only a few kilometres downstream the people of Xialing made a living from the sea.  ', Pacific Affairs, 66 (1993-1994, 577-8.
coast, gathered clams from tidal flats and trapped seawater in reservoirs to produce salt. In this world of contrasts, 'the sea bandit Huang Yuyi wreaked havoc on Jieyang city and forced [the residents of] 24 villages in Xialing to follow him'. 11 Huang arrived with several hundred ships in an area where people made a living from the sea, while agricultural wealth accumulated a short distance upriver in the county capital. According to the Longxi County Gazetteer of 1535, the residents of Xialing were forced to follow Huang. However, when bandits returned to the area in 1462, they once again found shelter among the residents of Xialing, which suggests that at least some residents were not forced but quite willing to cooperate with bandits. 12 Their presence was vital for bandits who relied on local residents for shelter and, perhaps more importantly, strategic information. The Rong river provided Huang's bandits with relatively easy access to agricultural wealth, but the residents of Xialing supplied the local knowledge and manpower that played a key role in this attack. Jieyang city attracted the attention of bandits such as Huang precisely because of the contrast between agrarian and coastal communities. The agricultural wealth and the close vicinity of coastal communities motivated Huang to attack this particular area, but this does not entirely explain why people such as Huang turned to banditry in the first place. Malcolm Greenshields, who studied violence in early modern France, argues that criminals are difficult to understand because they rarely record their personal motivation for violent crimes, which turns them into a 'dark figure'. 13 However, this 'dark figure' can be 'outlined' when accounts of victims, witnesses, officials and kinsfolk illuminate aspects of the individual criminal. In other words, the 'dark figure' can be understood through their impact on society. Greenshields uses this approach to uncover the factors that led individuals and groups to escalate violence in an 'endless round of provocations and retaliations' that he called the 'economy of violence'. 14 Robinson expands this term to refer to the 'administration or management of concerns and resources related to violence in society'. 15 In other words, the 'economy of violence' covers the reasons that people resorted to violence as well as the means by which they engaged in violence.
Ming officials assumed that peasants did not turn to violence unless the stable agrarian order was disturbed, for example when famine followed natural disaster or state failure. 16 In this worldview, banditry was an indication of state failure. Widespread banditry implied that the emperor himself, or even his entire lineage, had lost the mandate to 17 rule, in which case the ruling dynasty was replaced by another, worthier dynasty. According to Harriet Zurndorfer, this concept of the dynastic cycle also informed the early works of western Sinologists, which limited their capacity to understand collective violence as anything other than political failure. 17 This view of banditry as a consequence of state failure is still present in current scholarship. Dardess, for example, argues that widespread banditry on the south coast of the Ming Empire in the fifteenth century represented 'a crisis of law and order'. 18 Tong reaches a similar conclusion when he points to famine and lack of state control as the main reasons for collective violence during the Ming period. 19 This generalised explanation for collective violence closely reflects primary source bias, echoing Ming officials who believed that peasants did not turn to violence unless the stable agrarian order was disturbed.
The focus on peasants, famine and the absence or presence of the state places unnecessary limits on the study of banditry. Zurndorfer argues that people engaged in collective violence to adjust the reciprocal relationships between different segments of society. 20 Robinson similarly treats violence as part of the regular order, not as a challenge to the ideal social order. He also moves the discussion from motivation to means and opportunity. According to Robinson, 'men are more likely to consider the use of illegal means to attain their goals when the risk of punishment is low'. 21 Approaching banditry as an integral part of life on the south coast of the Ming Empire has several advantages over the notion that banditry occurred as a result of famine and lack of state control. First, this approach seeks to explain the diversity of banditry in relation to specific historical contexts, rather than as expressions of a singular underlying mechanism, which helps to understand why individuals turned to banditry in response to one failed harvest, for example, but not another. Second, the approach to banditry as an integral part of life is not as closely bound by agrarian society as the notion that banditry was a natural response to famine. This is of particular importance for the study of banditry on the south coast, because fertile land was scarce and many people made a living from the mountains or the sea rather than farming. Finally, focusing on means and opportunity provides understanding as to why some individuals participated in acts of banditry even when their personal motivation is unclear. In short, treating banditry as inherent to Ming society is a productive way to outline the 'dark figure' of banditry. The productive potential of this approach can be illustrated by a brief discussion of three cases of banditry in Longxi, Zhangzhou, which was the home county of Zhou Xuan. While the biography of Zhou does not elaborate on the specific reasons why Huang Yuyi attacked Jieyang city in Chaozhou, these cases from neighbouring Zhangzhou illustrate the broader social context that encouraged people on the south coast to engage in banditry. Furthermore, all three cases highlight the contrast between agrarian and coastal communities that facilitated conflict, as well as the inability of Ming officials to secure long-term solutions. Longxi city was located on the intersection of the Jiulong and Cang rivers. The city housed the civil offices of the county magistrate as well as the prefect of Zhangzhou and garrisoned the Zhangzhou Guard. The census of 1502 shows that Longxi county had a registered population of around 95,000 people, or about one third of the population of Zhangzhou prefecture. 22 The county also had about one third of the agricultural fields in Zhangzhou, around 250 square kilometres, as well as nearly 40 per cent of the markets in the prefecture. Over the course of the Ming period, the population of Longxi county continued to grow, while the equivalent figures for all other counties in Zhangzhou dropped. This suggests that people moved from rural counties to the urbanised heart of Longxi. Furthermore, the 2.5 to 1 male-female ratio in Longxi suggests that the county mostly attracted men, presumably looking for work in the flourishing, urbanised, administrative capital.
Longxi was an agrarian county caught between the mountains and the sea, like Jieyang. People on the banks of the Jiulong river and numerous islands along the coast made a living from the sea as fisherfolk and maritime traders, and sometimes as bandits: The residents of Haimen Island frequently went to sea as robbers. When they encountered passenger ships, they plundered them until nothing was left, and even worse, they killed people. 23 The residents of various places near the sea, like Yue harbour and Haicang, frequently traded goods with foreigners and excelled at robbing. 24 The area flooded. As of old, many people became bandits. 25 Haimen Island was only 30 kilometres downstream from the county capital. The island was strategically located in the river mouth and the people of Zhangzhou were afraid to pass the island for fear of being killed. Although this brief description does not elaborate on the islanders' motivation for banditry, carefully reading this description against the grain provides insight into their means and opportunities. The frequency of their bandit activities suggests that the islanders had easy access to ships to serve as vehicles for banditry at sea. Yet the Zhangzhou Prefecture Gazetteer of 1573 identifies them as the official residents of Haimen Island, not a group of bandits hiding there, which means the islanders were not recognisable as bandits unless they were caught in the act. In other words, the island residents engaged in opportunistic acts of banditry because their occupation and geographic location provided them with the means and opportunities to do so.
This ambiguity between island residents and bandits complicated the task of the Zhangzhou prefect in 1437. He was unable to identify individual bandits and decided to 26. Qingxiao, Mingdai fangzhi xuan, juan 4, 83a. move the entire island population to abandoned agricultural fields some distance inland from the southern bank of the Jiulong river. This geographic relocation also entailed a change in occupation and the cultural environment with which this occupation was intertwined. People with generations of maritime skills and knowledge of local currents were expected to work the abandoned fields and become 'good citizens'. In other words, they were to become farmers who conformed to dominant norms of acceptable behaviour in Ming society. The prefect displayed a disregard for a coastal lifestyle that can be interpreted as part of a wider conflict between the agrarian Ming Empire and a maritime world. When faced with a situation he could not control, he decided to uproot the entire community and forcefully incorporate them in agrarian Ming society.
The maritime merchants of Haicang, on the river bank north of Haimen Island, and Yue harbour, on the river bank south of Haimen Island, engaged in acts of opportunistic banditry like the residents of Haimen Island. As maritime merchants they were familiar with the vulnerabilities of trade along the south coast. They also had easy access to ships, and their occupation justified extended periods at sea. The Zhangzhou prefect in 1453, like his predecessor in 1437, struggled to identify individual bandits. He therefore implemented a policy that required every ship to return to port each evening. This policy, like the forced relocation of island residents, showed little consideration for the lifestyle of people who made a living from the maritime world.
While the islanders and the maritime merchants engaged in opportunistic acts of banditry, the flood victims of 1479 turned to banditry because their livelihood was destroyed. The rivers that deposited fertile soil also brought floods. The Zhangzhou prefect in 1479 immediately distributed relief grain, which was a sensible response in agrarian society, but banditry did not abate. Only five years earlier, in 1474, another Zhangzhou prefect handed out relief grain after a flood wave came from the mountains, and the people of Zhangzhou erected a stele for him in gratitude. 26 Why then did people continue to engage in banditry after the flood of 1479? The difference between the floods of 1474 and 1479 is that the first was a freshwater flood from the mountains and the second was a saltwater flood that came from the sea. Salt water ruins the soil and makes the flooded area unsuitable for agriculture for several years. The Zhangzhou prefect handed out relief grain to avoid famine, which normally works in agrarian society, but he did not understand that local conditions in coastal areas are different. People continued to act as bandits because their livelihood was destroyed for several years and relief grain would only see them through one winter. The Zhangzhou prefect displayed a fundamental failure to comprehend local conditions. Only after he consulted with captured bandits did he understand the issue. He immediately began the renovation of sea dykes, which his predecessors had neglected, to protect the area against future saltwater floods.
Reading descriptions of banditry in isolation gives the impression that exceptional officials effectively brought banditry to an end, but reading these descriptions as episodes in a longer story shows the recurring nature of banditry on the south coast of the Ming Empire. Furthermore, these case studies highlight the internal frontier between an agrarian empire and a maritime world. Access to the maritime world offered some people 27. Lin and Liu, Longxi xianzhi, juan 8, 652b. the means and opportunities to engage in banditry outside the reach of Ming officials. This was the case for the residents of Haimen Island and the maritime merchants on the Jiulong river banks. Huang, who arrived in Jieyang with a significant fleet, most likely had a similar background. Others resorted to banditry because the sea destroyed their livelihood, which also means that when bandits pierced sea dykes, they potentially increased the number of their followers. The Ming officials in these case studies responded to the unfamiliar maritime world with incomprehension and forceful attempts to either strictly control or completely cut off access to the maritime world. Ming officials and locals in coastal communities lived in two different worlds, and the ideal social order envisioned by Ming officials did not match with the social conditions on the south coast of the Ming Empire. Their solutions achieved temporary success at best, and their inability to understand the differences between agrarian and coastal communities perpetuated coastal banditry.

Zhou Xuan transforms bandits into 'good citizens'
The life of Zhou Xuan provides a striking example of the internal frontier between an agrarian empire and a maritime world, as well as the poor fit between a social order envisioned by Ming officials and the social conditions on the south coast. Zhou grew up in Longxi, an area where sea banditry repeatedly emerged. He was only 22 when the residents of Haimen Island terrorised the area, and he had only been an official for eight years when maritime merchants in his home county engaged in banditry at sea. Zhou was aware of the risks that the sea presented and he brought that experience with him when he took up the office of Chaozhou prefect in 1458. Among the first projects he initiated was a renovation of the sea dykes. This measure helped to prevent saltwater floods, which secured the agricultural livelihood of the people in Chaozhou. However, since famine was not the only cause of banditry, sea dykes did not completely prevent bandit activity in Chaozhou. When Huang attacked Jieyang, Zhou was forced to respond. Senior officials in the capital would have judged him on his performance. Had he left banditry unchecked, his political career would have been over. Furthermore, banditry interrupted tax collection, which was another important aspect of the performance review. Zhou took several steps to restore order in his jurisdiction: [Zhou] Xuan supervised soldiers in capturing and killing Huang Yuyi, and then he published an announcement. He ordered rural Confucian Chen Ji to go among the bandits to put up the announcement.
[Zhou] personally went to the bandit camp to instruct and pacify the bandits. Every bandit begged [for the chance] to surrender. [Zhou] Xuan took the opportunity to travel to each village [in his jurisdiction] and return the men and women who had been captured. . . .
[Zhou] Xuan stated that the people who had been forced to follow [the bandits] were not properly governed. He begged forgiveness for them and they avoided punishment. He restored a total of seven thousand households to their occupation. 27 The strategy of Zhou can be divided in three parts: he first identified and eliminated the bandit leader Huang; he then offered the remaining bandits a chance to surrender; and finally, he returned the men and women who had been captured to their homes. In other words, he eliminated the bandit threat, pacified his jurisdiction and made sure the people resumed their work. That final step was crucial for the restoration of regular tax collection in his jurisdiction. Zhou singled out Huang, not just by name, but as the one responsible for the entire issue of banditry in Jieyang. He did not offer Huang a chance to surrender with the other bandits. When Zhou had just arrived in Chaozhou in 1458, there was a local strongman named Lu who terrorised the countryside with violent behaviour. Zhou did not punish Lu, but sat down with him and pointed out how this behaviour violated accepted social norms. Zhou instructed Lu in proper moral behaviour. According to the Zhangzhou Prefecture Gazetteer, Lu subsequently changed into a 'good citizen'. 28 In this way Zhou policed the boundary between 'good citizens' who behaved according to the accepted norms of Ming society and those whose behaviour did not conform. While Ming officials such as Zhou excluded people from the social order with labels such as 'bandit' and 'strongman', they also offered a way back into the fold. Zhou created a personal connection with Lu to strengthen his moral instruction. Zhou also personally offered the remaining bandits of Huang's group a chance to surrender, but this offer did not extend to Huang. His execution was a reminder to the remaining bandits of the alternative to surrender: The remaining bandits did not dare come out.
[Zhou] Xuan stated that the bandit leader had already been caught, so the remaining [bandits] could be reassured and surrender. 29 The idea of moral instruction as a solution to banditry depended on the assumption that an example of superior morality had the power to transform an individual into a 'good citizen'. Underlying this assumption is the Confucian concept of rule on the basis of moral merit, rather than law, because law only instilled fear of repercussions, while a virtuous ruler inspired a sense of duty in his subjects, a duty to perform their roles in society. 30 The responsibilities of each role, and the relations between roles, were articulated through such values as respect for superiors, but also benevolence towards subordinates. These and other values created a moral framework that delineated acceptable behaviour in Ming society and consequently excluded individuals whose behaviour did not fit in with this social order.
At the top of this social order was the Emperor whose behaviour was an example to everyone in the realm. His failure to set the right example gave others the space to deviate from acceptable social norms, which in the end could collapse the entire social order. Officials, and the broader class of literati from which they were drawn, shared the same responsibility as the Emperor, but on a smaller scale. Their behaviour in their offices and homes across the empire transformed local society in the same way that the Emperor civilised the world. Commoners made up the bulk of this social order. They had the potential to be 'good citizens', but they required guidance from officials and literate elites in general.
Local gazetteers often liken this guidance to the way parents guide their children. They refer to officials as metaphorical parents of the people, or 'parent officials '. 31 According to John Walter, the language that officials use to explain why the social order is in the best interest of the people also limits their authority. 32 In this case, while the people were expected to follow official guidance, the metaphor also tasked officials with the moral obligation to take care of the people, for example in case of famine or when bandits attacked. Hilde de Weerdt similarly argues that the relationship between the Emperor, his officials and the people was reciprocal. The Emperor would be unable to rule without his officials or the people; the officials relied on imperial authority to govern the people; and without the Emperor or his officials the people would merely be people. Together the Emperor, his officials and the people constituted the body politic. Each group represented a distinct part of the metaphorical body, but they all relied on each other for the continued survival of the body as a whole. 33 The system of values that produced descriptions of acceptable behaviour also provided justification for the exclusion of unacceptable behaviour. Ming officials excluded those engaged in banditry whose violence and greed did not conform to norms of acceptable behaviour. Although sources may present bandits as those who '[placed] themselves outside the political organization of the community', 34 they were excluded precisely because officials labelled them bandits. Furthermore, this process of exclusion also worked in reverse. Officials relied on moral instruction to transform those labelled 'bandits' into 'citizens', thus allowing them to become part of the social order again. Although that transformation was presented as the individual choice of those who first followed evil men and then repented in front of the morally upright official, the ultimate decision lay with the officials who guarded the permeability of the social order. In other words, officials used moral instruction to include the excluded and thus removed the source of conflict between inside and out.
Moral instruction made sense in theory from the viewpoint of Ming officials, but the practical implementation as well as the actual interaction with bandits was more complicated. A key figure in the practical implementation of moral instruction in Jieyang was Chen Ji. The Longxi County Gazetteer describes Chen as a 'rural Confucian', someone who studied Confucian classics but did not have an official position that was recorded in any gazetteer. The clearest indication of Chen's status was the examination success of his son, Chen Shibao, who obtained the highest degree in the civil service examination system 35 at the age of 30 when he became Metropolitan Graduate in 1464. 35 Chen Ji had the wealth to support the extensive education of his son and raise at least three more sons, as well as an unknown number of daughters. Furthermore, Chen had the cultural capital that allowed him to communicate with Zhou and thus establish a relationship with him. In other words, Chen belonged to one of the influential households in Chaozhou and was therefore expected to take a leading role in local society. Chen established initial contact with the followers of Huang. He prepared the way for Zhou to visit the camp and deliver moral instruction. The content of the message that Chen and Zhou brought to the bandits is unknown, but likely appealed to a perceived desire to be part of society. An example from the sixteenth century asked villagers who had sons, brothers, friends or neighbours, among the bandits to deliver the following message: Among the thieves are numerous people who were captured and forced to follow. . . . Who would not want to return to their original village, to be together with father, mother, wife, and children, their entire kin, to secure together the sacrifice to their ancestors? . . . I regret that these people are not part of the body. . . . Persuade them to return home and live in safety, to avoid the tragedy of capital punishment. 36 This message emphasises family connections and a larger sense of community, combined with the thinly veiled threat of capital punishment for those who continued to act as bandits. According to the Longxi County Gazetteer, Zhou personally visited the bandit camp and he personally returned the people who had been captured. This personal connection was important in the creation of community. Ming officials imagined a community of 'good citizens' who performed their appropriate roles in the social order. James Scott argues that officials reduce social complexity to simplified schemes that aid them in governance, and since officials wield state authority, their simplified schemes have the power to shape society. 37 This simplified scheme of society informed the way in which Ming officials categorised and treated the people in their jurisdiction. Zhou identified Huang as the leader and executed him, but Zhou also distinguished a group of people who were 'forced to follow' the bandits, and he argued on their behalf. Another official proposed complete extermination to solve the problem, but Zhou overruled him and offered amnesty instead. Zhou returned people to their homes in an effort to repair communities. Personal connections, not only with him but with local elites such as Chen as well, helped to solidify that sense of community, which was further strengthened by family connections.

The bandits return to Jieyang
Zhou seemingly ended banditry in Jieyang. He was promoted to the rank of vice-minister in recognition of his swift and efficient intervention in Chaozhou. Zhou was successful 38. Chen, Jieyang xianzhi, juan 7, 1218. 39. Guo Chunzhen, Chaozhou Fuzhi (Jiajing) [Chaozhou Prefecture Gazetteer (Jiajing Edition 1547)] (Reprinted Beijing, 2009), juan 5, 294. in the eyes of central government officials. The task of any prefect was to maintain social order, or restore order when necessary, and gather tax revenue. Zhou pacified the bandits who followed Huang. He also returned the people who had been dislocated from their original households. These people resumed their occupation and started to generate tax revenue again. Perhaps more important was the opinion of the 'good citizens' of Jieyang. When Zhou received a promotion, 200 people blocked the road to keep him from leaving. Chen Ji and others petitioned the court to allow Zhou to stay, but a new prefect was already on his way from the capital. However, as soon as Zhou left, bandits returned. Wei Chonghui and Xu Wanqi besieged Jieyang city for more than a month in 1462. 38 They pierced the sea dykes that Zhou had repaired, thus flooding the area and ruining agricultural fields for years. The farmers who lost their livelihood became an additional burden on the county and some of them may have joined the bandits, like the flood victims in Longxi did. In many ways, the residents of Jieyang were in worse trouble than before. Zhou's approach to the issue of banditry in his jurisdiction relied on his presence. He developed personal connections with local elites, former bandits and rescued captives. Zhou tried to incorporate agrarian and coastal communities in his imagined community of 'good citizens' without acknowledging the occupational, cultural and socio-economic differences that gave rise to banditry in the first place. Zhou did not take advantage of his victory to consolidate state power in coastal society, for example through relocation of military garrisons to Jieyang city, or the establishment of additional state institutions such as police offices or coastal patrol offices. His approach depended on the guidance of a morally upright official, as well as the thinly veiled threat of capital punishment. But Ming emperors did not allow prefects and county magistrates to remain in one jurisdiction for more than a few years, lest they build up a power base to challenge imperial authority. Zhou was moved out of this problem area, and his achievement fell apart.
In the absence of Zhou, Chen Ji consulted with Jieyang magistrate Chen Jue (no relation). Chen Jue was a native of Zhangzhou, like Zhou, and he obtained the jinshi degree in 1457 at the age of 27, which meant that when he assumed the office of Jieyang magistrate in 1460, he was only 30 years old. Magistrate Chen donated his salary to purchase relief grain for the flood victims after bandits had pierced the sea dykes. As in the case of flood victims in Longxi, however, relief grain may not have been enough. Chen also strengthened the walls of Jieyang city in preparation for renewed bandit attacks. 39 Still, Chen was afraid the bandit problem was beyond his ability as a county magistrate to solve. Together with Chen Ji he composed a letter to Zhou and begged him to return. Zhou agreed and broke the siege of Jieyang city. The personal connections between Ming officials and local elites, and among officials as well, created a degree of freedom in an otherwise strict political structure. The relationship Zhou cultivated with Chen Ji allowed the latter to call on Zhou for support when the bandits returned, despite Zhou's formal appointment in Shanxi Province on the northern border of the Ming Empire. The political structure prioritised short-term solutions, which advanced political careers, but the social 40. Luo, Zhangzhou xianzhi, juan 17, 323a. 41. Kenneth Dean, 'Transformation of the She (Altars of the Soil) in Fujian', Cahiers d' Extreme-Asie, 10 (1998), 19-75. networks of Ming officials added a degree of continuity in the form of advice among colleagues or, in this case, physical reinforcement in a time of need. However, in the case of Jieyang, this degree of continuity still relied on the personal presence of Zhou. The swift return of bandits to Jieyang did not change Zhou's approach. He issued another announcement in which he called upon the bandits to surrender. They arrived in several hundred ships, dressed in luxurious clothing and holding their weapons at the ready. When Zhou asked why they still carried weapons when they came to surrender, the bandits explained the situation as follows: There was an illicit shrine at the seaside where the thieves asked the spirits through divination whether to comply with or resist [the order to surrender]. [The thieves] came to surrender on the appointed date, but they still carried weapons. [Zhou] Xuan asked them about this, and they answered: "The divination was inauspicious." [Zhou] Xuan arrived at the shrine at night and announced: "If the people want to follow the right way, then how can the spirits tell them it is not auspicious? These spirits do not know that there is [a difference between] obedience and resistance." 40 The importance of this 'illicit' seaside shrine for the bandits who attacked Jieyang city provides insight into the networks that held bandit groups together, and allowed them to regroup in the aftermath of state intervention. More importantly, perhaps, this incident illustrates that Zhou continued to approach banditry as a moral issue that could be solved through moral instruction. His inability to understand the importance of this seaside shrine for the social organisation of this bandit group, limited his ability to address the social context of banditry in Jieyang. Even though Zhou grew up the coastal county Longxi, he was a citizen of the agrarian Ming Empire.
Zhou felt confident in his challenge of religious authority because the founder of the Ming Empire had established institutions to regulate religious worship throughout the empire. Units of 110 households grouped together, also known as lijia units, were required to maintain an Altar of the Soil. Every lijia unit was responsible for the organisation of an annual ceremony at the altar where all members renewed their loyalty to the community. In principle the ceremony took place out in the open and according to official regulations. However, Kenneth Dean argues that in the case of Fujian, people gradually abandoned the official altars and replaced them with temples that more closely reflected established communities. 41 These temples continued to function as the spiritual focus of their community, though not necessarily within the administrative view of the state.
The seaside shrine performed a similar function for the bandits who gathered there, but was outside the reach of Ming officials. In this story, the shrine is labelled 'illicit' to exclude the shrine from the state-sanctioned system of religious sites, in the same way that the label 'bandit' excluded individuals from the social order. The Ming founder developed a system of state-regulated Altars of the Soil to simplify the complexity of religious worship in the Ming Empire. This illicit shrine in Chaozhou shows that individuals found ways to avoid or resist such simplified regulation. Bandits gathered at this shrine to connect with each other and discuss decisions that affected the entire group. The shrine thus became a focal point that strengthened cohesion in this group of bandits, while spirit divination facilitated decision-making even in the absence of clear leadership. Furthermore, the shrine may also have been a nodal point in a network that extended into coastal communities. The shrine was located at the seaside, which suggests a maritime spirit worshipped by people who interacted with the sea for a living, such as fisherfolk, sailors, maritime merchants, pearl divers, clam gatherers and a variety of travellers, as well as people who lived on or near the sea, or people who wished to secure the safe return of those who went out to sea. These are precisely the people with valuable maritime skills, local knowledge or connections to trade networks who could improve the success of a bandit group. In short, the seaside shrine strengthened cohesion among bandits and provided space for interaction between bandits and individuals from local communities. This, in turn, improved their ability to regroup in the aftermath of state intervention, even if their leader was killed.
Through the illicit seaside shrine we catch a brief glimpse of the rich maritime world that existed on, and beyond, the south coast of the Ming Empire. Officials such as Zhou, however, did not see this world. Ming officials tried to force residents of coastal communities to conform to the social norms of an agrarian empire. Their moralistic approach to banditry on the south coast ignored the many reasons why individuals in this environment turned to banditry in the first place. Zhou temporarily pacified Jieyang when he executed Huang, and his personal connections allowed him to come back and face the bandits again when they attacked Jieyang for the second time, but his unfamiliarity with the maritime world limited his ability to secure long-term peace. The illicit seaside shrine was only one node in a maritime network that allowed bandits to regroup and repeatedly emerge along the south coast, even though Ming officials repeatedly targeted their leadership.

Conclusion
The contrast between agrarian and coastal communities on the south coast of the Ming Empire attracted the recurring attention of bandits who suddenly arrived by sea and sailed upriver to loot the county capitals where agricultural wealth accumulated. Ming officials approached the issue of banditry from a moralistic viewpoint that ignored the geographical and socio-economical context that gave rise to banditry in the first place. Their attempt to incorporate residents of coastal communities into an imagined community of 'good citizens' often achieved temporary success, but failed to deliver a long-term solution. Furthermore, the political structure of the Ming Empire encouraged short-term political victories, although personal and professional connections among Ming officials and local elites did provide a degree of continuity. In the end, however, coastal residents continued to find the means and opportunity to engage in banditry along the south coast of the Ming Empire.
Zhou Xuan failed to convince the bandits at the illicit seaside shrine to surrender. County magistrate Chen Jue supervised the soldiers who attacked the bandit camp. He also ordered them to burn down the houses of residents in Xialing who provided shelter for bandits. Chen restored peace in Chaozhou in the autumn of 1462, as Zhou had done in 1459, but according to the local gazetteers of Chaozhou and Zhangzhou bandits emerged again in Zhangzhou in the 1470s and the 1490s, while other bandit groups attacked Chaozhou in the 1480s, 1500s, 1510s, and so on. The contrast between an agrarian empire and a maritime world continued to give rise to banditry on the south coast of the Ming Empire.

Author biography
Sander Molenaar is undertaking doctoral research at the University of Warwick, UK. His thesis examines the impact of coastal banditry on the interaction between state and society on the south coast of the Ming Empire during the fifteenth century. His research interests include Ming history, maritime history, travel writing and local gazetteers.