The Chinese State in Ming Society (Critical Asian Scholarship)

The Chinese State in Ming Society The Ming dynasty (1368–1644), a period of commercial expansion and cultural innovation, fashioned the relationship between the present-day state and society in China. In this unique collection of reworked and illustrated essays, one of the leading scholars of Chinese history re-examines this relationship and argues that, contrary to previous scholarship, which emphasized the heavy hand of the state, it was radical responses within society to changes in commercial relations and social networks that led to a stable but dynamic “constitution” during the Ming dynasty. This imaginative reconsideration of existing scholarship also includes two essays first published here and a substantial introduction, and will be fascinating reading for scholars and students interested in China’s development. Timothy Book is Principal of St. John’s College, University of British Colombia.

Critical Asian Scholarship Edited by Mark Selden, Binghamton and Cornell Universities, USA The series is intended to showcase the most important individual contributions to scholarship in Asian Studies. Each of the volumes presents a leading Asian scholar addressing themes that are central to his or her most significant and lasting contribution to Asian studies. The series is committed to the rich variety of research and writing on Asia, and is not restricted to any particular discipline, theoretical approach or geographical expertise. Southeast Asia A testament George McT.Kahin Women and the Family in Chinese History Patricia Buckley Ebrey China Unbound Evolving perspectives on the Chinese past Paul A.Cohen China’s Past, China’s Future Energy, food, environment Vaclav Smil The Chinese State in Ming Society Timothy Brook

The Chinese State in Ming Society Timothy Brook

LONDON AND NEW YORK

First published 2005 by RoutledgeCurzon 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN Simultaneously published in the USA and Canada by RoutledgeCurzon 270 Madison Ave., New York, NY 10016 RoutledgeCurzon is an imprint of the Taylor & Francis Group This edition published in the Taylor & Francis e-Library, 2005. “ To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to http://www.ebookstore.tandf.co.uk/.” © 2005 Timothy Brook All reserved. No part of this book may be reprinted or reproduced or utilized in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloguing-in-Publication Data A catalog record for this book has been requested ISBN 0-203-31133-7 Master e-book ISBN

ISBN 0-415-34506-5 (hbk) ISBN 0-415-34507-3 (pbk)

Contents List of illustrations

vii

Acknowledgements

x

Abbreviations Introduction: a grave in Nanchang PART I Space

xii 1 16

1 The spatial organization of subcounty administration

17

2 The gazetteer cartography of Ye Chunji

42

PART II Fields

60

3 Taxing polders on the Yangzi Delta

61

4 Growing rice in North Zhili

79

PART III Books 5 Building school libraries in the mid-Ming 6 State censorship and the book trade PART IV Monasteries

97 98 114 131

7 At the margin of public authority: the Ming state and Buddhism

132

8 Buddhism in the Chinese constitution: recording monasteries in North Zhili

150

Conclusions: states of the field

172

Notes

181

Bibliography

213

Index

232

Illustrations Cover A monk-demon is untouched by the arrows the official has ordered his soldiers to fire; illustration from the 1592 Hangzhou edition of the popular story by Luo Guanzhong, Sansui pingyao zhuan (The Three Sui quell the demons’ revolt). One could speculate on the tensions between civil and monastic establishments buried within this picture (see Chapter 8).

Maps

1

Ming China

xii

2

The Lower Yangzi region (Jiangnan)

64

Figures

1.1 Administrative map of Wujiang county, Suzhou

23

1.2 Administrative map of Taicang subprefecture

24

1.3 Map of coastal defences of Jiading county

25

2.1 Map of Huian county

49

2.2 Cartographic legend on the map of Huian county

50

2.3 Map of the Huian county seat

51

2.4 Map of Huian Township 2

52

2.5 Map of Zengcheng county, Guangdong

56

2.6 Map of the waterway network of Tongxiang county, Zhejiang

57

3.1 Small polders along the Yangzi River

63

3.2 The division of Changshu county by sector

71

3.3 Sector 34 of Changshu county

72

4.1 The paddy fields watered by Rice Canal

93

7.1 The Lamaist monastic complex on Wutai Mountain

140

8.1 Map of the seat of Hejian prefecture

167

Tables

1.1 Registered population of Anqui county, Shandong

21

1.2 Households per subcounty unit in five prefectures

22

1.3 Distribution of subcounty administrative units in a standard structure

26

1.4 Distribution of subcounty administrative units in an elaborated structure

26

1.5 The subcounty system in Huian county

27

1.6 Hundreds (li) per county, by province

34

1.7 Subcounty administrative units in the Ming

38

4.1 Number of counties and subprefectures for which rice is first reported as growing in North Zhili

84

4.2 Percentage of rice in Ming grain and hay taxes

88

5.1 Core texts in Ming school libraries

105

Acknowledgements To Mark Selden belongs the credit of helping me discover the book that lurked within these essays and encouraging me to fashion the parts into a coherent whole; the volume would not exist but for his prompting. I would like to name two others who have contributed to the thinking that went into this project. Ever since our graduate student days together, Bin Wong has read and criticized my work with the precision of a scholar and the patience of a friend, and did so once again on this occasion. Michael Szonyi has been a more recent influence, but his Practicing Kinship in particular helped me rethink some of the issues this volume addresses. For their more practical contributions, I wish to thank Trish McAlaster for drawing the maps, Nick Hawkins for editing the manuscript, and my research assistant Kevin Lu for being so patient and enthusiastic through the long process of getting from essays to book.

Abbreviations jr.

year in which juren degree was conferred

js.

year in which jinshi degree was conferred

r.

reigned

Map 1 Ming China

Introduction A grave in Nanchang By the time the dossier reached the Hongzhi emperor’s desk on 29 November 1499, the case had become complicated—which is what it had to be in order to get there. Wang Zhen owned a piece of land in the hills outside the city of Nanchang, the capital of Jiangxi province. In this hilly region south of the Yangzi River, population was dense, land scarce, and the locals often on the move elsewhere looking for work or land. “The hills are many and the fields few,” as a Nanchang county author noted by way of explaining why the local people were so lean.1 Even the hills, used for graves rather than fields, were at a premium. The most coveted bits of upland topography were those spots where professional geomancers judged that the lines of energy (qi) streaming through the landscape converged propitiously. Bury an ancestor within such an energy field and the deceased’s spirit will radiate fortune to his descendants. Jiangxi lineages competed for the best tomb sites and resorted to tricks and violence in their struggle to improve their fortunes at others’ expense. Grave land feuds were endemic to the province through the Ming and Qing dynasties. The case that went up to the Hongzhi emperor started because Zhang Yingqi buried a body on Wang Zhen’s grave land without his permission. Zhang was a student on stipend at the Nanchang government school. An aspirant to higher elite status, he was positioned to rise into the upper gentry should fortune, education, and wealth continue to conspire in his favour. Whom Zhang buried on Wang’s land, and why he had chosen to bury that person on land that was not his, are not stated in the surviving case summary that appears in the court digest, the Veritable Records of the Hongzhi Reign (Xiaozong shilu).2 Zhang appears not to have been driven by the usual goad of poverty. Geomancers must have declared this bit of hill as top grade for burial purposes, a place where Zhang might entomb his ancestor so splendidly that good fortune could not help but rain down upon the living, including himself. Wang Zhen, the owner of the land, was not a student, nor did he possess any token of official status. Yet even a commoner could take his case to court, if he were willing to deal with the exactions and interferences of the lesser functionaries standing between him and the presiding judge. This is what Wang did, filing a lawsuit with the prefectural government. Given the high costs of pursuing a case through the court and the impossibility of controlling the outcome, only the truly desperate surrendered their conflicts to official arbitration. But Nanchang people seem to have breathed a different judicial culture. The compiler of the earliest Ming-period gazetteer for the prefecture, produced in 1378 in response to the request from the Hongwu emperor (r. 1368–98) that local gazetteers be submitted to the court, praised the ardent passion for virtue and diligence animating the hearts of the local people. The chronicler also noted, though, that these ardent feelings—animated by the same energy (qi) that the geomancers detected in the land—could go to excess, breeding intolerance and propelling people into lawsuits. This was not news to the emperor, who in his final instructions to the people in 1398 singled out Jiangxi natives for being “prone to litigation” and complained that they “cannot endure even minor matters, and go directly to the capital to bring suits.”3 Nor

The chinese state in ming society

2

was a taste for lawsuits in Jiangxi uniquely an early-Ming reputation. The compiler of another gazetteer within the prefecture notes that during the Yuan dynasty, Nanchang people had a reputation for “relishing fights, losing their tempers, and enjoying lawsuits.” The subsequent edition of the prefectural gazetteer, in production at the time Wang Zhen’s case was being heard, observed that over-taxation, poverty, and land scarcity were making judicial tempers that much worse.4 Nanchang’s reputation for fractiousness did not dim with time. A gazetteer editor in 1565 complained that local people “enjoy instigating idiotic lawsuits and giving vent to their passions, all for the sake of small profit.”5 A few decades later, another Jiangxi commentator summarized Nanchang people as “avid in work and stingy in giving, indifferent to duty and happy to quarrel; cunning and glib, litigious and libellous.”6 The prefectural judge agreed to hear Wang Zhen’s case and decided in his favour. Presumably he ordered Zhang Yingqi to remove the bones he had buried there. Zhang, however, decided to fight back. He turned to a fellow student at the Nanchang school, Liu Ximeng, who had something Zhang needed: a connection to a higher authority. Liu had got a job tutoring the son of an assistant provincial surveillance commissioner, Wu Qiong (js. 1469). Liu had managed to ingratiate himself with the Wu family by handing out presents, thereby positioning himself as a small-time power broker between Wu Qiong and anyone who might want his help. The connection suited Zhang, for an assistant surveillance commissioner outranked a prefectural judge. Money changed hands, and a word from Liu in someone’s ear in the Wu household succeeded in getting Wang’s case against Zhang overturned. Zhang was unwilling to rest content with this victory. Presumably the disputed corpse was still in the ground and he was still vulnerable to counterattack from Wang. He chose to go on the offensive and put together a lawsuit against Wang. As the prefectural judge had already judged against him, Zhang found another patron, Assistant Education Intendant Su Kui (js. 1487). Su approved admissions to government schools in Jiangxi province and regularly tested the students who were admitted, responsibilities that placed Zhang in the position of being able to communicate with him legitimately. An issue such as this, however, lay outside Su’s proper jurisdiction. Su had a reputation for refusing to act on behalf of private interests, which makes his willingness to help Zhang, allegedly after receiving a bribe, puzzling. It is possible that none of this is true, or that Zhang misled Su into acting on his behalf without full knowledge of what was at stake. Wang Zhen meanwhile, seeing that Zhang had support elsewhere in the Nanchang bureaucracy, looked for a connection to another reservoir of state power, the eunuch establishment. How he got to Grand Defender Dong Rang, a eunuch of the imperial household whom the emperor had sent out to supervise regional security, is not known.7 Presumably the right sum of money could open any doors, so long as one knew which corridors to travel. Wang presented his case to Dong, and Dong obliged him by having both Zhang and Liu thrown in prison, where torturers could persuade them to withdraw Zhang’s claims. Until this point, there was nothing unusual about what the players in this little drama did. Two people caught in a struggle over land had looked up into the lower levels of the state bureaucracy in search of connections to aid them in their struggle, and mid-level provincial officials had been happy to oblige. Money had changed hands on both sides, and the justice system turned into a network for channelling bribes and influence from competing nodes of authority, not for resolving disputes. By acting as they did, the two

Introduction

3

litigants were conforming to the modus operandi of the political system in which they found themselves. This is how the Ming state—which may be defined as a coercive system of territorial authority and communication moving information, resources, and personnel in regular ways designed to ensure the wealth and security of the dynasty— worked. The apex of the communications system was the throne, and the channels along which information, resources, and personnel moved were the lines of authority and reporting through which percolated the throne’s capacity to act via the imperial household, the bureaucracy, and the army. By seeking influence with people embedded in the state’s administrative networks, Zhang Yingqi and Wang Zhen were simply responding to the opportunities available to them on the landscape of power. Their intention was not to send their dispute all the way up to the throne to resolve, of course. It was to send up a counter-flow of influence and information into the lower reaches of the state system, a modest capillary response to the percolation of authority downward, with the expectation of stemming (in Wang’s case) or directing (in Zhang’s case) its flow. The turning point in this communicative process—when a grave dispute between minor families makes its way up through the system to the emperor’s attention and leaves traces in the dynasty’s written records—came when Grand Defender Dong handed over Zhang Yingqi and Liu Ximeng to keen jailers. Dong intended only to intimidate the two students into backing off, but the torturers went too far. The hapless two readily revealed that they had been bribing state officials to back their side of the lawsuit. Once this revelation came out, what had been a local property dispute, best handled in that context, turned into a bureaucratic crime that had to travel up officialdom’s hierarchy to Beijing, first to the censorate, then to the Ministry of Justice, and finally to the throne. Dong Rang had gone too far, and now an emperor was looking down at the situation in Nanchang and demanding that the ministry investigate. The modest capillary action of rational bribery from below (what could be more rational than to appeal effectively to those who could produce a favourable decision?) overwhelmed the percolating gravity of state authority from above, in part because the information it carried was of a sort that interested this particular emperor. Hongzhi (r. 1488–1505), as it turns out, was passionate on the subject of corruption: his reign’s records are filled, more than those of any other Ming emperor (with the exception, as in all things, of the foimder, his great-great-great-greatgrandfather, Emperor Hongwu), with dismissals for official malfeasance and incompetence. The investigation uncovered that Eunuch Dong had already marked Su Kui as his enemy before Wang Zhen ever came to him for assistance. Indeed, the rivalry between them gave the bribes their traction. Dong felt that Su had insulted him over another matter, and so agreed to take Wang Zhen’s side as a way to get at Su. Did Su actually agree to support Zhang Yingqi? Or did Dong embroider the formal connection between them into something that made the commissioner look as though he were improperly involved in a land dispute, when he was doing nothing of the sort? It is impossible to tell. What is clear is that the dispute had spiralled out of the control of the two men who started it, turning into a case that had nothing to do with who had the right to bury his ancestors where, and everything to do with the political tension between the eunuch and civil bureaucracies in one provincial capital. Su returned Dong’s dislike, and was not alone in doing so. Dong’s highhanded activities on behalf of the imperial household inspired other officials elsewhere in the bureaucracy to petition for the man’s removal,

The chinese state in ming society

4

first to the Hongzhi emperor while he was alive, and later to his successor, though never to any effect.8 Dong was in the stronger position and managed to get Su thrown in prison on a corruption charge, probably in connection with a different matter. Students at the Nanchang school were so offended by the eunuch’s attack that a hundred of them stormed the jail and freed their superior. Su was exonerated and later promoted (everyone assumed he was innocent, though that is what they needed to believe), yet Dong was left untouched.9 Matters having gotten so far out of hand, the emperor could have dispensed terrible judgments on those involved, yet he chose not to. His hand may have been stayed by the need to protect his exposed eunuch servant, who after all was supposed to be in Nanchang to look out for imperial (i.e., his) interests. Or he may have wanted to avoid siding with one arm of the state over another, so as to keep his eunuchs and bureaucrats in dynamic tension with each other. Hongzhi explained his decision not to take harsh action by reasoning that no actual damage had been inflicted on any of the parties. He reprimanded Dong Rang and Su Kui for agreeing to adjudicate lawsuits they were not entitled by their positions to entertain, and he reprimanded Su Kui and Wu Qiong for taking payments. The burden of his judgment fell away from his officials, however, landing most heavily on the two students who started the affair. Zhang Yingqi and Liu Ximeng were not beaten or fined or sent into exile, which could have been their fate at the hands of an angrier emperor. Instead, they were stripped of their studentships and stipends and banned from ever again trying to climb that ladder of success—punishment enough in a status environment as competitive as mid-Ming China’s. The storm in Nanchang’s teacup happened to catch an emperor’s eye, and that it did suits my purpose, which is to frame the eight studies in this book by inquiring into the presence and power of the Chinese state in Ming society. The dramatic intervention of an emperor could be taken as a vivid example of the state’s capacity to control society: a demonstration that the Ming court could reach all the way down to the bottom of the realm and pull apart two men fighting over a grave. This is how Ming historians would once have told this story, when the emperor fetish that has long lurked around the field of Ming studies was still strong.10 The fetish is one that Ming historians inherited in the first instance from Ming officials, caught as they were within the operations of a public rhetoric that obliged them to refer to an emperor by the correct euphemism (Sagely Founder, one of Hongwu’s titles, would have done Confucius proud) and to glow when doing so. Our emperor fetish is also the product of a long historiographical tradition in Europe going back at least to the ghost of Georg Hegel, who could not conceive of China other than as a realm in which only the emperor had full individuality and every other person was his slave.11 If the remarkable man who founded the dynasty in 1368 had been the only emperor of the Ming, the fetish might well be justified, given his extraordinary energy and the mythic scale on which he created and destroyed. But he wasn’t, and it isn’t. Another way of understanding the Nanchang burial case is to remain within the state frame but reverse the equation between the emperor’s presence and the operation of the bureaucracy and regard what Hongzhi did as a momentary disruption in the routine functioning of state administration, the exception that breaks the rule of flawless state control rather than the one that proves it. For rarely did an emperor intervene in what went on in the bureaucratic structure beneath him; even less could he see into the social

Introduction

5

networks stretching out beyond the bottommost rungs of that system. As I shall note at several points in this book, an emperor had within his gaze only what his officials brought to his attention. Hongzhi put himself in the way of more information than most of his line. One of his first acts after his enthronement was to dismiss almost the entire staffs of the Ministries of War, Justice, and Personnel12 as a sign that he would not tolerate the corrupt and incompetent. This move prompted the zealous and the ambitious to forward more information to him about what was going on in the field bureaucracy than was usually the case. Even so, the range and depth of his knowledge was limited. As Hongzhi himself admitted in January 1499 in the edict of penitence he issued after the Qingning Palace within the nine-walled Forbidden City burned down, “I live deep within the Nine Walls, and though I stretch my thoughts over the entire realm, there are places my ears and eyes do not reach and where my grace has not been manifested.”13 This was not simply a matter of scale, however. The regular bureaucracy and the parallel intelligence operations of his eunuchs directed information his way, but both could block or distort information as well as transmit it. Dong Rang’s activities would never have come to his eyes or ears through the eunuch channel, for instance, nor were the alleged bribes that Wu Qiong and Su Kui accepted knowledge that the civil officials wanted relayed to the emperor if they could help it. An emperor’s communicative links to society were few, and easily closed when all his subordinates agreed that they should be. But this is not the crux of the problem of reading Ming history from the state side. That problem resides rather in our conceptualization of the Chinese state. Ming people knew they were subject to the authority of the emperor, but that is not how they experienced the state. The state exerted its presence in Ming society less because of what the man at the top did or wanted done—it must have shocked the litigants to have the emperor weigh in on their case—than because of interventions of state representatives further down the communications system. Even these interventions were exceptional, for most people knew the state only by distant proxy in the course of dealing with the systems through which their affairs were administered; specifically, the taxation, education, justice, and military systems that made the state present in society as more than an abstraction. The taxation system, with its regular grain taxes and labour levies, was the common context in which people interacted with the Ming state and its officers, which is why most of the studies in this book address issues arising from the taxation nexus between state and society. The education system, regularized as a bureaucratic operation in 1436 when education intendants such as Su Kui were appointed in the provinces, affected only a minority of young men, though the aspiration to gain a place at a government school was socially pervasive. The justice system held universal sway over the emperor’s subjects, ready to snare anyone who contravened his laws, yet the number of suspects and plaintiffs who went before a magistrate could not have been great. Most people managed to live their lives without getting tangled in the law. Least likely to touch the lives of ordinary people was the military system, which did not conscript civilians but drew its soldiers from hereditary military households. A military designation was a fiscal rather than martial category, though, and those who could escape it and move into the commoner population did so. The most likely way for commoners to encounter soldiers was when the latter were mobilized to deal with banditry or other unrest, which is why a eunuch grand defender such as Dong Rang was assigned to the unsettled province of hilly

The chinese state in ming society

6

Jiangxi. These agents of the imperial household were farther removed from the people than the regular bureaucratic systems, and more mysterious for their power because of the independence from bureaucratic oversight. The imperial household took an interest in military affairs early in the fifteenth century, when the Yongle emperor put eunuchs on military assignment. Best known from his reign is the Muslim eunuch Zheng He, who commanded the imperial fleets that sailed to the Indian Ocean. Military eunuchs were useful to an emperor, enabling him to keep an eye on security situations without having to see everything from his bureaucrats’ point of view. They were also an annoyance to those same bureaucrats, who distrusted them for not being accountable to regular standards and resented the relatively free hand they seemed to enjoy extracting resources on behalf of the imperial household—and themselves, of course. A year before the Nanchang grave case, the Ministry of War had gently suggested that Hongzhi reduce the number of full and associate grand defenders in the realm, arguing that their operations were a heavy financial burden on the common people. The ministry was tactful enough to blame Hongzhi’s predecessor, the Chenghua emperor (r. 1465–87), for escalating the scale of eunuch surveillance. Hongzhi turned the request down.14 When, just a few short weeks after the affair with Dong Rang, a county magistrate in the northeast indicted a eunuch grand defender there along with two military officials for misconduct that included the indiscriminate slaughter of border people, the emperor would not support his demand for a full investigation. Let the censor already in the region look into the matter, Hongzhi replied. A year later, he dismissed all charges.15 The eunuchs were simply too important to his strategy of control to leave dangling for civil bureaucrats to attack, however badly they behaved. The tension between them and the regular officials was a feature of the Ming constitution, and one that Ming emperors favoured as a device to retain some control over decision-making and policy implementation within the regular bureaucratic systems. The Nanchang burial case managed to entangle all these systems. Most visibly in play was the justice system, which Wang Zhen elected to engage through the proper channel, and which Zhang Yingqi sought to subvert by bribing the prefectural judge’s superior. Also prominent in this story is the education system, which provided Zhang with his access via Liu Ximeng to the assistant surveillance commissioner, and then with his superior in the educational hierarchy, Su Kui. The military system was not directly involved, though Dong Rang’s appointment to oversee regional security empowered him to weigh in and subvert the procedures for settling land cases. Finally although taxes did not come up in this case, the taxation system framed what was going on. Wang and Zhang were fighting over a plot of land that had not been put under the plough and so “brought onto the registers,” as the assignment of tax liability was phrased.16 It was land on which no tax had to be paid. Zhang’s eagerness to push Wang off it confirms this, as he would not have wanted to take over land for burial purposes were he obliged to pay an agricultural tax on it. The tactic for taking control of someone else’s taxable land was discreet encroachment, not forceful seizure, lest an official take note and transfer the tax burden. Zhang was not being discreet. He was “stealing a burial site,” as the court record phrases it in the language of brigandage. The court historian who summarized the dispute for the Veritable Records does not disentangle these systems, for they flow together in the same channel of imperial authority. He begins his account with Liu Ximeng’s relationship to the assistant

Introduction

7

surveillance commissioner, since this is the context that turned a land dispute, otherwise of no concern to the court, into a corruption case in which the emperor had an interest. This way of telling the story could be used to narrate Ming imperial rule as an autocracy in which the ruler served as the fulcrum on which the state had to pivot in order to function. This version would confirm the paradigm of despotism, the origins of which go back to Montesquieu and Hegel and the effects of which, thanks to Karl Wittfogel’s recycling of the trope of “Oriental despotism,” shaped the field of Ming studies as it came into being during the Cold War.17 Given the number of local players who became involved, however, the incident could sustain a different reading, one which tells the story not as the re-enactment of the paradigmatic relationship between an absolute ruler and his absolute subjects—between jun and chen, in the language of Hongwu18—but as the working through of some of the possibilities and constraints of bureaucratic administration. This is how the field of Ming studies began telling Ming stories in the 1970s, digging out from under the dynasty’s reputation as Oriental despotism’s ideal type by building up knowledge of how Ming government worked in practice.19 Consider, though, another reading, one which begins where I began telling this story: not with student Liu Ximeng’s connection with official Wu Qiong, but with the conflict between landowner Wang Zhen and burier Zhang Yingqi. These were the people whose actions set the story going, not the provincial officials, and certainly not the emperor. How each chose to act depended in part on the state systems accessible to him. Like Chinese in almost any period, both must have kept a weather eye on the state. But the conflict arose and took its shape because of the social networks to which Wang and Zhang were tied. Only when fellow student Liu comes into the story do the available state systems begin to direct the flow of events, bringing Wang and Zhang into the emperor’s view and turning a local land-dispute story into a national corruption narrative. State systems were important in this history, but the social was prior. Accordingly, I would like to propose that what was distinctive about Ming China was less its state than its society, since it was within society that the effects of demographic growth, expanding communication networks, rapid commercialization, and new critical thinking were most keenly felt. After the first highly interventionist Ming reign, the state more or less followed in the wake of these shifts, attempting to manage a realm of unprecedented complexity rather than remake what it found. Even when an activist emperor was able to impose organizational frameworks and limitations on local society, his agents could sustain them only by fitting them to the social networks that predated their imposition. Drawing ward boundaries and shepherding communities into lijia units as elements of a state-making program, as we will see the founding emperor doing in the first chapter of this book, were ambitious in what they were designed to achieve, yet underneath these state-directed interventions runs a very different process that political theorist Roberto Unger has coined “society-making.” “Society-making” denotes the process by which people interact with each other through structured networks and make the conditions of their social exis-tence on the basis of the resources available to them by virtue of their social position. “These resources include governmental power, economic capital, technical expertise, and prestigious ideals or the forms of argument that claim to show implications of these ideals.”20 The state may set or seek to influence how these resources become available and how they may legally be used, but the actual forms of the processes through which

The chinese state in ming society

8

the state intervenes to shape ranks and roles depends on the practices current within everyday life, regardless of its own legitimacy and authority claims. My purpose in invoking “society-making” is not to marginalize state effects, but to insist on an analytical distinction from the more familiar concept of “state-making,” the term we use when the state mobilizes resources to build up its administrative capacities and enhance its security. Unger’s purpose was to field an argument in favour of a radical reorganization of power to enable citizens to constitute a democracy not in thrall to the state: to make society prior to the state. He writes from within the European constitutional tradition dating at least to the eighteenth century, in which the rules of political life guaranteed the security of the person and his social relations from state interference. The foundational assumption of an eighteenth-century constitutional theorist such as William Blackstone was that the security and liberty of the individual were secured against state infringement by the law, and that no state representative had the authority to suspend that liberty except “when the state is in real danger”—with the strong proviso that “it is not left to the executive power to determine when the danger of the state is so great, as to render this measure expedient.”21 With these declarations, published in 1765, Blackstone celebrated the success with which the English had worked out their constitutional arrangements, which he regarded as a unique historical outcome (and which allowed him to direct his condescension not at some distant inadequate other, such as the Chinese, but at a group closer to home, the French). Blackstone’s way of conceiving of the just limitations on the state to protect the body and preserve the personal liberty of the individual was not how Ming law understood the relationship between the individual (and his social relations) and the state, nor how Chinese saw themselves within the state context familiar to them. This difference in the structures of legal meaning does not entitle us, however, to go on and argue that Chinese political culture was indifferent to individuals, or that society-making was alien to the Chinese constitutions, or that China was a state stronger than society. These ideas were popularized by Karl Wittfogel, who adopted the notion from Paul Milukov’s study of imperial Russia. Milukov published his book in Leipzig in 1898 at a time when pressure from elements of the Russian elite to dismantle that regime was growing and the impulse to frame the struggle in stark ideological language appealing.22 The Milukovian imaginary of a state stronger than society expressed the repressive character of late tsarist rule, and it was taken up by Cold War sociology to parse the totalitarian regimes of the mid-twentieth century. Whatever sense the argument made at the time, it says more about the gross abuse of the principles of law in the context for which it was formulated than it could possibly contribute to an analysis of the relationship between state and social power in Ming China. Rather than use as a baseline the concentration of state power Wittfogel feared in his own time and attributed to Ming China, a more productive theoretical approach grasps the fact that state-and society-making occurred in overlapping and interactive ways. A generation ago, Michael Mann argued in favour of getting away from the habit of reifying state and society as unitary and equal entities polarized in theoretical combat with each other. He advocated instead that we approach society as “a diversity of intersecting networks of social interaction,” of which the state was but one “interaction network” among many, and one to which fewer social actors are tied. Mann’s approach endows society with greater conceptual breadth and analytical power than the state,

Introduction

9

integrating the state more fully with society than earlier theorists had done. Though tangled in regrettable Eurocentrisms,23 he points us in the same direction as Unger’s shift from state-making to society-making as a more complete basis for building social theory. This parallels the theoretical moves I have found myself making over the course of interpreting the rich record of state and social life in Ming China. I have found state networks projecting imperial authority more deeply into society than was the case in Renaissance Europe, yet I have also detected a strong counter-flow of influence working upward in capillary fashion from social networks below that placed the state in a posture more reactive than formative in making Ming China what it was. This perspective was still inchoate in 1984 when I wrote the essay on subcounty administration that, in revised form, opens this volume. In the original version, my attention was divided between tracking the power of society to channel state intervention, and the capacity of the state to remake society, unable to decide between them and neglecting the extent to which both were occurring simultaneously and interactively The same duality of analysis ran through the original version of Chapter 4 on the spread of rice agriculture into north China, written in 1980, in which I found that the work of active dissemination promoted by state representatives succeeded only when linked to existing social, consumption, and economic networks—and failed when it did not. The simultaneity of state-making and society-making perspectives is most explicit in an essay I wrote in 1992 on the resilience of Buddhist monastic institutions in the face of state suppression, for which reason it appears here, as Chapter 7, largely unaltered from the original version. I have grouped the eight chapters in this book into four pairs based on the thing, activity, or social group over which state and social networks competed to assert control. In each case, it is tempting to view state-society interactions from the privileged position of state systems and state actors—to see society as reactive to the state—which is how most of the sources represent them. The import of each of the chapters, however, is to show the extent to which social networks obliged the state to adapt to local practices and institutions: in effect, domesticating the state into becoming yet one more resource which local actors could exploit in their competition for power. The first pair, “Space,” begins with a study of subcounty administrative systems the new dynasty imposed at the beginning. These systems were ambi-tiously designed to locate every community and household in Ming China and make them components of larger entities, yet they largely followed the boundaries that social and economic practices had already drawn. In Chapter 2, similarly, the county survey maps that Magistrate Ye Chunji produced could be taken as demonstrating the state’s capacity to configure space and thereby profitably control local social systems; yet they can also be seen as attempts to bring state vision into line with the socially visible. Ye drew his maps in relation to the larger process known as the Single Whip reforms, by which the state was obliged to revolutionize its fiscal systems in order to adapt to an economy that silver and commodity production had transformed away from the rural simplicity the founding emperor imagined when he imposed his systems. In both these chapters, the state’s organization of territory assisted in the formation of a distinctive Ming hegemony, arising from negotiations between state representatives and those for whom local social practices determined how they organized their lives.

The chinese state in ming society

10

If this all seems sensible, it wasn’t necessarily so to our Hegelian forebears, for whom China suffered a terrifying absence of private property and a vulnerability of individuals to arbitrary state power. Hegel is not the only ghost in the house. From a different perspective, Emile Durkheim worried most of a century later that the structure of rural life resulted not from the state machine, but from a mechanical spontaneity of communal organization that denied the independence and intimacy that for him constituted individual life.24 This fear of the local community was, however, a variation on the same anxiety that moved Hegel and Wittfogel: that the individual, and by extension his capacity to act autonomously, in China was profoundly compromised, whether by society or the state. In fact, the dialogue between state and society and the realm of possible social action that this dialogue in part defined, was far more complex and fluid than earlier observers of China were able to discern. The second pair of chapters, “Fields,” continues to consider the problem of the accessibility of land to state scrutiny, but in relation to the growing of rice. The two chapters are concerned with the dialogic interplay between state and society in two very different settings: the registration of rice-growing polders as fiscal units on the Yangzi Delta in Chapter 3, and the introduction of rice agriculture on the North China Plain in Chapter 4. Rice paddies posed particular challenges on state officials concerned to collect taxes but also to direct investment in agricultural infrastructure. The two cases diverge sharply in terms of how state systems interacted with local economies and societies between south and north. In the south, where conditions naturally favoured riziculture, the state’s challenge was to work effectively with local community leaders, who were better suited to managing local needs and resources in densely structured economies than were outside officials. In the north, the natural environment could sustain rice agriculture only with investments on a scale that was beyond the means of most communities. As Philip Huang has phrased the contrast, in Jiangnan the elite could intervene to ensure its access to levies on land and labour, even to take leadership in crisis situations when elites failed to act. These interventions produced a “frequent three-way interaction among state, elite, and peasant.” In north China, on the other hand, waterworks maintenance was on a scale that favoured state management, since converting dry land to paddy fields was too big for individual households or villages to undertake. Those were the projects that in the Yangzi delta drew the elite, the peasant, and the state into necessarily complex and changing relationships of a sort that did not exist in North China.25 Awareness of this complexity was the mark of a capable bureaucrat, and the means by which the state might not lose all contact with society. The third pair of chapters, “Books,” examines the interaction of state and non-state institutions in terms of another portable commodity, a collectible manufacture rather than a consumable crop, one accessed not primarily through ownership but through education, inclination, and opportunity. The wide availability of books meant that readers could read what they chose, not merely what the state put on the curricula of those seeking public office. At moments of anxiety, the Chinese state has worried about this promiscuity. It could respond by going into publishing and distributing the texts it wanted people to read, or by removing those it disliked from circulation. The Ming did both, and came up

Introduction

11

against social networks that enabled it to distribute certain titles and ban others without, however, ever gaining full control of books—or, for that matter, even conceiving of this as a task appropriate to what the state should regularly do. Chapter 5, which looks at the building of libraries at government schools in the midMing, shows that both state and social networks produced and distributed the books that filled the cabinets in these libraries, but that the latter outdid the former in volume and persistence. The libraries ended up owning texts that were not those the state wanted its students to read, as well as those that it did. Chapter 6, which examines the other side of book control, censorship, deals more with the Qing than the Ming because of the richer documentation for the later period. Two extended examples from the Ming are included, and these indicate, to me at least, that the Ming was not an active censorship state. If the Qing became one, it did so intermittently as a result of particular anxieties of the eighteenth-century emperors. Their somewhat misnamed “literary inquisition” was both driven and confounded, however, by the activities of commercial publishers, whose output was too prolific for them to counter. Officials exploited their bureaucratic networks to destroy books, but the capillary capacities of the networks in which book publishers and collectors operated ensured a counter-influence that saved many of the marked titles. A similar dynamic between state and society operates in the next pair of studies, “Monasteries.” Chapter 7 looks at the stringent regulations that the Hongwu emperor introduced in the second half of his reign to reduce the size of the Buddhist establishment and its power in local society. Much of the new legislation was immediately jettisoned after his death in the face of locally rooted opposition; other aspects were later reworked through subsequent renegotiations between Buddhist monks and their patrons on the one hand and the state and its local representatives on the other. The actual unfolding of the suppression provides a good opportunity to interrogate the social processes through which people, whether officials or not, accumulated and dispersed power in the administrative presence of the state. The draconian controls this emperor imposed could be cited as evidence of the state’s capacity for despotism, yet it was a despotism shortlived. Soon enough, the state had to negotiate with local interests concerned to protect institutions and practices salient to their own strategies of local control. Public authority was not the state’s alone to create or command in the face of dense social networks, such as those which bound many Ming monasteries to patrons of wealth and power. The state was not opposed to the presence of Buddhist institutions in Ming society so long as they did not serve to facilitate the emergence of a competing power base. Conservative state representatives often suspected them of this dangerous capacity, yet as Chapter 8 on the recording of Buddhist monasteries in local gazetteers shows, the legal recognition that Buddhism and its institutions enjoyed constrained Buddhism’s moral and ideological foes from doing anything more radical than condemning or demoting them in print. As I argue in that chapter, Buddhism was sufficiently protected by both custom and constitution that earnest Confucians in the local gentry could not call for its eradication and be listened to by anyone but themselves. No matter how strongly the gentry played the state’s hand, community boundaries, landownership, consumption patterns, literacy, commercial production and exchange, and religious devotion—among other factors—determined local outcomes.

The chinese state in ming society

12

In all these chapters, there appear—if not directly in view of the state, then at the edge of its vision, variously in the form of amalgamated monasteries that did not close, banned books that survived, or crops other than rice being planted—spaces for action the state could not colonize. Seen from below, the state’s dominance of Ming society was never unitary, nor its reach unobstructed. State systems and social forces might be in opposition when revenue or security was at risk, but this opposition does not provide a full analysis of the Ming. Better that we begin with this simple pair of assumptions: that social networks shaped the ways in which most people lived their lives, and that state systems could influence these networks but not, other than at exceptional moments, remake them. Active resistance was an option at both levels, whether to stymie state imperatives or to disrupt social networks, as the frequent outbreaks of low-level violence in the Ming attests,26 but the mutual interaction of these networks of authority was more diffuse: percolation from above with no guarantee of where the moisture of state influence might reach, and capillary action from below with potentially no limit on the capacity of society to remake the state. When Emperor Hongzhi looked down through the state systems beneath him at the fight over grave land in Nanchang, he addressed it as a series of errors that participants in state systems had made. As arbiter of those systems, he had to intervene to get them operating properly again. Where Zhang Yingqi’s ancestor got buried was not the emperor’s particular concern: let propriety follow property. Zhang saw it otherwise: abuse of state systems was not of interest except insofar as he could turn it to his advantage; what he cared about was the property in which he had buried his ancestor’s bones, and the social alignments or cohesions that that burial had served to express and reinforce. Burial was formative in the social processes of acknowledging relations to ancestors and therefore relations to living people, so getting the bones in the right piece of ground was critical to the process of individual and household reproduction. But the ground mattered as much as the bones in the everyday life of ordinary people, and that was what Wang Zhen cared about. The grave was on his property, and property was what the Chinese state, for all the exciting flamboyance and peculiarity of imperial interference, was committed to protect. The Ming regime, like any other, was rooted in the preservation of the inequalities of property and the hierarchies of social status among people who used property rights to reconsolidate their positions with every shift of power. The state nodded to the importance of the proprieties that ritual acted out, but for the most part it addressed property rights primarily, on the understanding that the satisfaction of ritual claims would follow. Once the Hongzhi emperor handed out his punishments, Zhang Yingqi’s attempt to steal Wang Zhen’s land disappeared from his sight, never to trouble him again. The conflict between these two men was not necessarily solved, however. Zhang may have carried on his battle by burrowing along other networks and dragging to his support other institutions to get him at last what he wanted. Wang for his part may have worried that he could not rest content with the emperor’s judgment, and that Zhang would still come back at him, obliging him to search among the same local networks and institutions for protection against the next attack. We have no way of knowing. They did not come to the attention of an emperor or a court historian again, nor did they rise to the notice of the local compilers of Nanchang gazetteers.

Introduction

13

Their heirs might still be at it, for all this Ming historian knows. Enormous political and economic changes have intervened in the last five centuries to alter the patterns of state-society relations Chinese find familiar. Yet it is not impossible that the conditions that led then to an eruption of intracommunity conflict—scarce resources, intense social competition, and a differentially responsive political system—are still present, albeit within vastly different political and economic circumstances, and that the approaches that the actors in this little drama took in 1499 might not be unfamiliar to Chinese caught in conflicts today. Within the particular mesh of state-in-society in China, their ghosts are still in the house. Best that ours not be, lest theirs invite ours in and their world begin to look suspiciously like our own. A note on sources The reports that court historians included in the Veritable Records, important as they are for glimpsing the concerns and controversies that swung into the view of the court, reveal only what an emperor heard, said, or chose to do. These voluminous annals are thus one chronicle in which the interventions of the state at the local level can be detected, though often what they record is limited to what state administrators wanted to do or needed to know. As such, these annals are not the best repository of primary materials for exploring how Ming people experienced the presence of the state in their local communities. They see the social from too great a distance, and mostly from the state’s point of view. Still, without the central state keeping occasional track of such matters as, for example, requests for copies of books published by the court printers (as we will note in Chapter 5), we would have no record that such request even went up to Beijing. More detailed, and more useful for analyzing most aspects of state-society relations in the period, are local gazetteers. These compendia of geographical, administrative, and biographical information were mostly produced under the supervision of the local magistrate. Though not official state publications, county and prefectural gazetteers were expected to chronicle the presence of officially mandated state institutions and those who interacted with them, and state representatives were the final arbiters of what their compilers included. As such, gazetteers reveal much about local reactions to and implementation of central policies coming down from the court. They also provide windows onto social and cultural changes flowing through society at a level beneath the operation of state systems. With the limited exception of Chapter 6, local gazetteers constitute the source base for the research in this volume. I think of gazetteers as particularly a Ming source. Although the genre did not begin with the Ming,27 this is the period when it was busily in formation, with the would-be compiler reading any and every gazetteer he could lay his hands on in order to refine and improve the principles of organization governing his own work.28 This process continued in the Qing, as we will see in Chapter 8 with the controversies surrounding the county gazetteer Lu Longqi produced in 1686. If the genre feels like a Ming production, it is also because this is the period when gazetteers began to be published on a scale that enabled copies to survive in large numbers. Close to 7,000 Ming and Qing gazetteers are extant. Most of these are of administratively defined units such as counties or prefectures, though close to a thousand were produced for institutional sites such as monasteries and

The chinese state in ming society

14

topographical sites such as mountains. Roughly 300 are cited in this book; these appear in a separate bibliography. Gazetteers are ideal for writing local history, which is how I use them in Chapter 2 to write a brief history of Huian county, Fujian. The consistency of format, however, gives their data a comparability across place and time that permits their use to detect countrywide trends, which is how I use them in Chapters 1, 5, and 7. Between these extremes of scale lie the regionally based studies in Chapters 3, 4, and 8, the first of which draws on gazetteers of the two provinces between which the Yangzi Delta was divided, and the latter two of which were written on the basis of the surviving gazetteers of North Zhili. As I have used them, so these gazetteers have used me. By virtue of what they are, what they include, and what interests those who compiled them had, gazetteers have played a large role in determining the historical topics I have chosen, the problems I have set myself, and the approaches I have taken. It is fair to say that this volume would not have been about the Chinese state in Ming society had these books not been about just that.

Part I Space

1 The spatial organization of subcounty administration* State administration in Ming China extended downward in an organizational pyramid whose apex was the court and whose base was all households. The connecting middle level was the county, where the state appointed a magistrate to administer local affairs on its behalf.1 The household-county relationship was not unmediated, however, for between these levels the early Ming state elaborated a complex structure of administrative units. This structure consisted of four distinct but interrelated systems, based on earlier precedents but compounded with new elements that the Ming brought into use. The first of these systems subdivided the territorial space of the county into a hierarchy of cantons, townships, and wards. The second system mapped the social terrain by grouping households for census and fiscal purposes into the lijia or hundred-and-tithing system. The third was the baojia or neighbourhood mutual-watch program. Some areas had a fourth system, the xiangyue or rural covenants. These hierarchies stood in parallel with each other and often overlapped, the boundaries of one set of units replicating those of the others. This replication contributed to making these units into a robust and integrated structure of civil administration. The structure continued, with modifications, into the twentieth century, and not all the old boundaries have been lost, even today. Together, these subcounty systems constituted a pyramid of stepped jurisdictions that funnelled resources to the centre and maintained security and surveillance over the people. These systems not only made local administration possible, but endowed the state with a measure of access and efficiency earlier dynasties could not match. Keeping these subcounty units in order was recognized to be a component of good administration,2 not least because they provided the spatial template for tax collection. The logic of earlier studies of local administration has been to distinguish stateimposed mappings from the “actual” boundaries of village society, to speak of “artificial” versus “natural” communities, of “administrative villages” versus “real, historical, social villages.”3 While the notion of a tension between state and society is intuitively appealing, opposing the artificial/administrative to the natural/social can have the effect of masking their interaction. Administrative organization did in fact have its impact on the shape of local communities, an impact that goes back several millennia. Given that the Chinese state was often successful in impressing its patterns on local communities, one is hard-pressed to identify “natural” peasant communities that have not been in some way absorbed into official hierarchies or at least tagged with the names they were required to use. On the other hand, the influence went the other way as well. Social entities consistently pressed upward and moulded state systems, even becoming them. The “community” (she) in the Yuan dynasty, for example, was recognized as an

The chinese state in ming society

18

administrative unit in the Ming, and the Ming “village” (cun) became an official unit in the Qing. When the state recast its administrative systems, it usually took care not to alter too radically what was in place. Drawing new geographical boundaries, even down at the village level, was a hazardous move.4 The cumulative inertia of administrative systems, units, and boundaries resulted in a “close and continuing interaction between decimal hierarchies and the natural divisions of Chinese society such as village, intervillage association, lineage, and market community,” as Philip Kuhn has noted of a later period.5 Social collectivities shaped administrative boundaries as much as, if not more than, they were shaped by them. The Ming founder, Emperor Hongwu (r.1368–98), consciously designed the lijia system to be an exception to this general rule of conformity to prior spatial organization, though there is little evidence that its implementation at the local level brought about the uniformity he envisioned. For the other subcounty systems, there is even less evidence that “natural” communities were squeezed into “artificial” ones. What at first looks like state imposition, when viewed close up, seems more a matter of the state formalizing the informal and deeming what was already there to be what should be there, allowing the inertia of reality to overwhelm the force of the ideal. However much the Hongwu emperor desired to use his subcounty administrative systems to disrupt old patterns and impose new, the bureaucracy could not afford to push aside the existing social networks and rebuild them into an entirely new structure. The main challenges to the system from the fifteenth century forward were in any case commercialization and urbanization, not an interventionist state, but their influences also tended to run along channels already established by long-standing administrative practices. The essayist Xu Yikui (1318–c.1400) eulogized the administrative work of the Ming founder, his contemporary by praising the setting up of the field administration that passed from the capital through the counties down to the locale and the creation of the lijia system. As Xu phrased it, “once he had achieved great stability throughout the realm, he ordered prefectures and counties [be set up] and established the localregistration system.”6 These two systems together cast a comprehensive net of administration over the entire country; both also implied an interlocking set of spatial units for organizing territory into a virtual map that made the realm accessible to the state and organized the flow of information, resources, and personnel that sustained it. This study focuses principally on these two universal systems, and secondarily on the neighbourhood watch and rural covenant systems, which were implemented regionally in the second half of the dynasty. Ming administrators recognized these various systems as interrelated but typologically distinct.7 Each of these systems had its own principles of spatial organization, though their units’ boundaries tended to coincide such that units at the same level in different systems became spatially indistinguishable. The net effect of these interrelationships was a high degree of integration at the local level. This integration may have facilitated governance for magistrates, but it has made it difficult for historians to see what was going on behind the labels. Our first order of business is to sort these out. In addition to the confusing overlapping of systems in a local area, Ming sources show diversity of terminology across regions. The second order of business is then to show that Ming administrative units, despite local variations, adhered to an identifiable template working within a reasonably consistent administrative vocabulary.

The spatial organization of subcounty

19

To achieve these goals, I have set aside most of the prescriptive regulations the state published and relied instead on records of actual administration in local gazetteers. Most of the data comes from subsections in the chapters on administration or taxation, variously entitled lijia, xiangli, xiangdu, xiangyue, fangli, or baoli. Most gazetteers were compiled under the name of the local magistrate, and in many cases the magistrate actually took an active role in editing the published version. Their orientation toward social reality involved a mixture of the concrete and the abstract: magistrates relied on local practices while compiling them, yet they were also obliged to shape that data so that county systems would appear to reflect state policy. This double orientation may have introduced some distortion into this study, since a desire to conform to official patterns could induce compilers to observe in their local administrative systems greater formal order and regularity than actually obtained in practice. But that in itself is evidence that state officials were effective in making these systems authoritative, even if administrative practice strayed from the official models. The subcounty administrative system At the end of its third year in existence, the Ming state claimed jurisdiction over 887 counties; by the end of the dynasty, that number had risen to 1,159.8 To an official serving in the central government, these constituted the lowest level in the state bureaucracy. Bureaucratic appointments went down only as far as these counties, and county magistrates alone were answerable to the central government for the implementation of policy at that level. The magistrate in his county seat was in turn perched on the tip of one of anywhere from 887 to 1,159 icebergs of local administration that, from higher governmental levels, were largely submerged from view, but were what he surveyed when he looked out around him. Not all counties were subdivided into the same set of levels and units. The standard structure (see the list below), common in central and south China, had three subcounty levels: cantons, townships, and wards. The abbreviated structure, common in north China and other areas of lower population density, had only two levels: cantons and wards. Certain densely populated parts of central China used an elaborated structure: a subcanton was added between the canton and the township resulting in a four-tiered arrangement. Terminology varied between structures and regions, especially central and south China on the one hand and the north on the other. The units I have identified as “canton” (xiang) and “subcanton” (li) have ancient pedigrees within Chinese administrative practice.9 This usage became so well established that by the Song the collocation xiangli came to mean “one’s native rural area.”10 I use “township” to translate du in rural areas, fang and yu in urban areas, and xianq (xiang in the fourth tone, to distinguish it from xiang in the first tone, meaning “canton”) in suburban areas. At the next level down, which I designate as “ward,” terminology divided between north and south: she and tun used in north China, tu used in the south. The subsections that follow examine each of these units consecutively from higher to lower levels.

The chinese state in ming society

20

Subcounty administrative system Standard structure county (xian) canton (xiang) township (du) ward (tu) widely found in Fujian, Guangdong, Jiangxi, and Shanxi Abbreviated structure county (xian) canton (xiang) ward (tu, she, tun) widely found in Guangxi, Henan, Huguang, North Zhili, Shaanxi, and Shandong Elaborated structure county (xian) canton (xiang) subcanton (li) township (du) ward (tu) commonly found in South Zhili and Zhejiang

Canton The Ming inherited named cantons (xiang)11 directly from Song and Yuan practice. This was the largest territorial subunit within the county. Aside from a few places in central and southeast China where xiang were dropped from official usage at the beginning of the Ming (which is why the final column for Zhangzhou prefecture in Table 1.2 is blank),12 this unit was universal. A county could have as many as twenty cantons, though the average was about eight. In the Tang, the xiang had been a demographic rather than territorial unit, rated at 500 households,13 but that standard collapsed over the Song and Yuan, when the number of xiang per county declined and forced cantonal populations up.14 By the Ming, cantons were strictly territorial, though they were still expected to have roughly equal populations (as they do in the case shown in Table 1.1). Between counties, the size of cantons varied widely. Table 1.2 shows a range from under 800 fiscal households (Changsha) to over 15,000 (Songjiang). Urban areas were usually left outside cantonal jurisdictions. The cantons were the countryside, hence the use of terms

The spatial organization of subcounty

21

Table 1.1 Registered population by canton of Anqiu county, Shandong, 1589 Canton

Urban area

Number of households

Registered population

Number of wards

Average population per ward

788

2,297

5

459

Wenshui xiang

2,630

6,608

24

275

Linhuai xiang

2,867

6,823

26

262

Anlie xiang

3,121

7,497

29

259

Renshun xiang

3,380

7,609

30

254

Paoquan xiang

3,420

8,510

31

275

Guangzong xiang

2,968

7,423

29

256

19, 174

46, 767

174

269

3,064

7,412

28

Totals Average per rural canton

Source: Anqiu xianzhi (1589), 8.55b–57a. Note: The gazetteer compiler warns the reader that the records have lost track of real population; these numbers are at least a century out of date, more likely two. Contrary to the impression given by the data in the flrst two columns, Shandong in the late Ming was said to have large households: one source suggests eight as the average number of members per household.

like xiangmin (“people of the cantons”) for country folk and xiangsu (“customs of the cantons”) for rural practices and attitudes. Cantons were mostly vestigial markers. Their only official use was to provide boundaries when new counties were formed,15 although cantons often served as the first unit of fiscal record below the county. In one Fujian county, we know that tax registers were kept by cantonal clerks (xiangshu), since they were discovered falsifying them.16 Cantons might also be used informally to coordinate administrative tasks, such as forming militia17 or organizing irrigation work.18 Cantons could be meaningful social units as well. In the gazetteer he wrote for Shouning county (Fujian) in 1637, for instance, story-writer Feng Menglong (1574–1646) notes that the annual mid-autumn parades for the leading local deity, Goddess Ma, were organized by canton, for he reports that “each canton has its own sheshou” or head of the local religious procession association (yingxian she) in charge of overseeing the event.19 Subcantons The subcanton (li), sometimes found between the canton and the township in the elaborated subcounty system, was even more vestigial than the canton. Since the Qin, but consistently from the early Tang, the li was the chief subdivision of the xiang, but it began to lose importance in the Southern Song, when in some areas the number of li per xiang sank to one.20 As lower-level units assumed greater administrative importance, the

The chinese state in ming society

22

li was dropped in favour of the township at the next level down, or pushed up to the canton level, with which it merged.21 As a result, some cantons came to be known by the old li names.22 In some areas of Jiangnan and the southeast, the xiang and li units

Table 1.2 Households per subcounty unit in five prefectures, 1492–1612 Prefecture

Date of Registered Households Households Households Households per per canton statistics households per ward per township subcanton

Zhangzhou (Fujian)

1612

34,917

139.6

529.0

N/A

Ghangsha (Huguang)

1532

63,801

191.6

N/A

N/A

787.7

Jianning (Fujian)

1492

124,932

134.7

886.0

N/A

3,203.4

Raozhou (Jiangxi)

1502

162,074

141.2

613.9

N/A

2,532.4

Songjiang (S.Zhili)

1512

203,826

144.0

1,772.4

4,076.5

15,678.9

Suzhou (S.Zhili)

1506

582,000

147.0

2,425.0

N/A

7,864.9

Sources: Zhangzhou fuzhi (1613), 8.15b–16b; Changsha fuzhi (1532), 3.3a–23a; Jianning fuzhi (1473), 7.4a–26b, 9.2b–6b; Raozhou fuzhi (1511), 1.7b–24a; Songjiang fuzhi (1512), 9.14b–119b; Gusu zhi (1506), 12.1 a–2a, 18.1 a–25b; Da Ming yitong zhi, 8.1 b–2a.

continued to be distinguished as they had in the Song, but both were increasingly regarded as vestiges. A sixteenth-century text from southeastern Zhejiang thus contrasts “the old names of canton and subcanton” and “the townships and wards in current use.”23 Other Zhejiang and South Zhili gazetteers also distinguish “the old li” as a redundant category, in contrast to “the current townships.”24 Subcantons became important only in the absence of cantons; that is, when they took the place of cantons.25 Township The township (du) was the middle level in the standard three-level structure of the subcounty system.26 The du came into administrative use in the 1070s as a part of Wang Anshi’s baojia system.27 By the twelfth century it had been integrated into the xiang-li system in south China, frequently replacing the li as the chief subunit within the canton.28 Many local gazetteers do not record the use of du, however, until the Yuan.29 By the Ming, the township was to be found throughout central and south China, where it served as the main unit in the subcounty administration. The township appeared only rarely in the north, and, where it did, often existed only on paper.30 Townships were numbered rather than named, following Yuan practice.31 Figures 1.1–1.3 are rare examples of maps on which numbered townships have been marked.

The spatial organization of subcounty

23

Figure 1.1 Administrative map of Wujiang county, Suzhou, prepared for the county gazetteer in 1548. The numbered du throughout the countryside are the rural townships. The numbering of urban townships (here called bao) within the county seat in the upper left is unusual. Source: Wujiang xianzhi (1561), zongtu 1b–2a. A canton could have anywhere from one to a dozen townships (the average is about three), a county anywhere from ten to eighty (the average might be about forty). This variation is conveyed in Tables 1.3 and 1.4. Turning back to Table 1.2 we see that there could be from over 500 to close to 2,000 fiscal households per township. The three lower figures—between 500 and 900—describe the more usual range, suggesting a population per township on the order of 4,000 to 5,000 people. In areas that experienced little population growth after the Song and in which townships were not regularly subdivided into wards, the population of a township could dip well below a thousand (see Table 1.5).32 The township was a significant administrative unit right from the start of the Ming, when the expression xiangdu began to compete with the older xiangli to mean “the countryside.”33 This term came into wide use in the south,34 though it can even be found in North Zhili in the late Ming, where townships rarely appeared, and

The chinese state in ming society

24

Figure 1.2 Administrative map of Taicang subprefecture, showing townships numbered from 1 to 29. Source: From the 1629 reprint of Taicang zhi (1548), tukao 3b–4a.

Note: It is even more unusual for township boundaries to be marked on county maps.

The spatial organization of subcounty

25

Figure 1.3 Map of coastal defences of Jiading county, printed in the 1558 gazetteer. Townships numbers appear in ovals, larger settlement names in rectangles. This map is unusually oriented with southeast at the top; the Yangzi Estuary on the left side is labelled “the great sea” (dahai). Source: Jiading xianzhi (1558), courtesy of the Nanjing University Library.

The chinese state in ming society

26

Table 1.3 Distribution of subcounty administrative units in a standard structure: Raozhou prefecture (Jiangxi), 1502 County

Poyang Yugan Leping Fuliang

Dexing Anren

Totals

Cantons (xiang)

20

13

11

10

6

4

64

Townships (du)

70

36

42

56

36

24

264

337

208

295

102

118

78

1,138

Registered households

47,289

30,182

41,560

17,660

11,891

13,492

162,074

Households per ward

140.3

145.1

140.9

173.1

100.8

173.0

142.4

Wards (tu)

wards per township

Averages:

4

wards per canton

18

wards per county

190

townships per county

44

cantons per county

11

Sources: Raozhou fuzhi (1511), 1.7b–24a; Da Ming yitong zhi, 50. 1b–2b.

Table 1.4 Distribution of subcounty administrative units in an elaborated structure: Taizhou prefecture (Zhejiang), mid-seventeenth century County

Linhai Huangyan Taiping Ninghai Tiantai Xianju Totals

Urban wards (tu, fang)

24

26

13

11

8

10

92

Urban townships (yu)

7

6

2

2

2

2

21

125

61

55

92

48

?

?

Townships (du)

69

45

26

53

37

45

275

Subcantons (li)

42

28

16

19

12

22

139

Cantons (xiang)

15

95

6

4

6

45

Wards (tu)

Averages:

wards per county

93

rural wards per canton

10

rural wards per township townships per county

2 49

rural townships per canton

6

rural cantons per county

8

Source: Taizhou fuzhi (1722), 3.60b–71b.

The spatial organization of subcounty

27

Table 1.5 The subcounty system in Huian county, 1573 Canton (xiang)

Subcanton name (li)

Township (du) number

Ward (tu) number

Households (hu)

Population (kou)

Cultivated land (mu)

Urban

1

(fang)

2

139

916

5,486

3

120

689

8,155

1

133

821

6,175

2

125

902

4,425

Chongde

5

125

661

10,163

Xiangfu

28

133

1,020

5,702

29

139

1,050

5,981

30

157

1,020

4,105

31

147

982

2,571

Wenling

32

125

1,054

2,283

Anren

33

124

907

7,265

1

140

1,025

6,628

2

151

960

5,465

140

772

4,815

21

124

1,310

6,160

22

147

1,248

2,671

23

143

1,359

2,093

24

128

1,162

4,430

25

128

1,160

4,589

26

135

1,201

5,423

133

647

8,321

210

696

8,426

172

1,088

3,937

Wenzhi

Pingkang

Yanshou

34 Xingman

Anmin

19

Chang’an

20

Taikang

Shoujie

27 Zhongshu

Deyin

6 7

1 2

8

1 2

The chinese state in ming society

Guangde

28

9

124

879

2,435

10

152

987

5,950

157

860

13,105

137

702

10,240

144

782

11,293

130

954

3,483

4,062

28,864

171,775

140

995

5,923

Daixian

11

Minsu

12 13

Xinyi

Guihua

3 4

Zunxian

14 15

Tongxin

16 17

Lidian Totals Averages per ward

18/16

18 34/27

40/29

Note: A blank in the three right-hand columns indicates that the subcanton, township, or ward on that line had been in use earlier in the Ming, but was not as of 1573. Split totals give the formal number of units first, then the actual number. Sources: Ye Chunji, Huian zhengshu, 4.2b–6.31a; Huian xianzhi (1936), 1.20b.

never as du.35 Townships were sufficiently in common parlance that the late-Ming traveller Xu Hongzu used them (“township number x of such-and-such a county”) in his diary to indicate where he was on his journey.36 The township was the principal unit by which a county magistrate organized the administration of taxes. When National Academy students were dispatched in 1387 to survey agricultural land and record the results in what came to be called Fish-Scale Registers, so named because of the visual appearance of the maps that appeared at the front of each, they did so on the basis of townships.37 The practice of organizing county fiscal records by township is reflected in the numbering system for registering land. Plots of land within a township had registration numbers or characters in a series that did not repeat within the township; the series then began anew in the next township.38 According to a description of the registration system in Haiyan county, Zhejiang, by the scholar Wang Wenlu (1503–86), the registration numbers assigned to cultivated land within the county were grouped by township, not only to simplify bookkeeping in the magistrate’s office but also to identify (and, it was hoped, discourage) people who bought land outside their own townships as absentee landlords. According to the procedure in Haiyan, someone buying land within his own township could have the registration for the land shifted from the vendor’s household to his own, thereby eliminating ambiguities of ownership. Buying outside the township in which he resided meant that he had to register himself as a dependant under the vendor’s household. The lack of clear title implied by

The spatial organization of subcounty

29

this registration status would have made some landlords think twice before extending their holdings beyond their home township. Of course, Wang adds, this rule was constantly flouted.39 The dynastic founder identified the township as the appropriate unit for setting up community granaries (shecang) in central and south China, and some counties did indeed build them at this rate.40 He also required that every township construct altars for conducting state-cult sacrifices.41 In the heavily irrigated Yangzi Delta, townships could serve as units for coordinating water control: when the governor of South Zhili at the end of the fourteenth century issued regulations for the use of treadle pumps, he required that village officers make them available for emergencies anywhere within their township.42 Landlords within a township might similarly rally together on the basis of their common residence. In Cixi county to the west of the city of Ningbo, for example, “the rural people of Township 5” jointly protested in 1587 against the occupation of land along the edge of a lake, which was being turned into paddy fields and adversely affecting their water supply.43 Lest the township become a sphere of mobilization to engage in more than local infrastructural projects, early-Ming regulations stipulated that village officers not go outside their home townships. The power of townships to shape local elite identity is suggested in Yongan county, Fujian. In 1403, the forty-four townships established in the Yuan period were reduced to thirty-five. The nine discontinued townships, all in peripheral hilly areas, were merged with other townships within their cantons. Sixty-seven years later, however, two mountain townships were reinstated in response to local demand. It seems that elites in the hills found it advantageous to resurrect their earlier administrative status in order to remain distinct from lowland elites, rather than be subordinated under them as junior partners.44 Ward The ward was the smallest unit and lowest level in the subcounty administrative system.45 In central and south China, where townships were in use, the ward was known as a tu.46 In the north, where the subcounty structure was usually abbreviated, it could be called either a she or a tun.47 Because the boundaries of the ward were identical with the lijia system (see below), wards in the north and the south were often referred to colloquially as li, here meaning “hundreds,” but hundreds and wards were components of separate systems, a distinction contemporaries recognized and gazetteer editors followed, at least through the sixteenth century. The standard term for ward, tu, meaning “map,” was adopted for administrative units in the Southern Song from the practice of mapping all taxable land and inserting these maps in the land register (tuce) for each du or bao.48 Thus one could speak of being in such-and-such a register as a way of indicating spatial location. The character is often written without the enclosing three-stroke box radical, as a way of distinguishing this usage from that for map.49 Consistent with its Southern Song origins, tu was not used in the north.50 There, she denoted townships populated by indigenous residents, and tun, those formed after the Hongwu era in areas where government migrants had settled.51 She had been the major administrative unit in north China in the Yuan, nominally flfty households though not held strictly to that number.52 Tun had been a unit for military and

The chinese state in ming society

30

civilian agricultural colonies since at least the Han. By the late Ming, the origin of the she/tun distinction was understood but significant only in one way: urban wards were never known as tun, always as she (e.g., fangshi she, “urban ward”), whereas rural wards could go by either designation. A county could have as few as a dozen wards, but most had several hundred. Exceptional in size are the counties of Songjiang prefecture: Huating in the mid-Ming had 801 wards and Shanghai had 614.53 The variable determining the original number of wards in a county was of course population, since ward boundaries followed lijia boundaries. The grouping of wards into higher units, however, followed no set pattern. A canton might have ten, twenty, or more, although some cantons in sparsely populated areas in north China had only one.54 So too at the next level down, the number of wards per township varied considerably. In Fujian and Huguang provinces, many had only a single ward,55 whereas some prefectures in South Zhili had close to twenty wards per township.56 Three or four per township was more the norm. Since the ward was contiguous with the hundreds in the lijia system north and south, it should have had something over the mandated size in that system of 110 households, yielding a population approaching a thousand.57 The averages per ward in Table 1.2, with the exception of Changsha (its figure of 192 may have been due to the frontier demography of Huguang), confirm this, showing only minor variation between 135 households in Jianning and 147 in Suzhou. The one case for which we have reasonably good census data from the late-sixteenth century, Huian, Fujian (the subject of Chapter 2), shows a range of population figures between 647 and 1,359 people per ward. The same county shows that wards, at least in coastal Fujian, could embrace from 2,000 to over 13,000 mu of agricultural land (see Table 1.5). In coastal Zhejiang, on the other hand, Haiyan county in the same period had an average of just under 1,600 mu.58 At less than 15 mu per household, compared with 42 mu in Huian, Haiyan peasants were living on the border of subsistence.59 Fish-Scale Registers for two wards in Changzhou county, Suzhou, show larger acreages per ward of 2,868 mu and 3,000 mu respectively.60 Roughly, then, we may think of a ward in south China as having a population of about a thousand people and a cultivated area of several thousand mu. Most counties had fewer wards at the end of the dynasty than at the beginning. The equivalence between the ward and the lijia hundred meant that a decrease in hundreds due to registration evasion from the mid-fifteenth century forward produced a decrease in wards. There is practically no county that did not have its wards reorganized at least once.61 Since wards were usually numbered consecutively within townships and not renumbered when changes were made, it is possible to reconstruct the process by which wards were combined in many counties. The reduction of hundreds could, however, throw the alignment of hundreds and wards out of whack, as appears to be the case in Linzhang county, Henan, which at the turn of the sixteenth century had seventeen hundreds, on the one hand, but twenty-two wards (eight she and fourteen tun) on the other.62 The ward was the lowest-level unit for registering land. Each ward had its own FishScale Register, in which were recorded all parcels of land within the ward, identified by sequential registration numbers. The landownership certificates and land tax receipts issued by the county office followed the same system of identification by township, ward, and registration number. As the author of a seventeenth-century handbook for county

The spatial organization of subcounty

31

magistrates advised, the best way to prevent fraudulent land claims was to “compile registers of the land owned by people in such-and-such a township and such-and-such a ward.”63 A late-sixteenth-century magistrate in Beijing observed, however, that “getting all the land in the county into all the wards of the county” had become the administrator’s impossible dream.64 As well as organizing land, the ward served as the basic unit for locating households.65 By the mid-Ming, when corvée labour was levied, this was done by ward; and when the lijia records were recompiled, it was by ward rather than hundred.66 Urban and suburban units The subcounty administrative system extended to urban as well as rural areas. The hierarchy of units in urban areas tended to be simpler, with at most two levels. The canton, characterized as a rural unit, was usually absent.67 Urban townships could simply be numbered consecutively as one or more du within a county,68 though the more common practice was to set them apart by separate numbering and different terminology. They were usually called fang, a traditional term for urban units in China since at least the sixth century.69 A common alternative term was yu or “quarter”; in the southeast, towns had been divided into four yu in the Song and Yuan, and the term was kept in the Ming.70 The joint expression fangyu was the colloquial term for “county town” in many parts of Ming China;71 another was fangxianq, combining the separate terms for urban and suburban townships respectively. A xianq was a township immediately outside the city walls.72 This term was used in south China in the latter part of the Southern Song, though it first appears in urban administrative contexts in the Tang.73 The term yu was used occasionally in central China for suburban areas, a usage one source dates to the Yuan.74 In principle, an urban or suburban township should have had the same level of registered population as a rural township. This is so in the case of the county seat of Jiaxing, Zhejiang, which in the mid-1580s was divided into nine townships (fang) and had a total registered population of 6,950 households.75 These numbers yield a pertownship population of 773 households, which falls roughly in the middle of the range of population we found for rural townships. But this statistic is naive. Official figures wildly undercounted actual residents, so the population of urban townships must have regularly exceeded a thousand households. The tendency for urban townships to have only one ward may signal the difficulty of maintaining an accurate record of urban residents. Few magistrates attempted to create new urban wards after the turn of the fifteenth century, so that as town populations expanded, so too did ward populations, rather than the number of wards. When they needed more units, south China administrators tended to acknowledge population growth by adding suburban rather than urban wards. In north China, where suburban sprawl was not sufficiently great to require the designation of suburban wards, new wards that on rare occasion had to be formed under the pressure of urban expansion in the north remained for the most part within city walls.76

The chinese state in ming society

32

The lijia system The lijia system located every civilian in a hierarchy of command extending from the court down to the household. Its initial declared purpose was to organize households for service levy The founding emperor held out the hope that these imposed communities might serve as the units within which other social practices could be organized, such as community religious rites and primary education, including training boys to memorize his Grand Pronouncements (Dagao).77 Formally installed throughout almost the entire realm in 1381,78 the system not only distributed fiscal obligations but, by grouping households into permanent registered units, made local society vulnerable, at least in theory, to comprehensive supervision. The decimal principle of lijia organization, which grouped households into interlocking sets of tens, stood in a two-millennia-long tradition of systems built from multiples of five and ten households.79 Wang Anshi’s baojia system in the Song may have been a prototype for Hongwu’s planners, though his own more immediate inspiration was the weisuo system of military guards, battalions, and companies, which he put in place in the early 1360s well before he started regimenting civilians into units of a similar, uniform size.80 Applying such military principles to civilians was natural to a ruler who desired his subjects to be as disciplined and diligent as his soldiers. Hundred and tithing For the two eponymous units of the lijia, English medieval history provides the approximate translations of “hundred” for li and “tithing” for jia.81 The tithing was defined as a group of ten fiscal households, among whom rotated the post of tithing head (jiashou). The larger hundred was a composite of ten jia plus an additional ten households known as hundred captains (lizhang), among whom the duty of leadership rotated, as it did for tithing heads. In urban areas, hundreds were called fang, in suburban areas, xianq, using the language of the subcounty system.82 The affairs of the hundreds were also overseen by men in less well-defined posts, called hundred elders (lilao). The li of the lijia is the same character as the li translated above as “subcanton.” They derive from the same origin, for a li in the Western Jin and Tang was set at 100 households. The Song and Yuan dynasties did not impose demographic limits on the li, allowing its size to grow as its importance in organizing rural society declined. When the Ming founder decreed that a li of 110 households be established, he was ignoring what li had in practice become and setting them back on their Tang foundations. In physical area, hundreds in practice were identical with wards, so that the ward usually served as the unit of fiscal record for lijia obligations. Jia is a term of less venerable ancestry. It first appeared in administrative usage in the eleventh century as a variant unit of ten to thirty households in the baojia system. It continued thereafter to be used intermittently as a small grouping of households, and by the thirteenth century was common in south China as the subdivision of a li or du.83 Although the hundred is the unit on which studies of the lijia system tend to focus, the tithing was its basic component. Contemporaries stressed the impossibility of working with hundreds if the component tithings were not firmly in place: “When taxable individuals flee, the household is burdened; when the households

The spatial organization of subcounty

33

flee, the tithing is burdened; when [the members of] the tithing flee, the hundred is burdened.”84 Tithings were numbered one to ten. Hundreds were also numbered, sometimes in one series for the whole county, more frequently within individual townships, like the wards whose jurisdiction they shared.85 Very occasionally, hundreds were named.86 According to figures published in 1461, China was divided into 64,854.5 lijia hundreds; a late 1660s source gives a total of 68,929.5.87 Unlike counties, which gently proliferated over the Ming as population and administrative complexity grew, hundreds slowly decreased in number, as households escaped tax registration and fiscal assessment shifted from households to land. The distribution of hundreds correlated at the start of the dynasty with population density. Accordingly, the crowded Yangzi Delta, split between southeastern South Zhili and northern Zhejiang, was the region having the greatest number of hundreds (see Table 1.6). The region with the lowest rate of hundreds per county was the far southwest, where Guizhou managed to average only two and a half hundreds per county. Between these two poles stretched a gradient of hundred-to-county densities decreasing concentrically outward from Jiangnan. Despite their 110-household definition, hundreds were not uniform in size. There could be fewer than the statutory 110, but usually there were far more. Extra households were included under the category of “attached households” (daiguanhu), and those too poor to qualify for lijia service were appended as “supernumerary households” (qilinghu). This is in part why wards averaged about 140 households.88 Hundredal jurisdictions were supposed to be resurveyed once a decade to maintain equity among them, and the results written down in what were called Yellow Registers. Every county magistrate was required to submit these registers for all the households in his jurisdiction when the decennial “great compiling” (dazao, short for da zaoce zhi nian, “great register-compiling year”) came round. These registers, compiled first in 1381, could be painstakingly detailed.89 In practice, though, officials did not forward the results of their surveys lest increases in population prompt officials higher in the system to raise county taxes. The state tacitly accepted this practice. Precise records were expected to be kept, yet the costs of regular reassessment were happily traded for the budget security of set quotas. Hundreds were seldom altered, their boundaries rarely redrawn, and their numbers almost never hiked up. Only when population decreased would local officials or elites actively seek to get the number of hundreds changed.90 After Nanjing’s population was halved by the removal of the court to Beijing, for instance, the prefect in 1437 could legitimately petition that the urban hundreds in the city be reconstituted and reduced.91 By the sixteenth century, the lijia system weakened as a method for organizing local social collectivities. Hundreds came eventually to be thought of as territorial rather than demographic units and were superseded in all but official documents by the ward (tu).92 (By the mid-Qing, southern Chinese knew the lijia by the mutated expression tujia.) Even in official discourse, hundreds and tithings became units of fiscal account bearing no relation to real population.93 Redefined by acreage, a tithing could cover from several hundred to several thousand mu, with the size decreasing over time.94 For example, in 1601 a tithing in Jiaxing was defined as 250 mu of cultivated land; by 1641, after several further fiscal reforms, it was fixed at a mere 120 mu.95

The chinese state in ming society

34

Table 1.6 Hundreds (li) per county, by province, 1461 Range of hundreds per county

Province

about 150

Zhejiang

148.9

102–118

South Zhili

117.9

Jiangxi

102.8

Fujian

69.4

Shandong

54.0

Guangdong

51.7

Shanxi

47.2

Henan

27.6

Huguang

26.3

Shaanxi

21.8

North Zhili

20.6

Guangxi

17.0

Sichuan

11.8

Yunnan

7.0

Guizhou

2.4

47–70

17–28

7–12 <3

Average number of hundreds per county

Sources: Da Ming yitong zhi, 1.5b–86.29b, reproduced in Liang Fangzhong, Zhongguo lidai hukou, pp.208–46.

Several factors encouraged the transformation of hundreds and tithings from demographic to territorial units. One was the inertia of existing boundaries, since individuals living within a hundred changed with birth and death while the hundred continued to exist and might better be measured by something, such as land, which did not change so radically or frequently as people. Without regular reassessments, there was no way to preserve the originally demographic character of these units. Another was administrative efficiency: it was simpler for a magistrate to use the same boundaries rather than redraw them as population grew and migration moved people around. This practice became the norm because of the government’s preference for fiscal quotas over actual assessments. Quotas made the upward revision of local figures unattractive, and the widespread evasion of registration after the 1381 census made realistic reenumeration impossible.96 The basic factor encouraging the territorialization of lijia units was the virtual identity of hundreds with wards combined with the superiority of the ward as a land-based fiscal unit. Hundreds existed as population entities that organized the levy of labour service, but the more important fiscal operation was the collection of land tax, and that was done by ward. Land contracts identify land only by subcounty units and never by lijia units. Even local officials by the sixteenth century had largely abandoned lijia terminology.97

The spatial organization of subcounty

35

Sector Overlapping the lijia system of tithings and hundreds was a system of tax captaincies (liangzhang). This system was installed only in South Zhili, Zhejiang, Fujian, Jiangxi, and Huguang.98 Although the lijia and the tax captaincies were initiated separately and not as linked components, they soon merged into a single structure of fiscal command and were recognized as such.” The spatial unit introduced by the tax captain system was the qu or “sector.” These sectors combined hundreds, the number depending on their cumulative tax assessment. Each sector was supposed to yield 10,000 shi in tax grain, though in practice collection levels were much lower and varied widely. For an earlyMing example, Shanghai county had 92 sectors and 620 hundreds, averaging 7 hundreds per sector, whereas Xiaoshan county (Shaoxing) had only 9 sectors and 149 hundreds, yielding an average of 17 hundreds per sector.100 A 1371 prototype of the tax captain and lijia systems directly paired these units with the subcounty administrative units: the tax captain’s jurisdiction was the township, and the hundred captain’s, the ward,101 but this neat one-to-one relationship was abandoned when the system was made universal in 1381. Sectors sometimes served as jurisdictions for other functions, such as bridge-building and temple repair.102 Having only cadastral definition, however, they did not become universal territorial units and tended to shrink in size.103 After the early Ming, they largely disappear outside of certain areas in Jiangnan, where they were revised in the sixteenth century to facilitate the supervision of waterworks maintenance, as we will see in Chapter 3. The baojia system The baojia was a mutual surety and village defence system built on the same decimal logic as the lijia. Unlike the lijia, the baojia was not put into effect on a nationwide scale. Regulations were issued centrally in 1548, but implementation was not mandatory: they were issued to guide local officials who chose to install baojia in their counties, whether to deal with such problems as coastal piracy along the southeast coast or banditry in the interior.104 Although the lijia might have handled such matters, by the sixteenth century it had become so fiscalized and abstract (“empty” is one term used)105 that it could not be guaranteed to access the real people needed to organize security By the end of the Ming, the baojia system had spread to half of China.106 Where baojia units were established, an attempt was made to have them more closely replicate the actual contours of local society than did lijia units. This was due in part to the preference for building up the decimal structure of the baojia from the single-hearth “family” (jia in the first tone, as distinct from its third-tone homonym meaning “tithing”) rather than the larger “household” (hu). A household was a fiscal unit that could include collateral relatives (usually unmarried). In a system where taxes were levied on the household rather than the individual, there was an incentive for the taxpayer to let his household increase in size rather than split it up into smaller households, such that in some places a fiscal household could expand to include an entire lineage.107 The family, on the other hand, was a social unit defined as those who shared the same hearth: usually parents, their children, and the surviving members of one set of grand-parents. The difference between the lijia composed of households and the baojia composed of families

The chinese state in ming society

36

is not remarked on by contemporaries, yet it is almost universally observed in the language of the sources that speak of lijia and baojia constituents. It would be inadvisable to lay too much emphasis on the distinction, given the rough equivalence of the two terms in the popular mind; in some areas the change from lijia to baojia was nothing but referring to the old units by new names.108 Yet the distinction could be a real one, reflecting the concern of those who framed baojia regulations that security could not be assured if able-bodied males hid behind their relatives’ households. By using families, local magistrates strove to make the baojia a working system that conformed to the existing social topography, not a system adhering mechanically to an artificial structure.109 Watches and tithings When Wang Anshi in 1076 established the first baojia system, its basic unit consisted of ten families, called a bao. In most Ming baojia systems, the ten-family unit was called a jia (tithing), and ten jia made a bao (watch).110 Each tithing possessed a placard (pai),111 which circulated in regular rotation among the members of the tithing. The family holding the placard at any one time served as the tithing captain (jiazhang) or tithing overseer (jiazong). The head of the watch was known as a watch supervisor (baozheng) or watch captain (baozhang).112 This was the system as it operated along the southeast coast, at least. A tithing could have as few as four families or as many as thirteen,113 and watches frequently had more or fewer than ten tithings.114 In some Jiangnan counties, an intermediate unit was introduced between the tithing and the watch. The term used was “compact” (dang), and it usually grouped thirty families.115 The officer in charge of this unit was called a compact supervisor (dangzheng) or compact captain (dangzhang).116 In other Jiangnan counties, one finds units above the watch. One prescriptive source speaks of this higher unit (ten watches?) as a compact;117 the more common term in the late Ming was the military “regiment” or tuan, overseeing an unspecified number of watches.118 Regiments could be headed by captains (tuanzhang) or overseers (tuanzong). Local magistrates instituted this unit to meet the need for superior militia coordination. Another system of terminology, proposed by Wang Shouren (1472–1529) and widely accepted in the Qing but used in the Ming only in the Nanjing area, moved the terms jia and bao up to the hundred-family and thousand-family units respectively, renaming the ten-family unit pai.119 Ming sources give little sense of the territorial extent of the baojia units. Gu Yanwu (1613–82) cites the Wuxi county gazetteer of 1574 as saying that regiments were established there on a scale of roughly one per canton.120 In less populated areas, the ratio fell to one watch per village. This suggests that regiments had well over a thousand families, though Gu does note that regiments in densely populated areas should be distinguished as “large regiments” (datuan). While there seems to have been some attempt to fit the new baojia units to the old subcounty units, a Jiajing-era writer in Huguang notes that regiments united several markets (shi) without regard for the subcounty system.121 Given the demographic principles shared by the baojia and lijia, the obvious place to look for contiguity is between baojia guards and lijia hundreds, and indeed this was the case. As the functions of the latter became narrowly fiscal, those of the former expanded

The spatial organization of subcounty

37

to deal with matters of security organization. In She county, Huizhou, the two systems worked as one, for we read of the magistrate in the early 1580s verifying landholdings in the county in consultation with lijia officers by referring to both the ward-level tax registers and the baojia registrations certificates.122 He was able to use the systems simultaneously because, at the level of the hundred, the watch, and the ward, they had merged. An early-seventeenth-century Fujian author says that it was common practice to maintain exact equivalence between them; if any boundaries could be crossed when drawing up baojia watches, it was those of the old lijia hundreds.123 In other words, the baojia has effectively replaced the lijia and merged with the subcounty administrative system. Gu Yanwu observes this convergence among systems in his essay on the lijia when he describes its structure as: “The county has jurisdiction over the canton, the canton over the watch—sometimes called a township—and the watch over the tithing.”124 In less densely populated regions such as Huguang, though, the lijia performed the security functions usually associated with the baojia.125 Indeed, officials in Yuezhou prefecture combined hundreds there into baojia-style regiments to deal with bandits.126 The covenant system The system of rural covenants (xiangyue) was neither mandated nor implemented universally, but was brought into use by energetic county officials undertaking local initiatives to reform taxation or security. The system involved forming groups of 100 households into covenants (yue), like the lijia hundreds, and requiring them to meet on a regular basis in a Covenant Recitation Hall (jiangyue suo) to recite Hongwu’s admonitions and perform rites reaffirming community membership and loyalty to the dynasty. I have found no evidence that covenants were formed during the first century of the Ming. The early-Ming institution for moral exhortation was the Pavilion of Exhibition (shenming ting), on which the names of the good and bad were posted and where informal judicial hearings were carried out. Kuribayashi Nobuo has observed that these pavilions were erected at a rate of roughly one per township in the Yangzi valley, suggesting that they were associated with fixed territories.127 The covenants emerged later in the dynasty, local magistrates instituting them in the hope of reinstilling the sense of community identity and responsibility that the lijia system was supposed to foster but by the turn of the sixteenth century could no longer be counted on to do so. For instance, when the magistrate of Changzhou county, Suzhou, ordered rural covenants to be set up in 1526, he not only declared that they were adjuncts of the lijia system but handed the responsibility for forming them to his hundred captains. The captains were not to run them, however; that duty was to be given to virtuous elders. The covenant leaders were then charged in turn with the task of setting up a community school in every covenant, in an attempt to revive Hongwu’s original plan to have every hundred establish a community school.128 The more common context for the creation of rural covenants, though, was the baojia, which was also on the rise in the sixteenth century.129 The association between the two was initially a loose one, but by the turn of the seventeenth century it had become a close and deliberate synthesis in the administrative theory of such statecraft officials as Lü Kun (1536–1618).130 The two systems also merged in practice.131 Once the watches had been

The chinese state in ming society

38

organized for guaranteeing public security, these entities could be convened as covenants for the purpose of organizing public education and disseminating government notices. Covenants tended not to enjoy any institutional independence, coming in as they did on the coattails of the baojia. One set of baojia regulations for Guangdong shows the two systems operating jointly as early as 1549.132 Both baojia and xiangyue were introduced in Shandong in the late 1580s and 1590s, though information about their implementation is distressingly scanty.133 In one Shandong county, a magistrate in 1590 set up 200 places for covenant recitation and appointed a covenant ofFicer for each one.134 Given that this county had 180 hundreds in 1461, it would seem that the 200 covenants (probably intended as an approximate number in any case) were established in correspondence with wards or watches. A late-seventeenth-century gazetteer from southern Anhui records that covenant recitation halls were distributed at a rate of one per urban township and two to twenty per rural township; that is, at the ward level.135 The overall “system,” to overstate the orderliness of these arrangements, appears in Table 1.7.

Table 1.7 Subcounty administrative units in the Ming lijia

baojia

Covenant system

Subcounty administrative system

Urban/suburban subcounty system

County (xian) Canton (xiang) Sector (qu)

Hundred (li)

Subcanton (li) Regiment (tuan)

Township (du)

Township (fang, xianq)

Guard (bao) Covenant (yue)

Ward (tu, she, tun)

Ward (tu, fang)

Compact (dang) Tithing (jia)

Tithing (jia)

Contemporaries confirmed that watches and covenants duplicated each other in their laments that covenants too readily came under the personal control of baojia officers. Their advice was that covenants be established before the baojia system was instituted so that the civilizing effects of the former might prevent the corruption that so often flowed from the latter.136 A quarter of a century later, another Fujian scholar argued again for the need to make the baojia officers subsidiary to the rural covenants officers, but this order would have gone against practice and seems never to have been followed.137 Local records from other areas suggest that, by the late Ming, the covenant system was at least formally in operation in much of China proper, though its ill-defined and improbable responsibilities prevented it from having the presence the baojia system came to have in the Qing.

The spatial organization of subcounty

39

Community, state, and society In the first years of the Ming, the court imposed its various subcounty systems of administration, levy, and control in order to make local society visible and accessible to the centre. It posited two administrative central places (Beijing and Nanjing), between 887 and 1,159 lower-level central places (the counties), and hinterlands for each that were precisely defined by population and land. The totality and legibility that the design created were impressive, and even more so, the state’s capacity to enforce implementation, at least in the early years. The parallel with William Skinner’s influential central-place model of marketing hierarchies is striking, as he himself has recognized. Central-place theory is based on the free operation of market forces and their relatively unhindered influence in bounding and linking social space, and is persuasive for modelling the formation of spatial systems in the context of societies with commercialized economies operating on market principles. Gentral-place theory assumes societies in which the economy is sufficientiy independent and powerful to override such prior shaping factors as community or kinship. In the same way, administrative centralplace theory must assume that the Ming state was able to override the influences of community and kinship to extend its administrative reach across the realm. It imagines a polity in which administrative forces operate on an unbounded national plane. Just as economic central-place theory may overshoot the realities of Ming economic life, so too the assumption of extensibility may not be the best way for us to conceive of the operation of sovereignty in the Ming, even if it was how Ming emperors saw things. A more reasonable starting point for both models might be to accept a higher degree of interaction among society, polity, and economy, so that the boundaries of lived communities are understood as running along contours created by all three: the societymaking effects of field boundaries, kinship collectivities, irrigation projects, and ritual circuits; the state-making capacities of existing administrative patterns, new state systems, and imposed units of demographic and territorial definition; and the economymaking factors of production, commodity networks, trading centres, and systems of resources access. When William Skinner and Keith Schoppa showed that, in the nineteenth and early twentieth centuries, administrative units were being created and reproduced in correspondence to marketing areas, and that marketing networks were the fundamental consideration when drawing boundaries for new subcounty units,138 they were testifying to the resilience of all these systems and their capacity to run together. This is not to suppose that there were no ruptures when the state drew boundaries, but it is to acknowledge that at subcounty levels, administrative units, markets, and social collectivities were usually closely aligned. Whether administrative units set the channels within which commercial networks ran, or vice versa, may not be a question for which there is a uniform answer, other than that both are true. Adapting to local conditions did not necessarily blunt the effectiveness of state supervision. It enabled officials to ensure that state administration was felt, at least nominally, in every village in China. The subcounty systems in fact provided official positions—clerks in the canton, watch captains in the baojia, tithing heads in the lijia, covenant supervisors in the xiangyue—that local leaders could use as political resources. On the other hand, as long as these systems extended down to the household level, these officers stood as proxies, however many times removed, of state power—if also

The chinese state in ming society

40

simultaneously of their own. At this point of interface, it becomes impossible to decide where the state ends and society begins. The oft-reiterated stories of malfeasance and coercion by village officers in the Ming and Qing testify to the extent to which power by proxy was turned to social (personal) rather than state interests. Baojia officers were particularly notorious for linking up with yamen runners more than with their magistrates, who were in any case too involved with the local gentry to be much concerned with lesser community representatives. The residents of peripheral cantons might well never see their magistrate, but they did see his proxies on a day-to-day basis. It has been suggested that the growth of population from the Tang through the Qing, by reducing the proportion of officials to population, caused a decrease in the state’s local presence. As Skinner put it, “a unified empire could be maintained into the late imperial era only by systematically reducing the scope of basic-level administrative functions and countenancing a decline in the effectiveness of bureaucratic government within local systems.”139 This view of administrative decay underestimates the proliferation of subcounty units and offices between the Tang and the Qing, the manpower for which was entirely local. As a mid-sixteenth-century writer exaggerated, “Up into the Zhengde era [1506–21), one out of every ten commoners was in office.”140 Without straining the resources and communication limits of the field administration, the state was able to augment the staff performing basic-level functions. Sixty-odd thousand units, each with ten captains and ten tithing heads, plus an indeterminate number of elders—hundreds of thousands of local officers appointed by the state, in other words—were working at least nominally to guarantee state interests. By building up an increasingly dense structure of subcounty units, the state was able to keep pace with, perhaps even overcompensate for, trends toward the dissipation of its power at the local level. The various subcounty systems may have “failed to attain the results which they were theoretically capable of producing,”141 as Hsiao Kung-ch’üan argued for the Qing, but exaggerating the shortcomings of the systems runs the risk of minimizing their local social presence. For these were not just state communities, after all, but social communities as well, which the state did not disrupt.142 Indeed, they may have strengthened the resilience of both state and social control: every lijia officer was a proxy for state power as well as a hostage to local interests. A hundred captain who was not already a village headman, or the head of a lineage, stood little chance of amassing the political weight he needed to carry out the functions the state assigned to his post. Formal power could not afford to rest across a gap from real, but had to let those who had real power at the local level represent state interests. Nor could the state withhold its proxy power. It had to accept the structure of power that was in place and intervene against village leaders only when key state interests were threatened. Thus, although local communities were not autonomous, they did have sufficient internal cohesion and social vitality to prevent them from being mere creatures of government authority. As long as the minimum requirements imposed by the state were met and the forms of the administrative systems made visible, local communities might run their own affairs as they liked. The universal presence of formal institutions and government officials—no matter how many levels above the household in the hierarchy they were situated—was usually sufficient to ensure that resistance to central authority could be tracked, though whether it was crushed depended on how well the networks of resistance were socially embedded. The state was careful accordingly to hierarchize local

The spatial organization of subcounty

41

units and create a structure of command that allowed for coordination only from above. While individual communities might achieve internal cohesion through the hierarchies of kinship, age, sex, wealth, and education, they had no mechanisms for inter-village coordination other than by combining units under state supervision.143 Ming society thus existed within an array of tensions among social community, economic network, and administrative state that has reproduced Chinese society into the present. Each was limited to what it could impose on the other, and each has relied dialogically on the other for its shape and substance. The ubiquity and functions of state systems at the subcounty level should alert us to the importance of knowing how these systems worked, and how they influenced administrative and social decisions in later periods. The state’s administrative presence in local communities, the social manipulation of that presence to accommodate and serve local needs, and the concern of officials to limit lateral linkages among rural communities not only formed the core of the Ming subcounty system but have been major elements of the Ming’s bequest to the present.

2 The gazetteer cartography of Ye Chunji When a magistrate arrived to take up his post in an unfamiliar county, he faced the daunting task of mastering sufficient knowledge of his new jurisdiction to govern effectively. In his popular handbook for magistrates compiled in the 1690s, Fuhui quanshu (The complete book concerning happiness and benevolence), Huang Liu-hung advised the new appointee to read the local gazetteer. “When the magistrate makes a thorough study of the local gazetteer, he will be able to have a clear picture of its geographical layout, the amounts and rates of taxation, and the vital statistics and degree of prosperity of its population,” he explained. “This information is indispensable in planning his administration.”1 Ye Chunji, a Guangdong juren of 1552 who pursued a modest career as a local official in the last third of the sixteenth century, would have given the same advice.2 When he arrived in the coastal Fujian county of Huian early in 1571 (or possibly late in 1570) to take up the second posting of his career, and his first magistracy, Ye Chunji went looking for the local gazetteer. He was fortunate in finding copies of four: the 1530 and 1566 editions of the county gazetteer (Huian xianzhi), and the 1525 and 1568 editions of the gazetteer of Quanzhou (Quanzjiou fuzhi), the prefecture in which Huian was located. That Huian should be chronicled in two county and two prefectural gazetteers as early as 1570 is not remarkable. By the late Ming, the local gazetteer had become well established as the official genre for recording geographical, historical, and biographical data of significance to the public life and administration of a county. It also regularly included maps of the county and the county seat, and usually illustrations of important state institutions. The genre began to take form in the Tang, and gained a measure of standardization in the Yuan after the central government issued compilation guidelines in 1296.3 Not all the conventions that came to govern how a gazetteer compiler should organize his material were fully in place in the Ming dynasty, as we shall note again in Chapter 8; that process would continue for at least another century Nonetheless, the gazetteer by the late Ming was a well understood genre, communicating knowledge of local administration within reasonably stable categories. Between the pre-Ming formative phase and post-Ming formalization, the sixteenth century was the period when published gazetteers became universal, when most counties and prefectures in China, as well as many mountains and monasteries, produced their first editions. Magistrates liked to sponsor their publication, and scholars with an interest in local geography and history considered gazetteers, and not just those of their own locale, worth owning. Of the four gazetteers available to him, Ye found the 1530 Huian xianzhi most useful for learning about local conditions, though it was forty years out of date.4 He discovered that the prefectural gazetteers simply borrowed their data from this gazetteer, and that the 1566 edition of the county gazetteer was little more than a reprint of the 1530 edition. It was therefore on the first Huian xianzhi that Ye relied for his data. The person who produced this gazetteer was Zhang Yue. Zhang was Huian county’s most prominent

The gazetteer cartography of Ye Chunji

43

native son in the sixteenth century serving incorruptibly and energetically in office. At home to observe mourning in the late 1520s, Zhang took on the task of compiling the county’s first gazetteer. He not only oversaw the project but wrote substantial portions of the text in a flne essayistic style that makes it one of the most nicely written of mid-Ming gazetteers. Ye was fortunate in having a good gazetteer to use “in planning his administration,” as he put it, yet he soon came upon disturbing gaps and discrepancies. The disturbance was not purely a matter of his Confucian commitment to complete knowledge. What drove him to probe, and later emend, the errors in the textual record was his equally Confucian commitment to statecraft, which in those post-Wang Yangming days regarded perfect action as predicated upon perfect knowledge, and the impairment of knowledge as threatening the effectiveness of action. We read in the 1573 preface, which the eminent Fujian expert on border affairs, Guo Zaoqing, wrote for Ye’s own survey of Huian county, that what first distressed Ye was the lack of maps in the 1530 gazetteer. Guo does not say whether Ye’s distress was as a reader or an administrator, but whichever was the case, the absence of maps inspired in him what became a professional absorption in the problem of recording precise knowledge of locations and boundaries. Maps were of vital concern to the Chinese state, as they are to every state that legitimizes itself through the delineation and control of territory. States require maps of borders to mark where jurisdiction ends; maps of the northern border were a particular obsession of the Ming court, and naturally so, given the constant raiding and instability in that region. States also require internal maps, both to chart the places where security must be guaranteed and to designate the location and extent of taxable land. Although domestic maps came to circulate extensively throughout the Ming realm, their ownership was a potential liability should their owners be suspected of seditious intent. As we shall see in Chapter 6, the discovery of “printed texts and maps” in the possession of another Guangdong native, who stirred up trouble in 1481, was taken as corroborating evidence that he had planned to incite a rebellion.5 Ye’s own concern was not with organizing security but, as we shall see, with getting an unimpeded view of the territory that it was his job to tax. The need to get taxable land into view and onto the books was a standard task for the local magistrate. Just at the time that Ye was posted, this need gained a particular urgency because of the nationwide move to reallocate fiscal levies from labour to land, a trend that would culminate in the next few years in a set of changes known as the Single Whip reforms. Making cultivated land bear the brunt of tax assessment demanded that a clear record of who owned how much land be available in order to fairly distribute the tax burden of a county. The Hongwu emperor’s cadastral system, outlined in the previous chapter, required that a full census be taken once a decade on the “great compiling” (dazao) year in order to reassess levies. This practice had been discontinued as anything but a formality until the recent pressure to redistribute taxes in the mid-sixteenth century The coming year, 1572, was a dazao year, and many of Ye’s contemporaries found themselves dealing with the problem just as he was. In Hezhou, South Zhili (see Map 2), magistrate Kang Gao drew up a detailed list of ten regulations for carrying out a thorough survey and re-registration of the land and population in the subprefecture. His first regulation states that the officers should actually measure the land this time, rather than simply resubmit the old figures on the tax records, as they usually did, for the long accumulation of misreporting and other abuses

The chinese state in ming society

44

had created a complete disjuncture between what the tax registers claimed and who really owned the land. Kang proposed that the hundred elders in every township take charge of the process by going out and resurveying all the cultivated land within the township. The second step was then to compile new land registers showing every plot in each hundred. At the front of each register there should appear a new composite map of the hundred based on these surveys. Given that Hezhou had forty-one hundreds, the new records would yield forty-one new maps.6 Kang’s maps do not survive, but Ye’s do. Like Kang, Ye decided that his first priority was to get proper maps drawn. Shortly after taking up his post, he called together a council of thirty-odd “local elders,” the same people on whom Kang would rely the following year, to discuss county affairs. At this first meeting he put the project of drawing maps before them.7 They agreed, but when the results came in, Ye was disappointed by their inaccuracy. They had given him exactly what Kang was afraid he would get in Hezhou. Coming up with good maps, Ye discovered, would have to involve a more complicated process than issuing a simple order. He describes what happened: When the elders’ maps were submitted, they did not tally [with each other]. Just at this time, Guo Zaoqing, a native of the province, visited me on his way up to the border. I showed him their maps and he questioned me sharply about their accuracy before departing. After examining the prefectural and county gazetteers, I believed what Guo had suspected [about their inaccuracy] and ordered the clerks to take compasses and go over them three or four times. Only after a year’s work were the maps finished. Meanwhile I collated the discrepancies regarding mountains, rivers, and administrative boundaries [among the local gazetteers, which generated] rather excessive documentation. When Guo returned and saw my notebooks, he said, “Every gazetteer compiler gives a different story. If you don’t write a text [to accompany the maps], who will realize that the present version is correct?”8 The outcome of this conversation was Ye’s idiosyncratic variation on the local gazetteer genre, a set of survey summaries which he entitled Huian zhengshu (Administrative records of Huian county). The manuscript was completed in 1573 and is current to that year. The copy I have used is the version printed in the 1672 re-edition of Ye’s collected works, Shidong wenji, of which the sole surviving copy is held at the Tōyō Bunko library in Tokyo.9 Maps were the inspiration for Ye’s work, and would prove to be his most significant contribution to the production of geographical knowledge in the late Ming. By recuperating the map as a succinct mode of organizing knowledge of place, Ye was returning, unconsciously it seems, to the Han-dynasty practice of recording geographical information on cadastral survey maps. It was this practice that led to the development of books of maps (known variously as tuji, tujing, and tuzhi) in the Sui and Tang, sometimes running over a hundred juan.10 Out of these simple atlases (i.e., books of maps explained by appended texts) grew gazetteers (i.e., books of texts illustrated by maps). This shift in the representation of knowledge from visual to textual was undertaken because of a need to record types of knowledge in greater descriptive detail than a visual summary allowed.

The gazetteer cartography of Ye Chunji

45

Precise knowledge of the status or quantity of things (such as land, crops, and taxes) came to outweigh precise knowledge of their location relative to other things. Henceforth, in gazetteers, location would be generalized, and maps accordingly reduced to meaningful though simple depictions of how the county looked from the vantage point of the magistrate’s yamen. By the Song, the term tu disappears from gazetteer titles (except when adopted as a conscious archaism) and the efforts expended on creating the maps in gazetteers decrease. The maps inserted in the prefatory material of Ming gazetteers are mostly pictorial summaries rather than exact renderings of precise knowledge of spatial dimensions and relationships. The reader would be hard-pressed to extract from such maps information about where places actually were, or even how one could get from one place to another; but that was not their purpose. Although maps lost their primacy as a mode of organizing local knowledge in gazetteers, they were revitalized by imperial fiat early in the Ming dynasty as a format for organizing fiscal knowledge of land at the village level. In 1387, the Hongwu emperor, displeased with the persisting evasion of fiscal registration by wealthy households, ordered his tax captains to compile booklets of maps for their areas showing the boundaries and ownership of all agricultural land. Every plot was to be paced out and measured.11 The resulting Fish-Scale Registers (yulin tuce), so called because the pattern of plots on the summary map at the front looked like fish scales, were sensitive to precise topography, since they created the official public record of who owned what land and who paid the taxes on it. The drawing of the Fish-Scale Registers for every village in China was the most exhaustive mapping program any government in China undertook prior to the twentieth century.12 Little otherwise appears to have happened in the field of local cartography during the Ming prior to the middle of the sixteenth century. The field was revitalized when Luo Hongxian (1504–64) revived and developed the grid system for drawing maps. Luo was a respected scholar of the mid-sixteenth century. Like Wang, he was a man of action, though being barred from the civil service between 1541 and 1558 for criticizing the emperor, he could not act out his sense of Confucian responsibility through the usual channels of officialdom.13 Undaunted, he remained home in Jiangxi and worked instead on local and regional problems in an informal, advisory capacity His work included resurveying local land to enable his county magistrate to adjust inequitable tax burdens, organizing defences of the county seat during bandit attacks, and compiling information to help regional officials defend the coast during the piracy surge at mid-century. His experience with land surveys, combined with his recognition that national defence required precise knowledge of coastal and border areas, led him to map-making, and to the revival of the grid system. The grid system is an isometric projection in which territory is mapped according to equal squares, each square in the grid representing an equivalent distance on the ground. The technique is first described by Pei Xiu (224–71). The earliest surviving example of grid mapping is Yuji tu (Map of the tracks of Yu), incised into stone in 1136, which shows all of China on a grid of uniform squares. It bears the inscription “each [side of a] square is equal to a ground distance of one hundred li (meifang zhe di baili).”14 In the Yuan dynasty, Zhu Siben (1273–1337) devoted a decade to composing a national map of China using this grid system. His Yudi tu (Map of the terrestrial realm), considered to be the most accurate cartographic rendering of China to that time, did not substantially alter

The chinese state in ming society

46

Chinese cartographic practice until Luo Hongxian, after three years of searching, found a manuscript copy and revived the system. Luo reports that the original, a chart 7 chi (over 2 metres) square, was drawn according to “the method of counting li and dividing into squares” (jili huafang zhi fa). What Luo did was to use the grid method himself to convert this great chart into a book of forty-five regional and provincial maps. Luo’s Guangyu tu (Enlarged terrestrial atlas) was first published 1555. The book was well received and went through at least five editions over the next quarter-century (1558, 1561, 1566, 1572, and 1579). The atlas became the standard for all subsequent cartographers of China, including Matteo Ricci (1551–1610), who relied on the most recent 1579 edition to draw the Chinese portion of his famous 1584 map of the world.15 Ye Chunji had already acquainted himself with Luo’s cartographic techniques before he took up his post in Huian county. In fact, he was acquainted with Luo himself. During the period he was studying for the jinshi examinations in the 1550s, Ye travelled several times with friends up to Jiangxi to meet with Luo.16 Luo’s seriousness of purpose and his scholarly interest in administrative geography impressed Ye, who adopted Luo as his intellectual mentor. When in the 1570s, after Luo’s death (and his own temporary banishment from public life), Ye chose the studio name Shidong (Stone Grotto), he did so to honour Luo, who at a similar stage in his career withdrew to his retreat-cum-school, Shilian dong (Stone Lotus Grotto). There is no evidence that Luo personally instructed his junior visitor from Guangdong in the craft of making maps. It was Ye, impressed with Luo’s example, who took up the study of his mentor’s work and taught himself. Repeated references through the first two chapters of Administrative Records display Ye’s familiarity not just with Luo’s method but with other recent work in cartography. He cites the remarkable atlas of the northern regions, Jiubian tulun (Maps and commentaries on the nine border regions), which Xu Lun (1495–1566) completed in 1534, presented to the Jiajing emperor in 1537, and published the following year. Luo relied on the Jiubian tulun for his renderings of the northern border areas and may have been responsible for introducing the work to Ye.17 Ye also refers to a book by a man surnamed Li entitled Yudi tuxu (Maps and descriptions of the terrestrial realm), also of the jiajing era (1522–66) but apparently no longer extant. He learned much studying both books, he says, yet neither had the level of practical detail that set Luo’s maps apart from all other cartographic work available to him. Using Luo’s method to draw a county map was not a straightforward matter, however. Luo himself had never applied his method on this fine a scale, for his atlas did not go down below the county level. He had revived the grid system only with a view to systematizing existing cartographical knowledge at the provincial level. What Ye did was to take Luo’s method down below the county and develop a technology of drawing accurate local maps on the basis of on-the-ground surveys. In his preface to Administrative Records, Guo Zaoqing draws attention to this innovation, which is Ye’s contribution to Ming cartography: precise mapping at the local level.18 Ye’s order to the elders to draw up local maps specifies that they were to use the grid method he had taken from Luo’s atlas. As we have noted, the project did not produce satisfactory results. Possibly Ye left the application of Luo’s principles to the local elders to figure out. If so, it seems that they met his request in good bureaucratic fashion: by simply sketching pre-existing drawings or maps of their areas onto the grid to give the

The gazetteer cartography of Ye Chunji

47

appearance of having used it, rather than actually surveying their areas in square sections and transcribing the results onto the grid. Ye’s second attempt “to apply Luo’s ingenious method to my county”19 was done under close supervision. He recruited “the well-read among the scholars junior to me” to help compile documents for the project, and explained to them the grid method so that they could work on the maps.20 Then he mobilized local elders and village leaders to help with drawing maps at the level of the lijia hundred. Altogether 155 local maps were made. Ye took these and used them as raw material for constructing maps of every township (du), the principal administrative unit between the county and the hundred. Huian had had thirty-four townships earlier in the Ming; by the sixteenth century, seven had been amalgamated to others, for a total of twenty-seven. Ye therefore drew twentyseven township maps. To draw these maps, Ye set up working grids whose dimensions he established through the following calculations. According to the 1530 gazetteer, Huian was ninety li by eighty li in size (Ye later found these figures to be incorrect). Since the county consisted of roughly thirty townships, he divided the total area of the county (7,200 square li) by that number, which yielded an average township size of 240 square li. No township was likely to be a perfect square, which meant that he could not be content to make do with a grid whose side was simply the square root of 240. To allow for irregular shapes, he instead divided the total area by ten and set up a grid that was twenty-four li on each side. At one li per square, his drafting board for tracing out the boundaries of each township was twenty-four squares across and twenty-four squares down, a total grid of 576 squares. These he would truncate or add to as the need arose.21 Following this method, Ye produced the twenty-nine maps that appear in Administrative Records: twenty-seven township maps, an urban map, and a map of the entire county.22 Printing his maps of townships of different shapes and sizes on pages that were all one size was Ye’s next problem. He could not simply draw them all to the same scale and hope to produce maps of uniform size. Ye had to vary the scale for each map, which he did not by putting them on grids of different scales but by reducing or expanding the size of the squares on paper. The squares on the printed maps will thus be found to range in size from 0.7 cm to 5 cm. What remains constant, for all but two maps, is the distance that one square in the grid represents. With two exceptions, every map bears the notation: ”[the length of the edge of] every square is one li” (meifang yili). As Ye phrased his method, “The grid of squares (fangce) has to be filled (biying)” uniformly at the same rate, but the size of the grid did not have to be the same on every map.23 The two exceptions to this uniform variable scale are the map of Townships 14 to 17, which uses a scale of two li per square, and the overall county map, drawn to a scale of five li per square. Ye states that his goal in making these maps was to compile a precise record of topographical features and settlements.24 Still, he did not choose to produce only a county atlas. He recognized from the start that certain categories of information, such as arable land, were difficult to represent visually. Despite the care he used to make his maps, his goal was to use them as one technology among several to efficiently organize knowledge. Rather than try to produce detailed cadastral maps, he chose instead to append knowledge in two other formats. A one-page text summarizing the main characteristics of the township prefaces each map, and five charts (biao) follow it: human settlements and

The chinese state in ming society

48

structures (post houses, villages, altars, pavilions, schools), lijia household categories, population figures (by male and female, adult and child), arable land, and taxes. As he did for the maps, Ye compiled these charts working from actual data collected by his junior associates rather than using material on file in the county office. One of his discoveries when he assumed office was that the yamen records had not been revised for forty years—in other words, since the time when Zhang Yue compiled the 1530 gazetteer. His own data is current to 1573.25 The grid design of the charts makes them as easy to consult as the maps, and the charts follow the same format for every township, again like the maps—a point that Ye himself notes.26 Ye thus treats knowledge about different places as no longer specific to where it occurs. He has standardized it into a single, infinitely repeatable form which can nonetheless be adapted to circumstances by modulating the mesh and scale of the grid on the maps and the size of the blanks on the charts. The first map in the book, which shows the entire county (Figure 2.1), is on a grid of 300 squares using a scale of five li per square. Like all the maps in the book, the county map is territorially bounded: that is, it leaves the space beyond the boundaries of the territory it maps blank, except to indicate adjoining settlements or administrative jurisdictions. His attention to detail focuses in this map on the larger features of topography and boundaries. The boundaries appear on the county map as thick lines, and the township numbers are marked (in most cases) as white numbers in framed black rectangles. Determining township boundaries was the biggest difficulty he says he faced, especially in places where neither mountains nor rivers served as natural boundary markers. In this task, the 1530 county gazetteer was of surprisingly litde use. It lists embankments, for instance, but according to Song subcantons (li) rather than Ming townships.27 Determining which natural feature or public work lay in which township required on-site surveys. Even then it could not finally be ascertained until township boundaries were set, since even local residents disputed about which pieces of territory fell within which township. Ye notes with regard to siting mountains that, by comparison, Luo Hongxian faced an easier task, since his national map shows only the five sacred mountains. Rivers were even more difficult, for whereas the actual shape and location of a mountain does not have to be indicated precisely when using the Chinese pictorial convention for mountains, the constandy turning course of a river could not similarly be fudged.28

The gazetteer cartography of Ye Chunji

49

Figure 2.1 Map of Huian county. Ye Chunji, Huian zhengshu, 4.1b–2a. Source: This and the following illustrations are reproduced from the 1672 edition of Ye Chunji’s collected works, Shidong wenji, courtesy of the Tōyō Bunko, Tokyo Most of the features Ye chose to put on the county map appear in the legend set in the lower left corner of the map (Figure 2.2). Luo Hongxian had used symbols in his Enlarged Terrestrial Atlas and grouped these symbols together in a legend, but he put the legend in the book’s preface, not on the map itself. Ye’s map of Huian appears to be the first Chinese map on which a legend appears in a rectangle on a blank area of the page. Only the first two entries in the legend, for mountains (shan) and rivers (shui), are pictorial; the rest are symbolic. Boundaries (jie) appear as thick lines, roads (lu) as broken lines, and city walls (cheng) as double encircling lines, although on the map the wall around the county seat is distinguished as a crenelated circle. The remaining eight symbols indicate structures: military camps (bao), stockades (zhai), beacons (dun), post houses (pu), villages (cun), pavilions (ting), altars (tan),

The chinese state in ming society

50

Figure 2.2 Cartographic legend on the map of Huian county. Huian zhengshu, 4.2a. and bridges (qiao). Some of the symbols Ye uses (such as those for stockades and beacons) come straight from Luo’s legend. Others, like the empty rectangle, are redeployed: for a prefecture on Luo’s maps, for a post house on Ye’s. Yet others are of his own choosing. In actual fact, the county map does not include all these structures, only the coastal beacons and post houses. The legend, which is not repeated on any of the township maps, was intended to serve for all the maps. The next map in the book, the county seat (Figure 2.3), was drawn on a grid of only twelve squares, each of which stood for one li. Because of the low scale, Ye was free to employ conventional pictorial elements more extensively in this map than in his others, notably the crenelated city wall, the towers and protective curtain walls of the city gates,

The gazetteer cartography of Ye Chunji

51

and Phoenix Pond in the southeastern area of the town. Otherwise, he repeats the use of symbols given in the legend—with one intriguing variation: an open rectangle was used to designate pu, but in the township a pu is a post house, whereas on the town map it is an inn. This map is striking when compared with county seat maps that appear in other gazetteers in terms of the detail it gives regarding the locations and names of the main streets, intersections, and public buildings in town, particularly the last.29 One could actually walk through the town with Ye’s map in hand and find everything marked on it. Reading right to left down the main street running from the north gate of Chaotian to the south gate of Tonghui, one encounters Aozhen Inn on

Figure 2.3 Map of the Huian county seat. Huian zhengshu, 4.10b–11a the south side of the street (just past Longqiu Sluice), then on the north side the main county post house, the suboffice of the Provincial Administration Commission (Buzheng fensi), the City God Temple (Chenghuang miao), the county school, Dengyong Inn, the sub-office of the Provincial Surveillance Commission (Ancha fensi), and Xiange Inn. Further west past the Longjin River, one finds the prefectural hostel (Fuguan), Longjin Inn, the county courier station (Jintian yi), across the street Leshan Inn, and down near the south gate, Qingquan Inn. The map of Township 2 (Figure 2.4) is drawn on a scale between the county and city maps. The ratio of the grid is one square per li, as in the city map, but the size of the township requires a much denser grid (of 120 squares) to accommodate its greater expanse. Township 2 lies immediately southeast of the county seat, the city wall of which is marked by the curving double lines on the left-hand side of the map. The land is

The chinese state in ming society

52

reasonably flat and featureless, which means that the topography is defined by its rivers. Longjin River and Longqiu Sluice flow out of the county seat left to right across the middle of the map and converge before reaching Yangzhai River, which flows across the bottom of the map and joins its tributary half-way up the right-hand side. Around each river Ye has drawn a continuous thin line, which seems to indicate the river’s spillway or floodplain. The meaning of this line is suggested by the depiction of Yanshou Bridge in the lower right-hand quadrant of the map. Unlike the four other bridges shown on

Figure 2.4 Map of Huian Township 2. Huian zhengshu, 4.23b–24a. Dotted lines indicate roads. the Yangzhai River, Yanshou is depicted not just spanning the riverbed but stretching from the line on one side of the river to the line on the other. This bridge, built in 1366, consisted of 129 arches and was over two-thirds of a kilometre in length:30 a bridge designed to keep the east road open even when the river flooded. This use of lines to mark floodplains—contour lines gauging topographical relief—is a striking feature I have seen on no other Ming map. Besides rivers, bridges, and floodplains, the map displays many other categories of information: twenty villages (each marked by a black dot), three altars, three pavilions, one post house (pu), and a lock (dai). In addition, in the area of the township close to the county seat are marked such institutions as the school archery compound (shepu), the charitable cemetery (yizhong), and the hostel for the destitute (yangji yuan), though in each case the name appears without an accompanying symbol. The only institution listed in the charts for Township 2 but not shown on the map is schools, of which five are listed. There was certainly no precedent for marking schools on maps, and Ye apparently

The gazetteer cartography of Ye Chunji

53

saw no purpose in starting one. Where the map is particularly good is with regard to the two networks of roads leading out of the north and east gates of the county seat. The main official road (daguan lu) running out of the north gate soon forks, and then further subdivides along its eastern branch. The road out the east gate immediately divides into three, with a fourth route branching off from the southeastern road and heading south. All are not only marked but named. Ye notes in his first chapter that he paid special attention to marking roads on his township maps consistently to ensure that the point where a road is shown as exiting at the boundary of one township is the same as the point on the map where it crosses into the next.31 Ye does not reflect on problems that may have arisen in the course of transferring his hand-drawn maps to woodblocks. The sole reference to publishing appears in the preface, where he says that he entrusted the work of having the woodblocks cut to his friend Guo Zaoqing, though the comment may politely indicate that Guo covered the costs. Looking at the maps closely, it becomes clear that the craftsman who engraved the map of Township 2 was not the engraver of the county and town maps. The obvious visual clues are the difference in character styles between this and other township maps, the use of black dots rather than empty circles to locate villages, and the sharper hatching of the dotted road lines. Whether the different hand introduced more than stylistic ticks that materially alter (or distort) the character of the map is impossible to determine. This difficulty is compounded by the fact that we are looking at a 1672 reprint of the book, not its original printing, so we cannot be sure of where the differences crept in. The maps and their appended texts and charts constitute about two-thirds of Administrative Records of Huian. In terms of the chapter structure, however, they fill only the middle five of the twelve chapters or “administrative records” (zhengshu) into which the book is divided. Three chapters precede them. The first, “Questions on Maps and Annotations” (Tuji wen), presents twenty-eight questions and answers that review in some detail the difficulties entailed in record-keeping and map-making. The second, “Investigations on Geography” (Dili kao), is a series of entries on mountains, rivers, bridges, and major buildings, in which Ye carefully notes discrepancies in previous gazetteers. His list of sites is not comprehensive, being restricted to those for which previous records are incorrect. The third, “Investigations on Cadastral Registers” (Banji kao), reads much like the chapter on county finances in any gazetteer. The last four chapters of the book, which follow the chapters of maps and charts, are reports on four subcounty institutions that Ye, following in the tradition of Luo Hongxian,32 sponsored as magistrate: rural covenants (xiang yue), community rites (lishe), community schools (shexue), and the neighbourhood watch (baqjid). Taken together, these twelve “administrative records” add up to something other than a regular county gazetteer. The book resists this classification, for four reasons. First of all, it highlights maps and charts over text to organize data. Secondly, it does not follow the customary arrangement of chapters, nor does it include such standard gazetteer information as records of former officials, lists of degree-winners, biographies, and local writings. Administrative Records reads more like the first third of a gazetteer, and only then one in draft. Thirdly, unlike most county gazetteers, Administrative Records does not aggregate all its data to the county level, but leaves much of it at the township level. In this way it is more localistic that the average county gazetteer. A fourth characteristic that sets the book apart is the presence of the compiler as narrator. Ye does not disappear

The chinese state in ming society

54

behind the fagçade of objectivity that the gazetteer format provides, but shows himself at work, analyzing, organizing, advocating. His book is thus revealing about the processes that informed the creation of the book and the maps. The gazetteer format was not what he wanted. Ye was not alone among magistrate—compilers in choosing to compile an administrative record of his county rather than a gazetteer. Others did the same, among them Ye’s elder and more famous contemporary, Hai Rui (1513–87). Hai’s justly famous record of his work as a county magistrate, Chun’an xian zhengshi (Administrative affairs of Chun’an county), completed eleven years before Ye’s Administrative Records, similarly discusses the tasks and procedures of local administration and does not aspire to be a complete account of the county.33 By comparison with Administrative Affairs of Chun’an, Administrative Records of Huian stands slightly closer to the gazetteer genre by focusing more on geographical knowledge and less on budgetary and judicial matters. Like Hai, Ye was a conscientious magistrate who used the knowledge he amassed and organized to administer the affairs of the county with a strict hand. He alleviated tax burdens that fell disproportionately on the poor, and sought to better regulate county life by revitalizing the baojia and rural covenant systems. These efforts were appreciated by the common people, but not by the local powerholders. The biography of Ye in the Quanzhou prefectural gazetteer of 1763 observes that he did not defer to the powerful. While he was in office, they were obliged “to stay their hands and did not dare to break the law.” Clearly, his strict administration of county affairs—and more to the point, his resurvey of agricultural land—made him enemies. When at the end of his tenure in Huian he received a promotion to a subprefectural magistracy in Sichuan, one of these enemies absconded with the official order for his transfer. Without it, Ye could not proceed to his new post, making him liable for punishment for desertion from duty.34 Cornered effectively, Ye could do nothing but plead illness to his superiors and retire to Stone Grotto. The blow was both vicious and effective. For failing to take up his post, Ye was banished from public life for the next eighteen years. He was being made to relive the fate of his mentor, Luo Hongxian. Ye’s forced retirement would have its beneflts for scholarship, however, for it released his talents from administration and led him to become active as a gazetteer compiler. When he wrote Administrative Records of Huian, he was not trying to compile a full gazetteer, but a gazetteer was reasonably the next stage to proceed to in summarizing and expanding the data on which he concentrated in this first experiment with local knowledge. In the 1580s, he compiled three gazetteers for places back in his native Guangdong: for Shunde county Guangzhou prefecture, in 1585; for Yongan county in his home prefecture of Huizhou in 1586; and for Zhaoqing prefecture in 1588.35 After he returned to official life in 1591, he compiled no more gazetteers.36 To the maps he drew for the 1586 Yongan gazetteer, Ye brought his cartographic innovations from Huian. He again deputed assistants to use the grid method and pace out the actual terrain, and he again drew them over gridlines for publication.37 The maps are immediately recognizable as Ye’s handiwork. The detail that was possible because of the scale he used for the Huian maps has been compromised in the Yongan maps. The gazetteer includes only four, one of the county and three of townships, compared to the twenty-nine in the Huian handbook. Whereas the Huian maps were drawn on grids of between one and five li per square, the Yongan maps use grids of fifteen and twenty li. At

The gazetteer cartography of Ye Chunji

55

this scale, much of the unique detail Ye could put on the Huian maps, such as the contours of floodplains, has disappeared. There is, however, one improvement in terms of the amount of geographical information presented, and that is the inclusion of roads, rivers, and occasional topographical markers in the empty spaces beyond the boundary of the unit being mapped. The Yongan townships seem less to float in abstract cartographic space. Ye also alters the legend to adapt to the circumstances of the areas he had to map. Rather than copy the Yongan legend from the Huian legend, he has reduced the number of symbols by two and changed the types of sites included in the legend, presumably to take account of the different institutions and terminology in inland Guangdong as opposed to coastal Fujian. The principal symbols remain unaltered between the two legends, but two are put to different uses and three new symbols introduced. Ye was flexible in adapting his work to the conditions he found. Local reality had greater authority with him than precedents, even his own. Ye’s maps, with their legends and grid lines, have the seductive appearance of modern cartography. The similarities, at the time that Ye was drawing his maps, were accidental. Chinese were unaware of the European cartographic practice of drawing lines to indicate latitude and longitude before Matteo Ricci printed his map of the world in 1584. Nor did they understand that the European grid of meridians and parallels was devised according to a spherical rather than a flat projection. Nonetheless, the coincidence had its impact. Ricci’s grid echoed Luo Hongxian’s atlas in such a way that Chinese readers were inclined to regard it with favour, to believe that this was a carefully made map. It is also possible that Ricci, ever alert for ways to bridge cultural differences with his Chinese audience, himself recognized the value of mapping on a grid precisely for this reason. After all, he understood Luo’s atlas to be the best there was when he drew his mappa mundi. This brief moment of mutual misrecognition between Chinese and European cartography did not promote any change in Chinese practice. Despite the obvious strengths of Ye’s method, local gazetteer cartographers continued to use the older pictorial conventions for conveying spatial information. A casual survey of Guangdong, Fujian, and Zhejiang gazetteers after Ye’s time has turned up no grid maps of counties or townships before the nineteenth century. The earliest gazetteer with a grid map is of Zengcheng, Guangdong, and dates from 1820 (Figure 2.5); others soon followed there and injiangnan.38 By the 1870s, the grid method was being widely used in gazetteers from Fujian up to the Yangzi Valley,39 though this was the decade in which Guangdong gazetteer editors started abandoning the Chinese grid in favour of latitude and longitude.40 One of these, in fact, is the 1876 gazetteer of Zhaoqing prefecture—for which Ye Chunji compiled the 1588 edition. It is difficult to argue for any sort of internal influence between these two editions. The revival of the grid system was clearly a response to the arrival of European maps in China, for grids gave Chinese

The chinese state in ming society

56

Figure 2.5 Map of Zengcheng county, Guangdong. Zengcheng xianzhi (1820), 1.1b–2a. This is the earliest Qing gazetteer map I have found for which the grid system was revived. The scale is 10 li per square. maps a specious likeness to them. The transitional nature of the revived grid is demonstrated in the 1887 gazetteer of Tongxiang county in Zhejiang (Figure 2.6), in which the cartographer has exploited the coincidence between Chinese and European grids by drawing his county map on a Chinese grid while citing longitudinal position.41 The grid system would never have been revived in the nineteenth century were it not for the apparent confirmation the method received from the outside. Its revival was also its demise, since the interest in making maps in the style of Ye Chunji died as soon as Chinese cartographers were trained to draw them according to European methods. The grid method, in other words, provided a convenient, but quickly dismantled, bridge to European cartography In genealogical terms, Ye Chunji’s contribution to the history of Chinese cartography is important nonetheless: he applied Luo Hongxian’s method at a level of detail Luo never attempted, and he left behind in his writings rich accounts of the process by which he produced his maps. In practical cartographic terms, though, his impact was limited. His work

The gazetteer cartography of Ye Chunji

57

Figure 2.6 Map of the waterway network of Tongxiang county, Zhejiang. Tongxiang xianzhi (1887), juan 1. This map is noteworthy for being drawn on the grid system (two li per square) but also correlated to longitude. The caption inserted at the upper right states that the county is 3°13′ east of the capital. In fact, the caption is in error by 1°. Tongxiang is located at 120°37′, and Beijing at 116°22′, a difference of 4°15′. The practice of running the prime meridian through Beijing began with Jesuit cartographers in the Kangxi era. The caption provides no indication of latitude, noting only that Tongxiang is 1,850 li south of Beijing.

The chinese state in ming society

58

did not circulate widely: Administrative Records of Huian was almost completely unknown at the time and has survived in only one reprint copy. His other three gazetteers are almost as rare. It seems that there was no demand for the sort of map Ye produced until a competing cartography arrived from Europe and altered existing visual authority, casting pre-grid methods into disrepute. The accuracy he strove for as a magistrate in the midst of the Single Whip reforms was not continued beyond that time, when it seems that magistrates were content to make do with what was on the books, whether it bore any actual relationship to the features and boundaries of the territory they were charged to rule. Ye’s place in the history of Chinese systems of geographical knowledge more broadly conceived is less easy to evaluate than his cartographic work. Administrative Records of Huian was an impressive and innovative piece of work that did not go anywhere. It did not lead even Ye to generate new ways of organizing geographical knowledge when he went on to compile conventional gazetteers. What motivated and shaped his work was his point of view, which was that of an incumbent magistrate whose overwhelming concerns were administrative. He did not compile or organize information for its own sake. He did so because he understood the value that precise information had for an official like himself “in planning his administration.” Maps were for the purpose of producing state order, and so a better map or a better compendium of documents was wanted to the extent that it contributed to better statecraft. These documents needed to be bettered only to the extent that the precision with which they registered certain categories of knowledge that the state required was improved. It seems ironic that the best cartography and the best summaries of geographical knowledge in the latter half of the Ming dynasty were produced by scholars such as Luo Hongxian and Ye Chunji who were forced to spend much of their careers out of office. Still, they did so at the edge of the state, pursuing knowledge for the sake of statecraft rather than for the sake of knowledge. It would never have occurred to Ye Chunji to think of his mapping work in any other fashion. But then, it would never have occurred to him to produce the maps he did were he not in need of the type of knowledge to which he was bound.

Part II Fields

3 Taxing polders on the Yangzi Delta Sometime about 1620, Ding Yuanjian (1563–1628, js. 1586) was boating through the southern part of his native Changxing county. One of the five counties making up Zhejiang’s Huzhou prefecture, Changxing bordered the southern shore of Lake Tai, the vast body of water that served as the hydraulic heart of the Yangzi Delta. Together with Wucheng county on its eastern side, Changxing was known as the wettest part of the prefecture, a water world of low-lying paddy fields. Ding was gliding past a village called Lüshan when he caught sight of an inscribed stone on the bank. He made this entry in his diary: I noticed a stone marker over two chi [2/3 m] high on which was inscribed “such-and-such-character polder, totalling so many mu in area.” The writing was almost indecipherable. The thing must have been over two hundred years old. Every sector and every polder once had them. Given their great age, some have been broken, others submerged in the water, others carried off by people. I didn’t have the opportunity to search them all out in all the sectors and polders. Thinking about this kind of system, which combined polders into sectors and sectors into counties, I found that, as a registration method, it is simple and clear.1 For those of us so many centuries in the future who pore over old inscriptions to understand how people lived in the Ming dynasty, it is a pleasure to see a counterpart in the late Ming having trouble making out the inscription on an old stone marker from his own dynastic past, just as it was about to recede into indecipherability. What was the stone doing there, and why by 1620 had it already become a fading memorial to a vanished past? The observer was not a naive tourist passing through the countryside, but a keenly responsible leader of the local Changxing gentry. Born into a prominent gentry family, Ding had converted the advantages of his privileged upbringing to pass the 1586 jinshi examinations at the impressively young age of 23. His youth, moral idealism, and unwillingness to compromise might have placed him at the centre of the factional conflicts that raged at the Wanli court, had he ever stayed in office long enough to become a major player. Every time he was given a post, however, he managed to get himself dismissed by alienating entrenched interests, notably the eunuch faction.2 Other than a brief posting in the provincial surveillance office in Guangdong, Ding gained no experience in the field administration, so his knowledge of polders was the knowledge he gained as a gentry activist in the Changxing region.3 And that was considerable. Committed to the statecraft vision that motivated so many reform-minded gentry of the

The chinese state in ming society

62

Wanli era, he believed that social problems were amenable to global solutions. The decay of polder walls in Changxing was one of these problems, and the stone marker he discovered at Lüshan seemed like one of the solutions. As it turns out, Ding was estimating on the long side of two centuries. It had been only 150 years since the stone had been inscribed and placed here at the head of the field facing the river. It happened in 1472 as part of a reorganization of the fiscal topography of Changxing county. Like every other county in Ming China, the fields and villages of Changxing had been divided into the subcounty administrative units that the Hongwu emperor had imposed on the realm in the 1370s as the basis for cadastral surveys and fiscal assessments. This fabric of lower-level administrative units was woven most densely here in northern Zhejiang and across the provincial border in southeastern South Zhili, where, as noted in Chapter 1, a county averaged well above the 150 lijia hundreds per county. Changxing, which lay in the heart of that region, had the remarkably high number of 257 lijia hundreds in 1461, and close to the same number of wards.4 In 1472, these wards were abandoned for land registration purposes in order to accommodate a reality that the Hongwu plan did not factor in. Hongwu’s lijia had laid an arbitrary fiscal map over the villages and fields; the 1472 reorganization chose instead to assess taxes by using the field boundaries as they actually existed. It was an approach that effectively reversed the Hongwu relationship between state administration and social practice. The stone markers Ding saw at Lüshan and elsewhere along the river he was travelling were the physical remnants of this change of system. Polders The 1472 reorganization was regional, implemented only in areas such as Changxing that had the poldered fields peculiar to low-lying counties. A polder (yu, wei), strictly defined, is an embankment or dike built around a field to protect it from inundation. The term is generally used, however, to signify the area of cultivated land within such embankments, converted from swamp or alluvium at the edge of rivers or lakes by a dike (see Figure 3.1).5 Given that the water level outside a polder was usually higher than the level of the fields inside, keeping the embankment in good repair is critical for the viability of the agriculture pursued within it. Typhoons were the great enemy of Ming polders, as gazetteer compilers frequendy note. Cultivators along the coast of the Yangzi River upriver from Shanghai experienced the consequence in July 1510, for instance, when “a typhoon breached the polders around the fields, driving people into flight and leaving countless starving, ill, or dead.”6 Less spectacular but more common was the gradual decay of embankments due to erosion and neglect.

Taxing polders on the Yangzi Delta

63

Figure 3.1 Small polders along the Yangzi River as painted in a handscroll by Fan Qi (1616-after 1694) entitled “Landscape along the Yangzi,” reproduced in Lothar Ledderose, Orchideen und Felsen, pl. 184. Source: Collection of the Museum für Ostasiatische Kunst Berlin. The Chinese state could react to polders in two ways—either by stopping them from being built to protect others in the drainage system, or by encouraging them to be built as a way of increasing agricultural output—and did both, depending on the circumstances. Polder construction brought with it such problems as dikes obtruding into waterways to obstruct boat traffic7 or forcing up water levels and either flooding other unprotected lowlying fields or overwhelming smaller dikes elsewhere in the local drainage system.8 But polders also turned uncultivable land into good fields, so that in times and places where polders were long established, where land was scarce, production barely keeping up with need, and tax income vital, the state might intervene to make sure that the embankments on which local rice production relied did not weaken. In a county such as Changxing, and indeed throughout much of Jiangnan, the latter concern predominated. Poldering was one technique among many that Ming cultivators used to increase arable acreage. Unlike hill terracing and the opening of frontier lands, which brought marginal lands into production through the Ming, polders promised to wrest rich land from nature that could reach a level of productivity ordinary fields could not reach. Polder construction dated back two millennia to the Spring and Autumn period, though they became an extensive feature of the Jiangnan landscape only as of the Song.9 Wang Zhen in his Nong shu (Treatise on agriculture) of 1313 gives the first systematic description of poldered fields, noting among other things that polder walls were a device powerful families could use to enclose agricultural land and control the fields within.10 Three centuries later, Xu Guangqi (1562–1633) in his Nongzjieng quanshu (Complete handbook of agricultural management) devotes three full juan to proposals concerning

The chinese state in ming society

64

their construction and maintenance, plus a fourth on the technology required for the efficient operation of polder irrigation.11 The size of polders varied gready, from as small as a few mu to as high as tens of thousands of mu.12 So too, the individually owned plots within polders, separated from each other by an often intricate system of ridges and channels controlling the flow and distribution of water, varied greatly in size. Since polder construction, maintenance, and irrigation were labour-intensive, polder communities tended to organize their labour communally.13 By the Ming, polders were the most conspicuous constructed field system on the Jiangnan landscape. “The land east and west of the county seat through which the rivers flow is nothing but polders,” a resident of Gaochuan county, west of Lake Tai, could write late in the sixteenth century “On all four sides of the county seat, left, right, in front, and behind, is polder land.” Given their ubiquity, here as elsewhere in South Zhili and Zhejiang (see Map 2), rural people were commonly referred to as yumin, “people of the polders.”14 Their lives were organized around the methods that polder agriculture required and the resources it offered as much in the sixteenth century as in the twentieth. Describing a village in the polder region of Lake Tai in the 1930s, Fei Xiaotong notes that villagers lived at the margins of four of the eleven polders where they held their land. These four polders met at the intersection of three streams that were bridged to allow for easy access from one part of the village to another, thereby preventing segmentation from occurring among the different sections of the village. Nine of the eleven polders were entirely controlled by members of the village, whereas some of the land in the other two belonged to members of other villages.15 Polder and village boundaries were thus not always coterminous, yet villages were shaped by the topographical arrangements polders imposed. When the new Ming regime set up its administrative systems and units in the 1370s, it did not look out over a landscape as uniform as the Canadian prairies. There in the 1870s, following the passage of the Dominion Land Act, government

Map 2 The Lower Yangzi region (Jiangnan).

Taxing polders on the Yangzi Delta

65

surveyors were able to lay out a perfectly uniform grid of square townships. Land surveyors in 1370s China could not dream of imposing such uniformity on the landscape. Landscape was not only variable by natural topography but marked by agricultural practices, polders most conspicuously. Surveyors had to take these variations into account to establish fair tax assessments on each plot of land, and yet the plots had eventually to be aggregated into wards of roughly equal size. The lijia regulations of 1381 include no special considerations for polders. Villagers there as elsewhere were grouped into hundreds and tithings, which were aggregated upward into sectors (qu) overseen by tax captains (liangzhang), initially at a rate of one per sector. For the purpose of registering land for taxation, their plots were registered by wards (tu), aggregated upward into townships (du). Sectors The sector figures prominently in the analysis of the system that Ding Yuanjian saw recorded on the stone marker, and so it is with this unit that we must begin. Sectors were designated on the basis of the amount of tax grain the land within them yielded. Ten thousand shi was suggested as an appropriate quota for a sector, but most sectors were smaller, with quotas in the thousands of shi. This meant that a sector in practice could stretch over as few as half-a-dozen wards, or as many as eighty. As the onerous duties of the tax captain devolved onto the hundred captains (lizhang) from the mid-flfteenth century forward, sectors in some counties were drastically reduced or done away with altogether.16 Where sectors were maintained, their jurisdictions were frequently halved into subsectors, for which the terms shan (ledger) or jiao (corner) were used.17 In Jiangnan, sectors tended to stay current because they were used for more than organizing tax collection: to them became attached subsidiary functions connected with the supervision of water resources. Sectors that disappeared as tax captain units could come back into use just for water management purposes, though usually the needs of tax collection and hydraulic infrastructure dovetailed. As a tax captain was personally responsible for collecting taxes within his sector in full, he had an interest in involving himself in projects to enhance agricultural production, especially irrigation.18 This involvement may have begun on an emergency basis, as for instance in Suzhou’s Wujiang county, where an early-Ming governor made tax captains responsible for coordinating the relief work of lijia officers within their sectors in the event of flooding.19 Already in the Hongwu era, though, tax captains across the border in northern Zhejiang were being assisted by water control officers bearing the title of polder captain (weizhiang).20 Wujiang native Shi Jian dates the supervision of Suzhou’s water resources by tax captains to the Yongle era (1403–24).21 By 1430 at the latest, every tax captain in Suzhou had half-a-dozen polder elders (yulao) assisting him in the work of water control, with a general polder captain (zong yuzhang) overseeing the work and lesser polder captains (xiao yuzhang) acting at the local level. These incomplete references do not indicate that polders were formally designated as subcounty administrative units, nor do the polder captains seem to have been charged with anything other than maintaining hydraulic infrastructure.22 Yet some time early in the fifteenth century, polders and polder captains began serving as informal agents of

The chinese state in ming society

66

state administration, and polders as their jurisdictions.23 The 1444 gazetteer of Chongming county, the heavily diked island in the mouth of the Yangzi River then under the jurisdiction of Taicang guard (later subprefecture), indicates that the sector was being used as an administrative unit between the canton and the ward—parallel, in other words, with the township—which could point toward a growing reliance on the polder to organize fiscal operations.24 A century later, in 1548, the gazetteer of Taicang shows that polder embankments (tangpu) were being used to mark boundaries in sectoral land registers for the subprefecture.25 These references are only suggestive of polders being used as units of state administration. Whether they were or not, whether such practices dated back to before 1444 or were a later innovation, whether it represented a solution to a local problem rather than a reorganization of fiscal operations, is not clear. The shift these observations are anticipating—from the subordination of polders to subcounty units, to the subordination of subcounty units to polders—appears to have occurred first in Ding Yuanjian’s home prefecture of Huzhou in 1472. The best data on this shift is not from his home county of Changxing, however, but from neighbouring Wucheng, with which it shared the same scale of poldered fields.26 Prior to 1472, Wucheng was divided into thirteen cantons (xiang), sixty-one townships (eight urban (jie) and fifty-three rural (du)), and 282 wards (li), of which 10 were urban wards. In 1472, twenty-three sectors were laid over this structure of subcounty units. The reform did not affect the grouping of wards into townships, but the imposition of sectors did alter the arrangement of townships within cantons. Frequently, a sector took townships from the same canton, but not all the townships of that canton. A sector could also combine townships from different cantons, disregarding the boundary between them. For instance, Sector 5 was created by combining one township from Sanbei canton and one from Zhashui canton with all eight urban townships, and sector 6 combined the other three townships in Zhashui with three townships from Lingshou canton and one from Jiuyuan. This shuffling of units aimed for a registration system based on a principle other than the one that shaped hundreds into wards at the start of the dynasty; that is, of actual territoriality rather than of community. As Ding Yuanjian notes, the stone marker on the riverbank recorded the number of mu of fields inside the polder, which meant that the sector into which they were aggregated was defined not by the number of households registered in it (the lijia principle) but by the amount of land there. Ding admired the 1472 arrangement: “polders could be combined into sectors and sectors into counties.” His county had twelve sectors as opposed to Wucheng’s twenty-three; beneath them stretched dozens of polders, tagged according to the sequence of characters in the Thousand-Character Text (Qianzi wen). It was a system, Ding declares with approval, “simple and clear.” In this part of Jiangnan as of 1472, fields were being permitted to determine how the state at the county level was organizing its fiscal control of the territory within its jurisdiction. Altering registration from hundred/ward communities to the field boundaries that defined polders was a significant shift. Polders were anything but an arbitrary number of households. They were fields that had been patiently constructed, in some areas over almost four centuries, and that had come to determine where households lived and how communities formed. The embankments that made possible the high level of agricultural production in the region were not simply physical infrastructures within which production was carried on; they were boundaries of property and boundaries of

Taxing polders on the Yangzi Delta

67

community. The lijia regulations had assumed that tracking and organizing households would enable officials to track and organize fields without having to take physical boundaries into account, a misalignment that the Fish-Scale Registers were compiled to double-check. That was only part of what determined how much a household paid to the state; the other part was the service levy, calculated on the basis of the Yellow Registers, through which households were corveed for labour. In the case of polder agriculture, the equitable distribution of the service levy was crucially important, since all who farmed inside the polder depended on the maintenance of its walls and canals and the communal operation of irrigation works, a breakdown in any of which spelled ruin for the entire polder community. The 1472 reorganization acknowledged these conditions, which determined the productivity of the rice fields that produced the grain tax and fed the corvéed labourers. Assessing levies on the basis of polders was carried out to adjust tax levels to resource capacities and ensure that the labour needed to keep the polders in good repair was accessible. This favoured the state by improving the collectibility of levies; it also may have favoured poorer polder households by distributing the tax burden more equitably and contributing to the upkeep of polder walls, which was otherwise beyond their individual resources. What was done in Huzhou prefecture did not remain unique. Local magistrates in other counties with poldered fields followed suit, in different ways, although evidence for this is scattered and fragmentary. Across the provincial border, in Songjiang prefecture, the sector was the main intermediate unit of fiscal administration above the ward by early in the sixteenth century, and probably before.27 What the use of sectors signified is not entirely clear, however. Mark Elvin has assumed that the reliance on sectors “emerged from [the] confusion in the sixteenth century to become the usual method of organizing all but the largest and smallest [water-conservancy] works.”28 Rather than confusion, the shift to sectors may signal adaptation, for these sectors were being used not just to mark the boundaries of tax captains’ responsibilities for water control but to assess the service levy; and it is more than possible that their importance goes back into the fifteenth century. In Songjiang’s Taicang subprefecture, the gazetteer of 1548, previously cited, reports that land registration there was by polder. The compiler observes that, prior to the 1520s at least, land survey registers included a separate map for each polder, and that later registers included a summary map for every sector. Polders and sectors were now the units within which land was surveyed. There would not have been registration by polder unless that unit had come to bear an administrative value. The shift to the use of sectors and polders for land registration was not exclusive. Ongoing confusion about what sectors were, among other effects, indicates that wards and townships continued to exist alongside the new units. The editor of the Huzhou prefectural gazetteer of 1649 dutifully notes that “hundreds are grouped by sectors,” as one would expect of the old lijia/tax captain system, but he then follows that observation with the remark that “sectors divide up mu”; in other words, that sectors were used to register agricultural land rather than just households, just as Ding had noticed.29 The sector had migrated to another system.

The chinese state in ming society

68

Polders and the Single Whip What we have here is not a redundant doubling of registration systems, but the emergence of a parallel system measuring land by acreage, yet designed to organize not the land tax (that was still apportioned within wards) but the service levy According to the original Hongwu program, a household’s service levy—how much unpaid labour time it owed the state—was calculated on the basis of the number of adult males in that household. In polder areas, this system came to be regarded as either unfair or unworkable. The labour requirements for keeping a polder in good repair were heavy, though the benefits that polder cultivators enjoyed from a well-maintained infrastructure correspondingly high. The problem was free (or low-paying) riders. Households with much land but few adult males carried a light assessment for labour service, yet reaped great rewards by virtue of the amount of land they owned. The Hongwu system assumed that households with greater labour power were wealthier and had more land than households with less labour power, but in an active commercial land and labour market, this correspondence, if it ever obtained, quickly disappeared. Shifting to a household registration system that assessed a household’s service levy on the basis of the amount of land that the household owned within a polder accepted the principle that the wealthier should provide more support for infrastructure than the less wealthy, and understood that land was a better index of tax-payability than household size. It was not the wards that the polder system was replacing, but the lijia. The next step in this logic was to convert the service levy of land-rich households to a money payment, since these households did not have the labour to meet a higher labour levy. The eventual result would be that every household holding land in a polder acquitted its service levy obligations with a silver payment based simply on the amount of land it owned. This reorganization of the polder service levy was an early stage in the development of what late-Ming contemporaries, and historians today, refer to as the Single Whip reforms, a process through which the service levy would be totally eliminated, and the lijia system, in form or in spirit, cease to exist; where any remaining poll taxes would be combined with the land taxes and the taxpayer would discharge his fiscal obligations to the state in the form of one single, consolidated silver payment.30 Ray Huang’s definition includes features developed only later, yet it captures significant elements of the 1472 reorganization: the abandonment of the lijia principle of community assessment and the substitution of land for adult males as the basis for determining the service levy. It also looks forward to the eventual folding of that levy into the land tax, which the new levy was in fact simply replicating under another name. Little of this was visible to Ding Yuanjian, as he gazed puzzled at the stone markers a century and half after they were set up. Not much more comes to light when we pore over Huzhou’s records. To fill out the picture, we have to look at what officials were doing elsewhere in the region to reform tax levies on polders. In his compilation of local texts

Taxing polders on the Yangzi Delta

69

from the mid-seventeenth century, Gu Yanwu includes proposals that two regional officials in South Zhili made in 1494 on these matters. Yao Wenhao (js. 1484) was a Ministry of Works official who served as assistant prefect in charge of water resources in Changzhou, the next prefecture northwest of Suzhou, from 1491 to 1494. On the basis of that experience, Yao promulgated regulations for polder land in the seven prefectures on the Yangzi Delta.31 His text indicates that land was registered by both its polder and its old ward. Yao wanted to impose uniformity. A ward having ten polders was henceforth to be subdivided such that each polder was a tithing. (That this correlation was feasible indicates that Changzhou had a topography of relatively small polders.) If a ward had fewer than ten polders, then the larger polders should be subdivided into tithings; if more than ten, then the smaller ones were to be combined into a single tithing. Yao’s purpose in aligning polders to tithings was to ensure an adequate supply of service labour. Using a modified lijia terminology, he assigned a “tax elder” to supervise the water control facilities of each sector, and alternate hundred captains (painian) to oversee the upkeep of the polder-tithings. And every polder in the system would be given an identification character and marked with a stone tablet—precisely of the sort Ding describes, though differently inscribed: these should give the township and ward within which it was located and the name of the alternate hundred captain in charge. Yao’s use of lijia officers to organize hydraulic maintenance on the basis of the old lijia tithings, and his registration of polders by township rather than sector, shows him to be adapting the traditional service levy to the particular needs of polder agriculture without moving fully to the newer system. Still, his suggestion that polders and tithings be made to correspond signals that the polder was regarded as the significant unit of fiscal registration. Jin Cao, a Songjiang sheng yuan who worked with Yao, presented his own set of proposals at this time concerning the maintenance of waterways in Songjiang, the easternmost of the seven Yangzi Delta prefectures. Like Yao, Jin sought to redefine the duties of the old lijia officers and establish new guidelines for assessing households for polder labour. Responsibility for the upkeep of polder embankments he invested with the alternate hundred captain from each tithing (paizhang), proposing that the captain’s name appear on a stone marker set up on the south embankment of each polder. Like Yao, Jin assumed that a polder was roughly the same size as the old tithing, and on that assumption assumed that the duties of the old lijia system could be transferred over to the polder unit. Above the polders, overall responsibility for maintaining the embankments within a ward still lay with the hundred captain (lizhang), but his subunits would be the real polders, not the abstract tithings. This shift is underscored by Jin’s recommendation that official land registers for a county be compiled on the basis of the records that the alternate hundred captain kept at the polder level.32 In other words, the dual system of recording households in Yellow Registers and land in Fish-Scale Registers would continue, but the polder would now be the unit within which both households and plots of land would be registered. This arrangement confirms what the Huzhou reorganization only implied: that the polder was being used to organize the service levy but also that it was now being used to levy the land tax, thereby transferring all of a household’s tax assessments, in Single Whip style, to its landholdings.33

The chinese state in ming society

70

Sectors and the Single Whip The challenge for the reform-minded magistrate was how to incorporate the changes being made at the level of the polder into his county’s fiscal operations. Suzhou’s Changshu county provides some evidence of how this was done in one county. Contemporaries liked to describe the perfect grid that intersecting waterways inscribed on the landscape of Changshu as looking as neat as the squares on a chessboard.34 We get the same impression looking at the eighty-six maps in Changshu xian shuili quanshu (Complete handbook of water resources in Ghangshu county), which magistrate Gengju published in 1606 as a guide for carrying out fiscal reforms in the county. Figure 3.2 shows the first map, a composite of the entire county. When Geng arrived in Changshu, he found the standard structure of canton, township, and ward; he also found a parallel structure, in which “the larger townships are divided into subsectors (shan) so as to group the individual wards, and subsectors are paired into what are known as sectors (qu).”35 His maps show both, some sectors occupying the area of an old township, others in more densely populated areas half or a third of a township. The eighty-three sectors outside the county seat were in turn subdivided into polders. Figure 3.3 is the map of Sector 34, which doubled as Township 29, on which at least twenty-three physical polders have been drawn. Geng was not the author of this arrangement, only its cartographer.36 The author of the Changshu system of sectors was an earlier county magistrate, Feng Rubi (js. 1532). In 1537, Feng had all the arable land in Changshu resurveyed and recorded by polder. He did not transfer the regular service levy to the new polder assessments. Rather, he left the older township/ward assessments intact to levy the regular taxes, and set up a separate system of accounting based on the polders when he needed to levy labour and funds to do repairs to the polders. Thenceforth Changshu’s fiscal operations flowed in two streams, and townships and sectors coexisted to keep both flowing. The same process appears to have gone on elsewhere in Suzhou prefecture in the late Ming, for the land records of Sheng’en Monastery in Suzhou’s Wu county show that the prior was keeping track of the monastery’s property both in terms of named polders and by the numbered units of the old township/ward system.37 This is the sort of administrative compromise that not just officials found themselves having to make on the way to implementing the Single Whip reforms.

Taxing polders on the Yangzi Delta

71

Figure 3.2 The division of Changshu county by sector, 1606. Gengju, Changshu xian shuili quanshu (1606), 3: zong, 7b–8a. Source: This and the following map are reproduced courtesy of the Tōyō Bunko, Tokyo.

The chinese state in ming society

72

Figure 3.3 Sector 34 of Changshu county. The wards are numbered, the polders marked out by the water channels. Changshu xian shuili quanshu, 5.74b–75b. Feng’s reforms failed to ensure that state levies kept up with the pace of physical deterioration, however, for the 1571 gazetteer of Suzhou’s Changzhou county reports that the navigable rivers in the region had been rising for the previous half-century, that the pressure on poor cultivators to keep polders in good shape had become intolerable, and that flooding had become endemic.38 When Regional Inspector Lin Yingxun (js. 1571) arrived in the province six years later, he found that the service levy was not keeping up with the hydrographic pressures on the delta. He proposed that local magistrates should in effect do what Feng had done (though he does not mention Feng by name in his recommendations): divide sectors into polders and organize the labour needed at the

Taxing polders on the Yangzi Delta

73

polder level. His plan equated a polder with a ward, not with a tithing as Yao Wenhao and Jin Cao had done eight decades earlier. Each polder should have a polder captain to supervise the work for keeping the polder walls in good repair, with an embankment captain overseeing the work at the level of the sector.39 Lin’s proposals were part of a larger trend, for there is ample evidence from the 1570s and 1580s that this sort of reorganization of the service levy by polder was being carried out throughout the Lower Yangzi Delta, from Yangzhou prefecture north of the river40 to Jiaxing to the south.41 They did not receive unanimous support, however. The provincial governor, Song Yiwang (js. 1547), decided that year to take a different tack and shore up the old system by reaffirming that fiscal levies be collected by ward, not by sector and polder. This was known as the “ward remittance” (tuyun) system. The only evidence I have found regarding implementation is from Changzhou prefecture, and only one county there, Yixing, put it into operation. The magistrates of the other four counties in the prefecture refused to cooperate and give up the polders they were using. One argued that wards were not equitable units, and that the variation in population from one to the next made it impossible to distribute the tax burden fairly.42 The order to levy taxes by ward remittance was cancelled in 1593. In that year, Wujin county, the prefectural seat of Changzhou, was divided into thirteen sectors.43 These sectors were used to aggregate local tax collection (twenty tax captains were appointed per sector) as well as to coordinate hydraulic maintenance. In order to make assessments for levying labour and funds to dredge waterways, he redefined hundreds into fixed, uniform units of 540 mu, and systematically recorded all polders, water channels, riverbeds, and grave land within each fen (or tenth part—effectively, a tithing) of every redefined hundred.44 Although the polder was acknowledged as the basic unit of survey and assessment, taxes were still formally recorded by wards. So Wujin did not shift completely to the sectors and polders, operating, at least formally, a double system something like what Feng Rubi introduced in Ghangshu county. The same reorganization went on elsewhere in Changzhou prefecture that year.45 In Wuxi county, the magistrate established thirteen sectors and twenty-six subsectors (here called jia) and appointed ten tax captains in each subsector to oversee tax collection.46 From county and canton gazetteers, it is possible to reconstruct the switch from Wuxi’s older townships to the newer sectors. The old structure was made up of sixty-four townships (sixty rural and four urban), arranged within twenty-two cantons. The new structure of units consisted of fourteen sectors, which took their names from fourteen of the old cantons. Twenty of the old townships were kept intact and redesignated as subsectors. The other forty-four were broken up and redistributed. Twenty-seven were split between the two subsectors of the same sector, but the seventeen were redistributed among different sectors. Take Township 1, for example: its eleven wards were divided between Subsector B of Tianshou Sector, which took four, and Subsector A of Jingyun Sector, which took seven. Subsector B of Tianshou Sector also received four of the nine wards in Township 2, while Subsector A of Jingyun Sector was enlarged to include four of the nine wards in Township 57, one of the seven wards in Township 58, three of the six in Township 59, all seven wards of Township 60, and two of the eight in the old First Urban Township. Among other effects, this way of constituting Subsector A of Jingyun Sector cut across the urban-rural boundary in the original township structure of the fourteenth century, presumably reflecting the urbanization of the intervening 220-odd years.47

The chinese state in ming society

74

The impression one gets from this translation of townships into sectors in 1593 is that tax administrators in Wuxi abandoned the old township system in favour of the new sectors. If local gazetteers continued to record the old system, as they did, it was purely as a set of spatial coordinates rather than as a working framework for anything anyone was actually doing. The editor of Wuxi’s eighteenth-century gazetteer explains that the context for this reorganization of the county’s fiscal system around polders was the promotion of the Single Whip reforms. As he put it, the new system “took land as the mother” of taxes and the household as the “child.”48 The other way around—which is what the old lijia system had done—was what Single Whip reformers were abandoning in their search for a fairer and more efficient tax system. The extent of boundary redrawing in Wuxi argues that the county magistrate did a thorough job of remapping the county’s administrative structure so that it meshed better with actual residential and field conditions. If Wanli-era magistrates hoped to make such remapping work, they would need to beef up the power of the local officers entrusted with adapting the tax system to the sectors. This is what Geng Ju did in Ghangshu county in 1606, to return to the case with which we began, which was to endow the local officers he called sectoral administrators (qu gongzheng) with the power to force everyone who owned land in his sector to sign a tax contract.49 That contract obliged the owner to pay for the labour needed to keep the hydraulic infrastructure in repair, transferring fiscal responsibility down to the local level where it was most needed but where tax moneys were least likely to return once they had been siphoned upward. It also effectively confirmed both that the sectors were the main fiscal units in a polder economy and that landownership was the basis of the service levy. Not every site of polder-based reform in Jiangnan achieved the thoroughness of the Wuxi and Changshu cases—which is why, among other reasons, the Single Whip reforms were so long in getting implemented. A county where boundary redrawing can be reconstructed to the same level of detail is Wucheng, next door to Ding Yuanjian’s home county of Changxing, but the results give a very different first impression. As already noted, the internal fiscal boundaries of Wucheng were thoroughly redone on the basis of polders in the 1472 reorganization. And yet when a Wucheng magistrate in the Wanli era had to shuffle these polders into sectors, he did nothing but keep all his old township boundaries intact and simply slide the townships out of their old cantons into the new sectors, without changing anything besides nomenclature.50 To be fair, he may not have needed to do more than a mild reshuffle to get his wards in line with the new structure, given the success of the 1472 resurvey. Yet the magistrate may have done what most local officials whom the state compelled to carry out any fundamental reorganization would do: switch around what was already there into an array that looked something like what his superiors wanted to see. Let compliance appear to have occurred, whether it had or not. This apparent sleight of hand is one of the reasons it is so hard to read down to social practices from Ming state documents. Are we looking at how people were really organizing their lives, or only at the orderly casing that some officials used to wrap social practices to make them look other than they really were? Are we seeing the state in action or in inaction? Only by digging more deeply into local history will we ever be able to answer that question, though even then the answer can only be local—which is probably all it should be.51

Taxing polders on the Yangzi Delta

75

State success/failure By 1620, Ding Yuanjian was already seeing the first innovation of reorganizing taxation by polders through the mists of time. If the admirable order of the original polder-andsector system filled him with nostalgia, it was in part because whatever simplifications and rationalities the system was designed to introduce had not been achieved or sustained. The 1472 reorganization had been an early step toward the larger fiscal reforms of the Single Whip designed to improve the state’s tax base in Jiangnan. Yet the stone marker at Lüshan appeared now to Ding as nothing but the detritus of an initiative that had failed. Why the nostalgia? What had happened between 1472 and his present? Clearly the polders did not disappear. Nor did the state revert to the old lijia system of registering land. What had changed, it seems, was not the polders but their management—for which the reform was both a means and a symbol. There are two places to look for the decline of polder management in Ding’s time. One is the withdrawal of the state from local social functioning. For this, Taicang subprefecture indirectly provides an intriguing example. Some twenty-five years after Ding’s boat trip took him past Lüshan, when the invading forces of the Manchus were sweeping southward toward Jiangnan, Taicang native Chen Hu fled with his father to a village in neighbouring Kunshan county. There he drew up a set of rules to regulate the local public management of polders around the village. His rules call for registers to be compiled polder by polder. These would then be forwarded to officers at the sectoral and county levels: a matter of reporting that was de facto a matter of expecting official authorization after the fact. The authority to supervise the management of the polders Chen gave to local officers he called field-tithing heads (tianjia). These heads were appointed not by the county office, however, which at this unsettled time was probably incapable of exerting authority in the countryside, but by the local landlords.52 In other words, his rules understood that the state was no longer maintaining the systems needed to keep polders in good repair, and that the maintenance of fields could only be assured by relying on the leadership of the local elite, who controlled the appointment of field-tithing heads. Although Chen drew up his plan in a time of particular disorder, it posits a condition of deteriorating state supervision that had been deepening for several decades in Jiangnan. To the extent that the reforms carried out in Jiangnan in 1593 were a response by local officials to the dwindling capacity of the state to raise labour and intervene effectively in rural society, Chen’s presumption was sound. A decade earlier, Taicang native Lu Shiyi had been advocating the abandonment of Yellow Registers in favour of tax assessment calculated purely from the Fish-Scale Registers for just this reason. Polders, he claimed, had become particularly troublesome sites for evading tax responsibility; only an assessment system based completely on actual landowning could end these abuses and generate the revenue needed to keep polders in repair.53 Lu’s assumption is that the state should be the agent for this work—as Ding Yuanjian imagined it was doing so well back in 1472—but his observation suggests it was failing to do so. A different explanation for this situation can be imagined, however: that the markers represented a time when polder society functioned on the basis of a mechanical solidarity in which rich and poor were not so desperately polarized, when locally rooted elites

The chinese state in ming society

76

accepted their responsibility to serve the needs of the community, and when the state was confident of leaving the public interest in local hands. This way of reading Ding’s nostalgia may neglect the fact that the markers were set up in a move to reorganize Changxing’s fiscal assessment, not to keep it the way it was, and it ignores the history of inequitable levies that must have prompted the move. What is plausible about this interpretation, however, is that it takes notice of the polarization of rich and poor which did indeed increase between 1472 and 1620. However illusory Ding’s paternalist image of fifteenth-century rural life, he could hold it up as a rebuke to the harsh conditions of rural life in his own day, when larger landlords were moving to the cities and taking with them whatever resources and leadership they might once have invested locally, and when the state seemed incapable of reordering community resources other than when an emergency struck. The intermittent attempts of local officials to revise levies in order to see the polders maintained could thus be as much about elite landlord absenteeism and the abrogation of community responsibility as about the incapacity of the state to intervene effectively in local affairs.54 We may not need to choose between these interpretations. Both ways of apprehending the problem—gentry withdrawal from liturgical responsibility versus state incapacity— were visible during Ding’s lifetime, locked in a dialogue of competition for resources and authority that neither could afford to win at the complete expense of the other. Where either the gentry or local officials took their Confucian duties to serve the public interest seriously, the competition need not come at a cost to local society. When it did impose that cost, however, polders were a public good that stood to suffer more conspicuously, and produce more disastrous immediate effects, than crumbling city walls or rutted roads. The history of polder taxation in the Qing contrasts with what happened in the Ming. Ming attempts to encourage the sustainability and taxability of polders were mostly initiated from below, at the county and prefectural level, moving up to the provincial level during the heyday of the Single Whip reforms in the Wanli era. The congeries of policies and systems that came about as the result of local initiatives, involving much lateral borrowing among local officials, was largely replaced in the Qing by a “strong” state operating under upper-level provincial and capital initiatives. This, at least, is the impression one gleans from a quick survey of Qing sources on polder taxation. The first major fiscal reorganization the Qing state imposed on the polder regions of Jiangnan in 1664 was the juntian junyi (equalization of fields and service levy) system. This system finally redefined all the old lijia units by acreage rather than by households, and then aggregated them in turn into new territorial sectors that did not conform to the sectors of the old tax captain system; in Songjiang’s Lou county (hived off from Huating county in 1656), local registrars were specifically warned not to use the “old sectors” when drawing up registers for the new sectors.55 The reassessment excited resistance from those landowners who had benefited from existing undetected evasions; in Lou county, the resistance was so fierce that the magistrate’s opponents managed to have him removed from office, though not the system he was mandated to put in place.56 This equalization was further deepened with the shunzhuang (village rotation) system, first developed by provincial officials in Ding Yuanjian’s native Zhejiang in 1727, and imposed as the registration standard throughout the region in 1733. Assessment by village required that tax registration be done entirely on the basis of actual residential villages. Each village (now administratively called a zhuang) was to organize families into baojia-style tithings.

Taxing polders on the Yangzi Delta

77

A register was to pass first of all among the ten member families of each tithing, then tithing by tithing through the village. Each family was expected to fill in complete information about its landholdings, wherever they might be, and circulate its report for public inspection. The system recognized polder communities as “villages” at long last, though it too fell short of the ever elusive goal of getting all taxable land acknowledged and paid for.57 In her study of rice economies, Francesca Bray has shown that conflicts within water control systems are better handled by codes of customary rights and local enforcement than by the state.58 The sources from which I have written this account of polder taxation record nothing about customary rights or about the communities that those who lived on polder embankments and culti-vated the fields within them constituted. Being records of state administration, their concern is to document the mechanisms administrators used to assess household levies and land taxes and see that the field walls on which polder agriculture relied stayed standing. These were the two goals of the conscientious local official: to incorporate polder cultivators into the mandated social units of Ming administration while ensuring that the fields in them were taxed adequately and fairly, and to do what he could to guarantee that the productivity of polder land was maintained while striving to distribute the cost burden equitably. In doing so, the magistrate was responding to the multiple existence of polders: as arable fields, as fiscal entities, and as lived communities. Each of these required different responses, depending on which the state perceived to be under threat. But magistrates were limited in what they could do. They came as outsiders, acting in the interests of the state more than of the local community, should these interests diverge. What they could do was circumscribed by the limits of their knowledge of and influence over real social relations, and success depended on the ability to work within local customs, local property relations, and the social networks through which economic opportunities were held and traded. It also depended on the willingness of a magistrate to take the problem of tax equity as seriously as he did the problem of meeting his tax quota, and to look across his borders at what others were doing to bring the two into tolerable accommodation. Perhaps what so impressed Ding about those stone markers was simply that they got put on the south bank of every polder around Lüshan village. It was something to be nostalgic about in more complicated times when neither the state, nor cultivators, nor the local elite had the upper hand. Accepting Ding’s nostalgia does not lead unerringly to the conclusion that the state was removing itself from Ming society, leaving its agents adrift either to do nothing as long as they could or else to come up with ad hoc, short-term solutions. This is certainly one theme struck by Wanli-era writers, who looked about them and, like one despondent contributor to a 1571 Suzhou gazetteer, worried darkly that if those above do nothing, then those below will have nothing to carry out, with the result that benefits cannot be provided and disaster cannot be alleviated. With tax receipts falling and the people in difficulty, how will it all end, I wonder?59 Contrary to those who liked to dramatize the decline and fall of the dynasty, it did not all end. Responsible officials continued to intervene and attempt to work out solutions to the

The chinese state in ming society

78

almost intractable problem of using state levies to maintain the viability of privately owned fields—which they had to in order to keep up the high tax income from Jiangnan flowing to the central government. Deciding whether their responses to hydraulic pressures are evidence that the Chinese state in late-Ming times was relatively weak or relatively strong depends on what we choose to place it in relation to. Relative to the Hongwu state, with its extraordinary capacity to devise and implement fundamental reforms, the Wanli state looks weak. On the other hand, relative to the complexities of late-Ming economy and society, which produced administrative challenges and moral conundrums far exceeding anything the Hongwu emperor had to deal with, the regional and local officials who worked to bring the dynasty’s fossilized fiscal operations into sync with the economic and social realities on the ground were engaged in huge tasks. Given the scale of their work, it seems invidious to dismiss the Wanli state as the hopelessly mired and inept operation that post-Ming accoimts of it have liked to depict. Such accounts do so because they are focused on a very different issue, which is to explain the failure to withstand the combined onslaught of rebellion, contagion, famine, and invasion three decades later, the reasons for all of which fell much farther from the Wanli tree than Qing conservatives were able to admit. Rather than struggle over whether the late-Ming state was strong or weak, it might be more productive that we recognize, more simply, that the Chinese state’s response to polder society changed over the course of the dynasty as it strove to adjust its administrative operations to the unprecedented scale and complexity of the realm, and that the capacity to change and adapt, however incompletely from our perspective, may be the best standard by which to measure what some officials actually did.

4 Growing rice in North Zhili Early in the fourteenth century, Wang Zhen proposed in his Nong shu (Treatise on agriculture, 1313) that Chinese agriculture divided at the Huai River: the land to the north of the Huai was suited to millet, the land to the south to rice.1 This was a reasonable characterization, and one that has continued to be accepted down to the present. During the Ming dynasty, however, attempts to spread rice cultivation into north China challenged the absolute validity of this tenet. At the close of the Ming, the prominent official Xu Guangqi (1562–1633) in his Nongzjieng quanshu (Complete handbook of agricultural management) criticized Wang Zhen for his view and insisted that rice could be grown in north China as well. “There could not be more than one or two places in a hundred where the—topographical character is unsuitable,” he insisted.2 The spread of rice ciiltivation into North Zhili, the metropolitan region around Beijing (known as Zhili in the Qing and Hebei province today) began on a significant scale during the latter half of the sixteenth century. Zou Yuanbiao (1551–1624), who served in Beijing in the 1580s and then again in the early 1620s, recalled during the latter period of service that “thirty years ago the people of the capital did not know what rice plants were, but now they are everywhere.”3 Xu Guangqi was in the capital at the time, and was himself experimenting with southern riziculture techniques on his farm in Tianjin.4 He hoped that this experiment would encourage northern cultivators to switch from dryland grains to rice, thereby increasing grain production in the north. Gu Yanwu (1613–1682) was an equally strong advocate of rice in North Zhili, believing it to be a viable way of easing the strain on food supplies for the large number of soldiers stationed in north China.5 The transport of military grain from south China, where it was mosdy grown, to north China, where it was needed to support the large garrisons on the border, was a large, long-standing charge on the imperial state. Altering the economic relationship of north and south would ease the great burden of supplying the garrisons defending the northern border. Advocating rice agriculture in north China became something of a tradition that carried on to Lin Zexu (1785–1850), better known as the special commissioner who negotiated with the British to cease their opium trade in Guangzhou. As he stated optimistically in the title of one of his proposals for irrigation, “The soil of Zhili is by nature suited to rice, and any place with water can be made into paddy fields.”6 By Lin Zexu’s time, though, rice was no longer spreading into new areas of North Zhili; indeed, it was disappearing from some areas where it had been cultivated previously. The low level of northern rice production in the late-Qing and Republican periods has obscured the incidence of rice cultivation in north China in the Ming and early Qing. That conceded, however, paddy was never a dominant component in northern Chinese agriculture, despite the efforts of Lin Zexu, Gu Yanwu, and Xu Guangqi. Why did these officials strive to make a region ecologically unfavourable to rice production a

The chinese state in ming society

80

rice-growing region? And what technical and social factors worked against the diffusion of riziculture and the transfer of rice technology into north China? Conditions of grain agriculture in North Zhili In his survey of Chinese agriculture in the 1930s, John Buck follows Wang Zhen’s division of China’s agricultural regions at the Huai River. He expressed the division in climatic rather than geographical terms, drawing the boundary at the 80 cm isohyet, the line of precipitation north of which less than 80 cm of precipitation falls. This line runs through northern Jiangsu (the Huai Valley), central Anhui, and southern Henan, extending finally to the Qinling mountain range. The region south of the line Buck called the Yangzi rice-wheat area, and that to the north he labelled the winter wheat-gaoliang area.7 For the latter region, North Zhili gazetteers consistently record that the grains grown there include not just wheat and gaoliang but large-panicled or proso millet (shu), small-panicled or broom-corn millet (ji), small-grained millet (su),8 barley (damai or moumai), buckwheat (qiaomai), and sorghum (shushu). Oats (yanmai or tiaomai) and pearl barley (yiyi or caoshu) also appear in gazetteers, though only in certain prefectures: oats in Hejian and Xuanhua, and pearl barley in Shuntian and Yongping. According to Qjng gazetteers, corn (yumi) was not grown outside the Tianjin area, and only then in the nineteenth century. North Zhili gazetteers also attest that a wide variety of rice was grown within the province. They record over a thousand different names, according to one commentator in 1824.9 In general these names can be reduced to four categories: wet and dry, and glutinous and non-glutinous. Wet and dry rice require different methods of cultivation. Wet rice (shuidao) is grown in paddy fields submerged from the time the rice is planted or transplanted. Dry rice (handao) is grown in dry fields, though it requires an assured supply of water for three to four months. Its yield per cultivated area is roughly twothirds that of wet rice. In Japan today, dry rice accounts for only about 5 percent of total paddy area, a percentage that is typical of its secondary importance in world rice production.10 Reports concerning the extent of dry-rice cultivation in historical North Zhili are contradictory. On the one hand, scholars such as Wang Zhen in the Yuan and Li Shizhen (1518–1593) in the Ming were of the opinion that dry rice was the only kind of rice one could find grown in north China. 11 A number of Qing gazetteers share this mistaken view.12 At one point in his Complete Handbook, Xu Guangqi appears to concur, noting that “today many people in the north grow it “and citing several counties in both North Zhili and Shandong.13 On the other hand, elsewhere in his book Xu expresses the opposite view—that dry rice was not widely cultivated—when he remarks that “perhaps northerners do not know how to cultivate dry-paddy fields.”14 In fact, the majority of North Zhili gazetteers emphasize wet paddy over dry. Only thirteen Ming or Qing gazetteers I have examined mention dry rice, and many of those stress that it was grown on a limited scale. Wet paddy appears to have been the preferred rice crop. Glutinous and non-glutinous rice are grown for different consumption purposes. Nonglutinous rice (jing and xiari) is a food grain, whereas glutinous rice (nuo) is used for making specially prepared foods and liquor.15 Twenty-two counties and subprefectures in North Zhili report cultivating both, though by far the more important variety was non-

Growing rice in North Zhili

81

glutinous rice because of its nutritional value as a higher-yielding food grain. Where glutinous rice was not grown, glutinous millet (shu) was used for making liquor.16 North Zhili was not an obvious place to think of growing rice. South of the 80 cm isohyet, average annual precipitation is 105.9 cm and the number of frostsafe days is 243. Buck found that 74 percent of the farms in this area grew rice at least once a year, on 56.8 percent of the total crop area. By contrast the area north of the line has on the average 59.2 cm of precipitation and only 191 frostsafe days. Buck found at the time of his survey that only 2% of the farms in north China grew rice, on a negligible 0.7% of the total crop area.17 The minimum average temperature throughout the life of the rice plant should not go below 20°C., yet in the northern half of the province the average temperature stays above 20°C for only three months: June, July, and August. Temperatures in the southeastern half of the province are milder, so the starting time for rice may be pushed back to 10 April, as is the case in much of Jiangsu, and the flowering time can go a little past the end of August.18 The gazetteer for the southward subprefecture of Cizhou reports that both glutinous and non-glutinous rice were planted in the fifth month (June) and harvested in the ninth (October).19 As one moves further north, the period within which rice can be cultivated narrows. June there is a reasonable planting time, but October is too late for harvesting. Early-ripening varieties are therefore necessary, and several are mentioned in gazetteers. A strain called xiaoqing dao (“small green rice”) is listed as an early-ripening rice in a 1544 gazetteer, and duo maichang (“first to the threshing ground”) is mentioned in 1612.20 The most commonly cited early-ripening strain is yudao (“imperial rice”) developed in one of the imperial gardens north of Beijing. The Kangxi emperor claimed that it ripened as early as the end of the sixth month (July) and advocated that it be used widely throughout the north, even double-cropped.21 More restricting than temperature on northern riziculture is precipitation. In the course of a normal year, the province begins to get rain more than ten days a month only in June, and heavy rains fall only in July and August. By September the level of precipitation drops sharply and remains low through the next eight months. The timing of the summer rains, however, is highly unreliable. They require the coincidence of two factors, moisture-laden winds from the southeast and cold air from the north, that on meeting cause the southern air to release its moisture. The only way to ensure the water volume essential for wet-paddy cultivation is irrigation. As we shall see, this was the single most important factor limiting the spread of rice in north China. The need for irrigation in turn means having to control the flow of water to the fields, a task that was not performed easily with the silt-laden rivers of North Zhili, as a Baoding writer explains: Water in the south is mostly clear, whereas water in the north is usually muddy. Clear water flows in a stable way but muddy water shifts its course unpredictably. Northern water is strong and northern soil is loose: when a strong current encounters loose soil, it erodes quickly, so that no benefits can be enjoyed for long…. All of the rivers and large canals in Zhili start in the northwest and cut channels across the terrain, down which they will flow through the drifting sand and the alkaline soil, one part mud for every ten parts water. In the deeper channels the water is shallow, and the shallow ones are filled to the top. The lower reaches are

The chinese state in ming society

82

even more readily filled up and blocked off, so that it becomes impossible to maintain stable drainage.22 The instability of drainage on the North China Plain was so severe as to cause vast areas of farmland to be inundated when a river spilled its banks. The lowlying land most suited to paddy-field construction was also in constant danger of waterlogging, since excess water is not readily drained off by gravity flow. As one seventeenth-century gazetteer compiler complained, “Rent and plant [paddy land], and nine times out of ten it gets waterlogged.”23 Soil quality is not a significant limiting factor on riziculture. With the exception of the southeast quarter of the province and the Ming prefectures of Baoding and Hejian, the soil of Hebei is quite fertile. The provincial gazetteer of 1735 praises the rice-growing regions of Baoan and Huailai for having soils rich enough to dispense with fertilizing altogether,24 though other gazetteers complain that the fertilizer requirements for paddy cultivation were too high. The main soil problem, especially along the Bohai coast and in the central southeast quarter, was not lack of nutrients but high alkalinity or salt levels. This is why the Cangzhou gazetteer reports that “the soil is rather unsuitable” and relegates rice to the end of its list of local products.25 Although certain varieties of paddy flourish in a slightly alkaline soil, excessive alkalinity can poison rice. Alkalinity can be reduced to cultivable levels by digging drainage canals and then flooding the land with fresh water to flush out the salts in the upper layers of the soil.26 Careless conversion from dry fields to wet-paddy fields, however, can damage them by increasing soil alkalinity. A poorly constructed drainage system can end up depositing the salts in the fields it irrigates back into the soil, leading eventually to salt poisoning.27 Also, drainage water that does flush salts from one field may soak them into another unless adequate care is taken in planning and building drainage ditches. These problems of climate and soil affected varieties of rice in different ways, and thereby influenced the selections cultivators made for local growing. In extreme cases, the physical conditions of North Zhili during the Ming and Qing made riziculture so difficult—requiring amounts of labour, water, and fertilizer far greater than millet or even wheat demanded—that it was not profitable to pursue. But in the majority of counties and subprefectures, investments could compensate to make rice a local crop. Statecraft activists believed in the value of such investments and called for them to be made. Rice cultivation in North Zhili before the Ming In his study of Chinese agriculture in Ming and Qing China, Dwight Perkins has accepted the observations of Wang Zhen and John Buck concerning the north-ernmost boundary of wet-rice cultivation. Varying only slightly from Buck’s map of grain crops in the 1930s, Perkins draws the boundary of eighteenth-century rice cultivation across the northern parts of Jiangsu and Anhui provinces. “If there was any extension of rice cultivation,” he says, “it involved the spread of paddy acreage in transitional regions where rice was already planted,” such as the section of Jiangsu north of the Huai River. As for substituting rice for loweryield millet and sorghum crops, Perkins judges, “it is clear that Chinese efforts in this direction have been quite minimal.”28 The history of North Zhili

Growing rice in North Zhili

83

agriculture might broadly confirm this general statement, yet it also shows that private and state efforts were made repeatedly to create and increase paddy acreage in the struggle to increase food production in the north. Even before the Ming dynasty, sporadic attempts were made to introduce rice. During the first half of the first century, 800 qing (80,000 mu) of paddy fields were opened in the border territory of Yuyang, now Miyun county northeast of present-day Beijing.29 Rice is said to have been grown along the Gulf of Bohai in the Jin dynasty,30 and famine victims there ate it in the Sui dynasty and again in the Tang, in the year 876.31 A Tang official, Pei Xingjian (619–82), initiated a large water control project using the water of the present-day Yongding River to irrigate several thousand qing of paddy, resulting in increased agricultural production.32 Similar projects were also carried out under the Song, in particular during the reign of the second Song emperor Taizong (r. 976–97), who advocated diversifying grain production in both the north and the south and called on the north to develop irrigation and grow rice.33 In 989, He Chengju converted flooded areas in the northern part of the province into irrigation ponds and used military labour to build paddy fields on a large scale, although problems with late-ripening varieties and the resentment of his soldiers at having to do agricultural labour brought the plan to an end.34 After a hiatus of several centuries when Khitans and Jurchens occupied the north, rice was again cultivated. According to a note in the gazetteer of Yongping prefecture quoted by Gu Yanwu, paddy agriculture in this far north-eastern corner of the province began in 1262, when a canal was completed and migrants were issued with two oxen each to plough their paddy fields.35 Several Yuan officials—Guo Shoujing, Yu Ji (1272–1348), and Tuotuo—put forward proposals to bring the northern stretch of the North China Plain and the Bohai coastal area under rice cultivation.36 The largest project within the vicinity of Beijing, advocated by Tuotuo, converted all state land, military land, and uncultivated land to the north and northwest of the city to paddy fields. Part of this project involved the draining of the Haidian Marsh, now the site of Peking University. This land was rented out to tenants, and a bumper rice harvest reported for the year 1352.37 A similar project was undertaken at the south erid of the province in the late Yuan, when new channels were dug for the Zhang, Fu, and Feng Rivers and paddy fields constructed to utilize their water.38 These records reveal that cultivation was concentrated in three areas before the Ming: Beijing, the Bohai coast, and the far south end of the province. These would be the areas where rice first appeared, or reappeared, in the Ming. The appearance of rice in the Ming and Qing periods To reconstruct the emergence of rice agriculture in North Zhili in the Ming and Qing, I have culled references to rice being either present or absent in 300 provincial, prefectural, subprefectural, and county gazetteers for places within the Qing boundaries of the province, dated between 1373 and 1900, plus the provincial gazetteer of 1910.39 In most of these gazetteers, the only indication of rice being cultivated is its inclusion in the list of grains at the head of the section entitled “Material Products” (wuchan) or “Local Products” (tuchan). In twenty-one gazetteers that did not have grain lists, sections such as “Customs” (fengsu) or “Water Control” (shuili) were consulted. Additional information has been culled from the dynastic history (Ming shi) and the writings of Xu Zhenming (js.

The chinese state in ming society

84

1572, d. 1590), Xu Changzuo (1558–1609), and Liu Tong (d. 1637), in addition to the essays of Xu Guangqi, Gu Yanwu, and Lin Zexu already cited. Given their dependence on the publishing cycle of local gazetteers (ideally, every sixty years), the sources do not date rice’s arrival in a county with any precision; a gazetteer’s report could be half a century out of step with what it reported. Additionally, compilers sometimes declined to correct or alter information that appeared in earlier editions. The compiler of the Lingshou county gazetteer, Lu Longqi (1630–93), prefaces the local products section by saying, “Many of the things listed in the old gazetteers exist now in name only”40 On the other hand, rice could have been growing in places where it is not named simply because no one had previously named it. The author of the 1877 edition of the Weizhou gazetteer, for instance, copies out the products list of the 1659 edition, then observes that he dare not add anything to the list, even though he believes some products are missing from it. We are left wondering what he might have included had he overcome his respect for past records, especially since rice was listed in the 1639 edition but taken out of the 1659 edition, from which he was respectfully copying.41 Within these limitations, and the understanding that the most randomizing factor is the happenstance of which gazetteers have survived and which have not, this data enables us to reconstruct the dissemination of rice in North Zhili. The findings on first appearance are summarized by reign era in Table 4.1.

Table 4.1 Number of counties and subprefectures for which rice is first reported as growing in North Zhili, by reign era, 1368–1900 Reign era

Dates

Incidences ofjirst appearance of rice per reign era

per decade

Hongwu

1368–98

2

0.7

Jianwen

1399–1402

0

0

Yongle

1403–24

1

0.5

Xuande

1426–35

0

0

Zhengtong

1436–49

1

0.7

Jingtai

1450–6

1

1.4

Tianshun

1457–64

1

1.3

Chenghua

1465–87

2

0.9

Hongzhi

1488–1505

2

1.1

Zhengtong

1506–21

1

0.6

Jiajing

1522–66

8

1.8

Longqing

1567–72

3

4.3

Wanli

1573–1620

16

3.3

Tianqi

1621–7

2

2.9

Growing rice in North Zhili

85

Chongzhen

1628–44

7

4.1

Shunzhi

1644–61

1

0.6

Kangxi

1662–1722

17

2.8

Yongzheng

1723–35

14

10.8

Qianlong(1st half)

1736–65

6

2.0

Qianlong (2nd half)

1766–95

0

0

Jiaqing

1796–1820

0

0

Daoguang

1821–50

2

0.7

Xianfeng

1851–61

0

0

Tongzhi

1862–74

1

0.8

Guangxu

1875–1900

7

2.7

The table suggests that rice first appeared in a few counties prior to the Jiajing era (1522– 66) only sporadically. The early cases of rice cultivation, in the south end of the province, were probably restorations of Yuan-period projects, as in Yongnian, or else expansions of these earlier projects, as in Cizhou.42 Other instances appear extraneous to patterns of local agriculture, as for example when southern soldiers stationed in Qingdu county early in the fifteenth century built a canal to irrigate paddy fields, which they did because they could not stand eating the local millet.43 These fields, totaling over 90 qing in area, were still being cultivated for rice in 1438.44 The sustained development of riziculture in North Zhili had to wait until in the Jiajing era. Thereafter the introduction of rice accelerated through the Longqing (1567–72) and Wanli (1573–1620) eras, reaching its peak in Chongzhen (1628–4). The early Qing, by contrast, was not a time when riziculture flourished in north China. As the compiler of the Ninghe county gazetteer notes, “at the start of the dynasty, fields watered by irrigation were abandoned because water and soil were not favourable.”45 The decay of irrigation facilities had a negative effect on rice production in the Shunzhi era, but both irrigation projects and rice cultivation revived during the Kangxi. For example, a state project in 1704 converted 100 qing of uncultivated land in the Tianjin area into paddy fields and arranged for several dozen families from southeast China to undertake the cultivation of rice in return for ownership of the land they worked and an ox to work it with.46 Private cultivation was also expanding at this time. In a memorial written sometime between 1699 and 1705, Li Guangdi (1642–1718) observes that the spread of rice was driving up land prices in Zhuozhou south of Beijing: easily flooded land that previously would not have fetched even 200 cash per mu was selling for 10 taels of silver after it had been made into paddy flelds.47 At the 1684 exchange rate of 800–900 cash per tael, this was a forty-five-fold price rise.48 The highpoint for northern riziculture came during the Yongzheng era (1723–35), when the emperor sponsored a vast program of water control and paddy construction throughout the province following a year of disastrous floods in 1725. While the first object of the project was flood control, one of its main effects was conversion to paddy To supervise this project, the Yongzheng emperor appointed his brother Yinxiang (1686–

The chinese state in ming society

86

1730), also known as Prince Yi, a man whose capability and dedication sustained the project and achieved remarkable results up until his death in 1730. The work began on a small experimental scale in 1726, then expanded dramatically in 1727 with the organization of four bureaux to supervise construction in four areas of the province. During the three years of 1727, 1728, and 1729, paddy acreage in thirty-nine counties increased yearly, most though not all of the increase being from the conversion of state land. In Zhuozhou, the project opened 1,952 mu of state paddy and 48 mu of private in 1727, followed the next year by another 751 mu of state and 255 mu of private. In Pingshan county, 6,011 mu, 9,954 mu, and 14,400 mu of state paddy were opened consecutively in 1727, 1728, and 1729. The parallel figures for private paddy were 2,202 mu, 907 mu, and 544 mu.49 The shadowing of state figures by private suggests that state riziculture stimulated increases in private investment. The success was so marked that, as one gazetteer put it, during these years “it became impossible to have a disastrous harvest.”50 The project collapsed in 1730 when Yinxiang died and his subordinates jockeyed to use the project to enhance their power. The bureaux nonetheless continued to exist and oversee further construction until 1733. Estimates of the total acreage resulting from the project vary from 6,000 to over 7,000 qing.51 In some counties into which paddy was introduced, rice production took root; well over a century later, some of the original paddy land was still being maintained as such.52 In other counties, it disappeared as quickly as it had come.53 During the Qianlong era, the spread of rice slowed its pace. Some counties report first having rice during the first half of the era, but 1756 is the last Qianlong date. Rice made no further new appearances between 1756 and 1831, when it was reported growing for the flrst time in Nangong county.54 Seven counties report rice’s arrival in the last third of the nineteenth century, a resurgence that may or may not relate to Western technology, though known pilot agricultural projects all postdate 1900.55 Reading the same gazetteer data for spatial patterns, one finds that the areas where rice was cultivated prior to the Wanli era were distributed along the eastern edge of the Taihang Mountains or else set on the flat, marshy plain between Beijing and the Gulf of Bohai. Most of the counties where rice made its appearance in the last seven decades of the Ming were also on the northern plain, with its flat terrain, reasonably accessible water supply and large tracts of uncultivated land. An important pilot project in this area was undertaken in 1602 by Wang Yingjiao (d. 1628) while he was serving as military governor for coastal defence in Baoding. Using the labour of soldiers, Wang diked and brought under cultivation 50 qing of land in Baoding, 20 qing of which were for paddy. Wang discovered that the salinity of the soil adversely affected dry paddy even after the salinity had been reduced, but did not affect certain varieties of wet paddy. The amount of land cultivated through the project was negligible, but its impact was considerable. As the provincial gazetteer reports: Thereupon soldiers and civilians began to believe that the cultivation methods of Fujian and Zhejiang could be put into practice along the north coast, and officials at all levels believed even more firmly that desalination could transform any alkaline soil into fertile soil.

Growing rice in North Zhili

87

An official who visited the area a generation later found Wang’s project still intact.56 More projects followed in this area in the Qing: those in 1704 and 1727 have already been mentioned. The last major Qing project, undertaken around 1890 by military labour in Xiaozhan, cultivated over 450 qing of paddy on coastal land. 57 In the opening decades of the twentieth century, the province’s rice producers were mostly in this area.58 Other areas in which rice was first planted in the early Qing extended the crop’s reach west and north of the Beijing-Tianjin corridor. More of southwest Zhili also reported rice at this time. The effect of the Yongzheng projects was both to fill in some of the gaps in the areas already planted and to push the rice boundary further northeast. Roughly the same trend continued later in the nineteenth century. The far north, with the notable exception of Chengde where paddy was grown on imperial hill estates, remained barren of rice right up until 1900.59 Other regions where riziculture made no inroads were the prefectures in the south of the province: Hejian, Baoding, Guangping, and Daming. Although reasonably well irrigated, these areas suffered from notoriously infertile soil, which rice could not tolerate. What rice there was in local markets came from provinces to the south and was an expensive luxury given to officials or offered to honoured guests at banquets.60 In general, rice never penetrated into these parts of North Zhili, even as a purchased luxury.61 The volume of rice production in North Zhili Although rice spread into more areas of North Zhili during the Ming and Qing dynasties, especially between the mid-sixteenth and mid-eighteenth centuries, its percentage in overall grain production remained small. Roughly one of every four gazetteers for counties reporting rice makes a point of saying that local production was limited. The county gazetteer of Dingxing, for instance, notes that rice, introduced sometime after 1779, was grown only at a certain village in one locality.62 Compared to other grains, reports the county gazetteer of Baodi, the amount of rice harvested was small.63 Another gazetteer points out that the yield of rice in its county was double the yield anywhere else, but admits that actually it was grown only in two villages.64 A rough sense of the comparative importance of rice as a local product can sometimes be gauged by noting its place in the gazetteer’s list of grains. Often the order is a formal or categorical one having nothing to do with levels of production, but sometimes position in the list is significant. When rice makes its appearance at the very end of a grain list, it has only recently been introduced into the county. It can also move up, though. Between the 1758 and 1877 editions of the Yongnian gazetteer, rice moves up from last position in the list, as it does between the 1624 and 1749 editions of the Dong’an gazetteer. Yet the opposite shift can also be found in the successive editions of the Wenan gazetteer, in which rice drops from third place in 1673 to last place in 1703. Estimates of the importance of rice relative to other grains can be derived from tax figures for a few counties and subprefectures in the mid-Ming before taxes in kind were converted into silver. Although tax figures were notorious for developing an increasingly distant relationship with real production levels, the percentage of rice in the grain and hay taxes may reflect its place in local production. Table 4.2 shows tax figures rounded to one decimal place for one subprefecture and four counties in the Ming, three of which can be

The chinese state in ming society

88

dated only loosely to the Chenghua era (1465–87). The data suggests that in the midMing, the percentage of rice in local grain production did not reach 5 percent, and was considerably below that level. Given the limited diffusion of rice before the Longqing era, the total volume of rice production in all of North Zhili at that time could not have been more than 1 percent of total grain production. Rice’s share of provincial production increased in the late Ming, but probably did not exceed 2 percent.

Table 4.2 Percentage of rice in Ming grain and hay taxes County

Date

Totalgrain tax (shi)

Paid in rice (shi)

% Total hay tax Paid in rice % (shu) (shu)

Changping

1453

21,637.3

674.2 3.1

276,711

12,140 4.6

Mancheng

1465– 87

6,048.4

42.0 0.7

77,384

784 1.1

Qingdu

1465– 87

7,414.3

181.0 2.4

95,887

3,383 3.5

Qingyuan

1465– 87

10,821.9

24.8 0.2

134,629

462 0.3

1549

15,518.9

53.5 0.3

199,967

1,000 0.5

Baodi

Sources: Changping xianzhi (1673), 6.3a–4b; Baoding fuzhi (1607), 18.7a, 18a–b, 2a; Tongzhou zhilüe (1549), 4.6a–b. Qingdu county was known as Wangdu in the Qing.

Comparable tax figures for the Qing are not available. The few regional and provincial estimates for the area of land planted with rice tend to underreport paddy acreage.65 We have already noted that, by the end of 1729, Yinxiang’s project brought at the higher estimate over 7,000 qing of land under rice cultivation. This figure excludes previous paddy acreage within the thirty-nine counties involved in the project as well as the acreage outside those counties, pointing to a total possibly twice as high as that figure. Lin Zexu’s estimate for the 1830s approximated paddy acreage at less than 2 percent of all land under cultivation; 2 percent of his overall provincial estimate of 640,000 qing puts paddy acreage at over 10,000 qing.66 If we were to correct Lin’s estimate according to the actual cultivated acreage in the 1830s (roughly 727,000 qing),67 paddy could be pushed up as high as 15,000 qing. Although imprecise, these estimates confirm that rice’s share of acreage fell after the Yongzheng era, though probably not below late-Ming levels.68 The decline continued into the twentieth century, for by 1916 paddy acreage in the province had contracted to a mere 2,443 qing.69 State involvement and local interests The only areas where rice made a significant contribution to local grain production were where a substantial water supply could be secured from sources other than wells. In most North Zhili counties, this meant constructing irrigation facilities. In Tang county, for

Growing rice in North Zhili

89

example, a mountain spring supplied water directly for one qing of paddy fields, but once a small canal network was constructed in 1599 the same source watered 10 qing.70 The well-known Donglin activist Zuo Guangdou (1575–1625) argued twenty years later that “once irrigation is constructed on a large scale, northerners will become familiar with the cultivation of rice.”71 Much of North Zhili, however, was faced with the problem of erratic or inadequate water supply. Even a county for which a fairly comprehensive water control scheme is recorded could not always guarantee that there would be enough water available. As a note in a gazetteer listing the sluice gates in Xingtai county says: Today some of these gates and dikes still exist and others have gone. Whether they are in good or bad repair follows no set schedule, since the rivers in Xingtai do not flow a long way from distant sources but come from low-volume spring-fed streams within the county. Sometimes the springs flow and sometimes they cease, and so sometimes the gates and dikes are in use and sometimes they are not.72 Constructing a system to bring water to the fields could not compensate for the difficulties of guaranteeing the constant supply of water required for paddy fields; larger hydraulic schemes were necessary, and hence the involvement of the state. Who led in the building of hydraulic infrastructure: state representatives or local investors? Both, but not in equal measure, according to the reports in local gazetteers. Local landlords and gentry are sometimes credited, such as Zhong Qiying in Huairou, who constructed a canal outside the county seat in the early Ming to reduce the alkalinity of the soil and make it suitable for paddy; or a gentry surnamed Sun, who built three stone sluice gates in Mancheng around 1570 to irrigate 10 qing of paddy.73 Such references are not widespread, however. Local infrastructural projects are usually represented in the gazetteers as having been undertaken, or at least supervised, by the county magistrate. The naming of magistrates as project benefactors may in part be an artifact of the genre, which obliged its compiler to present the locale as the site of good government, and the local magistrate as its best representative. Such representations could also extend into the local public sphere, in which the honouring of the local magistrate as a patron is appropriate as a discursive move, even if he did no more than give his blessing. That conceded, an active magistrate usually exceeded his merely discursive presence, and for good reason. Indeed, given the scale of project that was usually needed to rework the topography so that water would flow into paddy fields, it is difficult to imagine anyone but a magistrate being able to carry it off. Only with his access to capital resources and corvée labour could expensive projects be initiated to rebuild even local waterworks. Yang Yigui is a representative example of a magistrate who involved himself closely in local irrigation construction. Yang was posted to Tang county at the close of the sixteenth century In 1599, he arranged to build an irrigation canal that was able to supply over 10 qing of paddy fields. After one successful year, the county was hit by drought, and so in the winter of 1602–3 he supervised the construction of Guangli qu (Broad Benefits Canal). Yang has left detailed records of the estimates and costs of this project.74 His cost estimate put the stones, mortar, and stonemasons’ wages at minimally 100 taels of silver plus a grain allowance for 15,000 corvée labourers of 300 shi of grain. Yang

The chinese state in ming society

90

argued that the drain on the granary reserves would be brief and that the increased revenue resulting from the project would more than compensate for the amount of grain disbursed. The 300 shi was paid out during 1602 and the other expenses audited the following year. The total cost of the project ran to over 500 shi of grain, though the disbursements for materials and wages came in close to his estimate, at 72,000 cash, which at the 1611 exchange rate of 660 cash per tael amounted to roughly 109 taels.75 Other costs, such as wooden stakes and baskets, were covered by local levy and therefore outside the audit. The result was a seventy-li canal and twenty-five sluice gates that conveyed water to over 190 qing of fields in thirty-eight villages. Yang went on the next year to construct another canal along the course of one built in the Jin dynasty, irrigating another 10 qing of paddy fields.76 As magistrate, he had the funds to pay for this work, the authority to order it, and the connections to span the interests of several villages. Local officials such as Yang were interested in promoting irrigation as a way of strengthening production in order to ensure local food supply and a stable tax base. At higher levels of the state hierarchy, defence concerns came into play Xu Zhenming, a supervising secretary in the Office of Scrutiny for Works, was the first vocal advocate of large-scale irrigation and riziculture projects in the Wanli era. Xu argued that “just as southerners build polders” to enhance agricultural productivity, so northerners could reconstruct their topography to boost grain production.77 The concern with which he framed his proposal was military: to improve food supply to the garrisons on the northern border, and as well to ease the burden of the soldiers assigned to till military colony land. Soldiers were an asset because they could be mobilized to provide the labour for field construction, but Xu also sought to stimulate their interest by arguing that any land a soldier opened for cultivation would revert to his ownership after he was demobilized, so long as he continued to cultivate it. As many soldiers were southerners familiar with rice technology, this plan would transfer their technical skills to the north. Xu also argued for the positive effect this project might have on settling the large and unregistered migrant civilian population known to be living in the north: In former times [1515], when Seventh Liu and Eighth Liu rose up in rebellion, tens of thousands of people responded immediately to their call, and it was the drifters and good-for-nothings who joined them. Those who devote themselves to agriculture are tied to their lands and villages. Only good-for-nothing drifters think nothing of abandoning their native lands, hence they are susceptible to rebellion. Within the northwest region today there are drifting people and uncultivated expanses, about which the well informed are deeply concerned. If irrigation were introduced and the broad expanses put to cultivation, these drifters would have somewhere to go, conflict would be extinguished, and rebellion would be prevented.78 Xu’s concern with settling civilian labour, and not just motivating military labour, was an extension of his concern with securing the northern border. Enhancing military food production was one part of this plan, stabilizing the population that lived close to the frontier another. His vision of the agricultural north was in fact a return to the ownercultivator ideal of the early Ming, for he cites large-scale landowning and tenancy as

Growing rice in North Zhili

91

factors contributing to this instability in the north, and hoped that these state-sponsored projects could draw labour away from the large landowners’ estates. An intellectual lineage extending from Xu Zhenming and Wang Yingjiao in the latesixteenth century, through Xu Guangqi and Gu Yanwu in the seventeenth, and down to Lin Zexu in the nineteenth, saw the introduction of rice agriculture in North Zhili as a solution to a military need. Not all were able to turn their proposals into state-sponsored projects, but Xu Zhenming was. In 1585–6 he supervised a project northeast of Beijing that resulted in the opening of 390 qing of paddy fields. The following spring, however, the Wanli emperor suspended the corvée order that supported this project when “powerful elite families” objected to losing control of waste land that would have been cultivated by the project, waste land whose reeds and firewood they monopolized and derived considerable income from.79 The state thus might act in what were perceived as its interests, but might also retreat in the face of sustained opposition from local interests. Wang Yingjiao had better luck with his experimental military colony in the Tianjin area in 1602, whether because local interests were not as heavily threatened or because the military value of the rice in the wake of the Japanese attack on Korea was too great to forgo.80 This project did not last much beyond its founder, for an official in 1622 reported that the lands cultivated under Wang’s direction had reverted to waste.81 The central government’s sporadic attempts to build a northern base for rice production were often large in scale, but they tended to be less successful than local projects, whether undertaken by magistrates or county elites. The state could bring investment or corvée labour from above, but the maintenance of these projects could only be guaranteed by inputs from below. Just as Wang Yingjiao’s project faded within a generation, so also much of the paddy land irrigated by Yinxiang in the Yongzheng era reverted back to dry fields, some of it within five years.82 Subsequent to Yinxiang’s project, the state’s infrastructural investment in the North China Plain faded, and with it the possibility of stimulating extensive conversion to paddy. The latest date that I have found in any Zhili gazetteer for an irrigation project undertaken by the central government is 1760,83 which coincides closely with 1756, the latest date I have found in a gazetteer, prior to 1830, for the introduction of rice into a new county.84 The “strong” state of the eighteenth century chose not to sustain northern riziculture. The transfer of rice technology Critical to the spread of paddy agriculture in North Zhili was the availability of appropriate technology and the knowledge of how to apply it. Although northern and southern agricultural technologies had exerted complementary influences long before the Ming, northern cultivators still had an inadequate knowledge of wet-field techniques. Contemporary writers constantly return to this problem. “Northerners do not know about irrigation,” laments Zuo Guangdou in 1619.85 “The peasants of Tianjin do not practice wet-field planting,” observes a provincial gazetteer editor in a reference to 1727.86 One method for technological transfer was to bring southern “agricultural instructors” (dianshi)87 north to teach by instruction or example, as Tuotuo had done in the Yuan.88 Another was for the local official to give such instruction himself, as when Yongnian magistrate Gao Ruxing instructed the peasants of his county in the 1520s, despite the fact

The chinese state in ming society

92

that Gao was a native of Taiyuan in rice-barren Shanxi province.89 A more famous figure who initiated the same kind of intervention is the pragmatic moral philosopher Yuan Huang (1533–1606), who during his term as magistrate of Baodi taught the peasants how to plant rice and later popularized this kind of instruction in his writings.90 The practice of official instruction continued beyond the Ming.91 A northern official named Wu Bangqing (1766–1848) later took this task seriously enough to write a textbook for the purpose, entitled zenong yaolu (Essential records for hydraulic agriculture). In his preface, Wu writes that he inspected local riziculture practices in his home village in Bazhou and found some that did not correspond to the methods recorded in the standard agri-cultural handbooks, Qimin yaoshu (Essential writings for caring for the people, c.530) and Nongsang jiyao (Essentials of agriculture and sericulture, 1286). He decided to go out to the fields and lecture the peasants in their own language on best techniques. This experience prompted Wu to cull through all the information he could find in literary sources and compile this book. He urges his readers to peruse it at their leisure, then go out and instruct two or three elder cultivators who could disseminate this knowledge.92 His book is presented as a practical guide to improving the technical and technological level of all factors involved in paddy cultivation. It strongly recommends the development of irrigation and water-lifting technology, advocates constructing southern-style polders in the marshy areas at the south end of the province, and suggests draining uncultivated bogs in the north. Marsh conversion was a well-established technique in the north,93 but polder construction was not widespread (though see Figure 4.1 for an example from Yongnian county in the southern part of the province). Wu’s recommendations indicate that, as of 1824, the best riziculture techniques had yet to be exhausted in the north. Wu looked to the gentry to assume their proper paternalist role and lead the commoners in the arts of tilling, though who knows whether anyone ever adopted his instructions. The transfer of water-lifting technology was as much a challenge as the transfer of knowledge. Despite Yuan attempts to promote the use of treadle-wheels (shuiche), which involved issuing wood to those who could not afford the raw material, their use was not widespread.94 As late as 1771, Wangdu county magistrate Chen Hongshu complained that “the wild and uncultivated state of Wangdu’s paddy fields is half due to the people’s lack of effort, but also half due to the lack of treadle-wheel technology.”95 Chen saw that well-sweeps (jiegao), which was all the county had at the time, required too much labour to make wet paddy profitable.96 He considered hiring southern craftsmen to instruct the local people in how to build these machines, though he heard that Tianjin already had this technology.97 Half a century later, many of the standard components of pumping and water-lifting technology had still not reached Zhili.98

Growing rice in North Zhili

93

Figure 4.1 The paddy fields watered by Rice Canal (Daopu), as illustrated in the Yongnian gazetteer, were one of the county’s eight famous sights and the inspiration for Yongnian’s reputation as Little Jiangnan. Source: Yongnian xianzhi (1758), shou 15b–16a; courtesy the Tōyō Bunko, Tokyo. Despite their advantages, these pumps were expensive machines whose ownership prior to 1949 was restricted even in south China to landlords and rich peasants. Secondly, they were fragile and easily broken. A skilled carpenter was needed to build and maintain one, and skilled carpenters were not to be found everywhere in the north. Furthermore, although a treadle-wheel pump lifts water more quickly than a well-sweep, it is a timeextensive form of labour,” and in north China it was labour that restricted agricultural development. Like other commentators, Chen Hongshu averred that the people were lazy and did not make the effort necessary to cultivate paddy fields. Yang Yigui, whose irrigation projects were discussed above, expressed the problem more realistically: “Northerners who are engaged in agriculture find the labour of planting and weeding wet fields ten times greater than for dry fields, the result being that many abandon their fields: human effort is not equal” to the task.100 Relative to the south, northern farmers engaged in a type of agriculture that was areaextensive rather than area-intensive. The amount of labour time required to cultivate and irrigate paddy fields, absent external inputs, was beyond the capacity of most North Zhili cultivators.

The chinese state in ming society

94

Obstacles to the spread of rice For a concise list of the technical obstacles to the spread of rice in North Zhili we have the terse summary that Wang Zhiquan submitted to the throne in 1737, after the Qing’s most active decade of riziculture projects. He declared the north at a disadvantage for eight reasons: 1. Intensive cultivation quickly exhausts the north’s sandy soil (which may refer to the tendency for irrigation to increase soil alkalinity). 2. Northern rivers usually run dry in the spring and then flood in the summer, resulting in lacks and surpluses of water at the wrong times. 3. The terrain makes it difficult to store water and direct its flow. 4. Most of the cultivated land is flat, which complicates the construction of necessary reservoirs. 5. Zhili has no pre-existing irrigation system on which to build. This may have been one of the most retarding influences on the spread of rice, since everything from field boundaries to long-distance canals had to be built from scratch, hence the need for central government intervention. 6. Northerner cultivators lack the knowledge and the tools needed for paddy agriculture. 7 Northern dry fields already require heavy fertilizing, and paddy fields would demand even more. 8 Northern rainfall is insufficient and comes at the wrong time of year for growing the rice seedlings. Wang Zhiquan concluded this list by pointing out that, although it would be preferable to feed troops in the north with northern grain, the troop-supply argument was not in itself sufficient to justify central government investments in large-scale projects.101 He thus rejected the lineage of thought from Xu Guangqi to Lin Zexu arguing that proper investment would overcome these obstacles, and as we have seen from the fall-off of rice agriculture at mid-century, his view prevailed. A tendency to focus on technical problems could blind contemporary commentators, just as much as us, to less tangible cultural and social impediments to the spread of rice. These certainly existed, though they often slip unnoticed behind the more obvious technical challenges. We need to begin by recognizing that most northerners were used to eating millet and regarded rice with distaste. Zhu Yizun noted later in the seventeenth century, for instance, that northerners did not like rice and feared that eating it would make them sick.102 For many, the only reason to convert one’s land to rice production was to sell the entire crop for cash. That in turn meant relying on the market to realize an appropriate income and to purchase more palatable grains for household consumption. The Yongzheng emperor was aware of this problem. He stated in 1729 that he feared that cultivators unwilling to eat their new rice crops would sell off the harvest immediately to rice merchants. Unloading their rice would cause the price to fall, and that would discourage cultivators from growing rice again. Yongzheng’s solution was to order that the state buy up the excess rice at full price.103

Growing rice in North Zhili

95

Beyond the impediments of consumption or marketing, cultivators feared that physical improvements and higher productivity would threaten their security of ownership. As one Wanli-era official noted, hydraulic schemes undertaken to stimulate riziculture caused alterations in existing field divisions that worked to the disadvantage of the poor and the powerless.104 Two and a half centuries later, Lin Zexu was careful to include in his proposals provisions to protect cultivators’ land and compensate them for losses incurred because of hydraulic construction.105 There was in addition the more troubling anxiety that higher yields would increase tax bills106 and make land more valuable, and therefore more attractive to expropriation by powerful landlords. Gu Yanwu was aware that switching to rice could threaten security of tenure, citing an edict of the 1620s to the effect that as soon as paddy fields in previously uncultivated areas began to yield harvests, “the local powerful families and the managerial landlords” (shihao ji jingguan dizhu) took over the fields for themselves.107 This threat was a powerful disincentive for peasants undertaking paddy cultivation, and not one from which the state could adequately protect them. This is not a widely repeated observation, perhaps because most advocates of riziculture in North Zhili were themselves from the ranks of the powerful families and managerial landlords of whom ordinary people lived in some fear. Xu Guangqi, despite his reputation as one of the most forward-looking statesmen of the late Ming, was by virtue of his holdings in Tianjin—where he experimented on rice growing—inescapably one of the latter. He saw himself in a different category, however. He was a small, progressive landowner, and people such as him were not the problem. It was the larger, more conservative landowners, he insisted, who were hopelessly “content with the old practices” and unwilling to experiment with building paddy fields. This conservatism extended up into the ranks of officials, as he discovered when he tried to explain ditch drainage: “I once discussed this with the officials and the gentry, but those in charge did not understand the principle of it and gave up.” He also condemned the gentry in general for being indifferent to agriculture, and landlords for being unwilling to expend efforts on improving production.108 Landowners reluctant to follow Xu in building paddy fields had their own fears. They were anxious lest they lose control over the resources uncultivated land could yield, such as reeds and firewood, which as we noted led the Wanli emperor to shut down Xu Zhenming’s project in 1586. They also worried about having to rely on publicly shared investments such as irrigation systems, whose benefits might not always go in their own favour. This issue comes up in a memorial to the Yongzheng emperor reporting on a disturbance at two sluice gates in Cizhou. Those whose land was irrigated by water regulated by these gates agreed at the end of the agricultural year that, in the year to come, the gates would be opened and closed once every five days. When the time came to open the gates the following spring, a yamen employee and an “obstinate commoner” led a mob to the sluice gates to prevent them from being opened because they feared losing their share of the water. The emperor responded by demanding that both men be punished.109 Xu Guangqi, and everyone in the intellectual lineage of which his advocacy made him a part, believed that the limits on the introduction of rice into North Zhili were not technical but social. The solutions to increasing northern grain production lay not so much with the state, although its role as investor could be crucial, but with the local networks by which cultivators sustained, or abandoned, their expectations and attitudes,

The chinese state in ming society

96

and adjusted, or failed to adjust, to the new terms of production and consumption that rice introduced into their economy. Lin Zexu was one of the last in this line who believed that local effort could overcome regional ecology: “The reason why paddy fields are not extensive is entirely because people have not built them,” and not because they are technically impossible to build.110 No other prominent Qing official after Lin saw rice in north China as a solution to antiquated fiscal policies and a looming agrarian crisis. What the state and some of its representatives had done in the Wanli and Yongzheng eras in particular would not be repeated. The late-Qing state was not in a position to remove the social obstacles to rice production in the north, nor to invest the financial resources needed to demonstrate the value of changing social practice. Only after 1949 would some of those limits be removed, to the detriment of the ecology in many of the place where the new field regimes were imposed.

Part III Books

5 Building school libraries in the mid-Ming In his hugely popular General History of China, Jean-Baptiste du Halde singled China out for achievements that seemed remarkable from a European point of view. One item in the reports of his fellow Jesuits that captured his attention was, quoting from the first English translation of 1736, “the great number of Libraries in China magnificently built, finely adorn’d, and enrich’d with a prodigious Collection of Books.”1 Writing within a culture that had known book-printing for less than three centuries, du Halde regarded the widespread existence of sumptuous and well-stocked libraries as something to celebrate, even envy. Although libraries have been known in China for at least two millennia, large libraries of the sort that the Jesuits first saw at the end of the sixteenth century became a common fixture of Chinese society only during the century or so prior to their arrival in China. The growth of libraries in the Ming dynasty has to be linked in the first instance to the increasing availability of published books at this time. Books were being produced on such a large scale that they were also being consumed on something close to the same scale. Now that it had become possible to accumulate a large collection of books, book owners had to devise conventions for storing them in a way that protected the books as physical objects and organized them such that the contents of each remained as easily accessible as it would have been were it the only book the owner possessed. Libraries were one solution to the problem of what to do with a lot of books. The Ming did not invent libraries, of course, but it was only in the Ming that libraries in the thousands were built to house anywhere from hundreds to tens of thousands of fascicles (juan). The popular moniker Wanjuan lou (Library of ten thousand fascicles) was used in the Song period for libraries that claimed to approximate the ideal of comprehensiveness,2 but only in the mid-Ming did more than a few libraries hold books in this number, and only in the late-Ming did libraries begin to hold them in multiples of this number.3 The best known Ming libraries are those that housed large private collections, such as the justly famous Tianyi ge (literally, The Number One Library under Heaven) belonging to the Feng family of Ningbo, or the Wanjuan lou (Library of Ten Thousand Fascicles) owned by the Yu family of Shanghai. But libraries emerged at other points on the social landscape besides the wealthy book-owning families, one of these being the county school. Pavilions for Revering the Classics The widespread building of libraries in Ming government schools is a development that belongs to the dynasty’s middle century. Like the books they housed, school libraries

Building school libraries in the mid-Ming

99

were not a Ming invention, but did not become common until the Ming. School libraries existed in the Song, but they were rare and attached mostly to schools in provincial or prefectural capitals. One of the earliest recorded is the library of the prefectural school in Ji’an (Jizhou in the Song) in central Jiangxi, founded in 1032 and rebuilt in 1286–7.4 Other school libraries were founded in the Song,5 for at least five of which the great neoConfucian Zhu Xi wrote commemorative essays between 1176 and 1195,6 and more appeared in the Yuan. It was in the Ming that a school, even down at the county level, was expected to have a library. To be fair to pre-Ming bibliophiles, it was also only in the Ming, as I shall note below, that counties were expected even to have schools. Still, schools at that time did not imply libraries; these would come later, and for specific reasons. The standard term for a school library in the Ming is Zunjing ge (Pavilion for Revering the Classics). This name is first attested in the Yuan, being the name of the library built at the prefectural school in Jining, Shandong, in 1299.7 It was but one of many names in use. The prefectural school library built in Suzhou about 1060 was called Liujing ge (Pavilion of the Six Classics), then renamed Yushu ge (Pavilion of Imperial Texts) when it was rebuilt in 1187. Other names used in the Song included Jingshi ge (Pavilion for Classics and Histories), Jigu ge (Pavilion for Examining Antiquity), Cangshu ge (Pavilion for Storing Books), and Cangjing ge (Pavilion for Storing the Classics). The last would have been the most familiar to Song and Yuan ears, being the name for the library in a Buddhist monastery, which was better equipped in this regard. There it was a Pavilion for Storing the Sutras. The ambidexterity of jing, denoting both the Buddhist sutras and the Confucian classics, allowed the name to pass from Buddhist to Confucian libraries without cognitive dissonance. Song academies also built libraries and imitated Buddhist practice by calling them Cangjing ge.8 The name Cangjing ge fell out of use for school libraries early in the Ming in favour of Zunjing ge. This new naming convention was strong enough by 1440 to induce the rebuilders of the Cangjing ge at the Songjiang prefectural school (the original building collapsed in a typhoon in 1410) to bestow on it the new name of Zunjing ge.9 The author of the essay commemorating the rebuilding of this school library indicates that he was conscious of the difference between zunjing (revering the classics) and cangjing (storing the classics) by underlining the appropriateness of the new name. (I shall return to the connotations of “reverence” at the end of this chapter.) The preference for Zunjing ge over Cangjing ge may have stemmed from the desire of neo-Confucian educators to obscure the indebtedness of their institutions to Buddhist precedents, part of the general anxiety neo-Confucians felt about covering their tracks on the way to metaphysics. Whatever prompted this shift, Zunjing ge quickly pushed out competing terms. Of the other Song names for school libraries, Cangjing ge was reserved for monastic and academy libraries, Jigu ge continued to be used for some private libraries, and Yushu ge was specialized to designate pavilions housing a sample of an emperor’s script.10 A few other names were used idiosyncratically for school libraries in the Ming, like Chongwen ge (Pavilion for Honouring Literature),11 Jukui ge (Pavilion of the Constellation of Literature), and Wenchang ge (Pavilion of the God of Literature),12 but these are the odd exceptions in the process leading to the standardization of Zunjing ge. Evidence preserved in local gazetteers allows us to reconstruct the creation of Pavilions for Revering the Classics at county schools as a historical development

The chinese state in ming society

100

characteristic of the mid-Ming period. This evidence is not consistently recorded, however. In one gazetteer, one can find mention of a library having been restored, but not the date when it was initially founded;13 in another, passing reference is made to a library’s existence but with no report on any aspect of its history.14 Nonetheless, there are enough references to Ming school libraries in the large corpus of Ming and Qing gazetteers to make it possible to sketch a rough temporal profile of their emergence. On the basis of a random sampling of gazetteers, mostly from northern, eastern, and southeastern provinces, I have tabulated 102 instances of the founding, rebuilding, or restoring of school libraries between 1395 and 1631, plus a few earlier dates.15 Grouping this data by decade shows that the two most active periods of library-building and restoration fall in the periods 1458–68 (Tianshun/early Chenghua) and 1516–26 (late Zhengde/early Jiajing). Other decades showing significant activity are 1438–48 (Zhengtong), 1530–40 (mid-Jiajing), and the periods and 1599–1609 in the Wanli era. These dates suggest that interest in school libraries gained momentum late in the 1430s, was strong from the 1460s to the 1540s, and revived at a more moderate rate during the Wanli era. On a grosser scale, library-building in the Ming was strongest during the century from 1439 to 1539, with a significant revival during the half-century from 1570 to 1620. By the 1520s, accordingly, a Fujian author could look back and declare that the main reason why schools in the early Ming tended to lose their books was because they did not have libraries.16 This observation linking book preservation with library-building would not have occurred to an author writing a century earlier, when the school library was still a rare institution. The new notion that libraries were necessary for the proper functioning of a school (or at least for the proper protection of its property) did not mean that all midor late-Ming appeals to build or restore school libraries were successful; 17 many county schools did not get libraries until the Qjng. Nor did interest in founding school libraries entail that, once built, they were immune to neglect.18 Still, the great rarity of school libraries prior to 1395, and their increasing presence after 1439, indicate that the building of libraries on school grounds became conventional as an infrastructural investment in the mid-Ming, and justifies identifying the building of school libraries as a phenomenon of this period. Behind this data looms the obvious question of why the mid-Ming was a time when local officials chose to put up buildings consecrated to the housing of books. Like all causal questions, this one changes colour depending on what light we put it under. We might be tempted to assume that the rise in libraries reflects an opening of knowledge acquisition in Ming society; but this assumption, our own, has to be set aside until we know more about what Ming library builders thought they were up to. To obtain that knowledge, we have to work from a rather thin set of texts that have survived here and there, mostly in local gazetteers. My reading of these sources suggests five questions (and answers) that touch on five major aspects of library-building that might help make sense of this historical phenomenon: what got put in these libraries (books), where they were located (schools), what form they took (self-standing buildings), who built them (officials), and what they symbolized (the authority of textual knowledge).

Building school libraries in the mid-Ming

101

Official books We begin with books, of which the production of libraries assumes the prior physical production. The relationship between the two is not simple or direct, however, for libraries are equally inconceivable in the absence of certain political, cultural, and economic ideas about the value of books. Books can be regarded as the special bearers of political orthodoxy, for instance; as the vehicles of esteemed ancient wisdom; or simply as fancy objects that cost money. Which is to say that libraries depend on not just the physical production of books, but their cultural production as well. For libraries attached to state schools, it was both the physical and the cultural production of books published by the Ming state that contributed to the first wave of building. The Hongwu emperor (r. 1368–98) knew that a world at war could be won by stratagems and the force of arms, but learned that one key to ruling a post-conquest world was the control and dissemination of knowledge. Certain books he found useful. To his officials, he sought to impart knowledge of the laws, regulations, and rituals that they were to respect and see upheld among the people. Among students, he encouraged knowledge of the teachings of Confucianism according to the redactions produced by Zhu Xi (1130–1200) and other Song intellectuals in his tradition. For Buddhist monks, he demanded a level of understanding of the principal sutras. The common people he expected to read or listen to recitations of his simple rules for a reconstituted agrarian self-sufficiency and not bother themselves with any further reading. For these four categories of people, Hongwu had his advisors produce appropriate texts and have them published, often in large print runs. His first intervention in this area was toward the end of 1366 before he had declared himself emperor. He sponsored the compilation of two books. Gongzi shu (The princes’ book) was a set of readings that his princes should master in order to understand the difference between good and evil. “Most of them are unable to penetrate abstruse meanings,” he admitted, so his editors should keep it simple. For a much larger audience, he had Wunong jiyi shanggu shu (The book for cultivators, craftsmen, and merchants) produced. “Most sons of peasants, artisans, and merchants don’t know how to read, so it is fitting that their duties be explained to them in direct language.”19 Both titles were printed inside the court and distributed to the people. The Ming court issued first the Da Ming ling (The Ming statutes) in 1368, the year Zhu Yuanzhang assumed the imperial title of Hongwu. This slim text was the first of a series of ever larger judicial, administrative, and ritual compilations of regulations and admonishments that were produced during his reign to govern the practices of his princes and officials in particular. Among the earliest, a compendium of rites and sumptuary regulations, the Da Ming jili (Collected rites of the Ming dynasty), was ordered issued in October 1370, and appeared the following year; a handbook of administrative geography entitled Da Ming zhi (Gazetteer of the Ming dynasty) was published and issued injanuary 1371; and the first edition of the Ming Code (Da Ming lü) was ordered and distributed to the realm in December 1373, to name just three of the ever growing list of books coming out of the Hongwu court.20 The court also got involved in the production of textbooks for the examination system: the core Confucian canon—the Five Classics and the Four Books—were published and distributed in 1381 to every county school in the country For

The chinese state in ming society

102

monks, he sponsored the printing of the Tripitaka or Buddhist canon, the blocks of which were produced under the supervision of the Ministry of Rites in Nanjing and stored at Baoen Monastery. For the common people, he circulated his personal instructions, notably the four Da gao (Grand pronouncements, 1385–8) and the Jiaomin bangwen (Placard for instructing the people, 1398). The Yongle emperor (r. 1403–22) shared his father’s belief in the value of publishing what he wanted his subjects to read and obey. He had the Confucian classics and the writings of the major neo-Confucian philosophers fixed and reissued in the three collections in the official Daquan (Great compendium) series. He also initiated the work of re-editing and publishing the Buddhist and Daoist canons, though these palace editions did not appear until after his death. Yongle sponsored his own edition of the Buddhist canon, known as the Northern Tripitaka, which could only be obtained by petitioning for an imperial edict. His father’s, the Southern Tripitaka, simply required showing up in Nanjing with an ordination certificate and the money.21 To give moral instruction to the people, Yongle had several widely circulated didactic texts produced, notably Quanshan shu (Exhortation to goodness) issued under his wife’s name shortly after her death in 1407, Weishan yinzjii (Blessings of doing good secretly) in 1419, and Xiaoshun shishi (Testimonies to filiality and obedience) in 1420. After Yongle, Ming emperors produced and published texts less often and within a narrower range. The classics were already fixed, and simple morality texts seerned to lose their appeal, the Xuande emperor’s massive Wulun shu (On the five relationships) being the last significant title in this line, though too bulky to be popular reading. The palace turned instead to producing such reference works of administrative geography and history as the bulky Da Ming yitong zhi (Unity gazetteer of the Ming dynasty), published in 1461, and Zhu Xi’s historical digest, Zihi tongjian gangmu (Outline of the Comprehensive Mirror as an Aid to Ruling), published in the following decade. Emperor Jiajing (r. 1522–66) is the best known of the later emperors to sponsor palace publications, as he did several times in the 1520s to cow his officials into accepting his judgments on hugely controversial issues. The Dali jiyi (Deliberations on the great rites controversy), published in 1525 in imitation of Hongwu’s Xiaoci lu (Records of filiality and compassion), was a collection of documents from the court debate over whether to elevate his natural father to imperial status;22 it was much expanded and republished as Minglun dadian (Great compilation on virtue illuminated) in 1528. It was followed by Dayu lu (Records of the great imprisonment controversy) regarding a factional battle over a Shanxi official’s handling of a bandit leader. These books were printed and produced in the palace in the Classics Workshop (Jingchang) under the supervision of the Ministry of Works. They were known variously as “palace editions” (Jingben), “official books” (guanshu), or “imperially produced books” (yuzhi shu). Publication in a palace edition set the imprimatur of imperial sanction on a book. Unassailably authoritative, these editions generated a code for the bureaucracy as well as a canon for academe, and Confucian students who wished to qualify for state service had to master both. Gu Yanwu would later complain that Yongle’s Great Compendium editions tampered with the classics,23 though he did so without suggesting that a conspiracy was afoot to alter the texts in such a way as to favour the rule of the Ming house. Which recension of a classic an emperor or his advisors denominated as the correct version may have had less to do with the positioning of that text within academic

Building school libraries in the mid-Ming

103

debates about its textual history than with the simple need to fix on one version as authoritative. The purpose appears to have been to streamline and unify the textual basis on which the state rested for its legitimacy: the early-Ming emperors had to establish a canon of knowledge to reduce the ambiguity that obscure texts could generate. A canon not only set standards that could not be transgressed. It cut down on the mutability of texts, and hence reduced the indeterminacy of interpretation. Every copy in the hands of every student throughout China thenceforth would say exactly the same thing, and would always say exactly this same thing. Differences among texts could not then be mobilized by troublesome scholars to finger holes in state orthodoxy. Publishing texts was the first step in the imperial project of standardizing knowledge. This effort required a second step beyond publication, however, and that was to establish regular procedures to deliver these texts to their intended audiences. We do not know how the very first palace editions were distributed, but late in the Hongwu reign, state schools were being designated as one site for emplacing the emerging canon. Schools as book depositories Late in 1369, the Hongwu emperor ordered that “every prefecture, subprefecture, and county in the realm establish a school.” These schools were not basic educational institutions. They were places where ambitious young men were examined for the status of shengyuan, registered for higher exams, occasionally exhorted, and kept within the state’s sight. They were also ritual shrines to the founders and perpetuators of state Confucianism. Prior to the Ming, Confucian state schools had been mandated only as far down as the prefecture. Hongwu’s edict marks the first time that magistrates were required to operate schools at the county level to prepare students for the state examinations. The order was carried out almost universally in 1370, a year prior to the flrst sitting for the national jinshi examination in Nanjing. The operation of these schools was maintained so stringently that a Hanlin academician a century later was confident in declaring that “the schools of this dynasty cover the realm.”24 The creation of the canon came later. The schools edict says nothing about books or libraries. Hongwu did not institute this system of state Confucian schools for the purpose of receiving imperial editions. With it in place, though, the court had a framework for distributing the texts that the palace would publish. During Hongwu’s reign, the books that he ordered distributed to the realm began to be sent to these schools. The earliest record I have found of a county receiving palace editions, though the date of 1382 is clearly too early, is when the school in Dehua county, Jiangxi, received “classics, histories, the Ming Code, and the Pronouncements”25 How the deposit was arranged is not specified in the county gazetteer. The word used here and in other gazetteers is ban (“issued”), which implies that the books were bestowed free of charge. Indirect evidence that this was so is the speed with which schools acquired the palace editions. Had they been obliged to pay for publications of such high quality, one might expect a delay in acquisition. (Book prices for this period are not available, but by way of comparison, a copy of the Southern Buddhist canon in a massive 6,361 juan was set at between 64 and 290 taels depending on the quality of paper and binding desired.26) The catalogue of books in the county school of Cili, Huguang, preserved in the 1574 county

The chinese state in ming society

104

gazetteer, says that the Great Compendium editions of the classics were “issued” in 1414, The Blessings of Doing Good Secretly “issued” in 1419, and Testimonies to Filiality and Obedience “issued” in 1420—in each case, immediately upon publication. Cili county’s copy of On the Five Relationships in sixty-two fascicles was not “issued” until 1447, however, four years after it was published.27 The gap between publication and acquisition could indicate that time was needed to find the funds in the county budget or raise them from outside to purchase this work. So too might the fact that one school library in Shandong owned a handwritten copy of On the Five Relationships as well as a printed edition issued by the court, if we can assume that the printed version was acquired after the handwritten version.28 County gazetteers list this book as an “issued” rather than “purchased” item, 29 so it would seem that schools were receiving the palace editions free of charge. Some evidence for this can be teased out of a curious reference in the Veritable Records. At the end of 1448, the Ministry of Rites brought to the emperor’s attention the practice among teachers and students whose schools received a copy of On the Five Relationships of travelling to Beijing to express their gratitude for his generosity. The ministry condemned this courtesy as a poor excuse for neglecting their studies and secured the emperor’s permission that henceforth they be restricted to staying where they were and simply bowing in the direction of the palace.30 If schools were lining up to send delegations of thanks so often that they had to be stopped, this suggests that the book was going out in large numbers, probably automatically, to every school in the country. I have found no account in the Veritable Records of the procedure by which books were issued, other than the formulaic statement that the emperor “ordered it issued and distributed” (zhao ban xing shi), as Hongwu did for the Collected Rites of the MingDynasty in 1370. There is, however, an instance for which the mechanism by which palace editions were distributed is briefly outlined. In 1437 an official in one of the princely courts wrote to the Ministry of Rites to say that his establishment lacked a copy of the official compendium of rites. He does not refer to the Da Ming jili of 1370 by exact title, but his phrasing, “important collection of rituals of the dynasty,” suggests that this is the book he meant. The ministry relayed this request to the throne with the suggestion that the Ministry of Works should be ordered to reprint the book and send copies to all the princely courts. The answer that came back agreed that no princely establishment should be without this book, but since printing this flfty-fascicle publication would be expensive, the ministry should restrict itself to sending only as many copies as there were in stock.31 This exchange reveals that a request for a palace publication had to pass through the Ministry of Rites to the throne and then the Ministry of Works. Later entries in the Veritable Records record other instances of palace editions being requested. For instance, in 1526, shortly after his enfeoffment, a Ming prince in a different establishment received copies of the Great Compendium editions of the Four Books “in response to his request,”32 though nothing of the actual procedure is recorded. Nine years later, the National Academy in Nanjing received a set of Great Compendium editions and a copy of On the Five Relationships, presumably in response to a request to replace worn or damaged copies, though again nothing of these details is recorded.33 However the issuing process worked, it succeeded in getting most of the important palace editions into county schools throughout the country, and thereby determined in part what books a school owned. Schools that predated the Ming might already have books,34 but only with difficulty were schools able to shoulder the costs and find the

Building school libraries in the mid-Ming

105

sources to augment their holdings beyond what the state distributed. Otherwise, particularly in the first century of the dynasty, the books a school owned were the books it was given. From the limited number of school catalogues preserved in local gazetteers, it is possible to identify the core texts one would expect to find in a prefectural or county school in the mid-Ming. Those most commonly recorded are listed in Table 5.1: the Great Compendium classics with Zhu Xi’s commentaries; seven legal, administrative, and ritual texts; six moral tracts from the Grand Pronouncements to On the Five Relationships; the three palace editions of the Jiajing controversies; and three works on geography and history—all produced and distributed by the court. Commercial books These palace editions amounted to less than a thousand fascicles. If as the Songjiang prefect in 1440 declared of his school library that ideally it should

Tabk 5.1 Core texts in Ming school libraries no. of juan

The classics: Sishu daquan (Great compendium of the four books)

36

Wujing daquan (Great compendium of the five classics)

68

Xingli daquan (Great compendium on nature and principle)

70

Sishu jizhu ([Zhu Xi’s] Annotations to the four books)

26

Legal, administrative, and ritual texts: Da Ming ling (The Ming statutes, 1368)

1

Huang Ming zuxun (Ancestral instructions of the imperial Ming dynasty, 1373)

1

Da Ming lü (The Ming code, 1397)

30

Zhusi zhizhang (Handbook of government posts, 1393)

10

Sheli jiyao (Essentials of archery rituals)

1

Da Ming huidian (Comprehensive regulations of the Ming dynasty, 1503) Da Ming jili (Collected rites of the Ming dynasty, 1370, 1530)

180 53

Hortatory literature: Dagao (Grand pronouncements, 1385–7) Jiaomin bangwen (Placard for instructing the people, 1398),

1 35

Quanshan shu (Exhortation to goodness, 1407)

1 19

Weishan yinzhi (The blessings of doing good secretly, 1419)

2

Xiaoshun shishi (Testimonies to filiality and obedience, 1420)

2

Wulun shu (On the five relationships, 1443)

62

The chinese state in ming society

106

The Jiajing emperor’s deliberations: Dali jiyi (Deliberations on the great rites controversy, 1525) Minglun dadian (Great compilation on virtue illuminated, 1528)

4 24

Dayu lu (Records of the great imprisonment controversy, 1528) Geography and history texts: Da Ming yitong zhi (Comprehensive gazetteer of the Ming dynasty, 1461)

90

Daxue yanyi bu (Supplement to Exposition on the Great Learning, by Qiu Jun, 1506)

160

Zizhi tongjian gangmu (Outline of The Comprehensive Mirror as an Aid to Ruling, by Zhu Xi, Chenghua era (1465–87))

59

Sources: Anxiang xianzhi (1687), 4.10a; Changde fuzhi (1538), 9.4b–5a; Cili xianzhi (1574), 11.14b; Dingxiang xianzhi (1727), 2.13b; Hejian fuzhi (1540), 28.58b–59b; Ji’an fuzhi (1648), 15.7a; Jianchang fuzhi (1517), 8.1a–8a; Ruichang xianzhi (1568), 5.6a–7a; Ruijin xianzhi (1542), 3.3a–4a; Shenxian zhi (1548), 3.9a-b; Wenshui xianzhi (1673), 4.3b–4b; Wucheng xianzhi (1549), 1.38b; Yanping fuzhi (1526), 12.7a–8a.

hold 10,000 fascicles,36 he had to assume that more than palace editions should grace its book cabinets—as indeed they did. Imperial distribution initiated the collections of books that most schools had, but commercial circulation enabled them to grow beyond the canonical core. When a library was being built for the school in the Songjiang county of Shanghai in 1484, a contemporary could record proudly that the collection consisted of “the six classics, the imperially produced books issued by the Ming court, plus the philosophers of the hundred schools and the histories: there isn’t a title they don’t have!”37 Here was a collection that went beyond the palace corpus. Conversely, when Huguang Education Intendant Chen Fengwu (1475–1541) looked over the catalogue of books in the Wuchang prefectural school in 1505, he was dismayed to find “only the editions of the classics issued by the court, but neither the writings of the philosophers nor the histories. So I sent someone to Nanjing to buy classics, histories, the writings of the philosophers, and literary collections.”38 Clearly Chen did not regard a library consisting only of palace editions sufficient for his students’ education. He expected the commercial presses in Nanjing, where one went in the mid-Ming to buy the best editions available,39 to supplement them. The distinction between state publications and commercial printings is preserved in some sixteenth-century catalogues of school libraries, which note which books were “issued” and which were “purchased.” The titles shown within the “purchased” category range widely. They include unofficial annotated editions of the classics, dynastic histories, literary collections by famous authors, the writings of Song neo-Confucians, even the odd Daoist and Buddhist text. For example, the 1540 catalogue of the prefectural school in Hejian, North Zhili, lists seventy titles. Fifteen are Yongle and Jiajing palace editions; the other fifty-five titles (“title” here is a misleading term, since one of these titles was all twenty-one dynastic histories) were acquired in two separate purchases in the mid- and late 1530s.40 The 1517 catalogue of the prefectural school in Jian-chang, Jiangxi, reveals an even larger collection, twenty-three “issued books” and 186

Building school libraries in the mid-Ming

107

“purchased books,” and records as well another forty purchased books in the collections of county schools outside the prefectural capital.41 These titles include such practical pedagogical items as primers and vernacular translations of the classics. The Jianchang school library also owned the woodblocks for another fifteen titles. Some schools were also publishers.42 The acquisition dates of books in school collections are not often given, though those for the Hejian school are. These dates suggest that these books were commercial printings, for the 1539 purchase included the Great Compendium editions of the classics. Although the National Academy in Nanjing received a set in 1535, as noted above, it is unlikely that the court was still distributing copies to county schools. Most of the Great Compendium copies of the classics that survive today are in fact not palace editions but sixteenth-century commercial reprints.43 The commonest extant version of the Great Compendium on Nature and Principle, for instance, is the 1597 edition produced by the Huizhou publishing house Shigu zhai and sold through the Nanjing and Hangzhou book markets. Whether to correct the errors creeping into commercial editions or to capture a piece of the market, the prefectural school in Nanjing published its own edition of this book six years later.44 Schools faced the same limitations as most book buyers when it came to amassing their collections: price and availability. The Songjiang prefect may have professed that he wanted his school to collect up to the ideal limit of 10,000 volumes, but it is unlikely that he actually believed this to be possible at the time. Schools in major commercial centres stood a good chance of building a collection that ran into the hundreds of titles, which put them well beyond one of the great English libraries of the day, at Cambridge University, which in 1424 had 122 titles, a huge collection by English standards.45 County schools located further from the commercial networks through which books of any quality passed faced considerable obstacles to acquiring titles for their collections on this scale. In rural Jiangxi, the Ruichang county school claimed thirty-two titles as of 1568, and the Ruijin school forty-one as of 1542. The situation was worse in Shandong, which was distant from any major publishing centre: the school in Shen county had thirteen titles, and in Wucheng county, only eleven. The schools within Changde prefecture in Huguang report the same book poverty: as of 1538, all had between nine and eleven titles, and almost all the same ones.46 These examples are to make the point that while commercial publishing did produce the books that went into the growing collections of schools in more developed areas, it was failing to circulate books to schools in the many parts of the country that were a long way from the book production centres in Jiangnan and Fujian. Like any other commodity, scholarly books of quality were too specialized to travel far down the marketing hierarchy and out into the hinterland, even as late as the sixteenth century. We shall return to this issue with the case of Hainan Island below. Library buildings Constructing a building to house books was one solution to the problem of what to do with the books that schools were acquiring. It is, however, an expensive way to store books. Some officials in the early Ming may have felt the imperial aura of the books coming from the palace justified the expense of housing them at a level in keeping with

The chinese state in ming society

108

their status as imperial gifts. For instance, officials in Hangzhou declared in the 1410s that the prefectural school should build its first library expressly because it had just received the Great Compendiurn classics.47 For storing three sets of books, this was an extravagant solution that few other schools at the time were prepared to contemplate. Before library-building became common, and even thereafter, other solutions to the problem of keeping books safe, orderly, and supervised were tried. The simplest way to keep track of books in a school was to compile a catalogue. The catalogue reminded officials of what books a school owned, and it could be used to check that books were not carried off. Incising the catalogue in stone “so that the books might be passed down for all time”48 lent greater permanence to the record, and this was sometimes done, as it was also for inventories of other school possessions, such as furniture or ritual implements. Inserting the catalogue into the local gazetteer was another device against loss and time. Catalogues kept track of titles, but not of the physical objects themselves. The problem of storage was handled in the first instance by keeping books in wooden cabinets, usually under lock and key. As long as there were not too many of them, such cabinets did not require being placed in a special building but could be kept anywhere. Books stored in wooden cabinets were prone to moisture damage, however, and library builders frequently explained their work by observing that damp cabinets were ruining school books.49 For more substantial storage, a school could consider designating a small, secure room in an existing building as a book storeroom (cangjing ku). As schools often constructed locked storerooms to preserve their other valuable property, a book storeroom was simply an extension of this practice.50 This solution increased protection, but it impeded access. Another arrangement, which also reduced access, was to set aside an upper room in an existing building.51 Constructing a dedicated building was the most expensive solution to the problem of storing the ever growing edifice of canonical and commercial books that some schools were amassing. References to school-building in the early Ming rarely mention separate structures to house books. For instance, when the prefect of Beijing in 1435 requested the throne to pay for the restoration of the school in nearby Tongzhou, he listed as its actual or potential component buildings the Minglun tang (Hall of virtue illuminated), which was the ceremonial shrine to Confucius that served as the central hall of every Confucian school, plus the galleries that ran down the east and west sides of the main courtyard, the kitchen, the storeroom, and the archery compound—but made no mention of a library.52 Within a few years, however, libraries began to enter the consciousness of school builders. Some school libraries could be small, but others were almost as large as the main hall, behind which they were conventionally located, following Buddhist practice. Their scale was not purely a matter of storage. A few educators could cite the number of volumes in their collections to justify constructing a large building, but they could also argue that the library, being the repository of textual learning at the heart of Confucian knowledge, deserved to be physically elevated and visually impressive. The imperial origin of the collection’s core also militated in favour of building a grand structure. And so it was that the school library was often one of the most impressive buildings on the landscape. The Pavilion for Revering the Classics in Songjiang, rebuilt and renamed in 1440, was a magnificent square tower measuring 22 metres on each side and rising over 25 metres into the air. From its top, one could gain a panoramic view of the lakes and rivers of this

Building school libraries in the mid-Ming

109

piece of Jiangnan53—a comment that suggests that people went to this library to do more than look at the books. (Unfortunately, no source comments on who went to school libraries or why; indeed, I have yet to find a text describing a student actually going into a library and reading its books—although this activity is at least implied.) The Songjiang school library slipped into decline as the dynasty wore on, and was further damaged in the Ming—Qing transition. From the far side of that transition, Ye Mengzhu could look back to his youth and recall that “even though the building was somewhat dilapidated, still it stood erect in lofty eminence.”54 I have found no reliable data on the cost of building a library. One figure given for building a library in a Buddhist monastery early in the 1570s was 1,000 taels,55 though this seems too modest a sum for the construction of a large, two-storey building. Building and restoration projects were expensive not because of construction costs alone. Once built, a library prompted officials and other patrons to see that it was filled with books, and that was in itself an expensive undertaking. Certainly this was the effect when the Songjiang prefectural school was rebuilt in 1440, for the prefect, whose enthusiasm for the library I have already noted, decided that certain books had to be purchased to fill gaps and bring the collection up to the building’s standards.56 Patrons Almost all school libraries are recorded as having been founded or restored by local officials. This was in part a convention of the gazetteer genre, which dignified the conformity of the local to dynastic standards and heaped what honours it could on the incumbent magistrate. Officials might well canvas the local wealthy to help shoulder the burden, as did the Songjiang prefect in 1440, but in very few cases was the project to build or restore a school library spearheaded independently by a member of the local gentry. Given the advantages that a good school could bring to the sons of the gentry, their general absence from library patronage activities seems curious; on the other hand, schools were imperial institutions, and perhaps too far inside the official fence, and the official budget, to attract outside funding. To explore why library construction did not engage the local gentry, it might be useful to consider a rare exception when it was. The outstanding statecraft scholar of the late-fifteenth century, Qiu Jun (1420–95), started life as a native of distant Hainan Island. His family owned a good collection of books running to several hundred fascicles until his father’s death in 1426, when Qiu was only 6 years old. The collection was subsequently pilfered. As Qiu grew older and needed books to pursue his studies, he tried to get the collection back, but was able to recover only fragments of it. His alternative was to rent books from local book dealers, though he complained that “the books on the market were mostly low-brow, miscellaneous works, so those [of any academic merit] that I was able to acquire amounted to only a handful.” In the distant south, commerce was not circulating the canon. Although Qiu overcame this disadvantage and launched a successful career in the bureaucracy, when he returned to Hainan in 1469 to observe mourning, he rediscovered the disadvantage of living in a place lacking in good books. Although the prefecture, which was contiguous with the island, had ten counties, only one of these counties on the entire island had a school library. Qiu decided in 1472 to build a library for Hainanese students at the prefectural

The chinese state in ming society

110

school. To guard against the humidity of south China, he had the library built of stone; only the cabinets containing the books were made of wood. Stone Chamber, as he called it, survived for over a century until the books were moved in 1614 to a new building.57 Qiu declared enthusiastically that Stone Chamber was “a narrow place from which one can grasp the breadth of all within the four seas,” and its books were the means “to grasp the world for ten thousand li from within the space of one room.” He saw his project as giving future students on Hainan Island the power to transcend their location and participate in the communicative context of scholarly and political life. The two for Qiu went hand in hand. Stone Chamber was a library designed to make books accessible, not store them away. Qiu Jun was the exception to the rule that school libraries were built by incumbent magistrates. Rarely are gentry named as library builders,58 an absence that contrasts strikingly with their visibility in other local building projects, most notably as patrons of Buddhist monasteries from the 1550s onward. When the gentry patronized a public institution like a monastery, they tended to do so as a collective undertaking through which they associated with each other and presented themselves as the elite of local society.59 Library builders were officials, outsiders who belonged to the world of state power rather than local power. This would suggest that library patronage was viewed as situated principally within the field of state power, not local power. When gentry patrons took the lead, as Qiu Jun did, it seems that they did so from a structural position as individuals competing in the bureaucracy, not as members of the local elite. From this perspective, it could be suggested that Qiu built Stone Chamber in relation to his prestige as a metropolitan official and his position as the principal patron of Hainanese students working their way up the ladder of success, not as a gesture of solidarity with other local gentry. The same logic may be applied to the financial contribution that the hated Chief Grand Secretary Yan Song (1480–1565) made to the library at the Yichun county school, in his hometown in Jiangxi, for library purchases: acting as an individual whose power base lay outside his home county, not as a member of the local elite.60 The increase in the numbers of books that schools owned, while a necessary condition for building a library, was not necessarily sufficient to move either a member of the gentry or the local magistrate to organize such a project or raise the funds without other factors coming into play. Magistrates were far likelier to engage in this work than the gentry because of factors specific to their position and career. No Ming document addresses this issue directly but the circumstances suggest several factors that may have been significant. First of all, the magistrate was the custodian of imperial gifts, which amounted to state property; he would have been concerned to ensure that the books the palace issued were well protected. Secondly, as the local official responsible for education, the magistrate would have encouraged a project of this sort that might strengthen the chances of those registered at his school to pass the examinations. A third factor may be a consciousness of competing institutions that were not under the magistrates’ aegis, such as Buddhist monasteries, and even more, private academies. It was understood that an academy should have a library, as a Jiangxi author notes in his account of the founding of a school library: that “academies have pavilions in which to store books”61 is a good reason that the school should. One more factor may have tied local officials to library-building. It is suggested by the timing of the first wave of building to the decade from 1438 to 1448. The magistrate was

Building school libraries in the mid-Ming

111

responsible for keeping in good repair the institutions that the dynastic founder mandated. This could pose a heavy burden at certain times, since buildings were not permanent structures but had to be restored on a regular basis to stay standing. It was generally accepted that a large building needed to be restored once every fifty years to keep it in good condition,62 and more strongly believed that going beyond sixty years threatened a building with dereliction. Since the first round of school-building had occurred in the immediate wake of the Hongwu emperor’s demand for county schools in 1369, the cycle of decay was just swinging past many of these original schools in the 1430s. This was when they had to be restored, which was just when the first libraries were being built in any number. This coincidence may indicate that, as repair schedules were forcing school restorations onto magistrates’ workloads, the latter were responding by exceeding the original design and enhancing local schools with libraries. That in turn may have to do with cultural shifts that touched magistrates as much as it did the gentry in the midfifteenth century: rising educational levels compared to when the schools were first founded, an increasing intellectualization and bookishness in elite culture, the first signs of an independent philosophical sphere developing independently of the state curriculum, and of course the growing availability of books. All may have pushed school restoration projects in the direction of library-building. Critically minded magistrates who imbibed the new learning of Wang Yangming and his immediate predecessors might, like Qiu Jun, have hoped that libraries could open students’ minds in a way that the set curriculum did not. On the other hand, this may be the point at which this flow of reasoning runs dry, since the schools represented the state’s interest in perpetuating that curriculum. The conversion of the library and shrine of the Luling county school in central Jiangxi into Baiyang Academy in 157163 might just signal the heady vitality of the academy movement in that province at that time, but it might also indicate the poverty of the state curriculum and the ebbing of magisterial enthusiasm for school libraries. Reverence and teaching Even among patrons, what a library was intended to achieve—and thus why you would want to build one—was a matter of disagreement. The essays written by library patrons commemorating the founding or restoration of school libraries reveal subtle differences in this regard that revolved around the tension between book knowledge and state power. These texts often explain the reason for a library’s existence in relation to a popular construction of classicism prevalent in the mid-Ming. An essayist writing on the building of the Shanghai county library in 1484, for instance, justifies the cost by stressing the sacredness of the classics that were to be stored there. The classics were the vehicle for the Way, and as such had a special status that demanded reverence (zun). A Pavilion for Revering the Classics was a good investment because it was an appropriate way to do just that: to inspire the proper sense of reverence among students, who had to learn that this should be their relationship to the knowledge enshrined in the Confucian classics.64 Without reverence, students would not know how to approach the canon. A similar way of thinking guides an essay commemorating the repair of a school library in Quanzhou in 1535.65

The chinese state in ming society

112

A different approach can be detected in an essay by the great Jiangxi neo-Confucian Luo Hongxian (1504–64), whose cartographic work was introduced in Chapter 2. Luo wrote this essay to commemorate the building of the Jishui school library in 1546. He begins the essay by clearing away what he considers false notions about classicism: it and antiquity, he insists, are two separate things. Although the classics are used to teach the learning of the ancients, he notes that they are not all ancient texts. Furthermore, though the classics deserve to be revered, reverence of the classics was not an ancient practice. Classicism is a present construction that has only an attenuated link to the past. Luo’s point was to underscore Wang Yangming’s notion that the ancients had more direct ways of grasping their “fundamental minds” (ben xin) than by relying on fixed texts. Texts had become important only because of a falling away from the correct understanding, which the ancients had achieved through self-cultivation rather than sterile book-learning. In these later ages, men were obliged to rely on the classics as the medium through which to apprehend the teachings of the ancients because they had no other resource.66 Luo thus put himself implicitly at odds with the reverential pose of the Shanghai essayist, who saw the classics as constituting the only direct connection to the Way Like Qiu Jun, Luo did not object to treating the classics with reverence, but he saw the importance of a library in terms of its protection of the physical volumes, not of their symbolic power. The Way was sacred, the books themselves something less. Reverence (zun) was important only in the context of teaching (jiao).67 The act of reverence represents an orientation to knowledge that is closed and limits itself to established standards. Teaching, as Luo Hongxian understood it, should be an open process that extends and communicates knowledge, not a closed reinforcement of the authority of what is already known. Zhu Xi would have agreed with Luo. He honours the classics in his essays on school libraries, yet he consistently stresses the importance of teaching, and writes as though he expected that local students were actually supposed to read the books.68 Zhu would also, though, have been puzzled that Luo needed to fight this sort of rear-guard action against Ming academicism, an effect of the entrenchment of the examination system as the narrow gate to social advancement. The tension between zun and jiao was played out among some gazetteer compilers, who faced the choice of where to insert school library catalogues in their publications. In the 1517 Jianchang and 1540 Hejian prefectural gazetteers, the book lists appear with the bibliographies of works by local authors.69 This approach highlights the books as repositories of learning, particularly local learning, in a way that Luo Hongxian would have approved. The more usual place to insert the catalogue, however, was among the inventories of school property: with the lists of ceremonial vessels, musical instruments, sheet music, even land.70 In this position, books got represented less as texts to guide learning than as property to be evaluated and conserved. The stewardship of books was thereby placed within a field of reverence, not of teaching. From this point of view, libraries were sites where the classics were canon, not knowledge. We accept libraries as inevitable institutions in print cultures, yet no necessity leads from books to specialized buildings devoted exclusively to storing them. This happened at Ming schools, however, all over the country The spread of school libraries must be placed in the context of book production: initially, to house the growing edifice of canonical texts that the Ming court was busily producing and obliging schools to receive, and subsequently, to accommodate the growing volume of commercial books that came

Building school libraries in the mid-Ming

113

into school possession. As books accumulated, their storage and management became problems that had to be solved, and one of these solutions was to build a library. County schools were not unique in this. Other institutions and individuals were building larger libraries from the mid-Ming forward, as books became available for acquisition in greater numbers and the competition among bibliophiles to push to, and then beyond, the 10,000fascicle ideal and to find rare volumes intensified. The book culture of the era thus has to be taken into account to explain why magistrates got into the practice of building libraries for their schools. Yet library-building within schools was linked as well to bureaucratic and intellectual issues. Our explanation is made more complex because of the special position textual knowledge occupied in China, where the Confucian system of knowledge, codified as it was in a curriculum of fixed texts, was tightly bound to the state’s exercise of power. The posture of reverence, by subordinating knowledge to power, expresses the tightness of this bondage. For scholars such as Qiu Jun and Luo Hongxian, deeply committed to Confucianism’s more open notion of teaching, this relationship was subtly to be resisted. While both motivations were at work urging local magistrates to build libraries, the attitude of Qiu and Luo seems to have been the minority view among patrons, who were intent less on “grasping the world for ten thousand li” than on defining and limiting what can be known. When they built edifices to house books, most magistrates understood that they were participating in the state’s project to edify and control knowledge, not to open men’s minds. Our temptation to read the history of library-building in the Ming as reflecting a social movement to open the process of knowledge acquisition and spur its development may be only that: ours, not theirs. Still, for all the conservatism that shaped the state school system, some minds may well have been opened, as indeed many were in the sixteenth century, and some of these may have been pushed open just a little by having gone to the library to read.

6 State censorship and the book trade The history of the book in China tends to regard the Ming and the Qing dynasties as censorship states. The Ming is portrayed as the dynasty when censorship became a feature of state control: no longer an episodic occurrence, but something approaching a political movement: the beginning, as one study from Shanghai puts it, of the long night of state suppression of the expression of ideas in China.1 What the Ming began, continues this logic, the Qing continued and intensified. Its greater capacity to inspect and intervene in local affairs only spread the net of censorship even more tightly over society2 The evidence in this essay confirms that the late-imperial state did indeed prohibit books and ban writers. But the assertion that these initiatives constituted something that could be called state censorship is a modern construction. Judging the Ming and Qing to be censorship states may depend too much on looking back from the nightmarish blackouts that twentieth-century Chinese states imposed on Chinese readers and writers, and not enough on looking at what actually occurred when a book, or more often a writer, was brought to an emperor’s attention, or what an official argued for when he offered a text for imperial suppression. Our concept of censorship today is rooted in an understanding that ideas can be mobilized as a political resource: that books provide effective vehicles for disseminating those ideas and that the modern nation-state has a role in controlling and exploiting this resource. The history of the twentieth century is rich with examples of states both liberal and dictatorial that actively intervened in the production and dissemination of books when their leaders felt threatened. The state censorship of books in the late-imperial period had much to do with the ideological anxieties of the Ming and Qing states regarding the good order of society and dynastic legitimacy, yet neither dynasty employed censorship as a sustained ideological intervention as modern nation-states have. The Qianlong court (1736–95) came closest during the years from 1772 to 1793, in the process generating ample documentation on the state suppression of books. It is for this reason that I focus in this chapter on this period, though I do so for a purpose other than analyzing censorship as a mode of ideological discipline. My purpose, rather, is to examine this material for what it reveals concerning the conditions of book production and circulation. My approach is not to follow the ideas that books expressed, or even the authors who wrote them, but to follow the books. I will begin by sketching my approach in relation to two instances of book censorship in the Ming. Censorship may not have a unified history across the Ming and Qing: the differences in their modes of sovereignty and in the cultural understandings by which they framed their right to rule were great. In particular, Ming emperors were never beset by legitimacy threats based on ethnic or cultural difference, as Manchu emperors were. I have put the Ming and Qing together for another reason: whatever the rupture between the dynasties, the history of the book across them was continuous. Anxieties about dynastic survival changed in 1644; books did not.

State censorship and the book trade

115

New Tales while Trimming the Wick Chancellor Li Shimian (js. 1404) of the National Academy submitted a memorial in August 1442 in which he told the emperor about his distress over the popularity of writings that “dress up strange events in ungrounded language.” His particular example was a collection of fanciful and erotic fictional tales entitled Jiandeng xinhua (New tales while trimming the wick). The book bears a preface dated 1378 and was reprinted in 1420. A copy came to Li’s unhappy attention, probably because he found his students reading it.3 Li is not explicit about what he regarded as problematic about the book, beyond vague suggestions of impropriety. He seems more disturbed by the book’s readership than by the book itself, for he pictures “many professional scholars tossing aside their proper studies and devoting their days and nights to memorizing passages in order to use them in conversation.” This was not what an education in high literacy was for, and in Li’s dark imaginings it could only produce a domino effect. “Unless this sort of thing is strictly prohibited, I am afraid that heterodox doctrines and eccentric views will come to life every day and burgeon every month, exciting confusion in people’s minds.”4 What precisely constituted “heterodox doctrines and eccentric views,” and what was the hurdle scholars had to trip over before they could be judged as having transgressed “proper studies”? Li does not specify, nor does any other censorship memorial of the fourteenth or fifteenth century that I have seen. Li takes his claim as self-explanatory from a soberly Confucian point of view, and yet, as I have argued elsewhere, Confucianism is not a tradition that favours suppressing frivolous or alternative ideas. Not all books were equally sound vehicles of the truth, of course. As Cheng Yi warned of the writings of some of his eleventh-century contemporaries, bad books could be potent enough to “damage the Way,”5 and wasting one’s time reading them detracted from study, yet Confucian intellectuals did not go on to endorse the notion that the state should suppress such books. If the Way was damaged through reading, it was the fault of the reader, not the book. If the reader was to be corrected in his error, the task was his teacher’s, not the state’s.6 Nonetheless, some Ming officials desired to bring the state into play and mobilize its resources to track down copies of books they judged offensive, as did Li Shimian when he asked the Ministry of Rites to send out the provincial censors to look for copies of New Tales while Trimming the Wick. “Both those who print and sell it, and those who own and read it, should be punished according to the Code.” Li’s request is extraordinary in two senses. First of all, no mechanism of state surveillance was in place for carrying out this sort of program. It is doubtful that provincial officials would take up this task with much enthusiasm when, lacking established procedures for reporting and feedback in this sort of matter, they had no assurance that their efforts would be rewarded or even noticed. Thus, when the emperor’s advisors sent the proposal over to the ministry for discussion, the minister of rites sent back a somewhat cagey response that such an order “could be carried out.” No records survive indicating what happened when the order went out. Li Shimian buttresses his proposal for suppressing the novel by appealing to the Ming Code. As it happens, no article in the Code provides for such censorship. The official who wanted to find an analogy for the crime in order to punish the owner or seller of New Tales had two places in the Code to go to. The first appears in the section on ritual, where people are prohibited from possessing such exclusively imperial items as astronomical

The chinese state in ming society

116

implements, sky charts, imperial portraits, and jade seals. This article, in a lovely piece of circular reasoning that recurs in Chinese state logic, prohibits “writings that should be prohibited.” The circularity suggests either that the framers of the Code were keeping the category intentionally vague, or more simply that they regarded the applicability of this law as self-evident: nothing could be permitted to trespass on the sanctity and legitimacy of the throne. Since the state of the heavens was always potentially a sign of legitimacy (as it was in Europe)7, any means for reading those signs, such as astronomical instruments and sky charts, was prohibited. This anxiety did not lead to the category to “books” in general as objects of suspicion, nor to the proposal that books needed to be inspected. Rather, it was to ensure that nothing within the prerogative of the emperor got out into uncontrollable public circulation. The other article in the Ming Code to which an official might go to censor a book owner or bookseller is the law against composing or owning “seditious writings” (yaoshu), explained there as writings that can delude the masses. Coming a prominent third in the criminal section—after plotting treason and planning rebellion—this law was intended to extend the state’s protection of the throne into the realm of writing, where threats to state authority might get expressed. It differed from the previous censorship law in extending the reach of punishment beyond texts that transgressed imperial prerogative by virtue of what they recorded, to texts that threatened the security of the monarch’s rule by virtue of what they said, specifically, prophesying or advocating dynastic overthrow.8 This law could be stretched to cover texts that did not so much propose the overthrow of the Ming as imagine an alternative realm, such as heterodox religious texts with a millenarian tone, particularly those with Daoistic-sounding titles. The earliest censorship case I have noticed in the Veritable Records is precisely of this sort. In 1390, a resident of Kaifeng submitted to the throne for censure a thousand tracts with such titles as The Great Unity of the Nine Palaces (Jiugong taiyi), The Great Unity Enters the Cosmic Flow (Taiyi ruyuri), and Rough Calculations of the Great Unity (Taiyi caosuan).9 Given that the Great Ultimate was thought to be the mysterious origin of all things, and an important concept for prognosticating the future, it is not surprising that someone thought these books worth bringing to the court’s attention. The Veritable Records does not use the label “seditious writings” (nor does it record whether the emperor responded) but that would have been the appropriate legal category. It is the category most consistently used in subsequent entries in the Veritable Records concerning writings the state deemed unacceptable. For example, a Guangdong man who “happened to obtain copies of seditious writings as well as printed texts and maps” used these in 1481 to form a group to “delude the masses and stir up chaos.” A North Zhili man “made seditious writings and deluded the masses” before leading an attack on a county seat in 1498. Both were executed, though their possession of “seditious writings” appears to have been criminal because of the more serious offence of open rebellion to which each case was linked and would not have come to the state’s attention had rebellion not followed. Most of the case summaries in the Ventable Records do not give details of the books’ contents, though a title that surfaced in Huguang in 1578, Record of the Rise and Fate of the Great Dynasty (Dachao qiyun lu), was too unambiguously antidynastic not to get suppressed.10 To return to New Tales while Trimming the Wick, did Li Shimian mean to situate this minor piece of erotica within the category of texts capable of stirring the masses to

State censorship and the book trade

117

rebellion, and expect that anyone in reach of a copy be executed? By comparison with the threat of rebellion contained in Record of the Rise and Fate of the Great Dynasty, this seems to be a harsh ruling. I doubt that Li Shimian would have called for the emperor to hand out executions for frivolous reading, however morally irritating he found it. Still, his comment about “exciting confusion in people’s minds” and his belief in the domino theory of moral effect point to the implication that erotica was a first small step toward sedition. Three centuries later, the Qing Code does make the connection. It repeats both articles almost verbatim, but attaches new jurisprudence to the article on “seditious writing.”11 The four substatutes added to this article in the 1740 revised edition of the Code cover a far wider range of censorable texts than the original Ming authors intended. It itemizes three types of texts as “seditious writings”: handwritten posters, satirical songs (whether oral, in manuscript, or in print), and printed pornographic fiction. Li Shimian would have had no difficulty fitting New Tales into these categories. Memorials such as his may have contributed, in fact, to turning the ban on “seditious writings” into a jurisprudence of censorship of morally as well as politically offensive texts. Neither of the articles in the Ming Code addresses books as something to be concerned about by virtue of being books. This is why I translate the substantive noun in both articles, shu, as “writings” rather than as “books” (the Chinese word covers both). The original framers were not conscious of printing technology as a threat to imperial prerogative. This indifference begins to change in the later years of the dynasty, however. When a Shaoxing man named Chen Yinming in 1609 was charged with “fraudulently printing seditious writings that falsely claimed to have been imperially authorized,” the critical issue was the forging of an imperial edict, but the technology of printing was beginning to come into view as a problem, as the oddly phrased expression “fraudulently printed” (jiayin) suggests.12 When monk Fuben printed fan poems by Yangzhou prefect Liu Duo (js. 1616), which the monk had in his private collection, he got into trouble for including a poem by a clerk in the Embroidered Guard, which the Nanjing Minister of Justice in 1626 deemed as “covertly attacking affairs of state.”13 The fact that the brief case summary in the Veritable Records refers to “published books” (tushu, zhishu) four times in three lines suggests the possibility that circulation was becoming as troubling to the court as content, that printing was emerging as censorship’s sufficient condition. Not until the 1740 edition of the Code does the problem not just of printing but of commercial distribution come to the fore. Substatutes explicitly cite “publishers” (kanke), “distributors” (chuanbo), “booksellers” (fangsi), and “printers” (keyin) among those who could be punished when the state chose to ban a book. Changes in written law always lag well behind changes in social and even legal practice. It did not take four centuries for the problem of printing to come to the attention of those who would ban books. An awareness of commercial printing was already there in Li Shimian’s memorial when he argued that “those who print and sell” books of this ilk should be “punished according to the Code.” The framers of the Code had nothing to say about printers and sellers, but Li did, since it was probably a copy of the 1420 printed edition of New Tales that prompted him to approach the state as the appropriate agency for removing it from the hands of his students at the National Academy. Printing increased the chances that people who should not read such stuff did. Something more than taking his students’ copies away from them had to be done. Not much was.

The chinese state in ming society

118

Books to be burned and hidden away The most prominent case of state censorship in the Ming—the one most often cited as evidence that the Ming was a censorship regime—is Zhang Wenda’s impeachment of the controversial essayist Li Zhi (1527–1602), leading to the latter’s suicide. Scholars of the generation to come of age after 1644 often commented on this case, to praise Zhang’s attack as an apt judgment on all that they felt had gone wrong in the Ming, and to hold Li up as a scapegoat for the moral collapse that preceded, and led to, dynastic collapse.14 The burning of his books was not, for them, an inappropriate use of state power, given that his writings were a part of what they felt had put the survival of Chinese culture at risk. The generation immediate junior to Li, on the other hand, was appalled at the treatment he received in Beijing and refused to bow to the censorship that followed. Li Zhi began to formulate his criticisms of Confucian authority in the 1570s when he became part of a Nanjing-based coterie at the cutting edge of intellectual speculation on moral action and the potential confluence of Confucianism and Buddhism. After retiring from public office, he came to doubt the possibility of setting absolute moral judgments by external standards. Truth was not a matter of what was written down in the Confucian canon; it was available only to the individual, and only through conscientious introspection. Li gradually distanced himself from his former associates, taking Buddhist tonsure in 1588. Though he did not formally withdraw from the world, for the rest of his life he lived as a self-fashioned Buddhist Confucian. Controversy eventually drove him from his adopted home in northern Huguang. In April 1602, when Zhang Wenda (d. 1625) presented his memorial, Li was living outside Beijing with Ma Jinglun (js. 1589). Among other charges, Zhang alleged that Li’s books had become a danger to the younger generation. Since 1590, Li had been energetic in getting his essays, reading notes, and letters into print and into circulation, and giving them such baldly provocative titles as Fen shu (A book to be burned) and Cang shu (A book to be hidden away). Zhang demanded that these books be given the burning they called for. In his own defence, Li acknowledged that “the accused has written many books” but protested that his oeuvre “should be seen as benefiting the propagation of the teaching of the sages rather than damaging it.”15 This was too disingenuous, given the cold wind blowing at this time. We can detect this mood in the writings of a contemporary, Tu Long (1542–1605), who mentions in a preface to a collection of aphorisms he published in 1600 that a friend had warned him not to let a book of his “be distributed too widely or reach the capital.” Tu decided to follow this advice, stating that “for the present I have hidden it away in a trunk.”16 The cold wind that would sweep Li Zhi to prison two years later was thus touching others as well, though the fact that both were bold enough to openly advertise the controversial content of the writing they were hiding away suggests that they did not see themselves in immediate danger of suppression. Perhaps they claimed this need more as a declaration of intellectual autonomy from the world than as a prediction that state censorship would fall on them. Did the state regard the ideas of Tu Long and Li Zhi as dangerous? Zhang’s impeachment certainly damned Li for criticizing the Confucian canon and scoffing at Confucius, but the focus of his criticism was Li’s conduct, especially allegations of improper relations with a female student. This is the smoking gun which points toward the politics behind Zhang’s attack. The charge concerned one student in particular, the

State censorship and the book trade

119

widowed daughter of the capital official Mei Guozhen (js. 1583). Mei was in conflict at the time with the eunuch-backed faction surrounding Chief Grand Secretary Shen Yiguan (1531–1615), who was Zhang Wenda’s political patron.17 On a rumour that Li was preparing an attack on Shen, Zhang struck first. The rumour was not true, though it was widely understood at the time that Li was a pawn, and that the object of Zhang’s attack was Mei Guozhen. “People never thought Li would delude the world or advocate licentiousness,” his host Ma Jinglun observed shortly after his suicide. “Those who heartlessly stirred up this business were clearly not after this old man.”18 The judgment on Li was a suspended sentence and an order to burn the printing blocks of his books, but the file got stalled at court and Li chose to commit suicide rather than face the humiliation of being sent home in ignominy. Discussion of this case has tended to focus on the substance of Li’s ideas, the moral aptness of Zhang’s critique, and the nature of the intellectual conflict involved. But Jin Jiang has argued persuasively that the state did not go after Li for his ideas. These may have been objectionable, but the court would normally have given them no notice had Li prudently stayed away from the capital. She concludes that what got him into trouble was his heteropraxis, not his heterodoxy; that “it was his physical presence in the area, not his publications, that warranted the court action against him.”19 Fair enough; his heterodoxy could be ignored as a matter of individual misjudgment that need not invite state suppression; but I do not see his publications as quite so innocent of the charge. Think about what a book by Li Zhi was: the physical vehicle of Li’s ideas, certainly, but also a sign of his identity that could travel widely and disseminate his reputation. Zhang Wenda made much of the books, especially the fact that “recently he has once again published them,” with the result that they “circulate within the realm and excite confusion in people’s minds.” Zhang’s recommendation was that “they cannot not be destroyed.” If this instance of censorship was a matter more of praxis than of ideas, aggressive publishing was part of the praxis that created Li’s reputation and troubled those whose position his reputation might threaten. Despite Zhang’s call for the destruction of Li’s books, people continued to print, own, and circulate them for the next four decades. A collection of some of Li’s unpublished writings, pointedly entitled A Continuation of a Book to be Hidden Away (Xu cangshu), came out in 1618 under the editorship of none other than jiao Hong (1541–1620), the senior figure in the Nanjing scholarly world. As Li Weizhen (1547–1626) noted in his preface, “Master Li has disappeared, and yet his surviving books circulate in large volume.” Li also had the temerity to include at the end of his preface a list of all the prominent people who had ever associated with Li Zhi.20 Seven years later, Censor Wang Yaliang asked that the ban be repeated, but went a step further by targeting not just the books but the commercial networks that distributed them, asking that “it should be impermissible for bookshops to sell them.”21 The bureaucratic response appears to have been less than half-hearted once again, for a decade and a half after this second suppression, Qian Qizhong (js. 1628) sponsored a subscription campaign to engage a commercial printer in Ningbo to reprint Li Zhuowu zhiyi (Li Zhi’s regulated-meaning essays). This book was a collection of Li’s model examination essays written in the socalled “eight-legged” (bagu) or “regulated-meaning” (zhiyi) style.22 Despite the noisy harping of late-Qing commentators, Li Zhi, Qian Qizhong, and indeed the rest of the lateMing intelligentsia, including Tu Long,23 saw nothing wrong with composing essays in

The chinese state in ming society

120

this style. Qian was not unaware of the censorship campaign against Li, for he notes in his fund-raising appeal that “when the Master was arrested, the authorities burned his books and all copies in print of his regulated essays were also banned in order to eradicate his writings.” But clearly the suppression was of no concern to him, and had no influence over his own reading and publishing choices. “However much they wanted to bury his name, they could not throw all his writings into the flames,”24 he blithely observed. The fact that the court’s ban against Li Zhi was flouted outside Beijing suggests that intellectuals at the time felt no obligation to uphold an essentially factional political decision, or even fear it. It is a sign of something else too: that however much the state might wish to remove certain books from the public realm, it had little capacity to stop them from circulating when commercial printers kept them in print and book buyers wanted to purchase them. The rich sources for the literary inquisition of the Qianlong emperor will show how deep the state’s lack of control over the book trade could go. The Complete Library of the Four Treasuries In an edict issued on 11 December 1774, the Qianlong emperor warned his officials that books published back in the seventeenth century were liable to include “rebellious words”: “If the books contain language that is anti-dynastic, then the woodblocks and printed sheets must both be put to the flames. Heterodox opinions must be quashed that later generations may not be influenced” (p. 126).25 Six weeks later, the emperor once again inveighed against books of this era: “These seditious books, both those in printed form and their blocks, should be sought out and prohibited” (p. 136). The emperor’s concern over “seditious books” came in response to the flood of texts pouring into the capital for the massive state bibliographic project known as the Siku quanshu, the Complete Library of the Four Treasuries. This imperial book-collecting project was initiated three years previously to expand the holdings of the court so that it might include copies of all known works. By 1774 the bibliographers began turning up books that the emperor found offensive. What followed was an intermittent campaign against certain types of books that would drag on for fifteen years. At the core of Qianlong’s concern was not Confucian orthodoxy or sexual propriety, but how to represent the historical relationship between China and Inner Asia. The legitimacy of the Manchus depended on their being seen as rightful receivers of the mandate of heaven, rather than as barbarian interlopers from beyond the pale of civilization, which is how pre-Qing Chinese texts often represented them and their steppe antecedents. Since civilization meant, among other things, the recording of knowledge in texts, those who controlled that record held the keys to state legitimacy. Qianlong needed to dominate discourse about the past so as to be able to project certain historical interpretations that would justify Manchu rule in the present. Qianlong’s campaign was common knowledge in nineteenth-century Europe, to judge from a passing comment by Karl Marx on Prussian censorship in 1842.26 It first received Western scholarly attention in the 1930s from Carrington Goodrich who, working on the basis of a Chinese scholarship that had come of age under the new Republic and was steeped in anti-Manchu sentiment, explained what he termed the “literary inquisition” as

State censorship and the book trade

121

an expression of Manchu anxieties about their right to rule China. The roots for this view go back to the time of the inquisition itself, when foreign observers such as the Korean Pak Chi-won, who visited China in 1780, frankly stated that “instead of burying the scholars alive as the Qin dynasty did, the court buries them in labours of collation; and instead of burning the books as the Qin did, it scatters them in the Bureau of Assembled Pearls,” the publishing office at court. The sporadic harshness of the inquisition was in turn attributed to the ill humour of an aging emperor. According to Pak, Qianlong “becomes daily more violent; he is suspicious, brutal, and overly strict; he has unpredictable fits of glee and rage.”27 The inquisition has thus come to be characterized in terms of Qianlong’s personality and his suspicion of the scholarly community. In his study of the Four Treasuries, Kent Guy has questioned this emotional assessment. Without denying that ethnic conflict and the whims of the emperor had much to do with the drift of the inquisition, Guy has shown that it grew into a hydra of suspicion and denunciation because the Chinese (as opposed to the Manchu) elite found in the project’s hazy guidelines opportunities for pursuing personal vendettas. Officials began to denounce each other, both to settle old scores and to attract the attention of higher officials.28 In fact, as Guy points out, it was the Manchu elite that intervened to halt this spiral of denunciation, fearing it would shatter the working agreement between them and the Chinese. The literary inquisition was thus shaped more by Manchu-Chinese collaboration than by a desire on either side to rupture that collaboration. To understand why the Qianlong inquisition occurred at this time, when the Manchus were at the height of their power and unchallenged in their rule, one would be hardpressed to build an explanation from the ideas that get voiced in the documents, of “antidynasticism,” “heterodoxy,” and “sedition.” Consider, instead, another set of terms: “woodblocks,” “printed sheets,” “books, both those in printed form and their blocks.” These references draw our attention to the technical environment of book production and distribution within which the inquisition occurred and to which it was largely responding. If we choose to follow the books rather than the ideas, we can begin to develop a less ideologically trapped—and hence more historical—account of what the Qing court found itself up against, and what it found difficult to control. Lacking any institutions through which it could interact with the publishing world, it had to implement its surveillance from outside the profession, operating the inquisition as a bureaucratic exercise. This structure was to shape the inquisition in ways that had little to do with judgments about which books were acceptable and which were not. As it was practised, censorship in the eighteenth century unfolded in relation to the technology of book production and the technical constraints of government surveillance at least as much as it did to ideas of legitimacy or the state’s perception of its relationship with other sources of authority. The history of Reformation Europe provides comparative insights for developing a technical approach to censorship. The sudden development of print in the fifteenth century occasioned unambiguous responses. In England, Henry VII moved to license publishers within nine years of the arrival of the first movabletype printing press.29 The spread of printing technology was even more alarming to the papacy, which alerted secular leaders to the possibility that the new invention might be used to spread heresy. Sixtus IV in 1479 authorized the rectors of the University of Cologne to censor books, their printers, and their readers; Innocent VIII in 1478 issued a bull warning against heretical books, and banned specific works in 1483 and 1487; and Alexander VI in 1501

The chinese state in ming society

122

and Leo X in 1515 authorized universal censorship of the press by the church.30 The early censorship decrees, however, had little or no effect prior to the appearance of papal indices banning certain titles in 1559 and 1564. The papacy had reason to be concerned, for such controversial figures as Martin Luther were using the new invention to produce books and pamphlets critical of the church. By the time of the papal indices, both Catholic and Protestant states were censoring printed matter. Using printing to frame censorship history in late-imperial China is more difficult, as China had been a xylographic (woodblock) print culture for a millennium prior to the Qianlong reign. The earliest surviving samples of woodblock printing are Buddhist texts printed in the eighth century, produced in multiple copies to generate multiple merit. By the eleventh, the block-printing of books had become so widespread that the bibliographer Ye Dehui (1864–1927) could count twenty-one different expressions for “woodblock” in sources from the period.31 Printing was not something new to the Qianlong era. There was no sudden technological change to precipitate an intensification of censorship, no visible shock created by an alteration in the technical conditions of the period, as there was in Reformation Europe, where the application of movable type to the printing process has been singled out as the key to the social impact of books.32 The same logic cannot be applied to China. Chinese printers were using movable type at least as early as the ninth century, though for a variety of reasons this technology did not displace the xylographic production of books.33 The outstanding examples of Qing-dynasty movable-type printing happen to come from the palace itself, for the finest Four Treasuries editions were produced using movable type. Situated in a social and political context unlike that of Renaissance Europe, this technology did not upset the conditions of book production in eighteenth-century China. The differences between China and Europe in the character and role of the state (universal versus regional) and in the political and religious parameters of intellectual inquiry (state-dominated versus church-dominated) dulled the impact of printing’s challenge to the Chinese political order. I shall argue, nonetheless, that the history of printing in China was a major influence on the literary inquisition, just not in the way it was in Europe. What we must look for is not a sudden technological shift, but a slowly developing, cumulative impact. The logic of this approach is simple: by the eighteenth century, the number of books in society had gone beyond any precedent. The xylographic technology that had emerged by the eighth century, which by the eleventh was bringing books into print in great volume, did not undergo further technological change.34 Except for the painstaking process of using multiple blocks for colour printing, developed at the start of the seventeenth century, the processes of printing remained static and simple. What changed was the cost: by the sixteenth century, the cost of engraving woodblocks had fallen dramatically, as illiterate labourers mastered the necessary skills.35 The Jesuit missionary Matteo Ricci (1552–1610), for one, was struck by the ease and cheapness of book production in China: “The whole method is so simple that one is tempted to try it for himself once having watched the process,” he declared. “The simplicity of Chinese printing is what accounts for the exceedingly large number of books in circulation here and the ridiculously low prices at which they are sold.”36 The decreasing cost of producing a book was an effect of the extensive process of commercialization in the sixteenth century coupled with the simplicity of the technology. Entrepreneurs seized on the commercial possibilities of publishing, and a burgeoning commercial press resulted.37 So too did private collections.

State censorship and the book trade

123

The inquisitorial investigation of Buddhist monk Dangui in 1775 revealed that the family in Shaozhou, Guangdong, that had helped finance the publication of his writings in the previous generation owned 121 titles, plus the blocks to two more books. Their in-laws owned 68 titles.38 Neither collection was regarded as remarkable. Commercialization’s effect on censorship was not straightforward or even direct, however. As the memorials and imperial rescripts generated during the course of the Qianlong inquisition show, the books that came under proscription in the eighteenth century were more often produced by private individuals than by commercial publishers. Commercial publishers generally kept to politically innocuous mass-market items like almanacs and novels—though a series of edicts between 1652 and 1754 sought to prohibit bookstores from selling books, such as “licentious novels,” that were not “of benefit to neo-Confucian philosophy and good government.”39 (Li Shimian and Zhang Wenda would both have approved.) To a politically sensitized mind like Qianlong’s, privately published books were the greater worry because he supposed that they lay beneath the radar of his officials and circulated quietiy through relatively closed networks. Qianlong soon learned that this was not the case. Commercialization may not have printed all the books he found offensive, but it sustained the networks that allowed even privately produced books to circulate beyond the gift market.40 How much worse it was, though, when books were “printed for commercial distribution,” as one indictment bemoans,41 for then they took on the mobility of full commodities and travelled far from the hands of their original publisher. Take for example the case of Cai Xian (b. 1697), a juren degree-holder from the Shanghai area.42 In the decade preceding his arrest in 1767, Cai had blocks for seven of his books engraved, the last of which was Xianyu xianxian lu (Idle notes of an idle (fisherman). Idle Notes typifies the interaction between private publishing and the commercial book trade. Cai hired a local engraver by the name of Wen Zishang to cut the blocks for his book; he then arranged with a travelling book salesman (shuke) from Huzhou named Wu Jianqian to print 120 copies. Wu supplied his own paper and hired a local professional printer known as Engraver Ma (the inquisition documents do not record his given name) to print the book at Cai’s house. Gai received twenty copies as payment for the use of his blocks, and Wu kept the other hundred. Wu gave one copy to Ma as his fee (literally, “wine money”). Ma was illiterate, but the book was worth something to him because of its resale value. Wu took his remaining ninety-nine copies back to Huzhou to sell. He still had some in his possession when he was apprehended, but those he had sold the provincial authorities were unable to recover. Of his twenty copies, Cai gave one to Wen Shengyuan, the son of his engraver Wen Zishang, also illiterate, when he heard that Zishang had fallen ill and needed money. He distributed all but five of the rest to “subordinates and associates.” Some of the recipients treated them as saleable commodities. An affinal relative immediately took his copy to the town of Jiangning, just outside Nanjing, and sold it for an undisclosed amount. (The speed with which he got rid of the book saved him from punishment later.) Wen Shengyuan was intending to sell his to pay for his father’s medical expenses, but had not done so at the time of the investigation. (Had he lied to Cai Xian about needing the money?) Chen Mingshan, who bore no relationship to Cai, had acquired a copy through trade from one of the original recipients and hoped to make money on what everyone regarded as a “fashionable text” (shiwen), though he too was still in possession of it when

The chinese state in ming society

124

Qianlong’s censors took notice of Idle Notes. Like Wen Shengyuan and Engraver Ma, Chen was illiterate, so the book was valuable to him solely as a commodity. (Illiteracy would mitigate the severity of the punishment to which Chen, Ma, and Wen were sentenced: eighty rather than a hundred blows, and exemption from the three-year banishment otherwise mandated for selling prohibited books.)43 The copies Cai and Wu had given to friends and associates the investigators were able to trace, though those that had moved along commercial rather than personal networks they did not recover. The officials burned all the copies they could track down, along with Cai’s blocks. Officials again and again were frustrated by the difficulty of reconstructing all the ties in commercial networks. Book production was dispersed to begin with, and the peregrinations of the men who transported books around the country to supply local bookstores made the task of following books even more confounding. The movement of wholesale commodities was not random, of course, which meant that an energetic censor could track down many of the copies of a proscribed book if he chose to take the trouble. Take the case of literary critic Shen Deqian (1673–1769), posthumously suppressed in 1777 for compiling an anthology of Qing poetry that included ostensibly seditious verses.44 When the provincial governor of Guangdong received the order to scour his part of the country for Shen’s anthology, he sent out agents to check the bookstores of Guangzhou. They found no copies, but were told that book salesmen from Jiangsu who might be carrying Shen’s works stayed at the Nanjing merchants’ hostel. Investigation there turned up Zhou Xuexian, who had first come down to Guangzhou from Nanjing to sell books in 1760. It was discovered that he had also brought with him a manuscript copy of the poetry anthology in question, and had had blocks for it engraved in Guangzhou, taking advantage of the cheaper engravers’ rate there. Zhou did not print copies from these blocks, but took them back with him to Nanjing, where he had the printing done. The inquisition documents are not clear as to whether he brought any copies with him when he returned to Guangzhou. Three censored books To explore facets of the relationship between publishing and state proscription in the Qing dynasty, I have chosen three late-seventeenth-century books that fell foul of the censors, each of a different type—one a commercial production, one produced by an institution, and the third published by a private individual—each tangled in the operations of state power for different reasons, yet each with links to the others. Gui Qian chidie (Model letters of Gui and Qian) is a collection of letters by two prominent scholar-officials, Gui Youguang (1506–71) and Qian Qianyi (1582–1664). The works of both these men were banned by the literary inquisition, as was this title.45 The Qianlong emperor developed an intense personal animus toward Qian for having dared to serve both the Ming and Qing regimes as a high official. Gui, a creative but unsystematic essayist of the Wanli era, was suspect not for anything he wrote because his works had been annotated and printed in the 1670s by the Nanjing scholar-publisher Lü Liuliang (1629–83). Lü had been posthumously dismembered for anti-Manchu sentiments by order of the previous emperor, Yongzheng (r. 1723–35), in the largest sedition case of that reign. Anything he published was tainted by association. Model

State censorship and the book trade

125

Letters was edited for publication by Gu Yu, a native of Changshu county, which was Qian Qianyi’s home county A preface by an equally unknown person hints that Gu put up the money to have it printed.46 According to the title page, on which Gu’s name appears, the book was “distributed” by Wanwei tang (Hall of Courtesy). Carved into the blocks at the end of most of the individual fascicles are the imprints of two seals stating that the book is “in the possession of Master Gu” and that it was “printed by Ruyue lou (Moonlight Tower) of Yu Mountain.” Yu is a picturesque mountain adjacent to the city of Changshu and served as its literary designation. There is nothing in the book that explains the nature of the relationships among Gu Yu, the Hall of Courtesy, and the Moonlight Tower. Gu’s own postface is silent on this; it reads rather like advertising copy, informing the reader that the book would be a valuable addition to any household. It would appear that Gu Yu, who held no higher degree, was not a scholar but a publisher; that Moonlight Tower was his publishing house; and that the Hall of Courtesy was either a bookshop that arranged to distribute the book for him or, more likely, a publisher that bought up the woodblocks sometime after 1699 and reissued the book from the original blocks, adding its name to the title page but not bothering to remove the old colophons. In the eighteenth century, the commercial production of books was a complex industry. The ban on Qian Qianyi extended to other books as well, including Caoxi tongzhi (Comprehensive gazetteer of Cao’s stream).47 This descriptive record of the prestigious Nanhua Monastery in Guangdong was published in 1672 following a major reconstruction. Though compiled by a layman, the book was engraved and printed at the monastery. Such an institution was well endowed to publish, for it had the funds to print its publications and the space to store the bulky woodblocks, and, unlike a commercial press, could afford to produce one or two specialized books and leave the blocks idle until they were needed for a reprinting. There had been an earlier edition of the Nanhua gazetteer, compiled in 1598 and published in 1604 or possibly earlier, but Nanhua Monastery had changed greatly in the intervening years. The post-reconstruction gazetteer compiled by Ma Yuan was an entirely new edition. The 1677 edition was banned for several reasons. Most immediately, it was found to include writings by Qian Qianyi and other proscribed writers. Additionally, it was found that the Comprehensive Gazetteer of Cao’s Stream “in its language offends against taboos”48 concerning the use of imperial titles. That some books fell short of this standard by flouting naming and dating taboos caught many officials by surprise. They had assumed that the transaction between subject and ruler, the agreement that the Manchus had inherited the mandate of heaven, had been worked out long ago. We note this uneasy surprise in a memorial two gubernatorial officials (one of them a Manchu) submitted to the emperor in 1774. To cover their embarrassment at having failed to identify a “seditious” work circulating in the area of their jurisdiction, they wrote: “We did not suppose that scribblers at the end of the Ming were privately writing stuff that contained seditious words and sentences, polluting the stream of history” (p. 116). Even greater was the surprise of another Manchu governor who three years later learned that he had contributed to the publication expenses of a lexicon found to contain infractions against the naming taboos of the dynasty, for which he was placed under sentence of execution.49

The chinese state in ming society

126

Unmentioned, but equally damning, is the Comprehensive Gazetteer’s connection with Shang Kexi (d. 1676), the Prince who Pacified the South. Shang was a Chinese general who brought Guangdong under Qing control in the 1650s, and who remained there as a sort of regional military governor until his retirement in 1673, at which point he became implicated in the Rebellion of the Three Feudatories, in which the south threatened to overturn Manchu rule, and was subsequently condemned. Shang had been the principal sponsor of Nanhua Monastery’s restoration in 1667, and his contributions are explicitly documented in the first and third juan of The Comprehensive Gazetteer of Cao’s Stream.50 The mere appearance of his offensive name in the book should have been enough to get it burned. Initially, though, it was not. The extant copies of the 1672 edition incorporate material to 1680, which indicates that the blocks were cut after the Three Feudatories rebellion was crushed. The proscription came later, though it faded in time: the prohibited edition was re-engraved for a new edition in 1836. Our third example of a censored book is the privately produced collection of essays and letters by the prominent writer Ai Nanying (1583–1646). This book was published in 1699, the same year as Model Letters. Tianyongzi ji (The collected writings of heaven’s hired hand)51 was published in the author’s home county in eastern Jiangxi province.52 According to the author’s grandson, Ai Weiguang (b. 1632), Ai Nanying “published as he wrote,” producing an oeuvre of short, disparate texts.53 In 1661, fifteen years after Ai died in the service of one of the Ming loyalist regimes fighting the Manchus, the prefect in his home region had a local scholar edit these remnants and arranged for a local printer to publish a uniform edition of his works. The printer later fell on hard times and pawned the woodblocks to a local family. The county magistrate subsequently redeemed them for 20 taels of silver and deposited them with a trustworthy student at the county school. During the brief occupation of the region by bandits from the neighbouring province in the summer of 1674,54 the student’s in-laws, surnamed Hu, took advantage of the disturbances to gain possession of the blocks. A generation later, Ai Weiguang tried unsuccessfully to buy the blocks back from the Hu family. They not only refused to give up their monopoly but even declined to let him print one copy. He looked for support from other members of the local gentry, but to no effect. Ai Weiguang decided that he would have to undertake the project of publishing his grandfather’s works from scratch, and turned his attention to making money in commerce in order to meet the costs. In the meantime, he was able to persuade the county magistrate to put up the funds to buy a handwritten copy of his grandfather’s writings from a scholar in Hubei, the province to the northwest. The new edition, referred to subsequently as the “family printing,” was finally published in 1699. Eleven years earlier, and unknown to Weiguang, a scholar by the name of Zhang Liangyu from Jiangsu province had a slightly shorter collection of Ai’s writings printed, which became known as the “Zhang printing.” By the start of the nineteenth century, the blocks of the family printing were in decay. A member of the family wanted to have new ones engraved, but the project had to wait another generation, when Ai Zhou saw it to completion. Collating both the Zhang and family printings (the Zhang printing contained less material but was better edited and annotated), Ai Zhou produced a new edition in 1836 under the imprint of Jiuxue shanfang (Chalet of the Old Learning), probably the name of Ai Zhou’s residence rather than a professional publisher.55

State censorship and the book trade

127

There were good reasons why Ai Nanying, even though he lost his life fighting to restore the Ming dynasty, should have escaped proscription. The Qing emperors took a positive view of dying for one’s ruler, even a Ming one, so long as such devotion was in the past and did not explicitly invoke ethnic antagonism. In addition, the conservative Ai advocated a return to Song philosophy and literary styles and rejected the flamboyance of the Ming, themes approved by the Qing emperors. His vulnerability to suppression came not from what he did or wrote but from the intellectual lineage in which he stood. One of his most ardent posthumous disciples was none other than Lü Liuliang, the Nanjing scholar who published the letters of Gui Youguang. One of Lü’s editorial projects in the 1670s had been to produce a new annotated edition of Ai Nanying’s essays on the classics under the title Ai Qianzi gao (Drafts of Ai Nanying). The work would have been popular with examination candidates, since Ai, like Li Zhi, was renowned for his skill at writing the “eight-legged” or “regulated-meaning” essays required for the exams. Lü makes his presence known in this book, listing himself as editor both in the table of contents and on the first page of the text, and contributing a preface. Ai’s posthumous link with Lü, both intellectually and through publishing, meant that he was tarnished by his disciple’s alleged anti-Manchu attitude. The Collected Writings of Heaven’s Hired Hand thus suffered, as Model Letters in part did, because of a tradition of anti-Manchu sentiment with which the author had become identified posthumously. Model Letters was also censored because of the Qianlong emperor’s personal dislike of one of the authors. A Comprehensive Gazetteer of Cao’s Stream was censored because it included the writings of banned authors and referred to emperors using inappropriate titles. None of these books was placed under prohibition because of the ideas it expressed. Prohibition was to ensure that the historical relationship between the Manchus and the Chinese be coded in ways that showed due respect to the imperial family’s ethnicity. The main issue may have been doubt about Manchu legitimacy to occupy the Chinese throne, but in fact self-censorship blocked its expression. The Qianlong inquisition thus became a contest over codes rather than ideas. It was not language, but the formulas by which language was arranged, that alerted the state to political implications external to the content of a text and motivated it to suppress those that were wrongly coded. The power to publish Qianlong was conscious of the role printing played in spreading these incorrect codes. He realized that the destruction of printed books was fruitless unless the woodblocks were also located and destroyed. Time and again reference is made to tracking down both: “the woodblocks and printed sheets must both be put to the flames” (p. 125); “search everywhere for them, and for woodblocks too” (p. 198). The emperor often reminded his officials that they should be on the watch not only for the original woodblocks but the reengraved blocks as well, which might exist in other provinces (pp. 107, 162). The blocks could be in many places: sometimes at the home of the author (p. 164) or his descendants (p. 113), sometimes with a disciple (p. 171), sometimes with a publisher who had bought up the blocks at a later date (p. 177).

The chinese state in ming society

128

The emperor was willing to let the woodblocks of certain books survive so long as those passages judged offensive were planed off the blocks (p. 217). He was less happy when publishers tried to save their investments by altering the blocks by themselves. Evidence of such alteration abounds in Qing prints. The copy of Model Letters of Gui and Qian in the East Asian Library of the University of Toronto shows this kind of selfcensorship. The names of many of the people to whom Gui and Qian addressed their letters, men subsequently judged treasonable by Qianlong, have been blotted out, both in the table of contents and in the body of the text. This alteration was done on the woodblocks by cutting out the offending text and inserting small flat tablets of wood in the columns where their names originally appeared. Sometimes only the given name of the recipient is covered over in this way, but more often his entire name is obscured. By tampering with the blocks, the publisher hoped to go on printing eopies without incurring the expense of having them engraved afresh.56 By 1780, the emperor was suggesting something of the same procedure when he told Four Treasuries editors and provincial officials to remove offensive passages from confiscated woodblocks and replace them with more acceptable text.57 This method of making “simply a few considered emendations” (p. 147) put back into circulation texts in authorized versions, fighting fire with fire. Substituting new passages would also get around the embarrassment of having books in circulation with highly suggestive blackened or blank spaces. Inquisition documents indicate that the court was aware of the ease with which printed materials circulated: note the recurring phrases about books “spreading widely” (p. 165), “spreading throughout the province” (p. 170), and “circulating freely in the provinces” (p. 194). A region of particular concern was Jiangnan, the Lower Yangzi region comprising southern Jiangsu and northern Zhejiang. This was China’s elite cultural core, and the area that Qianlong regarded as most resistant to direction from the state.58 This resistance in book matters was due in great measure to the fact that the publishing industry was centred in Jiangnan. Qianlong notes of this region that “the number of their literary productions, of private libraries, and of bookshops is double that of other provinces” (p. 157); that Jiangnan “represents the converging point for all books” (p. 167); that “the province of Jiangsu is a centre of culture, private libraries and bookshops abound, and people who have hidden away rebellious books are not a few” (p. 188). In Qianlong’s mind, Jiangnan’s power to publish was a potent means by which the area could assert a high degree of cultural autonomy from the centre and become a site of ethnic resistance. Qianlong was not the first to single out the bookmen of Jiangnan for criticism. An official proclamation that circulated in Suzhou and Songjiang a century previously had scolded the booksellers in this region for their willingness to put into print anything that would make money.59 The commercial prosperity of Jiangnan thus came to be seen as the source of the region’s power to evade the watchful eye of state censorship. Such wealth was dangerous for the alternatives those who had it could imagine. The power to censor In a minor controversy known as the “Affair of the Placards,” Frangois I found his authority as sovereign of France being challenged on religious grounds by pamphleteers, and responded in October 1534 by placing a ban on all printing. A second censorship

State censorship and the book trade

129

edict of 1535 required that Parlement select twenty-four booksellers for special surveillance.60 Even though the ban was not enforced, it served as a warning to printers, at least in Paris. In a country the size of China where print culture was firmly entrenched and the book trade lively, imposing such a ban would have been even less realistic. Qianlong notes that bookshops were one of the sources of books arriving in the capital (p. 125); he tells the Four Treasuries editors that they should include “books in general circulation” and not just books from private collections61 and makes his officials legally responsible for informing booksellers which books were under proscription (p. 232). And yet the infrequency with which commercial publishers and booksellers are named in the surviving inquisition documents is striking. The acting governor-general of Jiangsu and Zhejiang reports to the throne that a deputy in Suzhou had obtained some banned books from Qian Jingkai (p. 128), one of the three largest book dealers in Suzhou.62 The governor of Hunan province in 1779 states in a memorial that his men found the blocks of a banned book, originally engraved in Jiangnan, in the possession of Eryou Hall, a bookstore in the provin-cial capital which had acquired them from the author’s penurious descendants in 1760 (p. 117).63 The names of a few other book dealers also survive in inquisition documents, but otherwise there is little evidence that the book trade was coerced. Dealers found selling suspect books were apparently deemed innocent unless they had personally published them.64 Such treatment may have been less a matter of lenience than a realization that harsh measures against booksellers would only drive banned books further from the censors’ reach. The Qianlong emperor lacked one advantage that English monarchs enjoyed in their attempts to censor offensive literature, and that was the cooperation of an organized corporation of publishers. In 1556, Queen Mary approved the creation of the Company of Stationers, to which almost everyone in the London book trade was obliged to belong, and invested it with the task of suppressing “seditious and heretical books” on the crown’s behalf through a licensing system. Books published without previously obtaining a licence from the company were automatically condemned.65 This system, in which commerce and the state happily colluded, remained in force until 1695, by which time printing was too widespread to be controllable through this mechanism. Having the book trade regulate itself was peculiar to England, for elsewhere in Europe printing guilds did not have much authority even over their own members. China had no comparable organizational framework that could oversee the trade throughout the realm by supervising it in one city. The Shunzhi emperor in 1653 might decree that bookstores should be allowed to print and circulate only books on principle, works on governmental administration, and texts that contribute to the literary enterprise; works with petty concerns or immoral language, and all literary exercise books or collections sponsored by literati associations, are strictly forbidden.66 It was an ambitious demand, but neither he nor his successors had a network through which these prohibitions could be transmitted. Bookstore chains were few, and largely operated under arrangements that granted their managers considerable autonomy.67 Booksellers’ guilds existed, but these did not extend in any formal way beyond the book trade in one city. What the emperor needed was a bookmen’s guild with national

The chinese state in ming society

130

connections, but such an organization did not exist. Beijing in the time of Qianlong had a guild for book dealers from Jiangxi province, who dominated the trade, but those who were not Jiangxi natives were excluded from the guildhall. In the mid-nineteenth century, their northern apprentices formed a separate North Zhili Book Merchants’ Guild, and built their own guildhall in the 1870s.68 Neither, however, could claim to represent all Beijing publishers, and neither had regular ties with publishers in other cities. Commercial London could dominate England, and the English state availed itself of that domination. Commercial Beijing was in no position to dominate China. Beijing was in any case neither the publishing nor the cultural centre of the country. The only national organization an emperor could use to dominate publishers was the bureaucracy—which is what Qianlong did, imposing on his officials the burden of tracking down and punishing them when they failed to turn them into the capital in satisfactory volume. His use of bureaucrats meant that they became vulnerable to the charge not of circulating “seditious writings,” to use the Ming phrase, but of failing to carry through his directives. The commercialization of the book trade only increased the prospect that they would fail. There were too many books to trace, and no way to guarantee that a banned title could be eradicated once it had slipped down commercial networks and disappeared among readers. In England, where the “mystery” of printing was a novelty Henry VII could not fail to notice it, and seek to control it, soon after it arrived. For Qianlong, on the other hand, printing was no mystery What he did not anticipate was the capacity of printing to create chains of textual reproduction that extended out beyond the reach of the state. Unable to rely on internal organization within the book trade, he could only call upon his bureaucrats to work against the natural process of dissemination. The inertia and limited reach of the bureaucracy made the pursuit of particular titles almost impossible. Qianlong had to escalate his search into a political campaign, as no Ming emperor ever did, to compensate for the fervour he believed his officials lacked.69 The majority of the printed books he banned, circulating through commercial networks, survived Qianlong’s campaigns to destroy them. The Qing may have been an anxious state, but it had neither the will nor the capacity to be a censorship state. Printing thus served history in ways that the emperor had feared, but was powerless to prevent.

Part IV Monasteries

7 At the margin of public authority The Ming state and Buddhism China is a realm of culture and ritual where people are compassionate and amenable to instruction. A monk who can attain the style of the Patriarchs need only explain Mahayana and the people listen and understand; he need only speak of karma and the ignorant are transformed. The Hongwu emperor, 1371 Those you find walking the Buddhist path in recent years are not wise men but corrupt characters without registrations who have changed their names and gone into Buddhism to evade trouble and save their own necks. The Hongwu emperor, 13941 Between 1371 and 1394, something happened to change the founding emperor’s mind. Zhu Yuanzhang as Emperor Hongwu launched his dynasty striking the pose of the sage ruler. His goodness and compassion, he believed, would transform his subjects, and his reign would be buoyed by all faiths, including Buddhism. His comment twenty-three years later indicates that that pose had fallen apart. Benevolence had been superseded by disgust, declarations of compassion replaced by threats and denunciations. The difference between 1371 and 1394 may have had something to do with the declining quality of the clergy, but this is doubtful. It had far more to do with the Hongwu emperor’s changing sense of the efficacy of the norms and systems governing his realm. He responded by changing the state regulations governing the institutional life of Buddhism. Between 1381 and 1394, the emperor put together the pieces of a legislative edifice designed to subordinate monks and monasteries to the authority of the state. Buddhism would spend the rest of the Ming dynasty adjusting to, and growing out from under, the burden of these regulations. The relationship between Buddhism and the state would shift early in the fifteenth century, and shift again in the last quarter of the sixteenth, as we shall see. These shifts were less violent than the early-Ming assault on Buddhism, and are only weakly reflected in state texts, but they are equally revealing of the instability in the state’s relationship to organized religion. Seen together, these three shifts provide an opportunity to explore the relationship between social institutions and public authority over the course of the Ming dynasty. More specifically, monastic Buddhism’s responses to shifts in state norms help

At the margin of public authority

133

us to probe changes in the variable constitution of public authority between the founding and late phases of the dynasty. The concept with which I wish to frame this discussion is “public authority” rather than the narrower “state control.” I find this distinction useful for redirecting the natural urge within Ming studies to assume the hegemony of the state—of which the subjugation of Buddhist institutions would seem like just another confirmation. It also helps in detecting change. The Chinese state remained relatively constant during the Ming, whereas, I would argue, public authority underwent redefinition. Consider the difference. The state as the main source of formal authority in China is represented or “made public” in the local arena by local officials. The concept of public authority, on the other hand, expresses the reception of the state’s authority in that arena: public authority exists to the extent that people are aware of and respond to the presence of acknowledged sources of power, most prominently the state.2 Not defined by the state alone, public authority takes form through the interaction of state and society. The inevitable instability of this dialogic interaction means that public authority is always vulnerable to redefinition, notably when powerful elements within society emerge to participate in, or alternatively to challenge, the dominance of state actors in the local arena. Public authority emanates from the state, but as it disperses into the public realm, it can suffer deflection or partial expropriation as new political possibilities take form. The power of elites to deflect state hegemony is not an issue in the early Ming, when the state could and did intervene to regulate social institutions when it chose to do so. But as the gentry developed a localist reorientation to power in the mid-Ming, they turned to monastic Buddhism in their project to redefine public authority. It is this understanding of public authority as vulnerable to redefinition, and of institutional Buddhism as implicated in this process, that sets the terms with which this chapter looks at the impact of changing state norms on Buddhism. Reading only regulatory texts, one might assume that the state’s relationship to Ming Buddhism was straightforwardly repressive: the Hongwu emperor formulating tough new laws to limit the size and influence of Buddhist institutions, and those institutions bending to the state’s will. The sense of an uncomplicated compliance with state regulations was so thoroughly worked into the historiography of Buddhism during the Ming that few modern historians, outside those working on the social history of Buddhism, are even aware of Hongwu’s assault on Buddhism, let alone its unprecedented scale.3 It is worth pondering why the Hongwu suppression is relatively unknown, for it anticipates what this essay seeks to expose. The suppression’s invisibility derives in the first instance from the manner in which it was announced and officially recorded. No Ming emperor declared Buddhism to be a threat to the state or to public authority; not even Hongwu said he was suppressing Buddhism. Rather, the regulations that he and subsequent emperors promulgated breathe with the logic of bureaucratic rationalization— a matter not of shattering delusion or even stamping out sedition, but more innocuously of adjusting resources to needs. If this logic passed scrutiny at the time, it was due in the second instance to Confucian complicity. Aspiring officials anxious to resecure their place in the political system after the difficult passage through the Yuan dynasty were content to see Buddhism moved to the margin of public authority after enjoying some gains from the patronage of the Mongol emperors. The sanctity of the Hongwu dispensation for the rest of the dynasty meant in the third instance that what the founder in his great wisdom had decreed could not be questioned or otherwise commented on.

The chinese state in ming society

134

The very notion that the Hongwu regulations on Buddhism amounted to a suppression disappeared behind the fiction of the wise ruler and the grateful, compliant subject. Chikusa Masaaki’s research on similar regulations in the Northern Song dynasty has shown their consistent failure to induce the intended effects.4 The failures are more interesting, and infinitely more informative of the state-society relationship, than the few successes. State regulations may be designed to impose norms, but they cannot determine how these norms will interact with the social practices to which Buddhism is tied. Subject to the supervision of the state, Buddhism as a social institution was vulnerable to severe restriction; yet by not being the official religion of the state, it remained relatively free to override the effects of state suppression, remake its social constituency and, in the late Ming, reposition itself in the matrix of public authority. Buddhism in the opening decade of Ming rule The founding Ming emperor understood the presence of Buddhism in the lives of his people. Not only was Zhu Yuanzhang abandoned to the care of a Buddhist monastery as a youth, but during his rise to power he was able to make effective use of Buddhist devotional networks to mobilize his followers for war. He was also aware that whereas these networks could survive wartime calamities, institutions were more vulnerable to destruction. The Yuan—Ming transition, particularly in the early phase in the 1350s, took a large toll on Buddhist monasteries. As a later Yangzhou writer noted, in the areas that warring armies overran at the end of the Yuan, only 20 to 30 percent of monasteries survived intact.5 Those that did survive found it difficult to hold onto the incomeproducing land that fed the monks who once lived there. For the first decade of the Hongwu era, Buddhism was regarded as a problem only to the extent that its dilapidated monasteries stood as mute symbols of the destructiveness of the dynastic transition. Hongwu responded during this decade by adopting the benign posture of imperial patron and sponsoring their reconstruction. He funded monastic rebuilding and bestowed large tracts of tax-exempt land in the capital region, and he brought monks into court circles, sending them on diplomatic assignments (to Turfan and Japan, for example) and appointing them as advisors in the courts of his princes.6 He called practically every eminent monk in the country to Nanjing for an imperial audience, and over the first five years of his reign asked them to conduct plenary masses at monasteries in the capital to rest the souls of the war dead. The Grand Plenary Convocations of the Dharma Flower (Guangjian fahua hui) held in Nanjing in 1371 and the following year were the greatest gatherings of leading Buddhist clerics since the Tang dynasty.7 The emperor’s Six-Dynasties-style imperial patronage cast Buddhism in the role of adjunct to a state-centred structure of public authority, almost an official religion. It was charged with regulating its own afFairs and also serving as a pillar of the state. The only significant legislation of the decade was promulgated early in 1373 when the emperor required monks to obtain ordination certificates from the Central Buddhist Registry in Nanjing, standard practice of the Chinese state for the preceding six and a half centuries.8 Hongwu’s intention in reviving the system was to end the dislocation, vagrancy, and mendicancy of the war years, when many adopted the guise of a monk to beg for

At the margin of public authority

135

sustenance or evade military service. It was also designed to halt tax evasion, since monks could claim exemption from the corvée duties that ordinary people had to perform. An innovation of the 1373 legislation when compared to earlier registration orders was instituting a system of universal registers (zhouzhi wence). These registers, which the emperor remandated in 1394, were to list the names of all certified monks in the realm. Copies were to be distributed to every monastery so that monastic officials could confirm the status of those who came claiming they were monks. Not surprisingly, there is no evidence that such unwieldy volumes were either made or distributed. The important feature of the 1373 legislation, however, is that no restrictions were placed on ordination, nor did the state arrogate to itself the authority to ordain. Ordination was still done by senior monks at ordination platforms in leading monasteries. This was simply a registration system, not a scheme to restrict the size of the clerical population. Buddhism after 1380 In 1380, Hongwu carried out a thorough purge of Hu Weiyong (d. 1380) and his faction at court and reorganized his regime so that he was personally in charge of the day-to-day running of the government.9 The transformation of the emperor’s images of himself and his realm which this crisis brought about affected his attitude toward Buddhism. It is unknown to me whether Hu Weiyong was in any significant way tied to the Buddhist establishment, or whether the new emperor had any other particular reason to give up his patronal relationship in favour of surveillance and coercion, but the anti-Buddhist suppression that grew during the . is too closely timed to Hu’s purge to be unconnected. Perhaps it is simply that Hongwu’s new orientation necessarily touched every element of the state, making the reform of Buddhism simply part of the sea-change affecting all aspects of his rule. After 1380, Hongwu no longer conceived of Buddhism as a component in the state-centred structure of public authority; it loomed instead as a potentially destabilizing, even autonomous, realm that, if left unchecked, could only undermine public authority. Buddhism was no longer a resource of rule but a threat to it. Monks were no longer men of wisdom but charlatans and tax dodgers whose very existence symbolized the failure of that authority to take hold. A series of ambitious policies and regulations over the years immediately after 1380 reorganized Buddhist institutions in such a way as to forestall that threat. The first step was to set up a bureaucratic structure under the Ministry of Rites to supervise the affairs of Buddhist monks and monasteries rather than leave these affairs in the hands of monastic officers who were responsible to none but their own institutions. This bureaucratic structure consisted of Buddhist (and Daoist) registries at the county, subprefectural, prefectural, and national levels, staffed by local monks. A few prefectural and subprefectural officials had set up registries in the early 1370s to manage religious affairs, for which there was a Yuan precedent.10 The office did not become universal until 1381, when Hongwu mandated a nationwide hierarchy of registries at every level of the field administration.11 These registries were known at the prefectural level as Senggang si (Office of Clerical Supervision), at the subprefectural level as Sengzheng si (Office of Clerical Rectification), and at the county level as Senghui si (Office of Clerical Convocation). The registrars at the three levels went by the titles dugang (supervisor),

The chinese state in ming society

136

sengzheng (rectifier), and senghui (convener). An identical structure of offices was instituted for Daoism as well. The Buddhist registries were usually housed in the most prestigious urban monastery in the local seat of government, which sometimes was restored when the registry was set up. They were founded throughout the country over the next four years,12 and it is a rare county that got away with not having one.13 The registrar was usually also the abbot of the registry monastery, although the prior (jianyuan), who was in charge of a monastery’s business affairs, might also hold the office.14 According to a summary in a sixteenth-century county gazetteer from Zhejiang, a registrar’s duties were four-fold: to supervise the Buddhist monks in his prefecture or county, to propagate the correct teachings of Buddhism, to report misdemeanours among the clergy for investigation and punishment by civil officials, and to conduct public rites.15 He could also be called upon to grade and register monastic land for tax assessment.16 Registrars in underdeveloped regions might bear yet other responsibilities. In the frontier areas of Yunnan province, for example, registrars were ordered to see that monks were distributed among the small chapels along main transportation routes (a rate of one chapel every 15 kilometres was suggested) so that they might extend a stabilizing influence over the countryside.17 The registry system thus served to supplement state supervision in a region where schools and academies were almost non-existent. These tasks indicate that the registrar’s function was to administer Buddhism on the state’s rather than on Buddhism’s behalf, and further, where the state’s presence was weak, to embody public authority. The registry system gave Hongwu the bureaucratic means to intervene in institutional Buddhism systematically His first move in 1382 was to have registry officers register and categorize all Buddhist institutions in the country. As part of this program, Hongwu redefined the sects of Buddhism as three: Meditation (chan), Doctrine (jiang), and Teaching (jiao). According to the edict of 1382, Meditation monks were to concentrate on meditational exercise leading to personal enlightenment. Doctrine monks were to study the scriptures to penetrate their meaning. The Teaching sect, created at Hongwu’s initiative, comprised monks who went out among the people to preach and conduct rites, especially funerary rites.18 There appears to have been no doctrinal reason for this bureaucratic simplification of the schools of Buddhism. It was simply a matter of imposing bureaucratic uniformity. By the mere presumption of defining and fixing sect categories, however, the state claimed a prerogative it had previously never enjoyed to dictate organization within Buddhism and subordinate its teachings to state uses. Buddhism was being shaped more and more into a creature of the state. Cutting even closer to the heart of institutional Buddhism than the control of sect identity was the control of property. Without sizeable tracts of land, large monasteries could not hope to collect the income to maintain their physical fabric or support their residents. Hongwu did not aspire to take direct control of the fiduciary property that monasteries owned, but he did desire to make sure that individual monks did not get that control, thereby enriching themselves and destabilizing the institutions to which they belonged. In 1382 he decreed that Buddhist and Daoist monks had no right to mortgage or sell the land on which they lived, and that the penalty for mortgager or buyer was the confiscation of all his landholdings.19 To prevent monastic land from being alienated, he ordered the Ministry of Rites, possibly in 1386, to require every monastery in the country to appoint a “foundation cleric” (zhenji daoren). This monk’s task was to oversee the

At the margin of public authority

137

monastery’s financial affairs and be responsible for the monastery’s fiscal dealings with the local magistrate. Although the emperor noted in 1393 that people were filing legal suits against foundation clerics for mismanaging monastic funds, he reaffirmed the post in an edict the following year.20 Hongwu’s intention in protecting monastic property was to ensure tHat the clergy could afford to stay secluded within their monastic confines and not have to go looking for support outside, where they might meddle in community affairs or build networks among the people. The seclusion edict of 1394, to be mentioned shortly, confirms the impression that the emperor’s prime concern in regulating property was to secure the means for segregating the secular and religious realms. In an economy in which land was a commodity traded on markets that the state could not control, however, this sort of regulation proved practically worthless. In 1391, Hongwu launched what amounted to the most severe suppression of Buddhist institutions of his reign. Through what was known as the “hundred-day edict” (because its provisions had to be carried out within a hundred days), he imposed an amalgamation order on Buddhist monasteries. This order required the majority of smaller monasteries throughout China to close and to transfer their residents and property to a limited number of larger institutions. These institutions were designated “abbeys” (conglin, a southern term for a large monastery). The only places that amalgamation left alone were tiny chapels with one or two resident monks, and convents for nuns. This edict amounted to a complete reorganization of the institutional life of Chinese Buddhism. Its implementation was largely completed within the hundred-day limit, though some amalgamations were still being carried out as late as 1415.21 A few institutions managed to avoid amalgamation, but they were the exceptions.22 Even abandoned monasteries had to be amalgamated to forestall their being revived at a later date as separate institutions.23 Thenceforth, only abbeys had a legal right to exist. The 1391 amalgamation order altered institutional Buddhism in China more thoroughly than any previous suppression, and there would be nothing like it again until the 1950s.24 Again, it was not presented in documents of the period as a suppression of Buddhism, but as a means of achieving a more efficient use of resources. This tack was not totally unreasonable, for many monasteries had been destroyed and abandoned during the Yuan-Ming transition and the number of monks qualified to reside in them had fallen since Mongol times. On the other hand, the campaign forcibly reduced the number of Buddhist institutions to perhaps less than a quarter of the number that had existed before 1391. Among other effects, the amalgamation order made available buildings that the state and local authorities could put to other uses, such as schools.25 Pressure was brought to bear to have the order countermanded as soon as Hongwu was dead. Even before his enthronement in 1402, the Yongle emperor allowed that recognized monasteries founded before 1382 could revert to independent status.26 The 1391 amalgamation order included yet another mechanism of state control, for it made abbeys eligible to receive an official name plaque (e) from the emperor if they did not already possess one. This was a wooden plaque bearing the name of the monastery scripted by the emperor or a calligrapher he deputed. It was in effect a legal document in that it signified state recognition of a monastery The hundred-day edict made possession of a plaque the test of an institution’s legality, and monasteries founded or rebuilt subsequently were ineligible to receive one.27 This amounted to a ban on the private founding of monasteries or chapels after 1391, and was incorporated as such six years

The chinese state in ming society

138

later into the Ming Code.28 The Yongle emperor reaffirmed the status of the name plaque as a legal licence in the edict of 1402 that partially reversed amalgamation. Later emperors could take it upon themselves to overrule Hongwu’s ban on the founding of new monasteries by presenting a plaque whenever they chose to do so, and officials and abbots could submit memorials requesting an edict of conferral. Submitting a memorial of application could be a politically delicate manoeuvre. A few Ming emperors preferred to issue blanket refusals so as not to deal with such requests, as the Chenghua emperor did in 1467, when he told his eunuchs, who were busily building monasteries around Beijing, that he would entertain no further requests from them for name plaques. He also reminded them that it was illegal to restore or enlarge a monastery lacking this official designation.29 The Hongwu emperor’s last major intervention in the realm of Buddhist affairs came in 1394, when he promulgated what could be called the edict of seclusion. By imperial fiat, monks were ordered not to beg in public, nor enter a magistrate’s office, nor communicate with the gentry, nor form friendships with officials or commoners, nor accept minors as acolytes. Most notably, monks had to live in large monastic communities (which amalgamation had brought about) where they could be better supervised. A monastery housing fewer than twenty monks might be permitted to exist only if it were at least twenty li (10 kilometres) from a settlement, well beyond the scope of regular contact. Monks who chose to engage in religious practice in secluded locations could do so on condition that they lived individually or in pairs; three monks together was too many.30 The goal of the edict was to seclude monks from secular life so completely that they could not gather to hatch plots to undermine state authority, if it is permissible to reconstruct Hongwu’s desperate imagination. The realms of religious and secular life were thus to be neatly separated so that the influence of the former on the latter be kept to a minimum. The state would no longer look to Buddhism as an aid to governing the people; instead, Buddhism was a competing source of authority that had to be watched. The edict of seclusion was as much a tirade as a policy statement, and I have found no evidence that any of its strictures was ever actually enforced. Nonetheless, it established a tone that poisoned official attitudes toward institutional Buddhism for decades. Hongwu’s initial posture of patronage had expressed a belief that Buddhism could supplement state institutions and supply the state with ideological and educational services. The amalgamation and seclusion edicts indicate that he had abandoned this belief. With the late-Hongwu reorganization, Buddhism was ruled out of playing any role in the composition of public authority, ideological or otherwise. It was now simply an object of that rule, representing a sphere of activity subject to state regulation and excluded from the realm of the political. It had been moved to the margin of public authority. State and Buddhism in the fifteenth century The editor of a 1566 county gazetteer from Shanxi province regarded the 1391 amalgamation order as the Ming state’s last restrictive legislation against Buddhism. After 1391, he notes, the state engaged in no further suppression of Buddhist personnel,

At the margin of public authority

139

literature, or places of residence.31 This is not precisely true. Hongwu’s heirs on the fifteenth-century throne did add restrictions in two new areas: the number of novices, and the amount of land monasteries could own. As I shall show, however, these restrictions lacked the force of Hongwu’s edicts. The Ming state’s relationship to Buddhism settled into one of toleration rather than suppression. The main intervention associated with the Yongle emperor was an edict of 1418 that capped the number of novices per prefecture (forty), subprefecture (thirty), and county (twenty), and required them to train for five years before applying to the Central Buddhist Registry for ordination.32 These caps were the first Ming attempt to impose a quota system on monks, analogous to the restriction on the numbers of students registered in county and prefectural Confucian schools.33 Hongwu had imposed some restrictions regarding age and competence in order to ensure that the ranks of the clergy were not filled by impostors, but not with the precise intention of reducing the number of monks.34 As monks were eligible for exemptions from the service levy, an increase in the number of adult males enjoying this exemption meant a decrease in state income. The corvée exemption also made getting tonsured and dressing like a monk an attractive tax loophole for men who had otherwise no intention of ever becoming monks. Yongle had already targeted these people ten years earlier, in 1408, when he needed to mobilize labour to reconstruct the new capital in Beijing and the Lamaist complex on Wutai Mountain in neighbouring Shanxi province (see Figure 7.1). He decreed on that occasion that those who were found to have gotten a Buddhist tonsure without proper authority would be forced to go to the capital region as fixed-term forced labourers along with their elders or masters (if they were servants), after which they would be required to stay in the north and open farmland.35 Yongle’s concern was clearly neither security nor ideology, but revenue and access to public labour. These were matters against which local social networks pulled strongly and irresistibly, which meant that Yongle’s orders and numbers went ignored. The mechanism for supervising them, the registry system, could not be relied on to enforce the new limits on novices. However, the notion that there should be quotas came to be accepted as at least a default policy In response to a memorial submitted late in the 1450s, for instance, the Tianshun emperor ordered that “every prefecture, subprefecture, and county fix its quota of Buddhist and Daoist monasteries and set its number of [ordained] Buddhist and Daoist monks” at current levels. These quotas were to constitute ceilings on the numbers of monasteries or monks permitted within each jurisdiction.36 The order failed, and others would follow at intervals. A memorialist in the 1530s suggested specifying limits of 200 monks per prefecture, 100 per subprefecture, and 60 per county.37 Although these numbers too went by the board, the desirability of quotas was still being voiced as late as 1585.38

The chinese state in ming society

140

Figure 7.1 The Lamaist monastic complex on Wutai Mountain, from a nineteenth-century reprint of the 1596 monastic gazetteer, Qingliang shanzhi (Gazetteer of Qingliang Mountain), 1b–2a. The other type of restriction imposed after Hongwu was on landholding. Here again fiscal concerns were uppermost. The Hongwu emperor had been concerned that monasteries lacked enough land to keep their monks within the walls. His heirs tended to the opposite fear, that monks were controlling too much land rather than too little, blocking other people from access to land and evading taxes. Attempts to limit monastic landholding were piecemeal, unsystematic, and short-lived. The Jianwen emperor (r. 1399–1402) accepted a memorial calling for a ceiling of 5 mu per monk in Jiangnan in 1401, but the Yongle emperor overturned it the next year before it could be applied. In Fujian, a limit of 60 mu per monastery was set in 1452 and then raised to 100 mu in 1480, though neither appears to have been enforced.39 A different approach to restricting monastic landholding was tried in 1448, when local officials were ordered to expropriate property acquired by monasteries after the Hongwu era and distribute it to landless civilians at a rate of 20 mu per adult male householder.40 This expropriation, applying only to land purchased after the Hongwu reorganization, was a compromise between limiting the economic independence of monasteries and respecting the institutions established under the dynastic founder; I have found no evidence that local officials ever carried out the plan. A wave of state interventions in monastic landholding did occur in Fujian in the latter third of the sixteenth century when the state needed emergency fimds to meet military costs, but this involved the expropriation only of rent, not of the land

At the margin of public authority

141

itself. By the sixteenth century, the state recognized that landowning monasteries enjoyed the same rights of private property and rent-taking as any other landlord, and it restricted its expropriation to rents on monastic lands that were no longer under active management. More to the point, as we shall see, the sixteenth-century state was no longer interested in regulating Buddhism. Taking over monastic rents was a fiscal expedient to pay for military costs, not a religious matter. As for the original regulations that Hongwu had brought in, while the letter of the laws remained in force, their spirit was soon compromised after his death in 1398. Only extraordinary bureaucratic vigilance could have ensured that his rules were observed, and such supervision was not available. Hongwu had hoped that the registries would serve as an internal supervisory mechanism monitoring Buddhism on the state’s behalf, but the design of the system did not favour success. With the exception of the prefectural registrar, incumbents were below rank 6 of the bureaucratic hierarchy and hence “not within officialdom” (wei ruliu). In lists of county officers in local gazetteers, the Buddhist registrar appears after the instructor at the county medical school: both posts were filled with local appointees rather than personnel from outside the province, unlike all the posts above county medical instructor.41 Registrars thus were yuan (functionaries) rather than guan (officials), and so earned little prestige for their posts.42 In addition to being local people without prospect of rising into the regular bureaucracy, registrars were monks and could be presumed to have interests that aligned them more closely with the clergy than with the state. Their official status meant that they could gain the ear of the county magistrate to argue on behalf of troubled monasteries—appealing for the return of expropriated monastic property, for instance—but it provided them with little other opportunity to affect local power alignments.43 From the state they could expect little more than their monthly stipend, and even that was suspended in the second quarter of the sixteenth century.44 One mid-Ming official who worried about such issues as court extravagance and the erosion of orthodoxy took aim at registrars as a waste of the state’s resources and an inducement to heterodox inclinations on the emperor’s part—and got dressed down by the Hongzhi emperor with the reminder that he was maintaining “the old institutions of the present dynasty” and the official did not know what he was talking about.45 It had been Hongwu’s intention that a Buddhist registry a Daoist registry, and a Confucian shrine in every county of his empire would constitute a tripod of ritual support for the state’s dominance of local society.46 By the mid-Ming, however, many registries were defunct.47 The rest survived largely in name only, the registrar being whoever happened to be the abbot of the registry monastery An individual monk might exploit the post of registrar to act on behalf of the local Buddhist community, as the registrar of Neixiang county, Henan, did in 1481 when he obtained state funds to rebuild Jianfu Monastery;48 or he might act in the conspicuous service of the state, as the subprefectural registrar of Yingzhou, South Zhili, did in the early 1500s when he conducted an annual ceremony wishing long life to the emperor;49 or he might act in the public interest, as Hangzhou prefectural registrar Liangjin did in 1542 when he prayed to Guanyin to end a drought.50 But in any of these capacities, the registrar’s structural power was limited. The registry system neither constituted an internal organization for Buddhism, nor was it integrated into the regular bureaucracy. Accordingly in many parts of the country by the turn of the sixteenth century registries either had disappeared51 or existed only on paper,

The chinese state in ming society

142

and registrars were no longer being appointed.52 Maintaining the system depended on local initiative. On the magistrate’s part, the main impulse for preserving a registry was to keep up the appearance of having in operation all the bureaucratic offices mandated by the Hongwu emperor; on the registrar’s part, it was to call upon whatever prestige still clung to the post in local eyes. Neither imperative was overpowering. Hongwu’s amalgamation order was similarly compromised. As the editor of the Songjiang prefectural gazetteer of 1512 notes, “When Buddhism was reorganized in 1391, smaller chapels and cloisters were amalgamated into what were called abbeys. In 1402 all the amalgamated monasteries were ordered restored to their former statuses.” This was not quite what Yongle had intended, but it was what happened. The editor continued to group the monasteries in his gazetteer under their 1391 abbeys “so as to preserve the old system,”53 he says, allowing the reorganization to survive at least as an official fiction. But its effects were not completely reversed: the amalgamation may have been cancelled in fact, but it hurried the demise of weaker monasteries that never came back into operation. De-amalgamated institutions had their status restored in 1402, but not necessarily their buildings or staff, for which local support was key Deprived of personnel and resources for over a decade and often converted to other uses, many proved impossible to revive. A scan of the section on monasteries in the 1561 gazetteer of Wujiang county, Suzhou, reveals that over half the monasteries and chapels that had undergone amalgamation in 1391 were defunct.54 Fifteenth-century emperors allowed most of the late-Hongwu legislation against Buddhism to become a dead letter. They occasionally even went back to reviving the early-Hongwu patronal pose, comfortable with the assumption that Buddhist beliefs could still be invoked to buttress public authority The revival was guarded, however. Yongle spoke favourably of the conspicuous power of Buddhism “to encourage the people to do good” in his preface to the widely disseminated Buddhist tract, Zhufo shizun rulai pusa zunzhe mingcheng gequ (Song of the names of the world-honoured Tathagata, bodhisattvas, and arhats); yet he rejected a memorial for an amnesty honouring the presentation of a Buddhist relic by a foreign envoy in 1406, recalling that the Buddhist excesses of the Tabgatch and Mongol emperors of the Northern Wei and Yuan had interfered with sound government.55 The Chenghua emperor had a reputation for being indulgent toward Buddhism; Tianshun and Hongzhi, for being strict—though even Tianshun handed out forty name plaques for new monasteries in the first year of his reign.56 In general, these emperors preferred the role of patron, prizing what value could be squeezed out of Buddhist ideology without paying the costs of organization. More to the point, the patronal pose allowed them to publicize their own virtue and their expectation of good conduct on the part of their subjects; it was inexpensive and noninterventionist. We need to be cautious about overinterpreting imperial patronage, for it could serve concrete political purposes. When Tianshun distributed those forty new name plaques, he did so as part of a program to confirm his dubious status as an emperor re-enthroned through a coup against his half-brother. When he conferred an edict of protection for monastic properties on Wutai Mountain (Figure 7.1) in 1458, he did so, he declared, because “for many years the monks have prayed for the fortunes of the state above and the people below.”57 This statement suggests a comfortable alignment between Buddhism and the state, yet other considerations were uppermost. Wutai Mountain was the Lamaist

At the margin of public authority

143

site in China most revered by the Mongols. As we have noted, Yongle sponsored construction there while he was having his capital in Beijing built. Tianshun, in his earlier identity as the Zhengtong emperor, had been captured by the Mongols in 1449, and as of 1458 had only recently forced his way back on the throne. The bland vision of Buddhist monks buoying up the state with their prayers probably masks a diplomatic initiative directed at his recent captors. In this case, the emperor was manipulating the imperial discourse of protective Buddhism and grateful state to project claims against Mongol challenges to his authority In contrast to their emperors, mid-Ming officials focused less on ideological claims than on administrative problems due to the declining regulation of Buddhist institutions. Through the mid- to late-fifteenth century, it became commonplace to complain about monks: their itinerancy, their non-registration, their parasitism, their alleged licentiousness, their affront to Confucian moral values.58 The tirades that interventionist Confucians directed against Buddhist monks through the mid-Ming suggest that they recognized not only that Hongwu’s seclusion order had failed to hold, but also that emperors had given up attempting to use Buddhism to gain ideological ground among the people at a time when other ground was being lost to the clergy: possibly moral (that was often their language of disapprobation), certainly fiscal. Considered another way, most of the fiscal problems in the mid-Ming that stemmed from the clerical service-levy exemption were state-induced. The Hongwu emperor had specifically forbidden the sale of ordination certificates when he revived the certification system in 1373, but later emperors, starting with Jingtai in 1451, resorted to this hoary device to raise revenue outside the regular taxation system. The certificates offered in 1451 were sold to raise grain to relieve a famine in Sichuan. The precedent was repeated in 1453 and 1454. Within a decade it had become a standard means for raising emergency funds. The largest sale, in 1466, involved the issue of over 130,000 certificates. At the turn of the sixteenth century, the cost of an ordination certificate was between 8 and 10 taels; by 1558 the government was selling it at a 4-tael discount; in 1572 the price was officially lowered to 5.59 Thereafter, Ming sources fall silent on the subject of certificate sales. The mechanism appears to have ceased functioning—or else ceased being worthy of notice. Scholars of Buddhist history have regarded the sale of ordination certificates as having corrupted the system of clerical certification, debasing the quality of the clergy and giving Buddhism a bad name.60 The assumption behind this judgment—that the men who bought the certificates went on to become incompetent monks—I regard as mistaken. The polite fiction surrounding the sale of certificates was that the purchaser was a monk. In fact, most if not all purchasers were simply paying the government a flat fee for a permanent tax exemption. The assumption that purchasers were monks rests on a misapplied comparison to the sale of Imperial Academy studentships (jiansheng), which also started in 1451 as an emergency measure to raise funds for the defence of the northern border after the Zhengtong emperor had fallen into Mongol hands.61 Some jiansheng did go on to take up positions in the Imperial Academy and seek to advance into the bureaucracy. Few, if any, who bought a monk’s certificate as a lump-sum prepayment on future service levies were interested in becoming a monk. Who would want to buy his way into such a non-lucrative profession? This was simply a way of raising relief grain by borrowing on future tax earnings, and seen as such at court. When

The chinese state in ming society

144

an official asked the Ghenghua emperor eight years after the 1466 issue for another 10,000 “empty-name certificates” to acquire grain, Chenghua turned him down on the advice of his minister of rites, commenting that “it is not suitable to be too profligate in issuing monks’ certiflcates; anyway, how much grain would it actually raise when the loss to the polity is so great?”62 The emperor not only understood the calculation, but pointed out that the sale of Imperial Academy studentships had far outstripped the capacity of the bureaucratic system to absorb such people. If the sale of certificates impinged on real monks at all, it may have been to make it possible for a novice who could not get a certificate through regular channels to buy one. The sole result of this practice for the Buddhist community, if such a result even obtained, was to provide novices with a legitimate, if expensive, means for acquiring certification, not to let loose a mob of impostor monks on society. Once the Ming began to sell certificates, in the words of an early-Qing commentator, “there was thereafter no fixed system for issuing ordination certificates.”63 With the certification system thrown open to other uses, this last surviving structure of control over monks had to give way. We should not overemphasize the impact of such policies on registration, however, for as early as 1436, prior to the commencement of sales, the minister of revenue regarded the certification system as no longer functioning.64 Some ministers of rites, on the other hand, tried to protect the certification system in the hope of having some mechanism to limit the size of a dangerously mobile population. In 1499, for instance, the large number of monks in Beijing prompted the minister of rites to ask that all uncertified monks be expelled from the city.65 Ministry records at the time showed that only 11,360 monks held certificates, which indicates that the system was out of touch with reality, the actual size of the clergy being at least ten times that number. It indicates as well that no one confused the sale of certificates with clerical status, for the hundreds of thousands that the state had issued over the preceding half-century simply did not enter the minister’s total. By early in the sixteenth century, clerical status was cut free from state certification, as had most of the regulated aspects of Buddhist monasticism. The selling of certificates in the mid-Ming simply indicated that the state was really only concerned with the fiscal consequences of monasticism. Fears about state security or clerical autonomy had faded. Like certification, most other aspects of state regulation went by the board. In 1530, a circuit censor in Fujian could declare that all edicts regarding Buddhist institutions and personnel issued between 1373 and 1455 were being disregarded.66 The state regulation of Buddhism had become a fiction. The state and Buddhism in the late Ming Through the course of the sixteenth century, complaints about immoral monks running riot across the countryside fade out of public discourse regarding good governance.67 With them fades the desire to restrict monks and monasteries. Still, the Ming state did not renounce its proprietary position. When the famous pagoda at Baoen Monastery outside Nanjing caught fire in a lightning storm in April 1566, for instance, the abbot and eighteen other monks were thrown into prison for negligence when the matter was reported to the court. Since Baoen had received significant patronage from both Hongwu and Yongle, its monastic officers could be held responsible for damage to imperially

At the margin of public authority

145

bestowed property. Only unrelenting efforts by Master Hanshan Deqing, who went back and forth between Baoen Monastery and the Nanjing Ministry of Rites for three months to negotiate a settiement, succeeded in getting the imprisoned monks released.68 As Baoen Monastery had a history of imperial patronage going back to huge grants from the Hongwu emperor and was still used as a site for state rites, this case cannot be taken to typify a generalized concern with regard to Buddhist monasteries. The state would have given little attention to the burning of a pagoda in almost any other monastery. Baoen’s destruction was the loss of a state-financed asset. It was also, and more dangerously, a calamity whose meaning could travel from particular damage to an ill omen for the royal family. One instance of general state intervention in Buddhist affairs in the late Ming is the ban on mass Buddhist ordinations in 1573, which the court issued in the first year of the child Emperor Wanli’s reign. With this ban we move from fiscal concerns back to the issue of the security of the state, for it was introduced as an attempt to stop sectarian organizing during a reign transition, always at least potentially a moment in dynastic time of instability.69 What could have been a major incursion into the autonomy of institutional Buddhism proved to be shortlived, however, for there is abundant evidence that public ordinations were conducted over the following decade.70 This effort to restrict one form of public Buddhist activity did not rest on a larger effort to reimpose thorough state supervision of Buddhist affairs. Monasteries by this time were effectively independent of that supervision. There would be no further attempts to legislate Buddhism for the remaining seven decades of the Ming. Memories of the Hongwu restrictions, particularly the amalgamation order of 1391, were nonetheless still strong in the latter part of the dynasty. The artist Wen Zhengming (1470–1559), writing in 1548, could look around him in Suzhou and describe its palpable effects: In ancient times, Suzhou had many Buddhist monasteries, but since the Hongwu-era. reorganization, many have been abandoned. Of those still surviving in the city today, seven out of ten are abbeys, and the rest are affiliated cloisters. There must have been a thousand chapels and halls that were amalgamated. The old grounds and abandoned sites were gradually taken over for private residences or made over into government offices. In some cases, the property remains intact but the name plaque has disappeared; in others, the name survives but the place itself no longer exists: in yet others, the site is overgrown and littered with trash and neither name nor building survives.71 All was not decay, however, for Wen goes on to use this gloomy portrait as a backdrop to highlight the change he could see occurring in his own day, as derelict old monasteries were being restored and monks were gathering again for religious cultivation. The long decline was being so surely reversed that he even allowed himself to wonder aloud, “What will it be like for those who come after?” The answer, as we now know, was a widespread revival of institutional Buddhism throughout China at the hands of gentry patrons in his and the following generations. However harshly the Hongwu regulations had affected Buddhism in the flfteenth century, the emergence of a new social base for

The chinese state in ming society

146

monastic vitality from the mid-sixteenth would cancel out their influence. The support of the gentry rendered those regulations of no account. The emperors of the late Ming followed suit, turning their backs on the regulations of the founder and taking up the patronal pose. The Wanli emperor, at the behest of his mother, the Empress Dowager Cisheng (1546–1614), as of 1579 was playing the role of patron in the style of a Six-Dynasties emperor, dispensing lavish gifts to dozens of favoured monasteries and clergy.72 Imperial support extended to all aspects of patronage. When a monk desiring to rebuild Baohua Monastery outside Nanjing submitted a memorial to the Wanli emperor in 1605 seeking a new name plaque to indicate official blessing for his project, he not only received the plaque but was given a large sum of money from the empress dowager, as well as a copy of the Northern Tripitaka, a Buddhist statue, and a banner proclaiming the emperor’s protection.73 Active as a patron of Buddhism, Wanli was not a legislator of Buddhist institutions. Imperial gifts were not acts of state, however, but the private dispensations of the imperial family. They signified the emperor’s (or his mother’s) personal piety, not state policy The founder’s laws were not revised. The formal relationship of the state to Buddhism in the late Ming—at least as expressed in the abandonment of quotas, the decay of certification, and the decline of registries—was otherwise one of relative indifference. Buddhist monasteries pursued their activities independently of the state and were left to their own devices so long as they paid their taxes and did not harbour vagrants or criminals. As far as the late-Ming state was concerned, a monastery was simply another fiscal household (hu), distinct only in that it was constituted from elective rather than kinship affinities. Independence from the state was not the theme that late-Ming patrons and apologists struck when they defined the relationship between Buddhism and the state, as their commitment to Confucianism required them to do from time to time. They preferred to picture an interdependent relationship, drawing on the language of earlier edicts of patronage to do so. Thus Li Bangyan (js. 1572) says of a place such as Nanhua Monastery in northern Guangdong that the teachings of the monks there “subtly support the life of the state.”74 The noted Buddhist patron Lu Guangzu (1521–97) in a text on a Ningbo monastery speaks of “comforting the human desire to look up to heaven, fortifying the imperial plan, and settling the realm within the four seas.”75 Li Zhi (1527– 1602) could justify the greatness of Nanjing’s Qixia Monastery by declaring it to be a place where “all pray that the present emperor will enjoy long life for tens of thousands of years.”76 And by the time a patron composed a text commemorating the restoration of Bochi Monastery in Huaian in 1677, he could write of Buddhism: When the Three Jewels [Buddha, dharma, and sangha] constantly flourish, the four quarters are forever at peace. Looking upward, [Buddhism] prays that the emperor’s life may be eternally lengthened. Looking down, it showers fortune on the lives of the people. Great indeed are the advantages it provides the state!77 Such extravagant claims were hardly to be taken literally—except perhaps by individual officials looking to harmonize their service to the state with their enthusiasm for patronizing the sangha. They served a distinct purpose nonetheless. By tying the

At the margin of public authority

147

patronage of the monastery to a statist logic, an author could defuse the ever lurking charge that to support a monastery was to put private concerns above public duties, to place religious needs as defined by the Buddha before the greater good as defined by Confucianism or the state. As always in the late-imperial period, the state set the terms of legitimation; hence the Bochi patron’s strategy of citing the state to justify supporting an institution that did not obviously contribute anything tangible to public authority as state officials might have understood the concept. Buddhism and the redefinition of public authority The rhetoric of Buddhism underpinning the state conformed nicely to the norm of imperial patronage that had appealed to Hongwu in the opening years of his reign. Even though this rhetoric had been set aside long before Wanli was willing to bestow his family’s attention on eminent monks and monasteries in the late Ming, patrons who invoked such phrases were not striving to work out any sort of new accommodation between Buddhism and the state. Rather, they were seeking to justify their involvement, as a state-oriented elite, with an institution that was central neither to the state’s agenda of rule nor to the reproduction of local society. Despite pronouncements to the contrary, Buddhism did not materially assist, or threaten, or otherwise matter to state hegemony Nor did the state desire to mobilize the religion for ideological values or intervene in its internal affairs. The failure of the Hongwu suppression to render Buddhist institutions permanently into more compliant form coincides with other failures to appropriate Buddhism or Daoism to the state’s cultic goals.78 Hongwu provided no material underpinnings to the castle of supervision he constructed in the air around Nanjing, and it dissolved as soon as he disappeared. His attempts to protect the property basis on which monasteries relied were of little defence against the workings of the market, and his decision not to incorporate Buddhist registries into the regular system of bureaucratic appointment impeded the creation of an organization that could be effective in realizing his goals. He lived long enough to regret his first policy of benign patronage, but not to see the futility of his second policy of suppression. Nor could he appreciate that monastic Buddhism, shunted to the margin of public authority, might there provide the gentry with means by which to pursue a quiet struggle for limited autonomy from the state-centred political realm. Having been banished to the margin of public authority Buddhism became available as a site for alternative constructions of public authority, in which the gentry could figure for themselves a prominent position. Marginality nurtured the monastery’s social existence as a realm somewhat separate from state institutions and processes. When that realm became integrated with local social structure, as it did through the widespread practice of gentry patronage in the late Ming, Buddhism provided a context for local elites to create an identity for themselves in terms that were to some measure independent of the state. The Buddhist monastery was attractive to the gentry as an object of patronage by virtue of being a respected non-state forum for publicizing their elite status. Unlike most other local institutions, the Buddhist monastery was hostage neither to the state’s interests nor to the particular interests of corporate groups. It was not public in the

The chinese state in ming society

148

sense of being a component in the structure of state authority; nor was it private in the sense of being closed to all but the members of a restricted collectivity. It existed between these arenas, constituting a limited but valuable space within which the local gentry could convert private wealth into public status and construct an identity at some distance from state mechanisms of status-conferral. Through their patronage of Buddhist monasteries, the late-Ming gentry were quietly reconstituting public authority in their favour.79 When local gentry patronized the same monasteries that the empress dowager funded, it seemed that both were adhering to the same norm of patronage—that the generosity of one was equivalent to the munificence of the other. Yet the social bases underneath this apparent congruence were different. The Ming state did not appreciate that gentry patronage was a matter of undertaking the construction of quasi-autonomous space within local society. This lack of perception has something to do with the intellectual repertoire available for thinking about Buddhist institutions. Traditional discourse provided a language to complain about monks failing to obtain ordination certificates, or about laymen engaging in sexual intercourse with nuns, or about bad characters “changing names to evade trouble and save their own necks,” to quote the aging Hongwu. But there were no rhetorical resources to connect the gentry’s pious giving, which looked as innocuously laudable as what the emperor was doing, and their self-differentiation from state systems; nor between the vision of a finely restored monastery and the construction of public space within the realm. The gentry were thus free to use Buddhist monasteries to deflect state control, and in so doing, to redefine public authority as an interactive fleld of rule and compliance within which they played a pivotal rather than passively supportive role. The gentry’s patronal involvement with Buddhist monasteries, and the new norm of toleration toward Buddhism that went with it, prompted a backlash among conservative elements, as we shall see in the chapter that follows this one. Guardians of the Hongwu dispensation liked to observe that the Ming dynastic founder had arranged Buddhist institutional life so perfectly that the monks could do nothing but “bring improvement to the people without corrupting our Confucian authority. Who but the greatest sage in the empire could have achieved this?”80 They suspected that the infatuation for Buddhism among their fellow gentry could bring the realm no benefits, and warned them against failing to devote themselves to their Confucian duties of providing charity and promoting education. This argument against the Buddhicization of the gentry transposed the more essential theme of rejecting gentry autonomy, which early-Qing advocates of a reconstructed Confucian order voiced. It did not, however, lead the Qing state to restore the supervisory mechanisms. The Qing was content to repeat the paper regulations for monks and monasteries laid down in the Ming and take no further action. It did not revive the registry system, nor impose quotas on monks, nor limit monastic property. Given the internal organizational weakness of Buddhism, which the Ming did everything to foster, the Qing state saw no need to police the clergy as closely as Hongwu did. The Hongwu suppression thus succeeded in a sense, for Manchu emperors could afford to adopt the patronal pose without worrying about what the Buddhists might get up to. In another sense, however, the suppression failed and, in failing, helped the gentry to redefine public authority This might have alarmed the Qing state, were it not for its own success in binding the gentry to its statecraft initiatives. These enabled it to recentre public authority on the state and block the possibility that monasteries could serve as a

At the margin of public authority

149

site for this function. The regime did not recognize what was afoot in local society in these terms, but it did believe that the late-Ming gentry, caught up with “mad Chan” and factional literary coteries, had failed its dynasty, and that for this reason the Ming house had fallen. It was the gentry, rather than Buddhism, that bore the brunt of this convenient condemnation. As for Buddhism, the Qing regime assumed that it remained safely and innocuously marginal to all forms of social power. With the Qing reassertion of a statecentred public authority, it did: its contribution to social change declined as the possibility of limited local autonomy faded. The Qing denouement does not diminish the importance of what Buddhism enabled in the Ming, nor of what it reveals of the possible variations that Ming society’s relationship to the state could tolerate. Patronage, suppression, and toleration, in that order, all had their effects on the social position and economic viability of monasteries, but the fate of the institution rested ultimately on its position within the local social networks that sustained it.

8 Buddhism in the Chinese constitution Recording monasteries in North Zhili Buddhism is not generally regarded as one of the elements constituting the socio-political structure of late-imperial China. It was, at certain earlier points in Chinese history, when devout emperors declared Buddhism to be the state religion and appointed monks to political office. By the Ming dynasty, as noted in the preceding chapter, this sort of patronal arrangement between the throne and the sangha was not a matter of state policy. Such memories of royal patronage that lingered or even momentarily inspired Emperor Hongwu were further buried by the harsh measures he imposed on Buddhist institutions in the latter half of his reign. Some of his successors revived the patron’s pose, bestowing gifts and recognition on favoured monks and monasteries, but they did so more as head of the imperial household, from which the funds for such gestures came, than as head of state. The state recognized certain popular cults for which temples could be built and at which sacrifices could be legitimately carried out, but it did not so authorize Buddhist sites. Buddhist monks were subject to state registration and their institutions treated as fiscal households, taxable as any other, but regular state systems did not otherwise engage with them. The Ming state was not interested in bringing the Buddhist establishment into its service, or into any other significant relationship. Though pushed to the margin of the state’s authority operations, Buddhism was not without a place in the Ming constitution broadly conceived. It is not conventional to speak of the Ming as having a constitution. The concept is usually restricted to the rhetoric of Euro-American political systems. There it is employed to enunciate an ideology of movement away from monarchical tyranny toward responsible government.1 Georg Hegel, an inhabitant of this way of thinking, declared that “we cannot speak, in reference to China, of a Constitution,” “for this would imply that individuals and corporations have independent rights—partly in respect of their particular interests, partly in respect of the entire State.”2 Hegel’s theory of the individual and the state did not permit him to accept that anyone but Europeans—or northern Germans, more exactly— could occupy a socio-political space in which legal custom and shared understanding protected the individual from arbitrary interference, and in which rights did not have to be legislated to exist in practice; least of all Chinese, who lived in a country that was for him a fantasy of absences confirming European uniqueness. Set aside Hegel’s parochial claims, however, and it is possible to imagine other histories and other interpretations. Philip Kuhn in Origins of the Modern Chinese State has identified the significance of changes in the relationships between elites and the state in nineteenth-century China by tagging them as “constitutional” in the sense of indicating “concerns about the legitimate ordering of public life.”3 That legitimate ordering need not cleave to notions of restricting state prerogatives or authorizing the individual to undertake political action in the ways worked out in the European context. Edward Farmer has gone further, arguing

Buddhism in the Chinese constitution

151

specifically for the Ming: “Taken together the codes and imperial rulings constitute a body of constitutional law.”4 Every political system has rulings and precedents governing the existing arrangement of power between the state/ruler and the individual in that system; in China they were codified and made available for magistrates and litigants alike to use. The meaning of the “legitimate ordering” that every political culture fashions may be enunciated in the language of freedom, or of morality, or of efficiency, as each culture goes to its own ideological ground to imagine the stability it desires; but that enunciation follows from the political reality it expresses: it does not precede it. Like any other stably constituted political regime, the Ming operated according to clear rules and precedents, marked out legitimate spheres of public action and prohibited those it considered deleterious to public order, and interacted with its elites in predictable ways to promote the public interest. And as is true of any regime able to exert its influence over local social organization, these constitutional arrangements extended to religion. Ming legislation on Buddhism One place to begin looking for the basic terms of the state’s relationship with Buddhism is the legal framework for the operation of the state enshrined in the Ming Code. Statutes relevant to Buddhism are found in two sections, taxation and rites. The section on taxation mentions Buddhism in two statutes, one banning the founding of new monasteries and the private ordination of monks and nuns, the other outlawing clerical marriage.5 The section on rites makes “heterodox practices” or “heterodox techniques” (xieshu) a capital crime. The relevant statute singles out Maitreya Buddha and White Lotus as Buddhist examples of the names under which heterodox groups could be organized, and mentions the revealing of hidden images and the collective burning of incense as indicators that heterodox practices were being performed.6 These laws barely begin to sketch the constitutional position of religion under the Ming state, but they do indicate two basic principles. The first is that a religious institution was a unit of fiscal account and therefore subject to tax regulations. The state banned private monastic founding to prevent the formation of unreported households over which it would be unable to exercise fiscal control; and it obliged monks and nuns to remain celibate, to back up the monastic rule of celibacy, certainly, but more importantly to block the formation of conjugal households inside the monastic corporation that might siphon off fiduciary property or transfer to itself the service-levy exemption customarily granted to monks. Ecclesiastics were expected to abstain not just from sexual relations but from property relations which might enable them to evade the rules of the tax system. The second principle addresses a different concern: the fear that religious communities might engage in behaviour that was heterodox, a term that is usually glossed in Confucian language, but which could also be expressed as what was considered not in the public interest. Behind this principle lie three implicit understandings regarding the constitutional status of religion: that the state reserves the right to determine the public interest; that what happens out of sight of the state is suspected of not being in the public interest; and that religion is a likely context in which ideas injurious to the public interest might gain currency. All three understandings impinged negatively on the reputation of Buddhism, notably in the eyes of conservative Confucians who, as we will see, often

The chinese state in ming society

152

chose to agitate against the power of Buddhism in local society. The concerns expressed in the Ming Code hint as much. Regardless of the state’s monopoly on the definition of the public interest, monks and monasteries occupied a liminal space between the conventional orthodoxy of the state realm and the hazier realm of unlegislated popular beliefs where other decisions about the public interest could get made, where much did indeed go on out of sight of state representatives, and where alternatives, if only of a mild sort, might find credence. The state understood that Buddhism could be more than its domesticated monastic presence. At no point does the Ming Code offer a general statement on the state’s relationship to religion in the ordering of public life; that was neither its purpose nor its textual mode. One can find something of that sort, very brief, in the main compendium of imperial legislation, the Da Ming huidian or Statutory Precedents of the MingDynasty. The second section of the 104th chapter in the 1587 edition of the Huidian contains all the important imperial edicts on institutional religion, and at the head of this section is this short introduction: Buddhism and Daoism have been popular among the people since the Han and Tang dynasties, and [would be] difficult to do away with completely All one can do is to be strict about [maintaining] the restrictions and agreements and not let the two spread further. The relevant regulations are all here, detailed and thorough in the extreme.7 It is hardly an understatement to say that this text is hostile toward Buddhism. The text assumes, first of all, that the ideal for the state would be to do away with Buddhism. It accepts, secondly, that Buddhism cannot be allowed to exist on its own, and that the state must have regulations in place to restrict it. It concedes, thirdly, that Buddhism has been hugely popular for well over a millennium, and that, were it left unrestricted, it would become even more widely practised among the people. Finally, it notes that the legislation put in place starting in the Hongwu era has been remarkably successful in restraining Buddhism’s natural propensity to grow and spread. The harshness of tone compels us to ask several questions of this text. First of all, working back from the end of this list, were the Hongwu restrictions effective? The preceding chapter testifies that Hongwu’s legislation did have a huge effect on the Buddhist establishment, although the answer as to whether it was effective in suppressing Buddhism involves both yes and no. Secondly, did Buddhism flourish in the Ming? Again, the preceding chapter suggests that it did, within the constraints imposed on it, notably among the late-Ming elite, although the evidence in the present chapter will go further to show the extent of its popularity in local society. Thirdly, did the state have to impose regulations in order to keep Buddhism under control? For this question there is no answer that is not itself prejudiced by the perspective of the person giving the answer. There was in fact a deep split, within elite opinion at least, concerning the necessity of state controls, evidence for which makes up much of this chapter. Some took Hongwu’s view and regarded Buddhism purely as a shelter for dodging one’s social and fiscal responsibilities and a scheme to dupe and defraud the ignorant; others accepted Buddhism’s proposal that it constituted an alternative social collectivity that marked out a realm separate from, but also complementary to, the social sphere dominated by the state.

Buddhism in the Chinese constitution

153

Finally, was the Chinese state committed to the eradication of Buddhism, as the Huidian commentator supposes? This question will come up in the course of this chapter. As we shall see, the answers will fall on both sides; and as we shall also see, the eradicationist view will be unable to best the preservationists’ argument that the state’s promulgation of regulations restricting Buddhist institutions amounted de facto to a clear acknowledgement by the state that Buddhism had a right to exist, that it was an ineradicable component of Ming society. What the Huidian editor does is not to describe the significance of Ming legislation on Buddhism, but to present a social analysis that chooses to isolate Buddhism within an alien category where it has nothing to do with the constitution of the Chinese state or with the reproduction of Ming society It argues for Buddhism’s extraneity to the Chinese constitution on the basis of a social analysis of Ming politics as a simple polarity between the state and the people. The people are Buddhist and the state is not; and so just as the state’s project is to control the people, so too it must be to control Buddhism. Missing from this analysis is any mention of the elite that the Chinese socio-political system generated, the gentry. Were the gentry in perfect accord with state desires on the subject of Buddhism and its ally in the war on Buddhism? Or were the gentry in the same category as the gullible people and therefore equally deserving of regulation by the state? The text assumes the first, and by so doing perfectly enunciates the Confucian/Hegelian ideal of a realm in which there is only the one jun/ruler and the many chen/subjects. It is a realm in which there is no work to be done to sustain the integration and coherence of Ming society through the reproduction of the political, social, and economic relationships (among others) between elite and non-elite that are basic to the functioning of any society The handing down of laws and punishments by the state is all that the model requires. This simplification points to the fundamental flaw in the Huidian text’s analysis: its refusal to acknowledge institutional religion as a component element of state and society as they actually existed. To dismiss Buddhism as nothing more than a parasitic growth on the body politic for which the options, to continue the medical metaphor, are prophylaxis (prevent it from spreading further) and medication (apply treatments to get rid of it) contradicts the claim on which the logics of eradication and regulation are based: that Buddhism is a potent force in society, and that it does have a direct relationship with the state. Making the gentry absent is essential for maintaining the fiction of parasitism. In fact, the gentry were much involved in Buddhist practices, even though they were under pressure not to own up to that involvement. For the anonymous official author of the Huidian text, the gentry cannot be shown in a relationship to Buddhism, since that repels them from the state side of the state-people polarity where their eligibility for state service places them. On the other hand, if the gentry are left inside the category of the state, their involvement with Buddhism implicates Buddhism with their relationship with the state. Contrary then to the claim of the Huidian text regarding the state’s desire to eradicate Buddhism, I propose that Buddhism was integrally involved in the ordering of public life, both as that life was lived in the context of local society and, as well, as it was lived in the context of elite politics, both local and national; and that almost everyone in the Ming understood Buddhist monasteries to be legitimate institutions within and around which other legitimate forms of social action took place. Even state documents could not tar institutional Buddhism with the epithet of “heterodox” (xie). I am, of course, positioning

The chinese state in ming society

154

the anonymous Huidian editor’s argument of Buddhism’s extraneity unfavourably in order to pursue my own for Buddhism’s constitutionality in the Ming socio-polity. If I engage in this bit of academic opportunism, it is to high light the difference in the social analyses underpinning his view and mine. Gazetteers as sites for representing Buddhism What this disagreement shows off to fine advantage is that Buddhism was a troubling and unstable element on the ideological landscape of Ming Confucians (if I may work with a social category somewhat smaller than “Ming gentry”). The materials that I present in this chapter attest that Buddhist institutions could be unstable points of meeting for many polarities: state and society officials and gentry, gentry and non-gentry, kinship group and kinship group. The most extensive and consistently confused evidence of this instability is to be found in local gazetteers. As a site for the dissemination of official knowledge about the life and administration of a county, the gazetteer was a text through which local officials and elites strove to project an image of local conformity to state requirements. It was also occasionally a forum for addressing aspects of public life or social practices that did not conform to state norms. The presence of Buddhist practices and institutions in the local setting was one of these sore points. How gazetteer compilers dealt, or refused to deal, with the presence of Buddhist institutions in the localities they chronicled reveals much about the presence and importance of Buddhism in the late-imperial socio-polity. Most, though not all, opened sections to contain information about Buddhist monasteries and monks. What a compiler put in that section required that he make some decisions about what was appropriate to record concerning all that was connected with the Two Masters (er shi), as the Buddha and Laozi were commonly known, or the Two Teachings (er jiao). The uncertainty from one gazetteer to the next resulted in a restless flow of political and social commentary as compilers circled around the problem of whether, in Confucian terms, Buddhism was “within the Way.” Did Buddhist monasteries uphold or subvert Confucian morality? Did they strengthen or weaken the authority of the state? Did their presence enhance or degrade the public interest? Did Buddhism contribute to or undermine the good order of society? The worrying of puzzled Confucian authors on these points should not inspire us to find the right solution to their bewilderment, but it may serve as an invitation to take their lack of resolution as evidence of a fundamental tension in the constitution of the late-imperial Chinese state. There were no fixed conventions on how a compiler should deal with information about monasteries, only a range of choices in practice. The first question he had to face was whether to include such information at all. If he did, he had to decide whether all the monasteries within the county should get listed, or only some to the exclusion of others. That decided, should he provide only the names of monasteries, or should he include other information, and if so, should he limit that information to addresses only or should he put in information about when they were founded or rebuilt, and by whom? If he chose to include more than the most basic information, should he also reprint the essays and epigraphy recording such information in greater detail, or should he keep his material to a bare-boned summary? The compiler then had to decide where to put this list: in its

Buddhism in the Chinese constitution

155

own section, entitled siguan (Buddhist and Daoist monasteries), or as an appendix to another section? Should it go in the chapter on ritual institutions, or in a separate chapter on Buddhist and Daoist matters, or in the residual miscellany that often made up the last chapter in the book? Having made these decisions, he then had to decide whether it was necessary to comment on any of his choices, since all were potentially contentious. This he could do in the preface (xu) or principles of organization (fanli) at the head of the gazetteer, if his acceptance or dismissal of Buddhism was either simple or stark. If he could not so easily dispose of the issue, he might compose a short essay to introduce or close the monasteries section in which he laid out his reasons for censoring or not censoring information on these religious institutions. No other section of the gazetteer gave a compiler so much trouble. The disturbance in the genre indicates not just that the rules governing gazetteers were unfixed, but that Buddhism was itself a matter of anxiety for state-oriented elites. To keep this study to a manageable scale, I have limited my survey to the gazetteers of one province, the metropolitan region in which Beijing lay: North Zhili in the Ming, Zhili in the Qing, and Hebei thereafter (the same region, and roughly the same sources, as in Chapter 4). The views expressed in North Zhili gazetteers are not necessarily representative of China as a whole, if such a generalization were possible, though my reading in the genre across the country suggests that they are not untypical. It would be erroneous to think that they should be taken to represent only the views of northerners, since many were compiled by magistrates from out of province. On the other hand, those who lived or worked in the north, particularly in proximity to the capital, tended to share a conservatism on cultural matters, compounded by the sense that they as smaller elites in less economically prosperous counties, should work harder to try to bring their locales into line with the strictest interpretation of state programs and rules rather than strike off into new ideological territory. North Zhili magistrates were anxious to keep their administrations within whatever expectations they presumed prevailed in the capital nearby; so too, compilers of North Zhili gazetteers were concerned that their county’s appearances conform as much as possible to the model mandated by Beijing. One component of that model was the Hongwu legislation against Buddhist monasteries. Almost every North Zhili compiler was sensitive to Hongwu’s amalgamation order and ban on the founding of new monasteries in 1391. This edict created two categories of monasteries: those that had official permission to exist, and those that did not, whether because they had been amalgamated or because they postdated the Hongwu era. An extreme compiler hostile to Buddhism could take Hongwu’s antipathy for Buddhism as an indication of best practice and exclude all monasteries from his gazetteer, but that was to go further than the emperor himself. The more reasonable conservative position was to record only those monasteries the state finds legal as sacrificial institutions, since they provide benefit to the people and have the authority of antiquity, and leave out anything built after the Hongwu era, which by definition must be places devoted to “licentious sacrifices.”8 Another compiler has phrased his editorial policy as including “only old monasteries,” to which he realistically adds “those that have been built during our dynasty by imperial edict.”9 One gentry author in 1576 stresses that compilers have no excuse for excluding an institution that the laws of the dynasty permit to exist; his only principle of exclusion beyond the Hongwu laws was one he imposed on those who would

The chinese state in ming society

156

produce future editions, and that was that “those not included now should not be added later.”10 Qing compilers followed the same rules and precedents. The extreme position of total exclusion found some favour in the early decades of the dynasty: as one declared in 1676, since “our kingly government has continued the [Hongwu] prohibition on the setting up of Buddhist monasteries and Daoist temples,”11 he decided it was best to treat all monasteries as illegal institutions and leave them out of his gazetteer. Most conservatives took a more moderate position, acknowledging the legitimacy of the state’s restrictions on monasteries, but sticking to the letter of restriction, which gave legal existence to some monasteries while excluding others. This is what a compiler in 1749 does when he declares that he has excluded privately founded monasteries, since these would fall within “the present dynasty’s meritorious ban on monastic founding.”12 More lenient again, especially when founding dates were confused, was to include any monastery founded before the Qing, to which could be added monasteries newly built but claiming a pre-Qing founding.13 Finally, Qing compilers usually accepted that they had to include any monastery with an imperial connection: one on which an emperor had bestowed an official name plaque or other imperial gift, or which an emperor had visited, or to which the court had given a tax exemption.14 These marks of imperial favour continued Ming practice.15 Where imperial patronage could strike local compilers as problematic was when it colluded, or at least coincided, with eunuch patronage. The eunuchs of the imperial household displayed a particular fondness for Buddhism and liked to extend their patronage to institutions within what was, for many of them, their home region. Sometimes these institutions were also privileged with an emperor’s notice, which a favoured eunuch might secure. Eunuch patronage was not only a fact in the Buddhist history of North Zhili, but an element in the calculations of the North Zhili gentry, who might condemn that patronage as a way of expressing antipathy for eunuch involvement in the operations of government. This antipathy could take more than discursive form. The gazetteer of Xincheng county records that “the local people” were so affronted by extravagant claims made on a stele erected at Huguo (“Protect the Nation”) Monastery on behalf of its eunuch patron that they cracked it.16 The brief comment does not indicate who these “local people” were, nor why they were contesting claims at an institution bearing such an egregiously state-oriented name, but by its tone the report sides with the righteous populace against the insufferable arrogance of the eunuch patron. Even if the Xincheng compiler felt obliged to include Huguo Monastery in his gazetteer, he could still press the local case against eunuch patronage by smuggling a reference to the incident into the public record. State restriction as Confucian prophylaxis Antipathy for eunuchs is not sufficient to explain why North Zhili gentry liked to complain about the power and influence of Buddhism in their gazetteers, as they did with untiring consistency. That complaint is of a piece with the northern gentry’s reputation as a dourly Confucian lot who were unsympathetic to the cultural and political enthusiasms of their southern counterparts. The taste for abbatial friendships and monastic patronage

Buddhism in the Chinese constitution

157

so strong among the Jiangnan gentry was not something most of them shared, except during the heady days of the late Ming.17 Their distrust of such cultural activities dovetailed with their grumpy attitude toward the richly non-Confucian world humming around them, a world in which popular religious practices went on out of their sight and made them anxiously dream of restoring a staunchly Confucian dominion of rites and deference. That dominion may never have existed, but appealing to it was a way of putting themselves between the people and the state, and giving themselves the illusion of having a more secure place in the order of things. The locus dassicus to which many a North Zhili gazetteer compiler appealed by way of supplying a Confucian justification for treating Buddhism in a hostile manner was the polemic that Han Yu (786–824) wrote to discourage a Tang emperor from receiving a Buddhist relic in the palace. In the Ming and Qing, citing Han Yu’s “Yuan dao” (On the foundation of the Way) was the easiest way to pitch a case against toleration for Buddhist institutions and practices.18 Han’s arguments against honouring the relic were more xenophobic than moral, but the logic of his rejection was of far less consequence than that he enjoyed a high reputation as a Confucian philosopher and therefore could be counted on as an unassailable authority for turning the state against religion. Beneath Han’s repudiation of the Buddha, however, lay a more dangerous challenge, which was to deny the emperor the right to do what he liked, which almost got Han executed. This implication may not have been lost to some of the anti-Buddhist polemicists in the Ming and Qing. In this way of looking at things, emperors who patronized Buddhists in a family rather than state capacity were nonetheless exceeding their constitutional limits. So explicit a constitutional challenge to imperial prerogative had to be kept utterly submerged when gazetteer compilers cited Han’s text. They did not dare command the emperor to stick to the Confucian straight and narrow, even though that is what they wished him to do. The anti-Buddhist comments that can be found in North Zhili gazetteers are generally phrased in the language of Confucian self-discipline and moral prophylaxis. That selfdiscipline committed the conservative gentry to act as the defenders of orthodoxy. It was a defence that most wanted to see extended over as much public ground as possible. Buddhism could not be left to the masses, who might mobilize it to promote their interests. It had to be marked as the first zone that lay in the path of the vigilant Confucian trekking his way into the cultural wilds of popular religious life. Among other things, that vigilance meant reminding other Confucians, whether they be local gentry or local officials, of what distinguished them from the common people. “We Confucians,” the compiler of the Jinzhou gazetteer of 1690 states, “do not talk about the Two Teachings and are strict about heterodox ways.” He cannot declare Buddhism and Daoism to be heterodox, since he knows they are tolerated by the state, but he can warn that they might become hotbeds out of which heterodox thinking arises. “We Confucians most certainly treat them with contempt,”19 another compiler assures his readers, as well he might, given the history of sectarian rebellion on the North China Plain.20 In general, compilers appeal more to Confucian values than dynastic regulations to justify restricting their records of the Buddhist presence in their counties, although the unfinished promise of the Chinese state to conduct itself as the exclusive patron of the Confucians never lurks far in the background.

The chinese state in ming society

158

This is the position that Magistrate Zhang Xun takes in the earliest surviving Ming gazetteer in North Zhili, the 1373 gazetteer of Zhuozhou subprefecture south of Beijing, when he observes, “Confucians do not talk about things related to Buddha or Laozi. Using their propaganda about sin and fortune to transform ignorant customs is like using a torch to brighten the sunlight.”21 In 1373, this pitting of Confucians against Buddhists and Daoists had a particular edge, for during the just deceased Yuan dynasty, Mongol emperors had treated all three alike as technicians of the invisible realm and equally worthy of support. Confucians did not like to think of themselves within the same category as Buddhist monks. Zhang’s objection may have had a doctrinal logic, but it more likely stemmed from his sense of the local market for patronage. At this early point in his reign, the Hongwu emperor was still Buddhism’s patron. Buddhists and Confucians may both have assumed that each would have to compete for opportunities to participate in local manifestations of the state’s presence. That could not have pleased Zhang. The Ming state was on Zhang’s side. It elected to re-differentiate the cultic streams in Chinese society and mark off Confucians for official state patronage and, after 1380, Buddhists for state supervision. It rehabilitated Confucianism as the conduit of state orthodoxy, extended formal recognition to cults connected to Confucians or exemplars whose moral virtue complemented Confucianism, and made these central to its official regulations governing sacrifices, the “sacrificial corpus” (sidian) or “ritual corpus” (lidian). This corpus did not include any Buddhist rituals. Accordingly, the compiler of the Wuqiang county gazetteer of 1694, who begins by announcing that “we Confucians spurn the Two Teachings,” concedes that Buddhist and Daoist rituals were in wide use, but can observe that these were not the rituals conducive to propagating the “moral teachings” that Confucianism, through state sponsorship, brought to the world.22 Many compilers make direct reference to Buddhism’s absence from the state’s sacrificial corpus, usually to justify their unwillingness to “indiscriminately mix them” with officially sanctioned state-cult institutions.23 Some use this absence to justify cutting down the scale of reporting of local monasteries,24 a few to cut them out of the published record altogether.25 Confucian competition thus combined with state restrictions to cast a shadow over the legitimacy of Buddhist monasteries, at least in the eyes of the gazetteer-writing segment of the local elite. It did so in the face of considerable popular support, as the compiler of the Xinhe county gazetteer of 1679 admits when he professes to be at a loss to account for the greater popularity of Buddhism over Confucianism: Buddhist and Daoist monasteries, chapels, and cloisters properly have no relationship to the official sacrificial corpus. Yet rural bumpkins and ignorant folk all go around to the monasteries to pray on their knees, and to the chapels and cloisters to burn incense and beg for good fortune. Why is this? Is it perhaps because the Teaching of the Sage isn’t as easy to comprehend as Buddhism and Daoism? Or is it because the sacrificial tablets aren’t brightly coloured as are their clay statues?26 The author chooses to wrestle with the fact that Confucianism had neither a reassuring message nor “curb appeal” for non-elites, yet does not reflect on the conditions of the lives of most people that left them vulnerable to disease, want, and unattended old age at

Buddhism in the Chinese constitution

159

a much higher rate than himself and his gentry friends. Instead, he parses their Buddhism as a lapse of intelligence, discrimination, and good taste. A linked argument against Buddhist institutions and rituals was that they were not just gaudy but wasteful. Some compilers found it an easy matter to jump from brighdy coloured statues to expensively decorated ones, and then declare the project of monastic patronage to be an unjustifiable drain on local resources. In his postface to the monasteries section of the 1604 Huairou county gazetteer, Magistrate Zhou Zhongshi launches his wide-ranging attack on Buddhism from this position: Today the realm has reached an extreme of poverty. If we wish to economize, nothing is better than cutting out extraneous expenses; among such expenses, nothing is more wasteful than constructing palatial buildings; and among palatial buildings, nothing is more wasteful than monasteries. Zhou goes on to concede that doctrinally Confucian arguments against Buddhism, though sound, get him nowhere with popular opinion: How can we use clear and readily understood principles to criticize them [Buddhism and Daoism]? The delusions of this generation cannot be dispelled. The first delusion is not respecting parents at home but respecting spirits and Buddhas outside the home. The second is not trembling before state regulations but secretly fearing to go against the Buddhist dharma. The third is not mending what is right in front of your eyes but instead trying to mend what is off in the next life. The fourth is fighting over wealth with kinsmen while giving riches to priests and monks. Why don’t even one or two ignorant men and women see this and return to orthodoxy.27 This diatribe goes on with quotes from Mencius and later Confucians, including Han Yu, to deplore the influence of Buddhist institutions over the people of his county By phrasing his concern in terms of Buddhist rules outranking state regulations, Magistrate Zhou signals his anxiety that, in our language, religion was replacing the state as the prime locus of public authority His only way to undermine this authority was by pointing out the costs involved—which, to those who willingly gave, would have made no dent in their devotion. The recording of the local institutions and history of Buddhism in local gazetteers was thus in the hands of an elite concerned to set themselves apart from the common people and stem the tide of heterodoxy they felt rising from below. Their appeals to state regulations and sensible economy were in that sense less against Buddhism than against those who identified with Buddhism, and therefore, by elite logic, against the authority of state Confucianism. This logic rests on the same false social analysis that the Huidian compiler used: the state versus the people, with the elite perfectly coterminous with the state. Yet gazetteer compilers were obliged to acknowledge that even the Hongwu emperor did not banish Buddhism from the realm. If that emperor did not deny Buddhism a place in the Ming constitution, how could they? There was, however, one eminent

The chinese state in ming society

160

early-Qing exception whose example created a troubling precedent that all subsequent compilers thereafter were forced to address. Lu Longqi and those who disagreed with him Lu Longqi (1630–93) was appointed to serve in Zhili as magistrate of Lingshou county in 1683. This was the second appointment of Lu’s career. His previous appointment had been as magistrate of Jiading county, just west of Shanghai. He was given that post in 1675 and by local account ruled virtuously,28 but lost it within two years, apparently for rejecting overtures from superior officials to do as they wanted. Forced into retirement, which lengthened out to six years while he observed the statutory mourning for his father, Lu taught and studied the classics, particularly the classics on rites. His reading notes from these years, completed in 1679–80 and published under two separate titles, marked him as an expert on the learning of Zhu Xi, particularly with regard to ritual, the choreography of correct conduct through which every act one performed in the presence of another person fulfilled his social obligations and confirmed his place in the moral order of things. In the unsettled decades following the collapse of the Ming dynasty, a careful adjustment of the personal to the social, and by extension to the cosmic, provided a reassuring modus operandi for the educated elite. Lu’s record for incorruptibility lent the authority of practical experience to his reputation as an authority on ritual, and he became one of the prominent neo-Confucian conservatives of his day.29 Lu Longqi brought with him his learning, his character as an upright official, and Zhu Xi’s hostility to Buddhist influences on Confucian practice when he went to Lingshou county. His interpretation of ritual would prove to be both an inspiration to administrative practice and a burden. “Ritual blocks what has not yet happened,” Lu conceived, thinking specifically of the prophylactic capacity of Confucian ritual; but it worried him. “By blocking what has not yet happened, people may not be able to see what its results would have been and will think it could have been put off.” The worry here was that people would not recognize what Confucian ritual was doing for them, the disasters it had saved them from, and think it worthless. They would then be tempted go off in pursuit of rites more attuned to short-term personal goals, which is what Buddhism’s promises of good fortune and salvation were. Properly speaking, Confucian ritual—of which the most important in the late-imperial period was the ritual relationship of jun/chen, of monarch/subject—was the only form that true ritual could take; no other ritual corpus did what Confucian ritual was charged to do: “to bring people to the Middle Way.”30 Lu initiated the project to produce a new gazetteer for Lingshou shortly after arriving. After two years of work, the gazetteer was published in 1686. Armed with more selfconfidence and rectitude than the average compiler of little scholarly reputation and modest Confucian learning, Lu set down strict rules for compilation. One of these was the radical decision to go one step further than the Hongwu emperor and cut the section on Buddhist and Daoist monasteries out of the book. In his editorial comments at the head of the gazetteer, he states categorically: “Suppressing heterodoxy and lifting up orthodoxy is the great prophylactic task of government. Therefore no Buddhist or Daoist monasteries have I deigned to list.” At the end of the section in the second juan on temples, which is where one would expect to find the standard list of monasteries, he observes that in the

Buddhism in the Chinese constitution

161

hills and upland valleys of his county, “the shrines of the Two Teachings are found in profusion, awe-inspiring and magnificent, where rituals are conducted daily without pause.” What distressed him was that people were visiting these shrines and carrying out their rituals, to the neglect of temples mandated in the official ritual corpus. He regarded this situation as a “great imbalance” and called on the educated to “promote the one and dispense with the other.”31 His own gazetteer would be a testimonial to that call by making all non-Confucian ritual sites disappear from the official record. Lu’s decision was much discussed, and much disputed, by other compilers in the region. Conservative Confucians cheered his uncompromising repudiation of the antiConfucian influences that some thought Buddhism exercised on local society. Few, however, were willing to follow him to this extreme. Chen Jinyan, the compiler of the Leting county gazetteer of 1755, sought to find a way down from Lu’s unassailable moral position. “When Lu Longqi compiled the Lingshou gazetteer, he did not record Buddhist and Daoist monasteries or monks. This could be said to be lofty wisdom,” Chen conceded, but he was not happy with the judgment. He felt that monasteries should be included, yet the earlier edition of the gazetteer from which Chen was working, possibly of 1622, had excluded them, as Lu had, on the grounds that they were “vulgar” institutions. While Chen accepted Lu’s position in principle, he could not get around the fact that, as he pointed out to his readers, “the county actually has them. Cutting them out of the gazetteer would be to go against the genre.”32 As a gazetteer compiler, he was tasked with recording whatever the genre endorsed. A monastic section was conventional to the genre, and Chen did not feel that he could follow Lu’s or his own predecessor’s example. In fact, some argued that including monasteries was not just by convention but by principle, since gazetteers were supposed to be complete historical records in the tradition of the great Han-dynasty historian, Sima Qian.33 The principle of maximum inclusion was dear to many compilers. As one writes in his preface to a supplementary chapter on monasteries and monks, “I wish to list in detail and without omission the events and places of the entire region.”34 This desire trumped his Confucian mandate to disapprove. We find the same logical switchback, expressed with less sophistication, by the compiler of the 1874 gazetteer of Fuping, the county directly to the north of Lu’s Lingshou: “One might ask whether it was right in the Lingshou county gazetteer to remove Buddhist and Daoist monasteries, I say, why not? Nonetheless, it does not do to be excessively narrow.” He goes on to note that there are people in Fuping who engage in legitimate religious exercises—legitimate from the state’s point of view by doing so in secluded spots out of the flow of public traffic—and that the gorgeous mountainous scenery has made the county ideally suited for building attractive monasteries of the sort that recluses love to occupy and literati to visit. He thumps in good state-Confucian fashion that monasteries should not be allowed to exist “in the villages and along the roads” where they can siphon off people’s wealth and lead them into heterodox ways, just as Magistrate Zhou Zhongshi did. But he does not take the next step and argue that Confucian prophylaxis justifies either their eradication or their removal from the gazetteer.35 Still, Lu’s unimpeachable Confucian authority in these matters was strong, so much so that one compiler tried to argue that, by just keeping to the rules about not recording recently built monasteries and listing only the older ones, he was living up to

The chinese state in ming society

162

Lu’s ideal, just doing so through a more moderate practice.36 It is hard to imagine that Lu would have been persuaded. Most Zhili compilers, regardless of their antipathies to Buddhism, chose not to adopt Lu’s position. Keeping the siguan section might require some discursive contortion for the compiler who wanted to get himself into what he felt was the right posture on the issue of religion’s relationship to the state, and this is what most chose to do. Here is an eighteenth-century example: The ancients had a saying: even a place that is set apart is not beyond the official realm. A religion may be different [from Confucianism], but it is still controlled by the Ministry of Rites, so old monasteries and remnant temples are not discarded by kingly rule. My intention here is to be brief and not compose a full section for siguan.37 The suitable editorial compromise between acceptance and rejection, as this example explains, was not to let the siguan section be a self-standing chapter, but instead to append it somewhere as a subsection of another chapter. The most positive arrangement was to open a chapter combining monastic notices with monks’ biographies and place it at the ritually low end of the book, its end. This is what the editor of the Baodi county gazetteer of 1745 does. Monasteries must be segregated there, he says, because they “are left out of the sacrificial corpus,” and monks’ biographies because they “are not to be listed in the biographical section.”38 The more common arrangement was to put siguan in the chapter on ritual institutions, usually as an appendix to the list of altars and shrines recognized by the sacrificial corpus (that section was variously entitled tanmiao (altars and temples), tanci (altars and shrines), or miaoyu (temple buildings)). One compiler notes that he checked other gazetteers to decide whether this was an appropriate arrangement, and decided that it was.39 Variations include putting siguan in with “Buildings” (jianzhi)40 or attaching it to “Ancient Sites” (guji).41 The most sanitary option, short of not having a siguan subsection at all, was to banish it to the back of the book together with other “miscellany.” This was where many compilers put material of dubious status, though often of much interest. Roughly one in five gazetteer compilers consigned monasteries to this grab bag. Chen Jinyan, the compiler of the Leting county gazetteer who struggled with refusing to follow Lu’s advice and his own predecessor’s practice, made this concession to those who wanted to take the high road, keeping his new section on monasteries out of the chapter on state-cult institutions and putting it at the end of the book. Interestingly, however, the next Leting compiler was unhappy with this compromise. In his list of editorial principles at the front of the 1877 edition, Chen observes that the convention of putting monasteries in the miscellany section was innovated by Han Jun (js. 1599) in his 1605 gazetteer of Jiading county. Significantly, Lu’s only previous appointment before Lingshou was as magistrate of Jiading. As a conscientious official, Lu would certainly have known Han Jun’s gazetteer, even though it had been superseded by a large, new edition just four years before he arrived at his post. The 1673 compiler rejected Han Jun’s placing the entries on monasteries at the end of the book apart from other structures. “How can the errors of the wise that have come down to us, however lamentable, be left out?”42 he asks in his editorial comments at the front of the gazetteer. Having thus rebutted Han, Chen puts

Buddhism in the Chinese constitution

163

them back in the “Ancient Sites” (guji) section. Lu does not record his own reaction to Han, though as his practice was not just to support Han but to go a step further in his own compilation for Lingshou, Chen’s reference to the Jiading precedent has to be read as a rejection of Lu. In his own defence, Chen cited the editors of the imperial book-collecting project, the Complete Library of the Four Treasuries, who criticized Han’s way of organizing information about monasteries. The Leting compiler declares that his method of dealing with the question has been to follow Kang Hai’s arrangement in the 1519 gazetteer of Wugong county, Shaanxi, which was to put ancient religious sites in the “Geography” (dili) chapter and current monasteries in the “Buildings” (jianzhi) chapter.43 Lu’s harshest critic was none other than the leading Qing authority on local history, Zhang Xuecheng (1738–1801). Zhang allows that Lu’s intentions may have been lofty, but he faulted him for failing to understand that the key principle of history is to record what has been done, even if it should not have been done. To exclude monasteries by way of criticizing Buddhism is “perverse,” he complains, especially since “centuries from now who will know any longer that [by leaving them out] he was condemning them as heterodox?” All that does is obliterate the historical record. Zhang lists three reasons why monasteries should be included in a gazetteer: monastery steles record significant historical texts; those who serve as monastic officials do so at the behest of state administration and should be recognized for doing so; and monasteries provide local officials with sites for distributing relief, caring for orphans, and enshrining local worthies. Monasteries “serve the great,” to use the language of loyal submission to the court. To leave them out would be a narrow chauvinism.44 The intellectual authority of Zhang’s critique of Lu inheres in that fact that it does not require taking a stand on Buddhism’s place in late-imperial society. It rests instead on the argument that, as Chen Jinyan has already put it, “the county actually has them. Cutting them out of the gazetteer would be to go against the genre.” This is the trump that the most strident Confucians could use against themselves, as one does when, after quickly disposing of Buddhism as silly superstition, he answers his own question “Why record this?” by saying “Nothing is to be left submerged.”45 There was also the powerful pull of textual precedent, as a compiler notes at the end of his prefatory comment to his siguan section when he observes: “It is said that what has been listed cannot be excised, so I have recorded siguan.”46 Anti-Buddhists could assuage their Lu-inspired guilt, if they felt any, by arguing that data on monasteries could have historical value in terms of preserving other information relating to state operations, which is what Zhang implies with regard to the people who held monastic office on behalf of the state. For example: The Ministry [of Rites] has appointed ordination officials and set up suboffices to record these [ordination certificates], such that officials can at any time inspect them to determine where [the monks] are affiliated. Therefore I have no choice but to list them in an appendix.47 Another historical value to preserving material on monasteries was in providing data on the fluctuations in order and prosperity, since the rise and decline of monastic buildings could be interpreted as a gauge of general administrative health. As one other compiler observes in his introductory notes, monasteries rise and fall like any other buildings and for that reason should be included in the “Buildings” section.48

The chinese state in ming society

164

Despite all these arguments, the pall of Lu’s extremism hung over Zhili compilers for the rest of the Qing. I have yet to find a compiler willing to justify the inclusion of information about local monasteries in other than defensive ways. Those who wanted to resist Confucianism’s bias against non-Confucian ritual sensed that they could take a stand against sectarian narrowness only at their peril. On the other hand, even stout antiBuddhists could argue their way to including Buddhist monasteries if they had to. The most ingenious I have come across is a sixteenth-century compiler who justifies including siguan so that readers of his gazetteer will see the follies of monastic patronage and the stupidity of religious superstition and thus be led, prophylactically, to shun heterodoxy.49 In fact, Lu’s notion of removing the section strikes me as a thoroughly Qing Confucian move. An extreme anti-Buddhist in the Ming might fantasize about a final solution for Buddhist monasteries, which would then obviate the need to record them,50 but until their statutory legitimacy was revoked, it was not within his prerogative to cut them out. What the Hongwu emperor allowed, no Ming scholar could forbid. As one long-winded Ming compiler put it, reading about Buddhist institutions in the local gazetteer allowed one to “witness the doings of the Great Sage,” i.e., the Hongwu emperor. “Therefore I have listed them in the pages of this book to show what existed in the past and is of no harm to the people.”51 “Dikes that hold back the flood” There was, however, a different argument to be made on behalf of recording Buddhist monasteries, which none of the compilers I have yet quoted has touched on. It is an argument that exploited the religious character of monasteries rather than setting it aside as the sort of thing a Confucian would not deign to notice; it also went so far as to find a legitimate way to accord Buddhism a role in the reproduction of state hegemony. This argument is that the teachings of Buddhism contributed to public order. It is an old theme in popular writings about Buddhism, and it was invoked by some gazetteer compilers, in milder or stronger forms, as a justification for allowing monasteries to enter the local historical record. As it is usually pitched, the argument combines a respect for the capacity of Buddhist teachings to create a pliant or submissive subjectivity among the pious, with the realization that Buddhist piety is far more widespread in society than Confucian doctrine. A mid-Ming compiler frames this thinking in terms of history: As for Buddhist and Daoist monasteries, it is because they are ancient vestiges that it is not permissible to destroy them. The reason we keep them is that in warding off disaster and protecting against harm, they help the people; and in enlarging the auspicious aura, they assist the state. And so I have recorded them all.52 He adds that he has put the monasteries subsection at the end of the chapter on state—cult temples, rather than at the back of the book, because they are “what popular custom reveres.” An early-eighteenth-century compiler similarly argues that Buddhism is both pacific and endemic. He asserts that “reverence of these doctrines [Buddhism and

Buddhism in the Chinese constitution

165

Daoism] can help in furthering the task of kingly transformation”; he then notes that Buddhism has proven to be a tough, inescapable presence in Chinese society: “Buddhist cloisters and Daoist temples, where people seek for good fortune and pray for blessings, cover the empire to an astonishing extent.”53 From monasteries issued ideological influences that enhanced social stability by encouraging devotion to the Buddha and piety toward the state. Even though, as an early-Qing compiler declares, “we Confucians do not talk about the Two Teachings and are strict about heterodox ways,” yet the ubiquity of Buddhism and Daoism can be turned to good effect: I say that if people did in fact constantly worry about life and death and about fortune and misfortune, then few would go against their superiors or make trouble. Thus the Way of the gods can firm up the proper Teachings [of Confucius]. For this reason have I recorded Buddhist and Daoist monasteries.54 One Ming compiler even goes so far as to suggest that this was all part of Hongwu’s plan from the beginning. “Our dynasty,” he claims, has allowed Buddhism’s “remnants” to continue to survive institutionally “simply to encourage the ignorant folk who believe in them to be good, not because it really reveres them.”55 This formulation saves Confucians from getting tainted by Buddhist doctrine while nonetheless allowing them to recognize that the ideological and organizational power of Buddhism was huge, and not necessarily at odds with their own political program. Common to these appeals is the sense of a world in decline from its golden Confucian age. Though inferior to Confucianism, Buddhism has its part to play in the historical solution to this loss. The compiler of the 1684 gazetteer for Beijing’s Wanping county offers this logic in order to banish all challenges to Buddhism’s legitimate place in the late-imperial socio-polity: When the Five Emperors and Three Kings ruled the world, there was only cultivating the self and keeping moral relations. With everyone cultivating himself and keeping moral relations, the Great Way was realized throughout the world. How could there be anything unreasonable or improper? After the Three Ages, sages no longer appeared. In the reign of Emperor Ming of the Han dynasty, Buddhism entered China. Its principal teaching was to do good and avoid the vices of killing, intoxication, and sexual licentiousness: simply to be pure and suffer patiently.56 He goes on to insist that Buddhism and Daoism both “are dedicated to teaching people to do good and refrain from evil” and therefore have “never been in contradiction with the Great Way of the Five Emperors and the Three Kings. Isn’t that the reason that people revere their teachings?” Not only are Buddhism and Daoism not in contradiction with the proper order of things, asserts another compiler three years later; they are essential to its maintenance: In the age of Yao and Shun, one had only to open sluice gates and let the water flow off. Today, if it weren’t for thick dikes and towering seawalls,

The chinese state in ming society

166

there would be no way to protect ourselves from the devastations of flood. Given the ways of this world, how are the ignorant lower classes any different from this? The Two Masters are truly the dikes that hold back the flood.57 This, he writes, is why he has included a siguan section in his gazetteer, contrary to the advice of Lu Longqi. His critics see only the ideal to which the world should conform; he sees its actual reality, and reality is the field on which Buddhism, not Confucianism, has the bigger role to play. From Ming to Qing These arguments ebbed and flowed in relation to the political and social changes that Chinese perceived themselves weathering over time. The highpoint of elite sympathy, or at least of their acceptance of Buddhist institutions as social facts too embedded in local networks to ignore, came in the late Ming. A 1618 compiler gives the strong version of this view when he opens his prefactory comments on siguan with the blunt statement, “The temples of Laozi and the Buddha are within the Way.”58 The conservative mood of the early Qing, however, was unable to tolerate such candid ecumenicism. Even as sympathetic an observer as the editor of the 1674 gazetteer of Kaizhou subprefecture felt obliged to declare that Buddhism and Daoism were not “within” the same Way as Confucianism. The only way to recuperate Buddhism was to declare that “the Three Teachings join together in having a common root in the Way.” Only from that formula could he proceed to justify why “scholars who have recorded temples, abbeys, monasteries, and famous mountains always do so in detail.”59 His pose as recorder was nonetheless the conservative one of registering the existence only of those monasteries that appeared in earlier gazetteers. In the early to mid-Qing, appeals to commonality between Buddhism and Confucianism as a reason to respect and record monasteries became more difficult to make. Those who invoked commonality had to do so indirectly and then back away from drawing radical conclusions about the legitimacy of a condominium between Buddhism and Confucianism. “For inquiring after the Way, the Three Teachings needn’t have come from the same source,” one cautious compiler avers, though he presses on nonetheless to suggest awkwardly that “in awakening the world they cannot at the beginning have pointed to something different.” Still, he is obliged to dampen the syncretistic implication of this comment by declaring himself opposed to the trick of mapping the concepts of one teaching onto the others in reckless combination, which was precisely what Qing Confucians thought Ming Confucians had done, and that Qing Confucians feared popular sectarian leaders were doing in their own time.60 Successive editions of a gazetteer sometimes provide a profile of the pressures on gazetteer compilers to conform to changing ideological fashion. Of the gazetteer sets I have tracked, only Fengrun county shows no perceptible change from one edition to the next in terms of the quality of coverage in the siguan section.61 In no other case were editors able to leave what their predecessors had done alone. At first glance, the three surviving gazetteers of Cangzhou subprefecture (1584, 1680, and 1743) appear to repeat

Buddhism in the Chinese constitution

167

each other’s editorial decision to include a section on monasteries, the only alteration being that each moved the siguan section back one juan from one edition to the next. A closer look at the 1743 edition shows, however, that that editor reduced the amount of information given in the 1584 and 1680 editions to just addresses.62 Other compilers are less discreet in their alterations. The two earliest extant gazetteers of Gucheng county (1594 and 1614) provide information about each monastery they list, whereas the 1727 editor not only reduces this information to addresses only but tacks on a short diatribe at the end of the section.63 The general pattern, as these examples illustrate, is for compilers to reduce the scale of reporting over time. Hejian and Zhending (Zhengding in the Qjng) are the noteworthy exceptions. In these two prefectures, some later compilers were willing to increase the space they devoted to Buddhist monasteries. The 1540 Hejian gazetteer lists no monasteries, although the cartographer for the project marked one on the map of the prefectural seat (see Figure 8.1). By contrast, the 1615 and 1677 editions both do.64 Similarly, the compiler of the 1549 Zhending prefectural gazetteer includes no Buddhist monasteries in his chapter on temples, whereas the editor of the Zhending county gazetteer in 1612 provides names and addresses and puts them in their own section, as does the 1762 prefectural gazetteer.65 One can find the same trend among gazetteers of counties elsewhere

Figure 8.1 Map of the seat of Hejian prefecture as it appears in the prefectural gazetteer of 1540. Although the compiler declined to list Buddhist monasteries, the cartographer marked Zisheng Monastery on this map, in the top right-hand corner of the city. Source: Hejian fuzhi (1540), tu. 6b–7a.

The chinese state in ming society

168

within Zhending prefecture. The 1617 gazetteer of Xinhe county has no siguan section, whereas the 1679 editor lists monasteries and adds some information. He appears to have felt himself to be going against the tide, for he adds the cautious proviso that “older gazetteers include them all without exception, whereas this I can’t do. Therefore I have edited [the list] down.”66 Either he was referring to the 1564 edition, which I have not been able to consult, or he is speaking more generally about North Zhili gazetteers in the Ming and sharing the censorious mood that would encourage Lu Longqi to exclude monasteries from his gazetteer in neighbouring Lingshou county six years later.67 Lu Longqi’s extreme position was shared later in the seventeenth century, even when it was not exactly followed. Gains in information about monasteries made in late-Ming or early-Qing gazetteers were often lost in their mid-Qing editions.68 The three gazetteers of Xincheng county follow this trend of disintegration step by step: the 1617 gazetteer includes a monasteries section with names and some information in its final chapter of miscellany, the 1675 edition reduces the information to addresses only, and the 1838 edition removes the section altogether.69 Occasionally, data that diminished and disappeared from local records in the early to mid-Qing reappears later. For instance, the 1445 Daming prefectural gazetteer includes a thorough list of monasteries at the end of his temple section; the 1506 compiler follows suit, though tacking on a comment indicating his sensitivity to objections; by 1672, his implied opponents have won out and the monasteries have disappeared. Monasteries return, however, in the Daming county gazetteer of 1790 and the prefectural gazetteer of 1853.70 The relaxation among later compilers may signal a healthy ecumenicism, or is it more likely that it indicates that Buddhist monasteries no longer had the power to disturb Confucian authority and so could be recorded without offending their sense of dominance? Buddhism in the Chinese constitution The growing disinclination among the magistrates and gentry of North Zhili to provide full reports on the Buddhist institutions and practices in their counties could be taken as a sign of increasing elite antipathy to Buddhism as a religion. Seen in terms of the gazetteer genre, however, this may not be a complete explanation of what was going on. The purpose of making a gazetteer was not to produce a documentary that transcribed the complex reality of local society, after all; it was to render that reality into a textual form that conformed to the principles by which the state governed the realm. Disagreement over how to report monasteries may reflect religious tastes, but I see it rather as a debate among state elites over the place of Buddhism in the Ming or Qing constitution. I can think of no better term than “constitution” to denote the set of norms and institutions that were understood as governing participation and control in the public realm. As a site of norms and institutions that could, at least in theory, compete with state authority within its narrow Confucian definition, Buddhism posed a challenge, especially in periods when state authority appeared to be under siege. There was no question of declaring Buddhism to be an illegitimate element in the authority structure of local society. Both the Ming and Qing states recognized Buddhist monasteries as legitimate religious institutions, so long as they met certain legal qualifications; so too they accepted Buddhist clergy as legitimate religious practitioners,

Buddhism in the Chinese constitution

169

so long as they followed the laws governing ordination. The question was whether the elite viewed Buddhism as protecting or threatening what they understood as the public interest. Therein arose different responses to the question of whether Buddhism should be fully represented in the local gazetteer. A compiler who suspected Buddhism of having a capacity to elude state regulation, and therefore the potential to erode the moral program of the state, would not wish to dignify its institutions by representing them in any but the most minimal way. A compiler who perceived the stability of local society as resting on more than official moral exhortation and correct administrative practice could take a different view and allow that Buddhist monasteries, as much as Confucian schools and state shrines, should be fully represented in an account of local society. Put starkly, these two positions represented two different ideas about the appropriate place of Buddhism in the Chinese constitution: the one as corroding the norms and institutions that should govern public life, the other as contributing to them. Members of local elites were keenly aware of the power and influence that Buddhism in fact enjoyed in local society. The compiler of the 1732 gazetteer of Wan county understood Buddhism as anything but inconsequential to the production of local order. He enumerates one dismaying sign after another that Buddhist institutions were in full flower in his county: Buddhist temples are ubiquitous and beyond counting, he complains; the “gold and azure sparkle” of Buddhist statuary dazzles onlookers; the amount of money fund-raisers are able to collect to support these institutions astonishes him. Most of all, he is horrified by the uncontrollable and vibrantly human activity that Buddhist institutions house: How do people dare illicitly build structures that are not on the list of canonical shrines? Still, lay people of this generation set up temples and make statues by the side of the high roads or at the edge of markets without justification, [places that] appear reverent and yet are rife with cacophony and confusion, without even a wall around the outside, while the stench [of incense] billows. How potent are their arts!71 This description is not of a religion on the margin or in decline. Buddhism was richly integrated into the networks of social participation and communication that extended from market to market along Wan county’s main roads. No wonder an anxious compiler a decade later stresses that he has excluded Buddhist sites “in the rural wastes and the village markets,”72 for these were the very places that the people who animated the network ties in local society crossed and thronged. The world of movement and activity beyond the capacity of the state to police in which this popular Buddhism was implicated was separate from the one the elite self-consciously inhabited, but it was not the irredeemable chaos that the compiler liked to imagine. Despite the impression of “cacophony and confusion” that Buddhist institutions may have given the Wan county compiler, these institutions were active in sustaining an order that was visible and reasonable to those who participated in it. The next slip down the slope of popular religious mobilization was the realization, unhappy though it was for conservative Confucians, that the Buddhist associational realm involved the local gentry as well. That involvement became something close to ubiquitous during the last century of the Ming, when the southern flirtation with

The chinese state in ming society

170

Buddhism spread north, to the dismay of some. A 1559 compiler comments on the emergence of this trend “in recent times” in his southward county of Nangong, just northwest of the busy Grand Canal port of Linqing, when he observes that “Confucians attired according to their status are all conferring their appreciation on the Two Masters, and even giving instruction in [the doctrine of] annihilation in order to appear eminent.” He closes his brief editorial with the comment “How distressing this is!”73 Not everyone thought so. As a local compiler stated somewhat defensively in 1576: Gentry go into retreat in the temples that Buddhists and Daoists have built on great mountains and in deep valleys, or stop there on their touring to catch the view and carve inscriptions. As the ritual regulations do not forbid local officials from supporting their doctrines or praying for long life, I record them to show all that has existed in the past.74 The taste for monastic visiting and the tolerance for official involvement fade in the early Qing, though even thereafter censorious editors complain about gentry infatuation with Buddhism. A 1676 compiler indicts his colleagues when he asks rhetorically, “How is it that we Confucians put ourselves under the Two Teachings?” 75 Such actions pressured the very identity of the self-proclaimed Confucian “we.” Another compiler four years later declares the Two Teachings to be “what Confucians should spurn,” yet notes that many literati do not, opting rather “to go and visit them in imitation of the ancients.”76 In the same vein, a mid-eighteenth-century compiler is exasperated to observe his fellow gentry going off to visit the county’s scenic sites that Buddhists and Daoists have taken over. “The gentry can go there to relax in their spare time to enjoy the pleasure of being beyond the mundane realm,” he allows, “but why do they have to lodge with them?”77 The problem with finding a lack of a barrier between these two social groups was that it required a different social analysis than the one offered by the Huidian editor. The gentry were not safely on the side of the state and its Confucian ideology, as that model insisted. They in fact were distributed unevenly across the polarity between state and popular society, a distribution that pointed to a more complicated social structure than Confucian sociology was content to admit, in which more than Confucian values glued society to the state, more than Confucian institutions mattered to the reproduction of the complex whole, and more than Confucian technologies produced the stable reproduction of rural society—though Confucians were not good at perceiving what they saw beneath them as stability. From any vantage point, however interpreted, Buddhism was conspicuously present on the local landscape of power. Conservative Confucians were unable to resolve the duality between the public order mandated by the state and the popular space in which most people lived, as their stream of offended commentaries suggests. The actions of emperors only made the situation worse, for while the state’s official position accorded with the separation of doctrines to which they subscribed, imperial patronage offended their sense of what Confucian emperors were supposed to do. If emperors wanted Confucians to keep up the bargain that they thought they had made in return for service, they should not dispense favours to Buddhist monks. Legal toleration and imperial prerogative were not matters the gentry could legitimately object to, which left them selecting the castrated, the pious, and the unruly as their targets of complaint.

Buddhism in the Chinese constitution

171

The tensions over the vital presence of Buddhism in local society reflect the fissures of class, privilege, and interest in the unequal power relations that animated political life from the county up to the court. These relations and fissures were central to the success with which the late-imperial socio-polity reproduced itself as something like an equilibrium: regulatory legal systems working from above, community networks brokering local interests from below. This array of power relations, rooted in the state system and filtered and acted out at the local level, is what early-modern Europeans were looking at on their own terrain when they spoke of having a constitution. The language of constitution allowed them to tell stories of individual freedom and elite privilege that confirmed Europe’s success in developing a manageable polity in which the privileges of the powerful were hedged, first monarchical power by aristocratic elites, then elite power by popular coalitions. State-oriented Confucian gentry did not enjoy the same relationship to their monarch. Lacking any franchise within the imperial state system, they were unable to mobilize for political action, although they did have cultural frameworks within which to pursue their interests in local settings with little reference to the state. When threatened from below by religious communities, they backed toward state systems and used them as bulwarks of enlightened conduct, fearful of a chaos that seemed poised to erode their local ground of privilege. For most everyone else, though, monasteries were part of the making and strengthening of the social networks through which they made their decisions about what to do, who to associate with, and where to invest in public and private goods. By furnishing them a place in the cosmic order that was also a place in the order of public authority and social exchange by which they lived, Buddhism grounded local social life as a reproducible equilibrium. It had a place, heterodox for some and normal for others, in the constitution of the late-imperial realm. Whether it was orthodox or heterodox depended on how they understood, or we understand, the composition and purpose of public life in that time and place.

Conclusions States of the field The administrative system the founding Ming emperor imposed in the latter part of the fourteenth century—powerful, well articulated, ambitious to penetrate social life—was an impressive structure, to read the documentary record. Whether in registering people, surveying land, publishing books, stocking libraries, suppressing religion, or imposing ritual codes on rich and poor alike—to mention some of the matters discussed in this book—the regime made its presence felt and mattered to how people organized their lives. These undertakings we recognize, from our time and place, as the work of a state. And yet the field of Ming studies grew up during the Cold War avoiding the word when referring to the power of the court and its political institutions, opting instead for such terms as “government” or “dynasty.” Tilemann Grimm was the first Ming historian, to my knowledge, to grapple explicitly with the question of whether the political operations of the Ming amounted to a state. In a 1985 symposium volume on the Chinese state, Grimm accepted that the Ming was indeed a despotism, and from there went on to take what he called “a sceptical view of the Chinese state.” Political life in the Ming was not governed by the organizational and legal norms associated with European states, he argued; it was a state that was not quite a state. Whatever the European state had, the Ming did not have enough of it: In order to qualify as a state, a regime must, in my opinion, have more than a power structure with certain executive agencies, and a certain number of bureaucratic procedures, and more too than a specific territory with armed forces to retaliate when “barbarians” happen to intrude. Lacking “fixed legal norms, which bind those who control it and those who depend on it alike,” the Ming was incapable of controlling “the steady competition for power and influence” at court and barely able to hold in check the “fluctuating sum of group conflicts” at the local level. Even so, Grimm chose to conclude his essay inconclusively: “Whether or not political power in such a system can be called a ‘state’ remains a subject for discussion.”1 It was a discussion that did not happen. Historians writing before 1985 who had refrained from using “state” continued to do so, while those writing after 1985, myself included, let the term quietly in through the back door. Ming China’s status as a state depends, of course, on what we think the word means. Should we rely on norms familiar to us from the history of the state in Europe since the seventeenth century and use these to judge the Chinese evidence? Or do we need to revise our assumptions and work toward a broader understanding of the state that is not dependent on the specificities of European experience? If the Ming is a state, is it so because it reminds us of European states, or because it persuades us to broaden our sense of what a state can be? Beyond the question of definition lies the more difficult question of comparative historical method.

Conclusions

173

If, as the philosopher of history Paul Ricoeur has suggested, historians proceed “from the classificatory term toward the explanation of differences,”2 should the anchor of our analysis be the differences or the classificatory term (“state”)? That is, do similarities between the operations of early-modern European states and the organization of politics within Ming administration enhance the analytical value of placing them in the same category, or are the differences so striking that we get better analytical leverage by putting them in separate categories? The European tradition of political and social thought has generally regarded difference as the thing to look for. The habit of differentiating polities is rooted in the Greek imagination of the world as divided between Greeks and Turks, which their rhetoric spoke of as a difference between Europeans and Asians. Despotism was the common language of that difference. Aristotle insisted in The Politics, for instance, that “Asians are more servile than Europeans, hence they endure despotic rule without protest”;3 and Hippocrates observed in the same vein that Asians “are not their own masters and independent, but are ruled by despots.” The word despotism lost currency subsequently though it was briefly revived in the fourteenth century when medieval philosophers rediscovered Aristotle.4 During the sixteenth and seventeenth centuries, “despot” was used almost exclusively as a technical term for the heir apparent or immediate subordinate of the Ottoman emperor. Only in the seventeenth century did the term began to acquire conceptual substance. Thomas Hobbes, for instance, posited a distinction between “despotical dominion” (conquest by force of arms) and “sovereign dominion” (the rule of a monarch).5 John Locke similarly distinguished “political power” and “despotical power”: that is, “an Absolute, Arbitrary Power one Man has over another, to take away his Life, whenever he pleases.”6 The older notion persisted, such that Samuel Johnson in his Dictionary (1755) gives the general meaning of despot as “one that governs with unlimited authority,” and then observes that “this word is not in use, except as applied to some Dacian prince; as the despot of Serbia.” Chambers’ Cyclopaedia (1728), on the other hand, pushed the concept up a more general level by giving “despotism” its own entry and defining it in Lockean fashion as “despotical government.” Only in the second half of the eighteenth century does despotism begin to stand clear of these earlier specificities, though it remains a concept defined by its opposite. When Thomas Jeffreys describes the condition of African slaves in his 1760 book on the French colonies in the Americas, he contrasts their miserable lives under the “despotic sway” of illiterate African rulers with the “civil” condition they enjoy under their American masters.7 It is an unfortunate comparison from our point of view, but it satisfied his sense of the superiority of European civil institutions and his condescension for a primitive Africa. That Jeffreys should call African tribal leaders despots indicates that the concept of despotism was not yet fully formed in mid-eighteenth-century England. But it was in mid-century France. The pivotal figure in the word’s transformation was Montesquieu, who famously divided the governments of the world into representative, monarchical, and despotic. For Montesquieu, despotism was a political concept rather than just a deplorable condition, and it could be given historical exemplification, since Russia, Turkey, Persia, India, China—in fact, every state outside Europe—fell into this category.8 Place these examples in a universal history that can sustain a teleology of political modernization favourable to Europe, as Hegel did (“China, Persia, Turkey—in fact Asia

The chinese state in ming society

174

generally, is the scene of despotism, and, in a bad sense, tyranny”), and the leap to full conceptual legitimacy for the concept (“Substantial forms constitute the gorgeous edifices of Oriental Empires in which we find all rational ordinances and arrangements, but in such a way, that individuals remain as mere accidents”)9 was but a short step. Beyond this step lay Karl Marx, Max Weber, Karl Wittfogel, and many other more recent, and not altogether innocent, theoreticians of Asian difference. This brief review of the history of the term “despotism” is intended only to suggest that the preoccupation with despotism in the field of Ming studies, however intellectually coherent it can be made to seem, is a product of a particular strand in the formation of Western identity. It does not resolve the problem of whether the Ming was a state or not, but it must alert us that buried assumptions about the state that should be left outside the analytical work we are trying to do can exert powerful influence on that analysis. To continue looking for China within Montesquieu’s residual category of the non-European despotic, or at Hegel’s dawn of history where “sensuous barbarism can only be restrained by despotic power,”10 is unlikely to produce an analysis of the Ming that does anything more than reproduce that intellectual tradition. In addition to this sort of discursive analysis, there is another way in to the question of the status of the Ming as a state, and that is to re-examine the history of the early modern state that underpins the transformations of language. This is not a problem that Ming historians have to face alone or for the first time. For several decades, historians of medieval Europe have worried about whether they were finding states when they contemplated medieval polities. Two decades ago, however, the French historian Robert Fossier argued emphatically against this doubt: “The idea of the State belongs to all ages,” he insisted.11 To deny the existence of the state is simply a trick of definition, rejecting the institutions and authority relations that were practised at another time and place because they do not conform to modern expectations about what a state is. The reasons some have given for ruling out Ming China, however, are interestingly unlike the reasons invoked to declare medieval Europe as lacking the state. Medieval Europe fell short of expectations; the Ming exceeded them, showing a precocity of surveillance institutions and disciplinary practices that Michel Foucault and others have taken as marking the onset of European modernity. Maintaining a narrowly presentist conception of the state as the body politic of civil democracy serves to emphasize the particular outcome of state-making processes in the West; allowing for a broader definition of the state as any centralized political organization governing through authority relations opens the range of difference that the concept can accommodate—and in so doing, widens the base for a comparative history of the processes through which authority is structured and the attributes that these processes generate, sometimes similar, sometimes dissimilar. “State” can mean different things in different contexts (when speaking of China, for example, scale is an important factor distinguishing it from other contexts), but these differences enlarge our sense of the possible norms of state functioning and the repertoire of ways in which complex political organizations can order (or fail to order) people’s lives.12 Rather than launch a comparative study by testing the Ming state for the attributes of the modern European nation-state, let us shift the comparison to the late-medieval period. For this exercise, I shall rely on David Levine’s recent synthetic analysis of the early history of European modernity, which he locates in the period from 1000 to 1500. Levine

Conclusions

175

identifies five master processes of social development in Europe during these five centuries. Conveniently for the comparison I want to make, the first on his list is state formation. This priority comes as a surprise, given that state-building is conventionally regarded as one of the things that the late-medieval polity was notably poor at promoting. Levine challenges this assumption and brings the work of Robert Fossier to his support. Fossier acknowledges that states during the recession of the long fourteenth century presented sad spectacles: “pontiffs who are honourable but contested, becoming dubious and hated; emperors swollen with projects, whose names we cannot recall; Western monarchies in full disarray; old men, minors, madmen.”13 But he also recognizes that the state was nonetheless in formation, despite this appearance of disarray. No longer was it possible to fall back on the idea that society was made up of three orders or estates, among whom there existed a stable condominium of power within which the state was a second-order actor. This “deranged notion,” as Fossier disdainfully calls it, was “already out of date in the thirteenth century,” though “this went on being said until 1789.” Instead, he argues, we should go back to the thirteenth century and attend to the novel developments that were enabling the state to emerge as an authority capable of insinuating its presence into social networks hitherto able to resist it. Levine extends Fossier’s admonition back before the Black Death, which many earlymodern historians have been unwilling to do. In this period, Levine argues, the contractual reciprocities of feudalism were being replaced by mechanisms favourable to absolutist rule, which he characterizes as “a centralization of power and authority in the hands of the ruler and his agents.”14 Centralization was slowed by the Black Death, yet the process was under way before that and afterward was strengthened by the desperation of those who lived on feudal rents and needed support against the enlarged independence of cultivators. What emerged between 1000 and 1500 was “a new kind of society”15 that was also implicated in a new kind of polity: one that defined sovereignty in territorial terms, used laws to regulate society, organized fiscal levies on the basis of quantification and surveys, and trained young men in what Levine helpfully terms “administrative literacy” and appointed them to govern. Pivotal to the success of these initiatives was an ability to impose a monopoly on the means of violence, a monopoly that often eluded medieval regimes. What made these changes possible were the growth of urban scale and autonomy, the power of money to organize the economy, the “magnification” of arable land (through such means as polder-building), and improvements in literacy and documentation amounting to “new disciplinary technologies,” which together “created regulatory grids within which the lives of all people would henceforth be framed—by money, the state, and the supervision of daily life.”16 In all this, Europeans were working uncertainly toward something they had never experienced before, a kind of state formation previously imagined at best, but never achieved. What was novel to Europeans, however, is, from the perspective of Ming China, terribly familiar. Once we adjust for the difference of cultural context, many of the attributes of an emerging modernity Levine traces in late-medieval Europe were achieved norms in Ming China, pieces of a mosaic of state authority that Chinese had already assembled. The absolutist state that Europeans later developed was already, on different terms, a reality in China, and on a continental scale beyond anything that Europe could match. These, then, are the comparative conclusions Levine’s work suggests to me: that the Chinese state developed in the Ming in ways that anticipated features that late-

The chinese state in ming society

176

medieval Europeans were developing fitfully and unequally at the other end of the continent; that China and Europe produced absolutist state formations whose attributes encourage a comparative perspective even when the processes creating them differed; but that the precocious development of the Chinese state occurred in partnership with the stubborn capacity of social forces to limit the state’s development toward absolute authority without, however, throwing up corporate structures that could force the state to renegotiate the terms of its authority. With these conclusions in mind, let me now turn back to the studies from which they were made. The four chapters in “Space” and “Fields” provide evidence of the capacity of the Chinese state to bring the realm into the court’s view by rendering both households and land legible to administratively literate officials working for the centre. They also show the power of society to shape and even thwart state policies. Hongwu’s extraordinary program to resurvey the entire realm into units of the subcounty and lijia systems in Chapter 1 enunciated a claim that the state’s sovereignty extended to defining and reordering the communities in which people lived. Not only that, but the claim was actually operationalized. The best evidence of universal implementation is the ubiquitous reporting of these units in the local gazetteers from which that chapter was written: local society had no choice but to absorb the intervention; deflection was not an option. We also saw, however, that implementation was gently bent to take account of community boundaries that were already in place. Further, the local officers assigned to perform the fiscal and security duties imposed on these communities could turn their posts into opportunities to pursue purely local goals that had nothing to do with the state’s operations or interests.17 The state’s goals were in fact more likely to be achieved when local elites found ways to integrate state systems into their own networks, not when the state overrode them. Thus, although Hongwu’s lijia system has the seductive appearance of Hegel’s despotic imaginary, it did not operate with a free hand. Its practice was mediated by capillary effects from below that worked to reshape the lijia to local preferences. The Single Whip reforms and the surveys required to implement them, of the sort that Ye Chunji carried out in Chapter 2 and local officials in polder regions attempted in Chapter 3, were not imposed with the uniformity and unwavering determination that propelled the lijia system into practice. These reforms originated in a different way, growing up out of local state practice rather than descending from the court. As magistrates struggled to reorganize the local interface with the state’s fiscal system under the pressure of a changing economy they came up with solutions that varied from one county to the next, and that worked best when they accessed knowledge of where the productive fields in the county actually lay and who owned them—knowledge that the state needed but that local elites had an interest in hiding. Staving off the state was just what the local elite of Huian county were up to when the “elders” handed in to Magistrate Ye their first set of maps. Ye was canny enough to check whether there was any substance behind their apparent conformity to his wishes and discovered that there was not. Their maps traced what was already in the land registers, not what actually lay on the landscape. A less ambitious magistrate would have let the matter drop and allowed local society to dictate to the state what its terms of access to local resources would be—which is what Ye’s contemporary in Wucheng county did in Chapter 3 when he simply shuffled his townships into sectors without trying to rebuild the tax system from the ground up. To

Conclusions

177

be fair to this magistrate, less was at stake in the Single Whip reforms compared to Hongwu’s Yellow Registers or Fish-Scale Registers; more exactly less coercion was mobilized to force compliance so long as he met his county’s tax quotas. It was an effort to approximate social realities of property, not an exercise in state geometry. Accordingly, almost everywhere, the implementation of the Single Whip was a matter of delicate politics between the reforming magistrate and the local elite, not the imposition of a triumphant state on a war-devastated population that the lijia was. Its more modest goal was to provide state administrators with new tax schedules based on more equitable assessments, not to redesign the matrix of communities in which people should live. Officials’ efforts to alter field systems by promoting polder agriculture on the Yangzi Delta in Chapter 3 and on the North China Plan in Chapter 4 demonstrate again the Chinese state’s capacity to alter local society, but these are not uniquely stories of state capacity. Other than at moments of extraordinary intervention, the ability to redesign local agricultural communities was hedged by local interests. State-ordered or stateorganized projects, whether to mobilize the labour needed for infrastructural maintenance in the south or to facilitate agricultural intensification in the north, had impact on actual practices only to the extent that they engaged with local elites and adapted to existing village communities. Step back from these projects and the local disappears behind a façade of state agency, a façade which the documents state officials wrote to justify their failures or celebrate their successes—and on which we in turn depend to reconstruct their work—encourage. Even Ding Yuanjian could gaze at the stone marker that a local elder had set up on a polder embankment in 1472 and be charmed into a nostalgia for a state system that had been devised in fact to correct the failure of the lijia system to ensure adequate labour to repair the polders, and that had long been inoperative. His example should caution us against slipping into a like nostalgia of our own and seeing more of the Chinese state’s capacity to order the countryside than was really there, or reading the ambitions that officials invested in state programs as having the substance they claimed. In both chapters, the resistance of the local communities was fierce. The claims of the state were almost always greater than what its agents were actually able to do; but those claims were made nonetheless, and respected when and where they were enforced. The two studies under “Books” suggest the same dialogic process between a state capable of imposing surveillance on readers and prohibiting books, on the one hand, and a society organizationally and economically capable of resisting a centralization of reading, on the other. Officials built libraries and stocked them with books approved and printed by the court, as we saw in Chapter 5, and they searched out and sent to Beijing books they thought the court would find offensive, as in Chapter 6. Depositing some books in libraries and banning others can be seen as components of a unified system of communications through which the state moved information and distributed resources; but these were also episodic interventions rather than customary or systematic acts to control knowledge. More saliently, as these studies show, social and commercial networks were equally effective in circulating non-mandated books into school libraries and out of sight of state officials. These networks created an environment for intellectual production that deflected the state’s claims to control the reading matter of its literate subjects, and that most officials understood well enough to live with. Hongwu’s imposition of draconian regulations on Buddhist institutions and clerics described in Chapter 7 offers more persuasive evidence of the state’s power to intervene

The chinese state in ming society

178

at will in Ming society. I have paired it with the much more ambiguous account of attempts by local elites and state representatives to diminish the discursive presence of monasteries in Chapter 8 in order to highlight the exceptionality of the Hongwu regulatory regime and the unrelenting capillary capacity of social networks at the local level to override whatever desiderata the state and its elite sought to impose on popular religious activities. In Chapter 8, the fault line was not between potent state and prone society, but between, on the one hand, the magistrate and a segment of the local gentry concerned to protect their privileges, and on the other, an equally significant segment of the local gentry and the common people who pursued their religious devotions, sometimes with guile, but usually in blithe indifference to the condescensions and anxieties of their superiors. As any state might, the Chinese state was concerned by religion’s potential to coordinate dispersed networks and bring into being a competing source of public authority, and arrogated to itself the authority to sanction and proscribe religious activities and communities as it saw fit. The Buddhist establishment in Ming China did not assert itself as a counter authority to the monarch, nor did it claim that right in principle, as the papal establishment in Europe did. The Chinese state legitimately monopolized all authority, secular and spiritual, yet that formal monopoly did not mean that state authority in China was unencumbered or unchallenged. So long as legal regulations were more or less respected and no political campaign was under way demanding root-and-branch conformity, religious entities could be selectively indifferent to state preferences in matters of religious practices or ideologies. They had to appear orthodox, but might otherwise deflect state controls, though through forms other than unauthorized public gatherings and fiscal evasion, which otherwise could attract the attention and intervention of the state. If the tension managed not to erupt into conflict for most people at most times, it was because reasonably stable social structures kept the dialogue with the state at a rough equilibrium. In large part, this was due to the work of the gentry themselves. Except when pressed to do otherwise, the gentry accommodated the realm of popular religion without having to surrender their authority at higher levels. And by investing their wealth in the many vertical and horizontal strategies of control and competition they developed through the sixteenth to eighteenth centuries, they turned economic opportunity into social gain without threatening the survival or legitimacy of the state. It was a delicate balance, but one to which the gentry were well used. This bargain between political and social power, paid for by trading wealth and autonomy for durable state tokens of status, held until the latter decades of the nineteenth century.18 Given the scale and range of the Chinese state’s capacity to rule, it seems only natural that the eighteenth-century Physiocrats would look to China for models of how to enhance the absolutist state against aristocratic and church privileges, which they did in matters ranging from taxation to the recruitment of officials by examination to court ritual. (François Quesnay was even able to arrange that Louis XV perform the spring ploughing ceremony that the Chinese emperor was required to conduct on Chinese New Year’s.)19 By that point, of course, looking to absolutism as the political ideal of the state was a backward glance in Europe; other processes were in play that would undercut the absolutist state with democratic institutions and force the monarch to share power with the new capitalist elites, if not hand it over altogether. But this is to come too far forward in time. Place the point of comparison at 1500 and states in Europe seem but weak

Conclusions

179

shadows of the Hongzhi emperor, who in that year, the thirteenth of his reign, commanded capacities and upheld commitments beyond anything a European monarch could match. Had Hongzhi wanted to address an order to every one of the 689,295 tithing heads in the country, according to one count, in theory he could have done so—and even have expected at least a pro forma response. Yet a Ming emperor was not the absolute monarch that Quesnay or Louis XV liked to imagine. He could do much, so long as communication systems ensured orders went out, security systems enabled responses to come back, and educational systems produced people capable of administering on the state’s behalf; but even he faced constitutional limits. Chief Grand Secretary Wang Xijue might flatter the Wanli emperor, as he did in 1590, by assuring him that “in every one of the affairs of the state, Your Highness may decide on his own,” but this was simply a rhetorical move to remind Wanli that this power to decide as he liked was precisely what he did not have in the matter closest to his heart: “only in the matter of the ritual authority for establishing the heir have previous dynasties (dai) all enjoined the Son of Heaven to cede authority.” Wanli did not want to follow imperial precedent for naming the heir apparent, and his chief grand secretary was delicately pointing out to him that he had no choice in so basic a constitutional matter. The emperor then did what anyone in a tight political spot does: he delayed. And so Wang Xijue found himself using almost exactly the same language two and a half years later. “Of the affairs of the state today, none is greater than determining the heir apparent,” he reminded his emperor, before going on to declare that “of the excellent points of Your Highness, none is more excellent than exercising the power to make all decisions on Your own.”20 Once again, Wanli was told he had the power to make any decision he liked, so long as it conformed to constitutional precedent. If Wang reversed the order of his observations this time, it was to make clear to the emperor that he could no longer put off the decision constitutionally required of him. However much the Chinese constitution constrained its emperors, the Chinese state was far more constrained by the commercial relations and social networks that it encountered in its interactions with Ming society. The material in this volume thus points, interestingly, in opposing directions: to the Ming as a developed state with considerable capacities to govern and intervene, and yet to the Ming as a state vulnerable to the capillary influences of social power, in some competition with elites and local communities over resources and excluded from many of the spaces that social practices constructed, whether through the deviance that religious devotion allowed or the conformity that kinship rules imposed. The social embeddedness of the state does not mean that it failed to exact submission to its policies, extract the resources it needed to reproduce itself, or sustain a legitimacy that elites and others continued to fmd plausible. This was a system reproduced by both integrated command and modulated resistance. Rather than see this tension as evidence of a compromised state, even as something that is not a state at all, I see it instead as a sign of this monarchy’s resilience. This state could allow the local outcomes of its policies to stray from their letter without teetering toward collapse, in part because it commanded coercive and persuasive means to ensure that the elite accepted their responsibility to buttress local social order and maintain a high measure of vertical integration with the state. The gentry could not imagine their

The chinese state in ming society

180

relationship to the state in any other way. State and social desires thus competed, but also mutually adjusted, to produce a resilient system of elite power, without the differences between them constantly breaking out into unmanageable conflict—as so often wrecked the early absolutizing states of Europe.

Notes Abbreviations used in the notes JFZ

Ge Yinliang, ed., Jinling fancha zhi (Gazetteer of the Buddhist monasteries of Nanjing, 1607)

HAZS Ye Chunji, Huian zhengshu (Administrative handbook of Huian county, 1573) QWD Qingdai wenziyu dang (Documents concerning the Qing literary inquisition, 1934) TJL

Gu Yanwu, Tianxia junguo libing shu (The strengths and weaknesses of the various regions of the realm, 1662)

Introduction: a grave in Nanchang 1 Jing’an xianzhi (1565), 1.18a. 2 Xiaozong shilu, 155.4b–5a. 3 Translated as “Placard of the People’s Instructions,” in Edward Farmer, Zhu Yuanzhang and Early Ming Legislation, p. 203. 4 Nanchang fuzhi (1588), 3.28a–29a, 30a. References to court battles over property in Jiangxi are legion. The gazetteer of Ji’an prefecture records the case of a fifteenth-century man who sold lineage property in order to buy off aggressive litigants; unusually, this story ends happily when the buyer generously tears up the deed of sale and returns the land to its coerced owner; Ji’an fuzhi (1776), 51.38a. 5 Jing’an xianzhi (1565), 1.18a. 6 Zhao Bingzhong, Jiangxiyudi tushuo, 2b. 7 Eunuch grand defenders (zhenshou) are mentioned in passing in Shih-shan Henry Tsai, The Eunuchs in the MingDynasty, pp. 59–63. 8 See Ming shi, p. 5351, regarding an impeachment memorial against Dong, for which the Hongzhi emperor punished the memorialist; and p. 4848, regarding an unsuccessful attempt to have Dong punished in the first year of Hongzhi’s successor’s reign. 9 Zhongyang tushuguan, Mingren zhuanji ziliao suoyin, p. 944; Jiao Hong, Guochao xianzheng lu, 90.9a. The incident is undated; I assume it was separate from the grave land case and occurred later, but have been unable to verify either assumption. 10 The emperor fetish is given an amusingly personal twist in Craig Clunas, “Oriental Antiquities/Far Eastern Art.” 11 “All that we call subjectivity is concentrated in the supreme head of the State”; G.W. F.Hegel, The Philosophy of History, p. 113. 12 John Meskill, Ch’oe Pu’s Diary: A Record of Dnfting Across the Sea, p. 114. 13 Xiaozpng shilu, 145.9b. 14 Xiaozong shilu, 144.6a. 15 Xiaozong skihi, 156.4b, 158.2b. 16 On the non-liability of uncultivated land for tax, see Lei Menglin, Dulü suoyan, p. 142.

Notes

182

17 Gregory Blue, “China and Western Social Thought in the Modern Period,” pp. 87–8. The first postwar generation of Ming scholars in North America bore the burden of having to respond to the charge of despotism that Karl Wittfogel levelled at this dynasty in particular; see, e.g., W.Theodore de Bary, “Chinese Despotism and the Confucian Ideal,” and Frederick Mote, “The Growth of Chinese Despotism,” notably pp. 18–35. 18 Zhu Yuanzhang, YuzJii dagao, Preface, 1a. 19 For example, Charles Hucker, The Censorial System of the Ming Dynasty; idem, Chinese Government in Ming Times; Edward Farmer, Early Ming Government; Ray Huang, Taxation and Governmental Finance in Sixteenth-Century Ming China; idem., 1587, A Year of No Significance. 20 Roberto Unger, Social Theory: Its Situation and its Task, pp. 151–2. 21 William Blackstone, Commentaries on the Laws of England, p. 132. I am grateful to Juliana Saussy for bringing this text to my attention. 22 Karl Wittfogel, Oriental Despotism, p. 50, citing Paul Milukov’s Skizzen Russischer Kulturgeschichte (Sketches of Russian cultural history). Milukov was a founder of the Kadet Party, which represented a liberal alternative in Russian politics after 1905. He served as Minister of Foreign Affairs in the first Provisional Government in 1917, then emigrated and taught Russian history at the University of Chicago. From his teacher Kliuchevsky, Milukov absorbed the idea that the strengthening of the Russian state in the pre-revolutionary period weakened Russian society. I am grateful to my colleague Lynne Viola for this information. 23 Mann faults Wittfogel for building his theory of state-society relations on “an underlying model of a unitary society” dominated by the state. A step in the right direction need not lead to conclusions that “Wittfogel’s characterization of China as an ‘Oriental Despotism’ is accurate,” or that Europe was the site of a “dynamism” and “miracle” missing in China (The Sources of Social Power, vol. 1, pp. 16, 95, 98, 500–1). Taking shelter in the prefabricated structures on the European intellectual landscape pushes us back down the slippery slope of reducing China to a crippled variant of European categories, rather than lifting us out of these Eurocentric conceptions and onto a more level analytical field. For a better method for the comparative history of European and Chinese states, see Bin Wong, China Transformed, pp. 71–2, 93–104. 24 Émile Durkheim, review of Maurice Courant, “Les associations en Chine,” p. 355. 25 Philip Huang, The Peasant Family and Rural Development in the Yangzi Delta, p. 40. 26 On the need to factor violence back into our histories of the Ming, see David Robinson, Bandits, Eunuchs, and the Son of Heaven, ch. 1. 27 On the earlier history of gazetteers, see my “Native Identity under Alien Rule: Local Gazetteers of the Yuan Dynasty.” 28 For example, the compiler of Guide zhi (1545) in Henan observes in his “principles of organization” (fanli, 2a) that he thought the previous editions too scrambled and went to other gazetteers in search of better models for organizing his own.

1 The spatial organization of subcounty administration Previously published as “The Spatial Structure of Ming Local Administration,” Late Impenal China 6:1 (June 1985), pp. 1–55; revised.

Notes

183

1 The county replaced the prefecture as the principal site of state operations in the Song as demographic growth pushed administrative functions downward to lower levels; Robert Hartwell, “Demographic, Political, and Social Transformations of China,” p. 396. 2 Huizhou fuzhi (1566), 1.30a. 3 For example, Liang Fangzhong, “The Tax-Captain System in the Ming Dynasty”; Tsurumi Naohiro, “Rural Control in the Ming,” p. 273, n.1. 4 When overseeing the setting up of a baojia system in a Fujian county, Zhen Dexiu (1178– 1235) was hesitant to impose community boundaries; the essay in which he mentions this is translated in Patricia Ebrey, Chinese Civilization: A Sourcebook, p. 186. 5 Kuhn, “Local and Self-Government under the Republic,” p. 259. 6 Xu Yikui, Shifenggao, 7.16b. 7 Fengrun xianzhi (1570), 3.15a. 8 The number 887 was reported to the court in January 1371; Taizu shilu, 59.2a. The dynastic history lists 1,159 counties plus 225 subprefectures (zhou), which had the same administrative systems as counties; Ming shi, pp. 881–1221. The figures given in juan 40 (p. 882) are slightly lower, reflecting what are probably sixteenth-century counts. 9 The xiang-li pair appeared in the Warring States period, was universally established by the Latter Han, and could be found in common use in the Song; Sogabe Shizuo, Chūgoku oyobi kodaiNihon ni okeru kyōson keitai no hensen, pp. 24–58, 175–6. 10 Lu You opens his diary of a journey to Sichuan in 1170: “I had planned to leave my xiangli at the beginning of the summer” (Ru Shu ji, 1.1a). For the same usage in the early Ming, see Xu Yikui, Shifeng gao, 7.17a, 8.5b; later in the Ming, HAZS, 2.9b. One finds xiangli in north China, where this pair of terms was not in regular administrative use; for example, Changzi xianzhi (1513), 2.33a. The more common northern expression was lishe, combining the lijia unit li with the ward unit she; for example, Zhucheng xianzhi (1764), 32.5a. 11 Almost all cantons were called xiang. I have encountered only two cases in which the term bao rather than xiang was used: in Changning county, Huguang (Hengzhou fuzhi (1593), 2.23a), and in half the counties in Henan prefecture, Henan (Henan fuzhi (1695), 4.7a–17b); also one case in which jing (“well”) was a local alternative (Gengju, Changshu xian shuili quanshu, 3.zong, 8a). In some contexts, xiang could simply mean a small village; for example, Gaoyang xianzhi (1730), 1.26b; Fuzhou fuzhi (1613), 3.7b–9a. 12 Jiayu xianzhi (1449) 1.19a; Zhangzhou fuzhi (1613), 28.1b–5a. 13 Hartwell (“Demographic, Political, and Social Transformations,” p. 435) cited the prefectural totals in Yuanhe junxian tuzji (Illustrated gazetteer of prefectures and counties for the Yuanhe era, 806–20) for averages of households per xiang close to 500. 14 In Wucheng county, Zhejiang, the number of xiang declined from forty in the Tang to eleven in the Southern Song to nine in the Yuan; Huzhou fuzhi (1649), 2.1a–2a. Forty Tang xiang in Wuxi county, South Zhili, were reduced to twenty-seven, then twenty-two, in the Southern Song; Taibo meili zhi (1897), 1.2a. In Yin county, Ningbo, xiang were increased from eighteen to nineteen in 990, then cut to thirteen in the 1070s, the number it had through the Qing; Yinxian zhi (1788), 2.4b. For examples from Jiangxi and Fujian, see Anren xianzhi (1543), 2.2a; Fuqing xianzhi (1747), 2.17a. 15 In 1430, Fengming canton was taken away from Chongde county (Jiaxing, Zhejiang) to create Tongxiang county; in 1469, Jiulong canton was taken away from Longyan county (Zhangzhou, Fujian) to found Zhangping county the following year; in 1512, Wanchun canton was separated from Yugan county (Raozhou, Jiangxi) to form Wannian county (Ming shi, pp. 1104, 1131, 1058). In the latter two cases, additional segments of territory were added to these cantons when they were elevated to county status. For a new county that overrode old canton boundaries, see Leif Littrup, Subbureaucratic Government in China in Ming Times, p. 46. When on rare occasions a county was dissolved in the wake of population decrease, it could be reduced to canton status; an early-Qing example from Sichuan is noted in Wu Zhenyu, Yangji zhai conglu, p. 320.

Notes

184

16 Ninghua xianzhi (1684), 3.50b. Tax rates could vary according to canton; for example, Mingbo fuzhi (1560), 11.3a; Dinghai xianzhi (1563), 8.3a–7a. 17 For “local militia” (tubing) in Wuxi county, see TJL, 7.52a. For “canton militia” (xiangbing) to fight bandits in Zhangpu county, Fujian, see Yinxian zhi (1788), 17.26a. 18 Concerning a water control project along Lake Guangde in Ningbo: “In 1391, Chen Jin, an elder of the county, proposed a water-control project. An official was dispatched to oversee the project during slack agricultural time. He ordered the families of Canton 7, who would be its beneficiaries, to provide the labour.” The project silted up, and people took advantage of this development by poldering silted areas along the edge of the lake to make paddy fields, blocking the water supply for the people of Canton 7. Forty years later, “when Wang Shihua of Xianshui was living at home during his official career, the people of Canton 7 asked him to take charge of the prqject, since he had fields among theirs. And so a stop was put to this” (Nanchang fuzhi fuzhi (1560), 5.17b). 19 Shouning daizhi (1637; 1983), 13. 20 For example, Yinxian zhi (1788), 2.5a. 21 A few counties in Huguang and Fujian used subcantons in place of the old cantons; for example, Lianzhou fuzhi (1637), 2.21a–24a; Ninghua xianzhi (1869), 1.5b; Xingguo zhouzhi (1554), 2.31a, confirmed in Hai Rui, Hai Rui ji, p. 204. The relationship between the two levels often became confused. In three of the nine counties of Fuzhou prefecture, subcantons were fully replaced by du; in one there were no du but only li and xiang, in another three where there were no du, li were subdivided into tu; in one, each canton had only one li, resulting in the combination of canton and subcanton; in another, xiang and li existed as equal units. Only in Fuqing county were xiang, li, and du operating as distinct levels; Fuzhou fuzhi (1613), 3.6b–31a. 22 For example, Wuxian zhi (1642), 2.1a–4b; Jinhua xianzhi (1598), 1.5a–6b; Gutian xianzhi (1606), 3.4b–7a. 23 Taizhou fuzhi (1722), 3.59b. 24 Xiushui xianzhi (1685), 1.10b; Quzhoufuzhi (1623), 1; Songjiang fuzhi (1512), 9. 25 Zhangping county, Fujian, formed in 1470 by taking a canton away from a neighbouring county, had only five li subcantons above the ward level-later four, after another reshuffling of territory These li accordingly attracted functions needed to coordinate activities above the wards. Two of these four had public halls (gongguan), both of which bore their subcantonal names; zhangping xianzhi (1935), 2.3a. In Fuzhou’s Yongfu county, li were replaced in 1448 by townships in response to the decimation caused there by Deng Maoqi’s rebellion: county records show a drop in registered population between 1381 and 1451 of 76 per cent, forcing a major administrative reorganization; Yongfu xianzhi (1612), 1.2b, 28b. Nonetheless, a monastic gazetteer of 1612 from the region refers to a township in relation to its former subcanton, suggesting that the latter were in popular use; Fangguang yanzhi (1612), 1.1b. The prefectural gazetteer of the following year shows the old li still on the books and makes no mention of townships (Fuzhou fuzhi (1613), 3.26b–28a), which points either to an editor’s lazy reliance on earlier gazetteers or to the tenacity of the old li, even after they had been officially discarded. 26 The use of du as an indicator of place goes back to the Zhou dynasty, according to Gu Yanwu, Rizhi lu jishi, 22.13b. Other terms for township were bao (Gusu zhi (1506), 18.10a; Nanyang fuzhi (1577), 2.9a–11b), fen, and jie (Hangzhou fuzhi (1475), 2.4a–8b). 27 Huian xianzhi (1936), 1.20a. 28 Sogabe, Chūgoku oyobi kodai Nihon, pp. 130, 171–83; McKnight, Village and Bureaucmcy, p. 78. 29 For example, Huizhou fuzhi (1613), 1.51a–55b, Quanzhou fuzhi (1613), 1.9b. 30 Lucheng xianzhi (1625), 2.26a. 31 Numbering was done in such a way that townships within the same canton were consecutively numbered, with numbers increasing the further one went from the county sect.

Notes

185

Very occasionally the numbering starts over with each canton, so that townships had to be identified by canton and number rather than by number alone; for example, Dinghai xianzjii (1563), 7.24a. Xiuning county furnishes an exception: the thirty-three townships established in 1386 all bore names; Xiuning xianzhi (1815), 1.7b. Some Guangzhou counties also appear to have had named townships, according to evidence in Guangxiao sizhi (1935), 5.1b–2a. Sometimes the old li names were attached to the townships when there was only one township per li; for example, Fuqing xianzhi (1747), 2.17b, also tu 1b–2a. 32 In less densely populated areas, the pre-Ming du could be quite small. Fujian’s Chongan county in the Yuan had fifty du and a population of about 500 families, averaging only ten families per du, though some of these “families” were actually large lineages that each controlled the land of several du; Yuan shi, p. 4373. 33 Xiangdu is used in Zhu Yuanzhang’s 1381 lijia regulations; Gao Jie, Houhu zhi, 4.1b. 34 For example, Huizhou fuzhi (1566), 1.40b; Suzhou fuzhi (1692), 27.4a. 35 For example, Fengrun xianzhi (1570), 3.15a. 36 Xu Hongzu, Xu Xiake youji, pp. 61, 112, 134, 161. 37 Kawakatsu, Chūgoku hōken kokka no shihai kōzō, p. 192. 38 See, e.g., Yinxian zhi (1788), 6.12b–14a. 39 Wang Wenlu, Bailing xueshan, 2.8a. 40 For example, Longxi xianzhi (1762), 3.8a. 41 HAZS, 2.19a. 42 Wujiang zhi (1488), 5.23b–24a. 43 Du Bai erhu quanshu (1805), 28a. 44 Yongan xianzhi (1798), 4.1b. 45 Some places in Guangzhou had another unit between the township and the ward, called a bao (with an earth radical underneath), as noted in Katayama, “Shinmatsu Kanton-shō Shūkō deruta no zukōdhyō to sore o meguru shu mondai,” p. 52. This “subtownship” emerged in Shunde county, for instance, when the county was formed in 1452 by carving off four townships from Nanhai. These townships in Shunde served effectively as cantons, requiring a middle unit between them and the wards at the bottom of the administrative system; Shunde xianzhi (1996), p. 87. In Fujian, I have encountered an administrative unit below the ward: Zhangping county in the Ming established she within its tu (zhangping xianzjii (1935), 1.2b). Since Zhangping was created from the canton of another county, the original subcantons were elevated to cantonal roles. Lacking townships, Zhangping’s wards were shifted up a level to take their place, making a subward (she) necessary. 46 A variant, also called bao, can be found in certain parts of central China; for example, Huizhou fuzhi (1502), 1.5b; Song Lian, Song xueshi quanji, 33.67a. 47 Other terms for the ward in north China were zhuang and bao (Yanzhou fuzhi (1596), 2.25a; Neixiang xianzhi (1485), 1.3a). The contrast between tu terminology in the south and she in the north led Hsiao Kung-ch’üan (Rural China, p. 548) to hypothesize that tu and du were antiquated nomenclature that survived in the southern provinces where the central government was less able to impose its administrative order. The difference is better explained not by the romance of an independent south but by the division of China during the Southern Song dynasty, when these terms came into regular use. 48 Haiyan xian tujing (1624), 1.29b. The same semiotic logic appears in other, less common terms. In Jiangnan, a tu might also be called a ce (booklet); Jiashan xianzhi (1894), 10.42a. So too a township might be subdivided into shan (registers); Wuxian zhi (1642), 2.2a. 49 This practice led to tu being glossed as a vulgar variant of the character bi, which is graphically similar to tu except for the removal of the outer wei (enclosure) radical (Mathews 7054) and the adding of the yi (city) radical (Mathews 3037) on the right; for example, Foshan zhongyi xiangzhi (1921), 4.4b–5a. A bi in the Zhou li was an administrative unit of 500 households, with flve bi constituting a county (Hsiao, Rural China, p. 26). Despite the impulse to give tu a decent ancestry in a classical source, no etymological connection can be

Notes

186

shown. Gu Yanwu argued that this logic was specious, that tu came into usage in the Yuan and had nothing to do with bi (Sogabe, Chūgoku oyobi kodai Nihon, p. 230). Even so, Gu’s publisher preferred to use the bi orthography (e.g., Rizhi lu jishi, 22.17a). Many gazetteer editors hedge their bets by using a version of the character without altering the radical, observing that it should be pronounced bi but was vulgarly pronounced tu. 50 A notable exception is Wanping county, within whose jurisdiction lay the western half of Beijing. There the circumstances involved in being the site of the central government led to some unique subcounty arrangements, one of which was the subdivision of one urban and four rural wards (i.e., she and tun) into two or three tu; Shen Bang, Wanshu zaji, pp. 13–14. 51 jizhou zhi (1524), 32a; Luan zhi (1618), 4.4a. 52 Matsumoto Yoshimi, Chūgoku sonraku seido no shiteki kenkyū, p. 94; Sogabe, Chūgoku oyobi kodai Nihon, pp. 197ff. The term she was used in other contexts as well. Zhu Yuanzhang used it to designate local school districts in 1375, to which were assigned fifty households in the old Yuan style (Littrup, Subbureaucratic Government, p. 171). It was also designated as the territory serviced by community granaries (shecang), though there was no attempt to defme its territorial extent. Chenghua-era regulations set only storage definitions, of 300–500 shi of grain (Yu Jideng, Diangu jiwen, 16.16a). She in these cases seems to imply a moral rather than administrative definition: “community” rather than “ward.” 53 Songjiang fuzhi (1512), 9.14b–19b. 54 Guide fuzhi (1753), 9.1a–10b. 55 For examples from Huguang, see Anhua xianzhi (1543), 2.2a; Liuyang xianzhi (1561), 1.2a; Xiangtan xianzhi (1554), 2.15b; Xiangyin xianzhi (1554), 2.8b. 56 Gusu zhi (1506), 18.1 a–25b. 57 When population could not be grouped into full hundreds, regulations permitted officials to create the half-ward (bantu) for grouping from forty to seventy households; for example, Anqiu xianzhi (1589), 8.52b. The half-ward could also be used when, on rare occasions, the boundary of a higher-level unit such as a canton cut across a regular ward, resulting in upper and lower half-wards; for example, Taizhou fuzhi (1722), 3.65b. The half-ward was avoided in the early Ming and by the sixteenth century was not introduced into a county where it was not already being used, although Hai Rui characteristically provides an exception to this general rule: while magistrate of Xingguo county in 1562–4, he created several half-wards (banli) in his effort to reconstitute old wards that had fallen into disuse (Hai Rui ji, p. 206). 58 Wang Wenlu, Bailing xueshan, 2.8a. 59 From his research on Fish-Scale Registers from seventeenth-century Suzhou, Tsurumi Naohiro (“Kōki jūgo-nen jōryō, Soshū-fu Chōshū-ken gyōrin zusatsu no dentō to kōsatsu,” pt. 2, p. 428) determined that the minimum amount of land needed in Jiangnan for sustaining a peasant household was 10 to 20 mu. 60 Tsurumi, “Soshū-fu no gyōrinsatsu ni kansuru ichi kōsatsu,” p. 5; “Kōki jūgo-nen jōryō, Soshū-fu Chōshū-ken gyōrin zusatsu no dentō to kōsatsu,” pt. 1, p. 317. 61 Hai Rui (Hai Rui ji, p. 206) records his attempt to reverse this atrophy of wards in Xingguo county. 62 Linzhang xianzhi (1506), 2.6a–8a. For other examples: the northern half of Township 2 of Wujiang county, South Zhili, in the Jiajing era had twelve wards and fifteen hundreds (Wujiang xianzhi (1561), 10.3a); and Wanli-era Xianju county, Zhejiang, had fewer hundreds than wards in five of six cantons, and more hundreds than wards in the sixth (Xianju xianzhi (1612), 1.2a–b). 63 Huang Liuhong, Fuhui quanshu, 20.8a. 64 Shen Bang, Wanshu zaji, p. 15. 65 For example, Shouning daizhi (1983), 13; Hai Rui, Hai Rui ji, p. 59. 66 Xiaozong shilu, 158.4a. 67 A few county and prefectural seats had urban cantons; for example, Gusu zhi (1506), 18.1b, 6b; Yuhang xianzhi (1808), 3.1b; Yanzhou fuzhi (1613), 3.28a; Guide fuzhi (1753), 9.7a.

Notes

187

Frederick Mote, “The Transformation of Nanking, 1350–1400,” p. 146, speaks of urban townships (fang and xianq) as administratively equivalent to rural cantons, but this is true only of all the urban and suburban townships considered together. 68 For example, Raozhou fuzhi (1511), 1.17b. 69 For the history of the term fang, see Sogabe, Chūgoku oyobi kodai Nihon, pp. 426–7. Variant terms in use in central China were bao (Gutian xianzhi (1606), 3.4a) and guan (Huizhou fuzhi (1566), 1.32a). Very rarely, she was used (Nanchang fuzhi (1588), 5.1a). 70 Jianning fuzhi (1493), 4.7a; Luochuan zhi (1545), 2.38b. The term yu originally meant a fire station, of which each county seat in the Southern Song was supposed to have four (Sogabe, Chūgoku oyobi kodai Nihon, p. 480). It was also used in Shandong (Yanzhou fuzhi (1596), 2.13a–21a; Zhucheng xianzhi (1764), 9.1a; Wenshang xianzhi (1717), 2.13a). Yuan yu boundaries were sometimes altered in the early Ming, but their names preserved; for example, Ruian xianzhi (1555), 1.4b–5a, 3.19b. 71 The term fangyu appears for instance on land purchase contracts from Dehua county, Fujian, dated 1551 and 1555, to indicate that the vendor was an urban resident; Fu Yiling, “Lun Ming-Qing shidai Fujian tudi maimai qiyue zhong de ‘yinzhu’,” p. 2. 72 Very rarely, urban townships were called xianq; for example, Wujin yanghu xianzhi (1886), 1.10b. 73 Sogabe, Chūgoku oyobi kodai Nihon, pp. 451–63, 490. 74 Renhe xianzhi (1687), 1.13b; Qiantang xianzhi (1718), 3.6b. Other terms for suburban townships are guan (Jinhua xianzhi (1598), 1.5b; Lucheng xianzhi (1625), 2.27a; Henan fuzhi (1695), 4.7a) and jie (Jinhua xianzhi (1598), 1.5b; Yinxian zhi (1788), 2.6a). 75 Jiaxing xianzhi (1909), 32:28a. 76 For example, Yuanshi xianzhi (1642), 1.22a. 77 Hongwu proposed that community schools be set up at a rate of one per hundred households in 1375; in Shandong, though, the initial practice was to adopt the old Yuan she of fifty households as a school district (Littrup, Subbureaucratic Government, p. 171). Although forced to cancel this scheme a few years later, Hongwu thought the plan a good one and hoped it might be reestablished; for example, Taizu shilu, 214.2a, where he mentions that a teacher in each li should bring his students to the capital triennially to be tested on their memorization of the Grand Pronouncements. An edict in 1447 reaffirmed the teaching of this text as a priority for officials, though without reference to the lijia system; Yingzong shilu, 37.3a. When Wanli-era officials such as Hai Rui and Lü Kun attempted to revive community schools, they set the school district at several hundred households. On the use and local manipulation of the lijia system as a context for community religious rituals, see Michael Szonyi, Practicing Kinship: Lineage and Descent in Late Imperial China, pp. 175– 90. 78 Xu Yikui in the early Ming speaks of mountain areas as “places not customarily subject to household registration,” i.e., outside the lijia system; Shifeng gao, 11.2a. 79 Mark Lewis, Sanctioned Violence in Early China, pp. 61–4. 80A company (baihusuo) had 100 (later, 112) men; a battalion (qianhusuo), ten companies; a guard (wei), five battalions. 81 The hundred was an administrative subdivision of the shire in tenth-century England. At least in the south, hundreds were organized on the basis of both households and land. Hundreds were further subdivided into tithings of ten men. Hundreds were responsible for catching criminals; members of a tithing guaranteed each other’s conduct. While the origins of these units were military their functions by the late Anglo-Saxon period were to support public security and judicial process (Lyon, A Constitutional and Legal History of Medieval England, pp. 66–8). By the eleventh century, the hundred was primarily a fiscal unit; it continued to exist between county and parish into the early-modern era. 82Gao Jie, Houhu zhi, 4.1b. The Ministry of Rites affirmed in the summer of 1381 that all the rules of lijia registration applied equally to fang and xianq; Taizu shilu, 203.5a–b.

Notes

188

83 McKnight, Village and Bureaucracy, pp. 35, 40; Sogabe, Chūgoku oyobi kodai Nihon, pp. 157, 175–6. 84 Yichuan xianzhi (1753), 8.11b, citing a memorial of the early Qjng. Yichuan is the only county I have found where hundreds were not divided into tithings. 85Dinghai xianzhi (1563), 7.24a. 86 Yichuan xianzhi (1753), 1.19a–21a. 87 Da Ming yitong zhi (Unity gazetteer of the Ming dynasty) and Gu Zuyu, Dushi fangyu jiyao (Notes on geographical matters from reading historical texts), in Liang Fangzhong, Zhongguo lidai hukou, p. 208. 88 Contrary to Hartwell (“Demographic, Political, and Social Transformations,” p. 378), the tendency of hundreds to have 30 per cent more households than the prescribed 110 does not reflect urbanization, since the lijia was also implemented in urban areas. 89 Thus far only eleven fragments of Yellow Registers have been discovered, but these testify to the exactitude with which information could be recorded, even if it was not forwarded upward; see Iwai Shigeki,” ‘Kasei yonjūichi-nen Sekkō Genshū-fu Suianken jūhachi-to kaichi-zu fueki kōsatsu zanbon’ kō.” I am grateful to Fei Si-yen for providing me with a copy of this article. 90 For example, a reduction in hundreds is registered for four counties in Yongping prefecture in the 1501 prefectural gazetteer: Yongping fuzhi (1501), 1.14b–16a. It is unclear whether these reductions, which ranged from 13 to 27 per cent, reflected population decrease or a successful bid to reduce the tax quotas for these counties. 91 TJL, 8.54a. 92 In lijia prototypes operating as early as 1370 before the inauguration of a nationwide system, the unit of 100 households was uniformly called a tu or a bao, but never a li; Su Boheng, Su Pingzhong wenji, 6.23a. Zhu Yuanzhang himself switches back and forth between the two in his famous “Placard for Instructing the People” of 1398; Zhang Lu, Huangming zhishu, vol. 1, p. 470. 93 For example, Gaoshan zhi (1877), 2.29a. 94 Dong Fu, Erlou jilüe, 3.1b–2a; this is also implied by the foregoing land statistics on wards. 95 Hamashima, “Minmatsu Sekkō no Ka-Ko ryōfu ni okeru kinden kin’eki hō,” pp. 172–3; Dennerline, “Fiscal Reform and Local Control,” pp. 105–8. The size of a hundred in Qiantang county, Hangzhou, was fixed in 1671 at 3,000 mu of cultivated land, making one tithing 300 mu; Qiantang xianzhi (1718), 3.6a. 96 Wei Qingyuan, Mingdai huangce zhidu, pp. 194–9. 97 Hai Rui (Hai Rui ji, pp. 59, 66, 113, 150) speaks repeatedly of lijia officers in terms not of the hundreds they lived in but of the wards; similarly, Ye Chunji, whose statistics form the basis for Table 1.5, speaks of lijia officers having jurisdiction by ward rather than by hundred. The same observation is made with regard to the Nanjing areain 1579; Niushou shanzhi (1597), 1.33a–b. 98 A similar post known as a “great household” (dahu) was used in Shandong, Henan, Shaanxi, and Sichuan; see Liang, Mingdai liangzhang zhidu, pp. 54, 58; Littrup, Subbureaucratic Government, p. 94. 99 For example, Wuxi xianzhi (1753), 5.5a. 100 Shanghai xianzhi (1588), juan 4, and Xuanzong shilu, juan 55, both cited in Liang, Mingdai liangzhang zhidu, pp. 40, 62. 101 Su Boheng, Su Pinzhong wenji, 6.33a. 102 A “tax-captain’s bridge” is mentioned in Gaoshan zhi (1877), 2.2b. A “tax-hundred elder” in Changshu county sought county approval in 1535 to repair Zhenwu Temple at his own expense (Jingzhao Gui shi shipu (1913), 3.84a). 103 Liang, Mingdai liangzhang zhidu, p. 85. 104 Lu Cheng (js. 1493), surveillance vice-commissioner of Fujian in the early sixteenth century, established baojia in coastal counties as part of his program for suppressing piracy

Notes

189

(Yinxian zhi (1788), 15.19a). The first use of baojia in Fujian was in the 1440s, part of a campaign to deal with illegal miners (Ming shi, p. 4467). Wang Jiadong (jr. 1606) tightened the baojia in Shahe county, North Zhili, a place “where several provinces meet” and through which bandits regularly travelled; the result was a lessening of bandit activity (Zhucheng xianzhi (1764), 32.3a). The initiative might come not from the magistrate but from a regional military official; for example, Jizhou zhi (1524), 1.52b, also in North Zhili. 105 The Minister of Revenue voiced this complaint in 1560 by way of advocating that local tax officials during the next great compiling year (1562) should make use of townships and wards, not the baojia, to track households; Shizong shilu, 489.2a. 106 References to the power of baojia officers increase in the seventeenth century; for example, Xu Hongzu, Xu Xiake youji, p. 149; Shiba, “Ningpo and its Hinterland,” p. 401. Miki (“Minmatsu no Hokken ni okeru hokōsei,” p. 67) asserts that the system was universal by this time, but empirical proof for this view is inadequate. Its universal implementation was achieved only in the Qing; see Ho, Studies on the Population of China, pp. 36ff. 107 Katayama Tsuyoshi, “Shindai Kanton-shō Shūkō deruta no zukōsei ni tsuite,” pp. 24–8; Szonyi, Practicing Kinship, pp. 73–80. 108 Littrup, Subbureaucratic Government, p. 168. 109 See Kuhn, “Local Self-Government under the Republic,” pp. 93–4. 110 Occasionally the ten-family bao was used in the Ming (Wenzhou fuzhi (1605), 1.28a; Kuribayashi, Rikōsei no kenkyū, pp. 260–7). 111 The term pai was used in the Guangdong area in the sixteenth century to designate the family unit within baojia organization (Huang Zuo, Taiquan xiangli, 6.1a). Pai came to replace jia for the ten-family unit in some areas in the Qing (Ch’ü, Local Government in China under the Ch’ing, p. 150). 112 Wenjuntian, Zhongguo baojia zhidu, p. 193; Xu Fuyan, Jinghe tangji, gongyi, 1.8b. 113 Zhang Huang, Tushu bian, 92.106; Suzhou fuzhi (1883), 147.32a. 114 Huang Chengxuan, Ming’ou tang ji, 29.6b. 115 Jiaxing fuzhi (1681), 18.18a. 116 The dang first appears in the Wei dynasty as a hundred-household grouping (Sogabe, Chūgoku oyobi kodai Nihon, p. 76). Supervisors are mentioned in Li Le, Xu jianwen zaji, 11.43b, and Xu Kaixi, Hanshan kao, 4.51a; captains, in Jiaxing xianzhi (1685), 4.2b. 117 Zhang Huang, Tushu bian, 92.38b. 118 Yinxian zhi (1788), 11.24b; Kuhn, “Local Self-Government,” p. 41. 119 TJL, 8.58a, though here the term li is substituted for bao. Pai was used idiosyncratically, e.g., in Yihuang county, Jiangxi (Xu Hongzu, Xu Xiake youji, p. 137). 120 TJZ, 7.52b. 121 Sun Yi, Dongting yuren ji, 53.7a. 122 Huangshan zhi dingben (1679), 3.81b. 123 Huang Changxuan, Ming’ou tang ji, 29.8a. 124 Gu Yanwu, Rizhi lu jishi, 8.13a. Gu’s equation of watches and townships is puzzling: such equivalence would obtain only in one-ward townships. 125 For example, Dangyang xianzhi (1866), 10.14b. 126 Sun Yi, Dongting yuren ji, 53.7a–b. 127 Kuribayashi Nobuo, Rikōsei no kenkyū, p. 76. 128 Wang Guoping and Tang Lixing, Ming-Qing yilai Suzhou shehuishi beike ji, p. 674. On Hongwu’s plan for community schools, see above, note 77. 129 Timothy Cheek, “Contracts and Ideological Control in Village Administration.” 130 Joanna Handlin, Action in Late Ming Thought, p. 49. 131 Miki, “Minmatsu no Hokken ni okeru hokōsei,” p. 89. 132Huang Zuo, Taiquan xiangli, 6.3b–4a. 133 Littrup, Subbureaucratic Government, pp. 168–70. 134 Zhucheng xianzhi (1764), 2.26a.

Notes

190

135 Xiuning xianzhi (1693), 2.18a-20a. This source also notes that covenant ceremonies were in the hands of the local gentry, more particularly the heads of lineages. 136For example, Xu Fuyuan, Jinghe tangji, gongyi, 1.7a–b; Cai Moude, “Yuebao fa” (A method for organizing covenants and baojia), in Chen Hongmou, Xueshi yigui, 2.2b. 137 Huang Changxuan, Ming’ou tang ji, p. 29. 138 Skinner, The City in Late Impenal China, p. 367, and Schoppa, Chinese Elites and Political Change, p. 82. 139 Skinner, The City in Late Imperial China, p. 21. 140 He Liangjun, Siyou zhai congshuo zhaichao, juan 3, quoted in Liang, Mingdai liangzhang zhidu, p. 128. 141 Hsiao, Rural China, p. 254. 142 Liang Fangzhong has noted that tithings in sixteenth-century Fujian were simply the preexisting villages, and that the imposition of formal units brought no change in the spatial shape of local society; “Lun Mingdai lijiafa he junyaofa de guanxi.” 143 Philip Kuhn (“Local Self-Government,” pp. vi–vii) has contrasted the “tinker-peddler” rather than the “nested-concentric” mode of coordination, by which sectarian leaders using routes and structures not part of the commercial-administrative hierarchy were able to achieve loose interregional coordination without being noticed by the state. This kind of spatial networking, though effective in linking voluntary groups, was not sufficiendy powerful to mobilize “long-term or large-scale coordination or defense.”

2 The gazetteer cartography of Ye Chunji Previously published as “Mapping Knowledge in the Sixteenth Century: The Gazetteer Cartography of Ye Chunji,” East Asian Library Journal 7:2 (winter 1994), pp. 5–32; revised. I am grateful to Bangbo Hu for pointing out errors in this earlier version. 1 Huang Liu-hung, A Complete Book Concerning Happiness and Benevolence, p. 129. 2 Ye Chunji was born in 1532 in Guishan county, some 300 kilometres west of Huian, and won his juren degree at the age of 20. He continued to study for the jinshi degree, but without success. For submitting a 30,000-word memorial on contemporary problems to the Longqing emperor on his ascension to the throne, Ye was awarded the post of county school instructor in Minqing county, Fujian, in 1568. He was promoted to the magistracy of Huian in 1570 or 1571, held the post for three years, and then was forced to retire from public office. He returned to official life in 1591, first as a subprefectural magistrate, then as a vice-prefect, and then as a bureau director in the Ministry of Revenue. He died in office, perhaps a few years short of the century’s end. See HAZS, Guo’s preface, 2a; Ye’s preface, 1a; Guishan xianzhi (1783), 10.13a, 14.8b–9a; Quanzhou fuzhi (1763), 27.61a. Ye does not appear in either the Ming shi or Goodrich and Fang’s Dictionary of Ming Biography. 3 Timothy Brook, “Native Identity under Alien Rule,” p. 241. 4 HAZS, 1.13a, 2.1b. 5 Xianzong shilu, 216. 5b. 6 TJL, 8.77a–b, citing the 1575 Hezhou zhi. 7 HAZS, Ye’s preface, 1b. 8 HAZS, 2.1a. 9 I was introduced to and given a copy of the text by Yamane Yukio, whose reading seminar at the Tōyō Bunko I had the privilege of attending in 1980.

Notes

191

10 Ogawa Takuchi, Shina rekishi chiri kenkyū, pp. 26, 41. 11 Wei Qingyuan, Mingdai huangce zhidu, p. 74. 12 Brook, “Geografia e cartografia,” p. 495. 13 See Stanley Huang’s biography of Luo in Goodrich and Fang, The Dictionary of Ming Biography, vol. 1, pp. 980–4. 14 Ogawa, Shina rekishi chiri kenkyū, pp. 48, 51. Pei’s account of the grid system is translated in Joseph Needham, Science and Civilisation in China, vol. 3, pp. 539–56; Yuji tu is reproduced on p. 549. 15Ogawa, Shina rekishi chiri kenkyū, pp. 59–62; Wolfgang Franke, An Introduction to the Sources of Ming History, 8.1.3. 16 Guishan xianzhi (1783), 14.8b. 17 Luo wrote a colophon to Jiubian tulun; see his Nianan wenji, 10.25b. Regarding Jiubian tulun, see Franke, An Introduction to the Sources of Ming History, 7.3.7. Portions of a probable late-Ming copy of this atlas are reproduced in Gao Wanru, Zhongguo gudai dituji: Mingdai, pl. 17–24. 18 HAZS, Guo’s preface, 6a. Guo was a friend of Luo Hongxian, and conceivably the person who introduced him to Ye. Guo argues for the feasibility of Ye’s approach by observing that Luo himself was capable of such precision, as he realized when he saw the genealogy Luo compiled for his family, which contained detailed diagrams for the performance of family rites. 19 HAZS, Ye’s preface, 2a. 20 HAZS, 1.12b; Ye names five of his assistants (see also 4.5b). 21 HAZS, 1.1a–b. 22 By Ye’s time, the original thirty-four townships of Huian had been reduced by amalgamation to twenty-seven, although the old township numbering system continued to be used. The four maps of the amalgamated townships (for Townships 3–4, 11–13, 14–17, and 19–20) show the original township boundaries (HAZS, 4.28b, 6.2b, 10b, 21b). 23 HAZS, 1.17b. 24 HAZS, 1.2a–b. 25 It is difficult to verify his population data. He observes that the county population records were completely out of touch with the actual households, and that household status and composition had been much manipulated over the years (HAZS, 2.1b–2a). Compared to the 1530 county gazetteer, Ye’s household figures are lower by about 8 per cent, and his population totals lower by over 25 per cent, decreases that do not reflect real demographic trends. See HAZS, 4.3b–4a; Huian xianzhi (1530), 6.2b–3a. 26 HAZS, 4.5b. 27 Huian xianzhi (1530), 3.3aff. On the other hand, mountains, when they are located, are located by township. 28 HAZS, 1.15a; regarding discrepancies among the gazetteers with regard to the location of mountains, rivers, and township boundaries, see 1.13a–b, 15a–b, 16b. 29For an urban map of comparable detail and slightly great abstraction in another late-Ming gazetteer, see Haiyan xian tujing (1624), 1.22a–b. 30 Huian xianzhi (1530), 3.11a. 31 HAZS, 1.16b. 32 Luo Hongxian was an advocate of organizing these types of institution to reconstruct local solidarity; for his views on the rural covenant system, see his Nianan wenji, 6.6a–7b. 33 Hai Rui ji, pp. 36–201. 34 Quanzhoufuzhi (1763), 31.35b. 35 Shunde xianzhi (1585), Yongan xianzhi (1586), and zhaoqingfuzhi (1588), as listed in his biography in Guishan xianzhi (1783), 14.9a, where the last is referred to as Duanzhou zhi, Duanzhou being the literary (pre-Ming) designation for Zhaoqing. All three gazetteers are extant; Zhongguo difangzhi lianhe mulu, pp. 685, 695, 709. Luo Hongxian was also a

Notes

192

gazetteer enthusiast; two gazetteer prefaces he wrote are collected in his Nianan wenji, 11.23a, 33b. 36 Ye’s one post-gazetteer publication was a comprehensive record of commercial levies on goods coming into Beijing, entitled Chongwen jueshu (Customs records of Chongwen Gate). It is mentioned in Guishan xianzhi (1783), 14.9a, but appears to be no longer extant. 37 The maps in the Yongan gazetteer are examined in Bangbo Hu, “Maps in the Gazetteer of Tung-an County” Having published his essay prior to seeing a copy of Huian zhengshu, Hu makes some arguments that are now superseded or are unnecessary, such as his elaborate attempt to prove Luo’s influence on Ye (pp. 96–7). 38 zengcheng xianzhi (1820), 1.1b–3a; see also Dianbai xianzhi (1825), 1.1b, 5b; Xining xianzhi (1830), shou. 1b, 8b–27a; Jiangyin xianzhi (1840), tu. 9b–49a; Xinhui xianzhi (1841), 1.1b, 9b–37a. Scales vary from 1 to 13 li per square; the latter three include township as well as county maps. 39 Fujian tongzhi (1871); Haiyan xianzhi (1876), shou. 5b–11b; Xiaofeng xianzhi (1877), 1.2b– 8a; Jiangning fuzhi (1880), tu; Shangyu xianzhi (1891). 40 Panyu xianzhi (1871), 2.1b–12b; zhaoqingfu fuzhi (1876), shou.1b–31a. 41 Tongxiang xianzhi (1887), 1.33–41.

3 Taxing polders on the Yangzi Delta 1 Ding Yuanjian, Xishan riji (Western Hills diary), juan 1, quoted in Hamashima Atsutoshi, Mindai Kōnan nōson shakai no kenkyū, p. 23. 2 An admiring biography of Ding Yuanjian appears in the Ming shi, pp. 6156–7. 3 For references to Ding’s interventions in regional fiscal issues, see Hamashima, Mindai Kōnan nōson shakai no kenkyū pp. 322, 325, 492; see also p. 605 regarding his role in putting down an uprising in 1624 that began with the assassination of the Changxing county magistrate. 4 Liang Fangzhong, Zhongguo lidai hukou, tiandi, tianfu tongji, p. 234; Changxing had 28,270 households. 5 There are two terms in Chinese for polder, wei and yu. The latter was the more common term in Jiangnan. Its modern pronunciation has been officially standardized to wei. Tang was also used variously for the embankments around a polder, the water channel outside it, or a water reservoir adjacent to non-poldered fields. 6 Louxian zhi (1738), 15.5b. 7 Changshu xianzhi (1537), 4.18b: “Since a ban on poldering fields” was imposed in the fifteenth century, “the watercourses are broad and without impediment.” 8 Peter Perdue, Exhausting the Earth, ch. 7. 9 For studies of polders in the Song, see Mira Mihelich, “Polders and Politics of Land Reclamation in Southeast China during the Northern Sung,” and John Steurmer, “Polder Construction and the Pattern of Land Ownership in the T’ai-hu Basin during the Southern Sung Dynasty.” On the spread of polders westward to the Middle Yangzi region in the Yuan, see Francesca Bray, Science and Civilisation in China: Agriculture, pp. 113–18. For the history of polder-building in that region through the Ming and Qing, see Perdue, Exhausting the Earth, pp. 197–219. 10 Wang Zhen, Nong shu, p. 186; this aspect of poldering is stressed by Nishiyama Takeo, “Ajia nōgyō no genryū,” p. 41. 11 Xu Guangqi, Nongzheng quanshu, juan 13–15, 17. 12 Ibid., pp. 17–18. The largest I have encountered in late-imperial sources, in Guangdong, was 50,970 mu; Gaoyao xianzhi (1826), 6.1b.

Notes

193

13 Fei Hsiao-t’ung, Peasant Life in China, pl. xi. When communities were also kinship entities, as so often the case in Hunan and Guangdong, polders could be lineage enterprises; Perdue, Exhausting the Earth, pp. 174–6. 14 Gaochun xianzhi (1881), 21.8b, 10a, 11a; I am grateful to Fei Si-yen for bringing this text to my attention. For a similar description, see Jiaxing fuzhi (Jiaqing), juan 22, quoted in Kawakatsu Mamoru, Chūgoku hōken kokka no shihai kōzō, p. 606. 15 Fei, Peasant Life in China, pp. 18–20. 16 Liang Fangzhong, Mingdai liangzhang zhidu, pp. 61–84. 17 Shan is a measure word for accounting records (see, e.g., Yu Jideng, Diangu jiwen, p. 194). In the Ming it was used to designate subdivisions of administrative units. Cantons (xiang) were sometimes subdivided into shan in Huzhou, e.g., Huzhou fuzhi (1649), 2.3a–b; rural townships were sometimes broken into shan in Suzhou and Hangzhou, e.g., Wuxian zhi (1642), 2.2a–5a; Hangzhou fuzhi (1475), 2.8a; so also were Hangzhou’s suburban military guard districts (weisi) (2.4a). Shan was also the term used for the eight salt administration districts in Yongjia county, Wenzhou; Yongjia xianzhi (1566), quoted in TJL, 22.133a. 18 Tsurumi Naohiro, “Rural Control in the Ming,” p. 262. 19 Wujiang zhi (1488), 5.23b–24a. 20 Kawakatsu Mamoru, Chūgoku hōken kokka no shihai kōzō, pp. 127–8. 21 Mark Elvin (“Market Towns and Waterways,” p. 450) translates a portion of Shi Jian’s text, which he cites from a 1973 essay by Kawakatsu Mamoru. 22 That polder captains dealt only with water control is suggested by the Wujiang county gazetteer compiled in 1548, which reports that “each sector has appointed to it one tax captain to collect taxes, one subsectoral clerk (shanshii) to oversee financial transfers, and one embankment captain (tangzhang) to maintain irrigation.” Here the polder captain has been renamed embankment captain in order to distinguish him from the new polder captains who were administrative officers at the very local level. The gazetteer reports that there were 3,158 polder captains in the county; Wujiang xianzhi (1561), 10.2b. The gazetteer does not indicate when this arrangement began, though the preceding gazetteer of 1488 records that sectors were the principal subtownship administrative unit; Wujiang zhi (1488), 1.4a. The post of tangzhang is described in Li Jixian, “Mingdai tangzhang shulüe.” 23 Here I follow the interpretation of Kawakatsu Mamoru, who likewise argues that the early posts bearing the labels of polder captain or embankment captain were for hydraulic maintenance rather than fiscal levy; Chūgoku hōken kokka no shihai kōzō, pp. 135–6. 24 Chongming xianzhi (1444), 2.4a. 25 Zhang Jia, “Da Xiaochuan taishi lun shuili shulüe” (Brief letter on water control in reply to Hanlin Academician Xiaochuan), quoted in TJL, 5.58b. This text appears in Taicang zhi (1548), 10.43b–50b, although page 49 on which the reference to sectoral registers appears is missing from the Tianyige copy of the 1629 reprint reproduced by Shanghai shudian. 26 My reconstruction of Huzhou’s subcounty administration is based on data from the 1491 (juan 4) and 1542 (juan 7) editions of Huzhou fuzhi, abstracted in Sogabe Shizuo, Chūgoku oyobi kodai Nihon ni okeru kyōson keitai no hensen, pp. 244–7. 27 Songjiangfuzhi (1512), 9.14b–19b. 28 Elvin, “Market Towns and Waterways,” pp. 450–1. 29 Huzhou fuzhi (1649), 2.1a. 30 Ray Huang, Taxation and Governmental Finance in Sixteenth-Century Ming China, p. 118. 31 Recorded in TJL, 7.59a; also quoted in Hamashima Atsutoshi, Mindai Kōnan nōson shakai, p. 22. 32 Recorded in TJL, 4.26a–27b; the opening section has been translated in Elvin, “Market Towns and Waterways,” p. 450. The identification of Jin Cao as an advisor to Yao Wenhao is made by Hamashima Atsutoshi, Mindai Kōnan nōson shakai, p. 23. I have been unable to determine where and how these proposals were implemented. The gazetteer of one of the central counties of Suzhou prefecture eighty years later notes that polder maintenance was

Notes

194

the work of embankment captains (tangzhang), not hundred captains; on the other hand, its polders were being adequately maintained up into the 1530s, which may indicate that these reforms, or something like them, reversed the decay of poldered fields in the 1490s; Changzhou xianzhi (1571), 2.9b. 33 The priority of landownership over household registration as the basis of taxation is further shown by Jin’s recommendation that the labour needed to repair polder embankments be levied from the owner of the land adjacent to the section of the embankment needing repair; TJL, 4.26b. 34 TJL, 4.11b. 35 Subsectors were also in use in Jiading, another Suzhou county, where they served as the jurisdictions of the alternate hundred captains, who were supervised by embankment captains charged with maintaining water control facilities; Jiading xianzhi (1605), quoted in TJL, 6.19b–21a. 36 Gengju, Changshu xian shuili (1606), 1.3a, 9b. Aspects of Geng’s reforms are discussed in Hamashima, Mindai Kōnan nōson shakai, pp. 84–92. 37 Dengwei sheng’en sizhi (1644), 7.13b. 38 Changzhou xianzhi (1571), 2.8a, 9b. 39 Lin’s proposals are quoted in Xu Guangqi, Nongzheng quanshu jiaozhu, pp. 345ff., and are translated in part in Elvin, “Market Towns and Waterways,” p. 452. 40 Jiangdu zhi (1599), juan 2, quoted in Sogabe, Chūgoku oyobi kodai Nihon, p. 249. Elvin (“Market Towns and Waterways,” p. 740, n. 11) notes that embankment captains were also to be found in northern Jiangsu, though the passage he cites from Gu Yanwu (TJL, 12.105b) suggests that their responsibilities were to maintain embankments along the Grand Canal, not support agricultural production. 41 For Jiaxing’s Xiushiu county, see Kawakatsu Mamoru, Chūgoku hōken kokka, p. 131. The county gazetteer notes that wards and townships were replaced by polders and sectors in the Jiajing era (1522–66), creating a hierarchy of nine cantons, sixteen sectors, and 453 polders; Xiushui xianzhi (1596), juan 1, cited in Hamashima, Mindai Kōnan nōson shakai, pp. 115– 17. Regarding neighbouring Jiashan county, see Jiashan xianzhi (1894), 10.6a, 41a, 42a; also Okuzaki Yūji, Chūgoku kyōshin jinushi no kenkyū, p. 98. Land in Jiashan was being recorded by polder and sector in the wake of the major survey (dazao) of 1582, at the latest. When jiashan native son Chen Longzheng (1585–1645) set up his system for registering relief recipients, he required his gentry surveyors in every one of Jiashan’s twenty sectors to draw up maps for each polder showing where villages were; Chen Longzheng, Jiting quanshu (Complete writings of Chen Jiting, 1665), juan 25, quoted in Hamashima, Mindai Kōnan nōson shakai, p., 537; see also Joanna Handlin Smith, “Benevolent Societies,” p. 329. Jiaxing’s hydraulic geography appears to have been transcribed in Jiaxing qixian tangyu shuili tu (Hydraulic map of the embankments and polders of the seven counties in Jiaxing), a book in seven juan noted in Huang Yuji’s bibliography, Qianqing tang shumu, 8.10b, but which I have been unable to locate. 42 TJL, 7.24a–b, quoting Wujin xianzhi (1605). 43 Gaoshan zhi (1877), 2.26b. 44 TJL, 7.5b. 45 The magistrate in neighbouring Jiangyin county followed his Wujin colleague’s method for assessing labour for dredging and found it effective in reducing tax evasion (TJL, 7.5b). Jiangyin was divided into sectors and subsectors in the late Ming; see jiangyin xianzhi (1878), 4.78a. 46 TJL, 7.5b; the magistrate was Sang Xuekui (js. 1592). 47 Wuxi xianzhi (1753), 5.5b–31a; Taibo meili zhi (1897), 1.2b–11a. None of Wuxi’s former urban townships was left intact. The urban wards were combined into four other subsectors (Subsectors A and B of Kaihua Sector, Subsector B of Xin’an Sector, and Subsector A of Jingyun Sector), each of which spanned the urban-rural divide in the original system.

Notes

195

48 Wuxi xianzhi (1753), 5.5a–b. Another metaphor for the new land-household relationship was “taking land as the lord” and the household as the “servant”; Haiyan xian tujing (1624), 9.25b, referring to the local magistrate’s “equalization of fields” reform of 1581. 49 A copy of this contract is reproduced in Xu Giangqi, Nongzheng quanshu, p. 376. 50 Wucheng xianzhi (1638), as tabulated in Sogabe, Chūgoku oyobi kodai Nihon, pp. 245–7. 51 The system of registering land by polder came into use elsewhere in northern Zhejiang, though not everywhere. Land that a prominent Hangzhou monastery owned in Qiantang county was registered this way in its gazetteer of 1646, whereas its holdings in neighbouring Renhe county were registered by township and ward; Shang tianzhujiangsi zhi (1897), 10.15b–16a. 52 Chen Hu’s program is outlined in Kawakatsu, Chūgoku hōken kokka, pp. 649–53. 53 Lu Shiyi, “Lun yulin tuce” (On fish scale registers), in He Changling, Huangchao jingshi wenbian, juan 29, quoted in Wei Qingyuan, Mingdai huangce zhidu, pp. 240–1. 54 This is the explanation Hamashima Atsutoshi gives for what he sees as the crisis in polder maintenance in the late Ming; Chūgoku Kōnan nōson shakai, pp. 523ff. Kawakatsu Mamoru on the other hand lays the problem at the feet of the state, which abandoned community governance and intervened on the side of landlords; Chūgoku hōken kokka, pp. 135–6, 653– 6. See also Philip Huang, The Peasant Family and Rural Development in the Yangzi Delta, pp. 35–8. 55 Louxian zhi (1788), 2.5b, 7.9a. The same insistence to distinguish the tax captain units of the old lijia system from the newer units is also to be found in other late-Ming fiscal records, e.g., Songjiang fuzhi (1630), 11.4b, which refers to the tax captain sectors as “the old system.” 56 Jerry Dennerline, “Fiscal Reform and Local Control,” p. 113. Mid-Qing land documents from Songjiang show that plots of land in Lou county were identified by sector and polder; Shanghai bowuguan tushu ziliaoshi, Shanghai beike ziliao xuanji, pp. 66–7. Land reclassification certificates issued by the Lou county office at the turn of the nineteenth century, held in the Tōyō bunka kenkyūju, Tokyo University, similarly specify location by sector and polder. 57 Kawakatsu, Chūgoku hōken kokka, pp. 583, 603–5, 615–16; Madeleine Zelin (The Magistrate’s Tael, pp. 248–9) dates its inception to 1723. The village rotation system was imposed injiangyin county in 1733, Jiangyin xianzhi (1878), 2.6a. 58 Francesca Bray, The Rice Economies, p. 64. 59 Changzhou xianzhi (1571), 2.10a.

4 Growing rice in North Zhili First published as “The Spread of Rice Cultivation and Rice Technology into the Hebei Region in the Ming and Qing,” in Explorations in the History of Science and Technology in China, ed. Li Guohao, Zhang Mengwen, and Cao Tianqin; revised. 1Wang Zhen, Nong shu, p. 6. This is not the same Wang Zhen we saw fighting to protect his grave land in the Introduction. 2 Xu Guangqi, Nongzheng quanshu jiaozhu, p. 628. 3 Lin Zexu, Jifu shuili yi, 24a, reprinted in Chen Zugui, Dao, p. 455. 4 Zheng Kesheng, “Guanyu Mingdai Tianjin de shuitian,” p. 102. 5 TJL, 1.49a–72b.

Notes

196

6 Lin Zexu, Jifu shuiliyi, 9a. 7 John Lessing Buck, Land Utilization in China, p. 27. 8 In some Shuntian gazetteers, small-grained millet is called gu, the generic term for any unhusked grain. 9 Wu Bangqing, zenong yaolu, 3.1b. 10 David H.Grist, Rice, pp. 163–5. 11 Wang Zhen, Nong shu, p. 57; Li Shizhen, Bencao gangmu, 22.8b. 12 One county gazetteer explains that wet rice is what southerners grow, whereas northerners grow dry rice; Ninghe xianzhi (1779), 5.10b. In Zunhua, dry rice was said to account for nine-tenths of the rice crop; Zunhua zhouzji (1676), 7.1b. Pingxiang reported growing only dry rice, albeit on a small scale; Pingxiang xianzhi (1878), 5.11a. 13 Xu Guangqi, Nongzheng quanshu, p. 628. 14 Ibid., p. 308. 15 The etymologies of the words jing, xian, and nuo are discussed in Matsuo Takane, “Rice Culture in China,” pp. 161–2. 16 Daming fuzhi (1672), 12.2a. 17 Buck, Land Utilization in China, pp. 111, 116, 188, 211, 215. 18 Grist, Rice, p. 11; Matsuo, “Rice Culture in China,” pp. 162–3. 19 Cizhouzhi zhi (1780), 10.3a. 20 Hejian fuzhi (1540), 7.7b; Zhending xianzhi (1612), 3.8a. 21 Jifu tongzhi (1735), 56. 2a. The increase in rice cultivation in north China after 1949 was due primarily to the development of short season varieties; International Rice Research Institute, Rice Research and Production in China, p. 41. Short-season paddy not only diminishes the threat of low temperatures but takes better advantage of the limited period of natural precipitation. 22 “Anlan zhi” (Quelling the floods), in Baodingfuzhi (1886), 21.45a. According to Wada Tamotsu, “Hoku Shina no tochi kairyo ni kansuru chōsa hōkokusho yōshi,” p. 7, a river in flood can increase its volume over a thousand-fold. 23 Baodmg xianzhi (1673), 1.43a. 24 Jifu tongzhi (1735), 46.22b. 25 Cangzhou zhi (1743), 4.28a. 26 The use of deep drains to flush the deeper soil layers was introduced only in 1911; Wu Lien Teh, “A Striking Example of Scientific Farming in North China,” p. 20. 27 The natural salt content of river water is usually less than 0.1 percent. While this is low in relation to the 3.5 percent salt solution in sea water, its cumulative effect can be disastrous (conversation with Professor K.S.Spiegler, Michigan Technological University). 28 Dwight Perkins, Agricultural Development in China, pp. 41–3. 29 Hou Han shu, p. 1098. 30 Jifu tongzhi (1910), 73.2a. 31 Hejian fuzhi (1615), 4.30b; Xin Tang shu, p. 1018. 32 Cefu yuangui (1005), quoted in Zhu Yizun, Rixia jiuwen, 24.11a. 33 Qju Jun, Daxueyanyi bu (1487), quoted in Chen Zugui, Dao, p. 96. 34 Song shi, p. 4264. 35 TJL, 2.42a. 36 Xu Guangqi, Nongzheng quanshu, pp. 283, 287. 37 Shuntian fuzhi (1886), 18.9a. 38 Yongnian xianzhi (1877), 6.2a. 39 These references are summarized in Chart 1 of “The Spread of Rice Cultivation and Rice Technology into the Hebei Region in the Ming and Qjng,” pp. 667–71, supplemented with additional rice-present dates from Yongping fuzhi (1517), 2.2a, Wuan xianzhi (1547), 1.5b, and Zhaozhou zhi (1498), 1.4b, 3.2b. 40 Lingshou xianzhi (1685), 3b. 1a; on Lu Longqi and this gazetteer, see Chapter 8.

Notes

197

41 Weizhou zhi (1877), 1.24a, 25a, 5.23b. 42 Yongnian xianzhi (1877), 6.3a; Cizhau zhi (1700), 9.1b. 43 Wangdu xianzhi (1771), 4.10a. 44 Yingzong shilu, 41.13b. In 1438 this land was reclassified for tax purposes as civilian land. 45 Mnghexianzfii (1779), 15.11a. 46 Qing huidian (1764), quoted in Chen Zugui, Dao, p. 233. 47 Yinxiang, “Yi xianqinwang shuchao,” 1b. 48 Peng Xinwei, Zhongguo huobi shi, p. 569. 49 Zhuozhou zhi (1875), 6.28b; Pingshan xianzhi (1854), 2.18a–b. 50 Yongnian xianzhi (1758), 13.1b. 51 The lower estimate was the official one, printed in Jifu tongzhi (1735), 46.2b. The higher estimates is from a text by Zhu Yunjin dated 1821, reprinted in Chen Zugui, Dao, p. 253. 52 Thirty years after the project first brought rice to Leting county, the irrigation and riziculture initiated by the project were still being maintained; Leting xianzhi (1755), 5.24b. A century later they were still in operation in Pingshan due to the efforts of subsequent magistrates; Pingshan xianzhi (1854), 2.19b. 53 Jifu tongzhi (1735), 47.8a, reports rice as present in Shahe county as of 1727, whereas Shahe xianzhi (1757), 39a, reports it as absent thirty years later. The first Yizhou gazetteer reports it present as of 1645, the next as absent a century later; Yizhou zhi (1645), 2.51, Yizhou zhi (1747), 10.6b. 54 Nangong xianzhi (1831), 6.7b. 55 Wu, “A Striking Example of Scientific Farming,” p. 17. 56 TJL, 1.51b. 57 Jifu tongzhi (1735), 47.9b. 58 Nōshōmushō, Shina no kome ni kansuru chōsa, p. 43. 59 Chengde fuzhi (1887), 28.2a. 60 Jingzhau zhi (1612) 3.38b. 61 The spatial distribution of rice agriculture corresponds inversely to the area hit by famine in 1743, as depicted in Pierre-Etienne Will, Bureaucratie etfamine en Chine au 18e siècle, p. 39, map 3. Those counties hardest hit by the famine (in Hejian prefecture) were largely the same counties that were rice-barren before 1900. The coincidence suggests only the obvious: that rice came first to those counties where agricultural conditions were most favourable to sustained production. 62 Dingxing xianzhi (1779), 10.10a; Dingxing xianzhi (1893), 13.6b. 63 Baodi xianzhi (1745), 7.6b. 64 Baoding fuzhi (1886) 21.40b. 65 A memorial of 1821 reveals that, in some areas, productive paddy land was not registered as such, but as land under hydraulic construction. False registration meant lower taxes for those who owned the land, and lower paddy acreage totals; Da Qing lichao shilu, Daoguang, 1st year, 2nd month, 13.4a–b. 66 Lin Zexu, Jifu shuili yi, 8a. 67 Sun Yutang and Zhangjiqian, “Qingdai de kentian yu dingkou de jilu,” p. 113. The cultivated acreage totals for 1822 and 1851 are both 727, 262 qing. Totals in the Jiaqing and Tongzhi eras are slightly higher. 68 Production figures for the Ming and Qing may be roughly compared by converting Lin’s figures into volume percentages. An Anhui gazetteer (Fengtai xianzhi (1814), 2.3a) suggests a millet-to-rice coefficient of between 2 and 2.5. Applied to Lin’s estimate of 2 percent, rice may have accounted for 4–5 percent of the volume of grain production. Though this figure seems too high, it suggests a late-Qing percentage higher than the late-Ming. 69 Nōshōmushō, Shina no kome ni kansuru chōsa, p. 4. 70 Tangxian zhi (1878), 3.29a. 71 Long Wenbin, Ming huiyao, p. 999.

Notes

198

72 Shunde fuzhi (1750) 2.27b. 73 Lin Zexu, Jifu shuili yi, 23b, quoting from Da Ming yitong zhi; Mancheng xianzhi (1713), 5.30b. 74 Yang Yigui, “Guangli qu ji” (A record of Broad Benefits Canal), in Tangxian zhi (1878), 3.32a–34a. 75 Peng Xinwei, Zhongguo huobi shi, p. 470. Given the higher cost in grain payments, it would be reasonable to suppose that the money cost was also higher than Yang’s original estimate, putting the conversion rate somewhere below 720 cash per tael. 76 Tangxian zhi (1878), 3.29a–b. 77 Mingshi, p. 5882. 78 Xu Zhenming, Lushui ketan, 4b; reprinted in Xu Guangqi, Nongzheng quanshu, pp. 292–3. Portions of his writings on northern agriculture appear in Ming shi, pp. 5881–4. 79 Ming shi, p. 2171. In Xu Zhenming’s biography in the Ming shi (p. 5885), eunuchs and enfeoffed imperial relatives are cited as Xu’s opponents. Examples of landholding by both groups are given in Zheng Kesheng, “Guanyu Mingdai Tianjin de shuitian,” p. 97. 80 Ray Huang, biography of Wang Ying-chiao, in Goodrich and Fang, Dictionary of Ming Biography, p. 1452. 81 Long Wenbin, Ming huiyao, p. 999. 82 Tangxian zhi (1878), 3.34b. 83 Zhengding fuzhi (1762), 4.33b. 84 Linyu xianzhi (1756), 2.21b. 85 Long Wenbin, Ming huiyao, p. 999. 86 Jifu tongzhi (1735), 47.1 la. 87 Fengtai xianzhi (1814), 2.4a. 88 Shuntian fuzhi (1886), 18.9a. Xu Zhenming repeated the proposal in the Ming, as quoted in Xu Guangqi, Nongzheng quanshu, p. 288. 89 Yongnian xianzhi (1877), 6.3. 90 Yuan Huang, “Quannong shu” and “Baodi zhengshu,” in his Liaofan zazhu. The former is quoted in Lin Zexu, Jifu shuili yi, 44a–b. 91 Li Guangdi proposed official instruction at the turn of the eighteenth century as part of a plan to build rice paddies in Hejian prefecture; Yinxiang, “Yi xianqinwang shuchao,” 1b. Yinxiang’s officials in the Tianjin area built wet seed-beds at Lantian for instructional purposes in 1727; Jifu tongzhi (1735), 47.11a. 92 Wu Bangqing, Zenong yaolu, preface, 1b–2b. 93 For example, Jifu tongzhi (1735), 46.20a, regarding such a project in Dacheng county in 1733. 94 TJL, 2.42a. 95 Wangdu xianzhi (1771), 4.16a. 96 On the treadle-wheel as a form of “time-extensive labour” (chōjikan rōdō), see Tanaka Masatoshi and Tsurumi Naohiro, “Ryūkosha to nōmin,” p. 12. 97 Wangdu xianzhi (1771), 4.17a. 98 Wu Bangqing, zenong yaolu, 5.2a. 99 Tanaka and Tsurumi, “Ryūkosha to nōmin,” pp. 5, 7, 8, 12. 100 Baoding fuzhi (1886), 21.45a. 101 Ibid., 4.14b. This list of the technical shortcomings reads like a reverse description of the rice-growing regions of central and southern China—as it was meant to. The south had a long tradition of rice cultivation and of irrigation systems under landlord control. In the north, the absence of these conditions and the technical obstacles to their quick development militated in favour of more state investment in the north than in the south. 102 Zhu Yizun, Rixia jiuwen, 38.bu.14b. 103 Jifu tongzhi (1735), 46.2b.

Notes

199

104 Jiang Yangwu, “Shuitian yi” (Proposal for irrigated paddy), quoted in Chen Zugui, Dio, p. 116. 105 Lin Zexu, Jifu shuili yi, 54a. 106 In the Yongzheng era, “unprincipled sub-gentry” (lieqin liejian) of Tang county sought to obstruct the introduction of rice into their area on the grounds that growing rice would increase the people’s taxes. Plans to build paddy fields were suspended until the central government intervened to punish the leaders of the obstruction; memorial of Yi Beixiong and Liu Shishu, quoted in Lin Zexu, Jifu shuili yi, 41b–42a. 107 TJL 9.9b 108 Xu Guangqi, Nongzheng quanshu, pp. 302, 300, 628. 109 Lin Zexu, Jifu shuili yi, 42a–b. 110 Ibid, 17a.

5 Building school libraries in the mid-Ming Expanded from an earlier version, “Edifying Knowledge: The Building of School Libraries in Ming China,” Late Imperial China 17:1 (June 1996), pp. 88–114. 1 Du Halde, The General History of China, vol. 3, p. 63. 2 For example, Huguang tujing zhishu (1522), 6.62a, referring to a Wanjuan lou built at the turn of the eleventh century. 3 I have derived this sense of the scale of book-collecting in the Ming from Wu Han. For examples of collections of 10,000 juan in the mid-Ming, see his Jiang Zhe cangshujia skilüe, pp. 10, 22, 145, 155, 176, 229, 232 (see also Ming shi, p. 5311); for collections in the tens of thousands in the late Ming, see pp. 20, 59, 89, 126, 138, 140, 201. One publishing business in early-seventeenth-century Nanjing took as its name Wanjuan lou; Lucille Chia, Publishing for Profit, p. 165. 4 Ji’an fuzhi (1648), 15.7a, 34.8a. 5 In addition to the five listed in the next footnote, the Suzhou school library was founded about 1060 and rebuilt in 1187 (Suzhou fuzhi (1883), 25.1b–2b); Songjiang’s, for which no founding date survives, was rebuilt in 1296 (Songjiang fuzhi (1630), 23.34b); Zhangzhou’s was founded in 1139 (zhangzhou fuzhi (1877), 7.1b); Yuanzhou’s dates to 1178 (Yuanzhou fuzhi (1874), 4.2a). 6 Four are collected in his Zhu zi wenji, vol. 8, pp. 3892–4, 3905–6, 3964–6, and 3979–80. His 1192 essay for the school library in Wuchang is preserved in Huguang tujing zhishu (1522), 2.wen.1b. In his essay on the new book collection in the county school in Jianyang, Fujian, he expresses surprise that a school here in a centre of commercial publishing was without books (p. 3905). 7 Jining zhouzhi (1672), 5.8a–b, 8.14b; restored in 1553 and 1564, the library was defunct by the time the gazetteer was published, to the compiler’s dismay. As he notes, the library’s books “were originally stored in the Pavilion for Revering the Classics. After the pavilion decayed, the books were not preserved. I chronicle this in the hope that hereafter a book collector will be moved” (5.10a). 8 On academy libraries, see Lai Xinxia, Zhongguo gudai tushi shiye shi, pp. 272–3. 9 Songjiang fuzhi (1630), 23.39b. The 1817 edition (30.12a) has corrected Wang Ying’s commemorative essay from 1443 (23.49b) to 1440, although the essay may postdate the reconstruction by three years.

Notes

200

10 Yushu ge was used on rare occasions for school libraries in the Ming (Yanzhou fuzhi (1613), 3.16a; Longquan xianzhi (1878), 5.2a–3a), though it tended to be replaced; the Yushu ge at the Longquan county school was called Zunjing ge by the eighteenth century if not earlier. 11 Tangzhou fuzhi (1733), 12.3b. 12 When the Xinxiang county magistrate restored the school library in 1582, he changed its name to Jukui ge, jukui being the constellation associated with Wenchang, the God of Literature. When another repaired it half a century later, he changed the name yet again to Wenchang ge; Weihui fuzhi (1788), 10.20a. As of the 1582 restoration, the library also served as a shrine to Wenchang, which may account for the magistrates’ preferences for names associated with this divinity 13 For example, Funing zhouzhi (1593), 3.24a, which records only that the library of the Ningde county school was rebuilt in 1525. 14 For example, Ji’an fuzfii (1648), 15. 7b. 15 Readers wishing to consult the data on which these fmdings are based will fmd it in the original version of this chapter, “Edifying Knowledge: The Building of School Libraries in Ming China,” pp. 96–101. Shaanxi, Sichuan, Guangxi, Yunnan, and Guizhou provinces were excluded from this sample for lack of comparable sources. 16 Huian xianzhi (1530), 9.10a. For a similar expression of concern about loss of books, see Cili xianzhi (1574), 11.14b. The loss is actually tallied in Dingxiang xianzhi (1727), 2.13b. Neither of these county schools had library buildings. 17 For a failed appeal from Chen Fengwu (1475–1541) to rebuild the twelfth-century Pavilion for Examining Antiquity at the Wuchang school, see Huguang tujing zhishu (1522), 2.wen.20b. 18 Some examples: (1) The library in Guangshan, Henan county, built in 1445, was derelict a century later; Guangshan xianzhi (1556), 1.27b. (2) The library in Ruichang, Jiangxi, founded in 1523, had sunk to such disrepair in less than half that time that its books had to be moved to another building; Ruichang xianzhi (1568), 5.4a. (3) Most of the county school libraries in Zhending prefecture, North Zhili, are listed as “abandoned” (fei) in the 1762 gazetteer, Zhengding fuzhi, 8.47bff. 19 Taizu shilu, 21.7b. 20 Taizu shilu, 56.14a, 59.2a, 86.6b–7a. 21 Jinling fancha zhi (1607), 31.19a. 22 Mingshi, p. 5310. The book’s distribution is recorded in Shizong shilu, 95.8a. 23 Gu Yzinwu, Rizhi lu jishi, 18.11b–12a. 24 Hejian fuzhi (1540), 5.5b, from a text by Peng Shi (js. 1448). 25 Dehua xianzhi (1872), 21.3b. This source dates the acquisition to 1382, but as the first set of Grand Pronouncements was not issued until 1385, 1382 is an error, perhaps for 1392. 26 Daoan, Dazangjing diaoke shihua, p. 127. 27 Cili xianzhi (1574), 11.14b. Jiaxing xianzhi (1637), 2.42b, notes that the Jiaxing school received its Great Compendia in 1417. 28 Shenxian zhi (1548), 3.9b. 29 For example, Jianchang fuzhi (1517), 8.1b. 30 Yingzong shilu, 172.6a. 31 Yingzong shilu, 37.2b. 32 Shizong shilu, 60.7a. This was the Prince of Nanwei, Zhu Yanbin (d. 1543), enfeoffed in 1525; Ming shi, p. 2741. For a bestowal on an official, see Shizong shilu, 96.10a. 33 Shizong shilu, 180.6b. 34 When the Songjiang prefect in 1364 discovered that over half the books in the prefectural school were gone, he arranged to buy annotated editions of the thirteen classics from a wealthy family in Suzhou; Songjiang fuzhi (1630), 23.34b. Many school collections were destroyed in the troubled mid-fourteenth century. When the school in Wanzhou subprefecture, Guangdong, was attacked in 1331, “the bandits burned all the books,” using a

Notes

201

phrase (qunshu) that suggests a large number; five years elapsed before any attempt was made to replace them; Qiongzhou fuzhi (1619), 6.26a. 35 For evidence of the continuing importance of citing the Jiaomin bangwen beyond Hongwu’s time, see Zhou Chen’s use of it to open his 1438 memorial on lawsuits in Yingzong shilu, 39.4a. 36 Songjiang fuzhi (1630), 23.35a. 37 Songjiang fuzhi (1630), 24.13a. 38 Huguang tujing zhishu (1522), 2.wen.20b. 39 On Nanjing and the book trade, see Lucille Chia, Printing for Profit, p. 185, citing Xie Zhaozhe’s (1567–1624) famous ranking of publishing centres. 40 Hejian fuzhi (1540), 28.58b–59b. 41 Jianchang fuzhi (1517) 8.1a–10a. 42 Chia, Printing for Profit, pp. 106, 140, 175. 43 Jianchang fuzjii (1517), 8.1a–10a. Of the classics that the commercial publishers of Jianyang, Fujian, printed in the Ming, roughly a third were Great Compendium editions. Lucille Chia (Printing for Profit, p. 222) comments that this “may seem surprising,” yet these were textbooks that students needed to acquire, as cheaply as Jianyang could provide them, to study for the exams. She notes elsewhere (p. 178) that much of this publishing clustered in the Jiajing era—which was when state publishing was in decline and schools were having to refurbish their libraries from other sources. 44 Tōyō bunka kenkyūjo, Tōkyō daigaku tōyō bunka kenkyūjo kanseki bunrui mokuroku, p. 452; Sonkeikaku bunko, Sonkeikaku bunko kanseki bunrui mokuroku sakuin, pp. 744–5; Luo and Hu, Guji banben tiji suoyin, p. 695. On the state’s concern for the corruption of commercially produced canonical texts in the latter half of the dynasty, see K.T.Wu, “Ming Printing and Printers,” p. 230. 45 Douglas Chambers, “A Catalogue of the Library of Bishop Lancelot Andrewes,” p.103. 46 Ruichang xianzhi (1568), 5.6a–7a; Ruijin xianzhi (1542), 3.3a–4a; Shenxian zhi (1548), 3.9a– b; Wucheng xianzhi (1549), 1.38b; Changde fuzhi (1538), 9.4a–11b. For a Guangdong example, see Lingshan xianzhi (1733), 5.10a–b. 47 Hangzhou fuzhi (1922), 14.28a. 48 Huguang tujing zhishu (1522), 2.wen.20b. 49 Chaozhou fuzhi (1547), 2.20a. By contrast, bibliophiles preferred to lay their books out on broad shelves for display and ready access, as Wen Zhenheng (1585–1645) advised; Clunas, Chinese Furniture, p. 89. 50 The gazetteer of Zhangde prefecture notes that the prefectural school, refounded in 1370, received palace editions and built a storeroom for its ceremonial vessels; its books were probably stored there until the library was built in 1496; zhangde fuzhi (1522), 3.4a. The Yazhou subprefectural school on Hainan Island had two separate storerooms, one for books and one for other objects; Yazhou zhi (1900), pp. 103–4. The Yazhou school did not get a library until 1667. 51 For example, Jiaocheng xianzhi (1882), 5.4b. 52 Xuanzong shilu, 110.8a. 53 Songjiang fuzhi (1630), 23.50a. 54 Quoted in Songjiang fuzhi (1817), 30.21a. 55 Chan monk Daoan began restoring Nanchang’s Yongning Monastery in 1567. Later he acquired a copy of the Tripitaka, and in 1573 raised the sum of 1,000 taels to build a library; Nanchang fuzhi (1588), 29.26b. In a fund-raising appeal Wang Zhideng (1535–1612) wrote to build the library at Hanshan Monastery, Suzhou, he declared that “vast sums” were needed to house the abbot’s Tripitaka; Hanshan sizhi (1911; 1986), p. 9. 56 Songjiang fuzhi (1630), 23.50a. 57 Qiongzhou fuzhi (1619), 4.63b, 6.8a, 9b; see 11.26b for Qiu’s record of building this library This text is reprinted in Qiongshan xianzhi (1917), 14.37a–40a.

Notes

202

58 Some examples: Weihui fuzhi (1788), 10.25b; Jinan Fuzhi (1840), 1.68b. 59 Brook, Praying for Power, pp. 215–23, 242–5, 271–7. 60 Yuanzhou fuzhi (1874), 4.20b. The Yichun school was founded by its magistrate in 1396. 61 Ji’an fuzhi (1648), 35.45b. 62 Brook, Praying for Power, p. 162. 63 Ji’an fuzhi (1648), 15.7b. 64 songjiang fuzhi (l630), 24.31a–b. 65 Quanzhou fuzhi (1763), 14.5a. 66 Ji’an fuzhi (1648), 35.5b–7a. 67 The importance of the classics for teaching (jiao) the Way is similarly argued by Zhang Jiuyi (js. 1553) in an essay on a school library in Henan in 1589; Xiangcheng xianzhi (1911) 9.35b. 68 Zhu Xi, Zhuzi wenji, vol. 8, pp. 3893, 3905, 3979, 3965. 69 Jianchang fuzhi (1517), 8.10a–24a; Hejian fuzhi (1540), 28.51b–59b. 70 For example, Yanping fuzhi (1526), 12.7a–8a; Ruijin xianzhi (1542), 3.3a–4a; Shenxian zhi (1548), 3.9a-b; Ruichang xianzhi (1568), 5.6a–7a; Ji’an fuzhi (1648), 15.7a; Dingxiang xianzhi (1727) 2.13a–b.

6 State censorship and the book trade Published in Chinese as “Ming-Qing shiqi de guojia tushu jiancha yu tushu maoyi,” Shilin (Historical review, Shanghai Academy of Social Sciences), 3 (2003), pp. 90–104. Part of this chapter has appeared as “Censorship in Eighteenth-Century China: A View from the Book Trade,” Canadian Journal of History 23:2 (August 1988), pp. 177–96. 1 Representative of such views is Tan Peifang, author of the section on the Ming—Qing period in Zhongguo jinshu jianshi, ed. An Pingqiu and Zhang Peiheng, pp. 162, 218, 227. 2 Xie Guozhen, Wan Ming shiji kao, 16.12a. 3 Lucille Chia, Printing for Profit, pp. 176–7, speculates that the book that came to Li’s attention was actually Jiandeng yuhua (More tales while trimming the wick), a supplementary collection that the county magistrate of Jianyang, the centre of commercial publishing in Fujian, authorized for publication in 1433. 4 Yingzong shilu, 90.5a, quoted in An Pingqiu and Zhang Peiheng, Zhongguo jinshu jianshi, p. 178. On Li Shimian, see Norman Kutcher, Mourning in Late Imperial China, pp. 68–9. 5 Wing-tsit Chan, trans., Reflections on Things at Hand, p. 42. 6 This view is more fully expressed in my “Confucianism,” pp. 570–2. 7 See J.L.Heilbron, The Sun in the Church, pp. 3–5, regarding the connection between papal legitimacy and the astronomical determination of Easter. 8 Da Ming lü, juan 12, p. 91, and juan 18, pp. 134–5. 9 Taizu shilu, 203.2b. 10 Xianzong shilu, 216.5b; Xiaozong shilu, 133.3a; Shenzong shilu, 76.9b. 11 Da Qing lüli, pp. 281, 368–9. 12 Shenzong shilu, 456.9b. 13 Xizong shilu, 70.23b. 14 The Veritable Records summary of Zhang Wenda’s memorial against Li Zhi was widely cited and approved by seventeenth-century writers. Gu Yanwu included it in his Rizhi lu

Notes

203

jishi, 18.30b. Wang Fuzhi was harsher; see his Du tongjian lun, p. 1111, and zhang zi zhengmeng zhu, p. 332. 15 Yuan Zhongdao’s biography of Li Zhi is reprinted in Fen shu, p. 4, and quoted in Ray Huang, 1587, a Year of No Significance, p. 190. 16 Tu Long, Shaluo guan qingyan, author’s preface, p. 1. 17 Jin Jiang points out that local politics were also involved, Mei Guozhen being a member of a prominent family in the Huguang county from which Li had to flee that was at odds with Li’s former patrons, the Geng family; see her “Heresy and Persecution in Late Ming Society.” 18 Ma Jinglun, “Zhi dangdao shu” (Letter to the circuit official), reprinted in Xiamen daxue lishixi, Li Zhi yanjiu cankao ziliao, p. 61. 19 Jiang, “Heresy and Persecution in Late Ming Society,” pp. 28–9. 20 Li Zhi, Xu cangshu, Li Weizhen’s preface, 1a, 4a. 21 Wang’s memorial is quoted in Gu Yanwu, Rizhi lu jishi, 18.31b. 22 Li Zhi, Fen shu, p. 117. On regulated essays, see Benjamin Elman, A Cultural History of the Civil Examinations in Late Imperial China, pp. 382, 396. 23 Tu Long, Baiyu ji, 4.7b. 24 Qian Qizhong, Qingxi yigao, 1.43a–b. 25 Parenthetical page citations in the text to inquisition documents are to L.Carrington Goodrich, The Literary Inquisition of Ch’ien-Lun; romanizations have been standardized to pinyin. 26 Cited in Norman Levine, “The Myth of Asiatic Restoration,” p. 74. 27 Quoted in Min Tu-ki, National Polity and Local Power, p. 14. 28 Kent Guy, The Emperor’s Four Treasuries, p. 158. Similar themes are struck, though without the detail Guy provides, in Zuo Buqing, “Qianlong fenshu,” p.176. 29 Donald Thomas, A Long Time Burning, p. 8. 30 Paul Grendler, The Roman Inquisition and the Venetian Press, 1540–1605, p. 71. See also J. L.Heilbron, The Sun in the Church, pp. 198–216, for the politics at work behind censorship cases touching on astronomy. 31 Ye Dehui, Shulin qinghua, p. 27. 32 Elizabeth Eisenstein, The Printing Press as an Agent of Change, ch. 2. 33 Lucille Chia, Printing for Profit, pp. 61–2, 202; see also K.T.Wu, “Ming Printing and Printers.” 34 A courtier in 1005 observed to the Song emperor that “today we are provided with printed books on a large scale, and the families of both gentlemen and commoners have them: this is a new dawn for scholars.” See Luo Mengzhen, Zhongguo gudai muluxue jianbian, p. 168. 35 Ye Dehui, comparing costs listed in a government publication of the fourteenth century and a privately printed book from the middle of the sixteenth, noted a drop from well over 1 tael of silver per page to an eighth of a tael; Shulin qinghua, pp. 178, 186. For further observations on the cost of printing, see Evelyn Rawski, Education and Popular Literacy in Ch’ing China, pp. 120–2. 36 As phrased in Louis Gallagher, China in the Sixteenth Century, p. 21. 37 In the estimation of historian Hong Huanchun, the publishing industry in central China was at its height from late in the seventeenth century through the eighteenth; Zhejiang wenxian congkao, p. 111. 38 QWD, 3: Dangui, 8a–15b. The investigators provide itemized catalogues of both collections. 39 Wei Jinxi, Xuezheng quanshu, 7.1a. Governmental and private documents concerning the censorship of fiction have been collected in Wang Liqi, Yuan-Ming-Qing sandai jinhui xiaoshuo xiqu shiliao. 40 Huang Jian had the memorials of his grandfather Huang Tingkui (1691–1759), a military official and grand secretary, engraved and printed in Shanxi province; he later sent copies to

Notes

204

his official superiors in Fujian when he held office there. See QWD, 4: Huang Jian, 3a. The book was censored for including imperial rescripts that Huang was not authorized to cite. 41 QWD, 1: Liu Zhenyu, 1a. 42 QWD, 2: Cai Xian, 1b–3a. 5b–7a. 43 Wei Jinxi, Xuezheng quanshu, 7.2b. The case of sedan-chair bearer Li Hao from Fuzhou (QWD, 2: Li Hao, 1a–2a) confirms that illiterates were active in the publishing trade, even as publishers. Li ran into someone in the streets of Fuzhou selling rubbings of a stele that, the vendor claimed, had miraculously appeared in Guangdong province to the south. Sensing an opportunity to make some money, Li bought a copy, took it to an engraver’s shop, and paid Engraver Fu eighty cash to incise the text onto a large block. Li got copies printed and took them up the coast to sell in Zhejiang, the next province to the north. There he was apprehended for distributing a text that, as it turned out, spoke of rebellion. Like the illiterates in the Cai Xian case, Li was excused from the capital charge of sedition because he could not read what he was selling, though he was found guilty on a lesser charge. 44 QWD, 7: Shen Deqian, 2b. 45 Wu Zhefu, Qingdai jinhui shumu yanjiu, pp. 462, 491. 46 Gu Yu, ed., Gui Qian chidie, Wang’s preface (1699), 2a. 47 For extant editions of Caoxi tongzhi, see my Geographical Sources of Ming-Qing History, p. 218. 48 Wu Zhefu, Qingdai jinhui shumu yanjiu, p. 347. 49 Guy, The Emperor’s Four Treasuries, pp. 176–7. 50 Caoxi tongzhi (1672), 1.12b, 3.8a. 51 Tianyongzi was Ai’s personal sobriquet, and can also be translated as “the man for whom heaven is just.” 52 A brief biography of Ai by Carrington Goodrich may be found in Arthur Hummel, ed., Eminent Chinese of the Ch’ing Period, p. 4. 53 Publishing history is taken from Ai Weiguang, “Yuanke fanli” (Foreword to the original edition; 1699), and Ai Zhou, “Chongke shulüe” (Account of the reprint edition; 1836), in the prefatory matter to Ai Nanying, Tianyongzi ji. 54 Fuzhou fuzhi (1876), 84.10b. 55 “Shanfang” was not an epithet commonly used by publishing houses (Ye Dehui, Shulin qinghua, 5.11a–22a). Wu Zhefu, Qingdai jinhui shumu yanjiu, p. 177, mentions a 1751 edition of Tianyongzi ji, to which Ai Zhou makes no reference. 56 The convention regarding subsequent editions of letters censored in this fashion was to replace the effaced name with mou (“someone”). This has been done in the 1911 Zhigu tang edition of Model Letters. 57 Goodrich, The Literary Inquisition, p. 192; Guy, The Emperor’s Four Treasuries, p. 13. 58 Philip Kuhn, “Political Crime and Bureaucratic Monarchy: A Chinese Case of 1768,” p. 94. 59 From the works of Tang Bin (1627–87), cited in Wang Liqi, Yuan-Ming-Qing sandai jinhui xiaoshuo xiqu shiliao, p. 90. The moralist Yuan Huang (1533–1606) was similarly disturbed that merchants were willing to “sell licentious books and erotic pictures as well as aphrodisiacs to turn a profit,” and warned booksellers who did so that their children and grandchildren would become actors and prostitutes as divine punishment (ibid., p. 178). 60 Francis Higman, Censorship and the Sorbonne, p. 34. 61 Yongrong, Siku quanshu zongmu, shou, fanli, 5a. 62 On Qian Jingkai, see Wu Han, Jiang Zhe cangshujia shulüe, p. 2; Ye Dehui, Shulin qinghua, pp. 250, 255. 63 QWD, 4: Tao Xuan, 2a. 64 Qianlong stressed time and again in his early inquisition edicts that “no questions will be asked and no charge brought against [those who have] clandestinely hidden away” proscribed books (p. 126). One is reminded of Mao Zedong’s promise prior to the AntiRightist Campaign in 1957.

Notes

205

65 Thomas, A Long Time Burning, pp. 9–15. 66 Quoted in Guy, The Emperor’s Four Treasuries, pp. 18–19, translation revised. 67 A chain of such bookstores in Zhejiang province figures in Wu Ching-tzu’s novel, The Scholars, p. 160. 68 Sun Dianqi, Liulichang xiaozhi, pp. 273, 280–3. 69 On the comparison of the inquisition to a modern political campaign, see Kuhn, “Political Crime and Bureaucratic Monarchy,” and Zuo Buqing, “Qianlong fenshu,” p. 159.

7 At the margin of public authority: the Ming state and Buddhism First published as “At the Margin of Public Authority: The Ming State and Buddhism,” in Culture and State in Chinese History: Conventions, Conflicts, and Accommodations, ed. Theodore Huters, R.Bin Wong, and Pauline Yü (Stanford, GA: Stanford University Press, 1997), pp. 161–81; slightly revised. 1 JFZ, 2.24a, 1.1b. 2 One could see in the distinction between state and public authority an analogy to the incomplete distinction in Chinese between guan (“official,” “of the state”) and gong (“public,” “in the public interest”), guan being narrower than “state” and gong broader than “public authority.” I return to this issue in the Conclusions. 3 The Hongwu regulations have been surveyed in English in Chün-fang Yü, The Renewal of Buddhism in China, pp. 144–69. The main Japanese studies are by Tatsuike Kiyoshi, “Mindai no sokan,” “Minsho no jiin,” and “Ming taizu de fojiao zhengce”; also Mano Senryū, “Mindai no bukkyō to Minchō.” 4 Chikusa Masaaki, Chūgoku bukkyō shakaishi kenkyū. Valerie Hansen summarizes Chikusa’s arguments in a review in the Bulletin of Sung Yuan Studies 20 (1988), pp. 97–108. 5 Pingshan tang tuzhi (1705), 6.13b, citing a text of the mid-Ming. 6 JFZ 16.6a; Xu Yikui, Shifeng gao, 11.8b; Taizu shilu, 53.3a; Zhu Hua, Hucheng beikao, 7.2a. 7 Mano, “Mindai no bukkyō to Minchō,” p. 246. Song Lian’s biographies of some of the Buddhist masters who took part in these convocations appear in JFZ 16.27a–29a. 8 For the history of ordination certificates through the Tang and Song, see Chikusa Masaaki, Chūgoku bukkyō shakaishi kenkyū, p. 17ff. There was one other significant edict on religion promulgated in 1373: an order that there should be only one Buddhist and one Daoist monastery per county, and that all Buddhist and Daoist monks should be made to reside in these two monasteries; Long Wenbin, Ming huiyao, p. 694. This unusually stringent edict was applicable only to the six prefectures nearest Nanjing, however, and even there seems not to have been enforced; see Yü, The Renewal of Buddhism in China, pp. 145–6. After the establishment of registries in 1381, the edict was interpreted to mean that one Buddhist and one Daoist monastery should be designated in each county to handle religious affairs on behalf of the state. 9 Regarding the purge of Hu Weiyong, see Charles Hucker, The Ming Dynasty, p. 41ff.; Edward Dreyer, Early Ming China, ch. 4. I have found no text linking the purge to Hongwu’s change in attitude toward Buddhism, yet the restrictive turn is consistent with the other shifts in policy following the destruction of Hu’s faction. 10 Two prefectural Buddhist registries were set up in Hangzhou in 1371 and “early Hongwu” respectively; Wulin fanzhi (1780), l.lb, 20b. A subprefectural Buddhist registry was set up in Quanzhou, Guangxi, in 1378; Xiangshan zhi (1682), 3.20b. County Buddhist and Daoist

Notes

206

registrars from the Yuan dynasty are mentioned in Jiaxing xianzhi (1685), 5.19b-20a; Liu Dunzhen, Beiping huguo si canji, p. 25; Cai Meibiao, Yuandai baihuabei jilu, p. 31. See also the conversation between the Hongwu emperor and a county registrar prior to the founding of the Ming, recorded in Zhou Tiandu, Tongsu bian (1751), 20.2b. 11 In 1382, county seats that also served as prefectural seats were excused from setting up a separate county Buddhist registry, their Buddhist affairs being delegated to the prefectural Buddhist registry; JFZ 2.7a. 12 For example, the subprefectural registry of Kaizhou, North Zhili, was founded in 1382 (Kaizhou zhi (1534), 2.4b); the county registry of Yanshi, Henan, in 1383 (Yanshi xianzhi (1504), shusi). 13 For example, Yichuan county in eastern Yan’an, Shaanxi, never had a Buddhist or Daoist registry; Yichuan xianzhi (1753), 2.3b. 14 For example, Caoxi tongzhi (1672), 3.13b. 15 Dinghai xianzhi (1711), 6.3a–b. 16 A case of a county registrar fixing land grades for taxation is noted for 1385 in JFZ 2.9b. 17 Chen Yuan, Mingji Dian Qian fojiao kao, p. 119. 18 JFZ 2.6b–7a. A second Hongwu edict on sect distinctions in 1391 substituted the term yujia (yoga, or tantrism) for jiao, though the content of the two categories remained the same; ibid., 2.15a–16b. The same text, however, expressly forbade esoteric yogic teachings. Vinaya (lü) monasteries were included in the Teaching sect; see, e.g, Dinghu shanzhi, 1.10a. 19 Da Ming huidian, 104.4a; JFZ 2.5b. 20 Yü, The Renewal of Buddhism in China, p. 168; Mano, “Mindai no bukkyō to Minchō,” pp. 270–1; see also JFZ 2.22b, 24b; Jingci sizhi (1888), 27.9b–10a. The term zhenji first appears in an imperial edict in 1372 requiring the land belonging to Nanjing’s Jiangshan (i.e., Linggu) Monastery to be recorded in a “foundation register” (zhenji pu); JFZ, 2.1b. 21 For example, the prominent Chan Monastery in Ningbo, Tiantong si: Tiantong sizhi (1811), 2.14b. 22 The gazetteer of Hangzhou’s Bianli Cloister makes a point of saying that the cloister was unique among the more important monasteries of Hangzhou in not having been amalgamated under the 1391 order; Bianli yuanzju (1830), Wu’s preface, 1a. 23 For example, Zhangzhou fuzhi (1613), 34.3b. 24 Parallels between early-Ming and early-Communist controls on Buddhism are striking. Both created monk-staffed bureaucracies to supervise Buddhism, intervened in landholding, closed monasteries and put them to other uses, concentrated monks in a few large centres, limited ordination, and restricted the free movement of monks. See Holmes Welch, Buddhism under Mao, pp. 29–31, 42–5, 73, 80–1, 117–18, 124–6. 25 In his study of early-Ming community schools, Matsumoto Yoshimi has estimated that 95 percent of these new schools were housed in converted monasteries; Chūgoku sonraku seido no shiteki kenkyū, p. 475. Most of these were monasteries closed in the amalgamation campaign. Matsumoto’s observation is attested in Hejian fuzhi (1540), 5.10b, which notes that rural schools had been made over from Buddhist and Daoist properties. 26 Da Ming huidian, 104.6b. 27 Yü, The Renewal of Buddhism in China, p. 146, reads the expression jiu e in the 1391 edict as “old quota” and argues that the emperor was banning monasteries that did not fall within the quota of one per county laid down in a 1373 edict. I read the expression as “old name plaque” and understand the Hongwu emperor to be banning recently constructed monasteries that did not have plaques; for corroboration, see Shaxian zhi (1701), 7.49a. The bestowal of name plaques in the Song is examined in Chikusa, Sōdai bukkyō shakaishi kenkyū, ch. 2. 28 Da Ming lü, ed. Huai Xiao, juan 4, pp. 46–7. 29 Xu Xueju, Guochao dianhui, 134.20b. 30 JFZ 2.24a–27b; the edict includes yet other restrictions besides those mentioned.

Notes

207

31 Guoxian zhi (1566), 8.9b. A similar observation from Fujian is made by the editor of Shaxian zhi (1701), 7.49a. 32 Da Ming huidian, 104.2a, 4b–5a. This regulation has been misinterpreted in the secondary literature variously as a quota on ordinations or as a quota on the total number of ordained monks in a county. It applied only to novices, who had to be between the ages of 14 and 20 and have parental permission for their noviciates. 33 The student quotas (40/30/20) were the same; see Dreyer, Early Ming China, p. 133. 34 Hongwu in 1387 restricted ordination to those under 20 years of age; in 1395, he defrocked all monks who failed a national examination; Da Ming huidian, 104.4a, 4b. 35 TJL, 3.35b; the attendant loss of state income is remarked on by Yu Jideng in his Diangu jiwen, p. 228. 36 Yu Jideng, Diangu jiwen, pp. 231–2. 37 Lin Xiyuan, Lin Ciya xiansheng wenji, 2.28b 38 Da Ming huidian, 104.8b. 39 Shimizu, “Mindai no jiden,” pp. 65–6; Wang Chongwu, “Ming Chengzu yu fojiao,” pp. 99– 101. A Fujian memorialist in the 1530s urged that the 100 mu limit be applied to every monastery in Jiangnan, noting that such a limit was needed to reduce the number of monks; Lin Xiyuan, Lin Ciya xiansheng wenji, 2.31b. 40 Chen Renxi, Huang Ming shifa lu, 39.41b–42a. 41 For example, Daming fuzhi (1506), 6.46a; Mianyang zhi (1531), 7.9b. 42 For example, Lushan xianzhi (1552), 4.19a. 43 For example, the Buddhist Registry of Haiyan county, Zhejiang, successfully petitioned to have the grounds of Zisheng Monastery restored in 1435; Haiyan xian tujing (1624), 3.60a. Registrar Liangjin of Hangzhou prefecture not only restored land that a monastery had lost in 1560 but erected a stele of protection outlining the history of its acquisition and its exact location; Yunju shengshui sizhi (1773), 2.3a. 44 Qingliang shanzhi (1661), 3.1b. The stipend was originally instituted in 1392; see Mano, “Mindai no bukkyō to Minchō,” p. 264. 45 Xiaozong shilu, 159.3b. 46 JFZ 16.10a. 47 For example, Hejian fuzhi (1540), 4.5b. 48 Neixiang xianzhi (1485), 2.19a, 4.70b. 49 Yingzhou zhi (1511), 4.15a. 50 Tianzhu shanzhi (1875), 8.57b. See also note 43 regarding Liangjin. 51 For example, both Ruian xianzhi (1555), 10.80b, and Shizhong shanzhi (1883), 3.7b, refer to the local registry as a long-abandoned site. The list of incumbents in the post of registrar in Jiujiang fuzhi (1592), 6.30a–33b, ends with the Chenghua era. In Changle xianzhi (1641), 5.58b, a record of temple restoration dated 1528 observes that as of that time there were no longer any Buddhist or Daoist registrars in the county. 52 For example, Guangshan xianzhi (1556), 3.5b, notes that “the county Buddhist registry of old was located in Baoxiang Monastery, but for a long time no official has been appointed.” 53 Songjiang fuzhi (1512), 18.1a. 54 Wujiang xianzhi (1561), 16.2b–26a. 55 Wang Chongwu, “Ming Chengzu yu fojiao,” pp. 93, 96. 56 Concerning the reputations of the Tianshun, Chenghua, and Hongzhi emperors with regard to Buddhism, see Noguchi Tetsurō, “Mindai shoki no bukkyōkai,” pp. 193–4, 212, 225. Regarding Hongzhi, see John Meskill, Ch’oe Pu’s Diary, p. 146. 57 Qingliang shanzhi (1661), 4.12a. 58 See, e.g, the complaints of Supervising Secretary Zhang Gu (1433) and of Minister of War Yu Qian and Censor Ye Luan (1450s), in Yu Jideng, Diangu jiwen, pp. 209, 216,227. 59 Da Ming huidian, 104.3b. 60 For example, Hasebe Yūkei, Min-Shin bukkyō shi kenkyū jōsetsu, p. 347.

Notes

208

61 Ho, The Ladder of Success in Imperial China, p. 33. 62 Xianzong shilu, 104.1a (1472, 5th month). 63 Wang Hongzhuan, Shan zhi, 4.22b. 64 Hu Lao, “Sengdao hudie shu,” in Chen Zilong, Ming jingshi wenbian, 19.9a; see also Yingzong shilu, 23.6b. 65 Xiaozong shilu, 145.5b. 66 Jianning fuzhi (1541), 19.58b. 67 In the 1530s, Huo Tao and Fang Xianfu (d. 1544) called for stricter controls on the clergy and the conversion of illegal monasteries into schools; Yu Jideng, Diangu jiwen, p. 303; Ming shi, pp. 5187, 5214. 68 Fuzheng, Hanshan dashi nianpu shuzhu, p. 18. 69 Da Ming huidian, 104.7b. For an earlier (1546) expression of concern about mass ordination in Beijing, see Daniel Overmyer, Folk Buddhist Religion, p. 171. 70 To mention a few examples: the great Vinaya Monastery in Hangzhou, Zhaoqing, started rebuilding in 1573 and resumed ordination in 1587 once the project was completed (Da zhaoqing lüsi zhi (1882), 1.15b); the ordination platform at the Chan Ordination Monastery in Suzhou, Kaiyuan, was restored in 1598 (Kaiyuan sizhi (1922), 14a); the monasteries on Putuo Island, where lay ordinations were popular, were restored starting in 1602 with the help of a grant from the Wanli emperor (Putuo shanzji (1611), 2.7a); and Hangzhou’s other great Vinaya Monastery, Jieshan yuan (Ordination Platform Cloister), was restored in 1613 and its ordination platform reestablished in 1628 (Longxing xiangfu jieshan sizhi (1894), 1.13a). 71 Canglang xiaozhi (1696), 1.19a. 72 The first of the empress dowager’s Buddhist projects outside the palace was to send 3,000 artisans to Wutai Mountain in 1579; Fuzheng, Hanshan dashi nianpu shuzhu, p. 45. 73 Johannes Prip-Møller, Chinese Buddhist Monasteries, p. 73. 74 Caoxi tongzhi (1672), 4.20a. 75 Ayuwang shanzhi (1619), 4b.8b–10a. 76 Li Zhi, Xu fen shu (1975), p. 97. 77 Bochi shanzhi (1920), 48b, from a stele by Tao Se. 78 In a similar vein, Kristofer Schipper (“Vernacular and Classical Rituals in Taoism,” p. 45) has observed that the imperial state could not control Daoism because it could not separate it from its social background in local cults: “the necessary economical and organizational foundations were never present to make such an institutionalization into an enduring historical reality.” 79 This argument is laid out more fully in my Praying for Power, pp. 311–25. 80 Songjiang Fuzhi (1512), 18.20a, though this is an early example.

8 Buddhism in the Chinese constitution: recording monasteries in North Zhili An early draft of this chapter was presented at the UCLA symposium, “From Late Imperial to Modern Chinese History: Views from the Eighteenth Century,” 14 November 1998. 1 My reflections on constitutionalism have grown in part from my reading of Scott Gordon, Controlling the State: Constitutionalism from Ancient Athens to Today.

Notes

209

2 G.W.F.Hegel, The Philosophy of History, p. 124. Hegel insisted that the notion of a constitution could be applied only to a state in which “all the members of the body politic have given their sanction” (p. 43); this is a constitution of a particular sort. 3 Philip Kuhn, Origins of the Modern Chinese State, pp. 1–2. In Soulstealers (p. 232), Kuhn is cautious about overreading curbs on imperial power as “constitutional,” since the emperor could not reliably “be held subject to the law, and there was no reliable civil protection for anyone who got in his way.” He suggests that officials were confined to invoking cultural rather than legal traditions to limit arbitrary imperial action; but perhaps this restricts our analysis too closely by anticipating that constitutional arrangements require European legal contexts. 4 Edward Farmer, Zhu Yuanzhang and Early Ming Legislation, p. 12. 5 Da Ming lü, pp. 46, 64. The fiscal implications of these statutes are suggested on p. 369 in the appended text, Wenxing tiaoli (Precedents offered in answer to judicial questions). 6 Ibid., p. 89. 7 Da Ming huidian, 104.2a–b. 8 Yongping fuzhi (1501), fanli. 2b. 9 Tongzhou zhilüe (1549), 12.4a. 10 Qiuxian zhi (1576), 3.14b. 11 Guangping xianzhi (1676), 1.28a; the Qing repetition of Hongwu’s ban on private founding is found in Da Qing huidian shili, 501.2b, 3b–4a, 7b–8a, 14b. 12 Nanhe xianzhi (1749), 3.7a. 13 Qingyuan xianzhi (1873), 18.siguan. 1a. 14 All these modes apply, for instance, to the monastic complex on Pan Mountain east of Beijing, which received an inscription from the Wanli emperor, a corvée exemption from the Minister of War in 1633, and several gifts and inscriptions from Kangxi, who also visited the mountain three times in 1675, 1678, and 1687; Panshan zhi (1696), 4.21b, 5.11a, 6.1b. 15 For example, Jigu Monastery, founded with the Wanli emperor’s blessing in 1579, was rebuilt without challenge in 1699 and listed in the 1739 county gazetteer; Tianjin xianzhi (1739), 8.21a. 16 For example, Xincheng xianzhi (1838), 18.3b–4a. 17 Brook, Praying for Power, pp. 94–6. 18 For example, Qingyuan xianzhi (1677), 12.13a; Luanzhou zhi (1810), 9.1a; Ningjin xianzhi (1679), 2.19a; Renqiu xianzhi (1762), 3.19a. 19 Jinzhou zhi (1690), 10.siguan. 4a; Qizhou zhi (1756), 2.32a. 20 On the ability of sectarians to organize a rebellion on the North China Plain, see Susan Naquin, Millenarian Rebellion in China. 21 Zhuozhou zhi (1373), 9.2b. 22 Wuqiang xian xinzhi (1694), 2.24b. 23 Shahe xianzhi (1757), 20.10b; also Gaocheng xianzhi (1698), 2.10a; Luanzhou zhi (1810), 9.1a. 24 Xinhe xianzhi (1679), 2.21a. 25 Lingshou xianzhi (1685), 2.6b. 26 Xinhe xianzhi (1679), 2.21a. 27 Huairou xianzhi (1604), 1.42b–43a. For another comment on the wasteful extravagance of monastic construction, see Nanpi xianzhi (1680), 3.14a. 28 For Lu Longqi’s interpretation of neo-Confucian ritual, see Kai-wing Chow, The Rise of Confucian Ritualism in Late Imperial China, pp. 52–3, 61–2. 29 Jiading xian xuzhi (1684), 2.19a. 30 Quoted from excerpts included in Chen Hongmou, Xueshi yigui bu, 4.38a–b. 31 Lingshou xianzhi (1685), fanli. 1b, 2.6b; the latter passage is repeated verbatim in the 1874 edition.

Notes

210

32 Leting xianzhi (1755), 12.28b. The term I translate as “genre” is ti, literally “body” or “structure,” but the sense seems to allow my translation. The author goes on to invoke other models that call for the inclusion of monasteries in written records. 33 Arguing that “the section on Buddhist and Daoist monasteries is intrinsic to the gazetteer genre,” a compiler makes the reference to Sima Qian to bolster his argument for inclusiveness as a fundamental principle of historical writing; Nanchang fuzjii (1588), fanli. 2a. 34 Luanzhou zhi (1810) 9.1a. 35 Fuping xianzhi (1874), 2.35a. 36 Renqiu xianzhi (1762), 3.19a. 37 Shahe xianzhi (1757), 20.10b. 38 Baodi xianzhi (1745), 15.1a. 39 Wanquan xianzhi (1742), 2.43b. Whether tanmiao should be divided into tan (altars) and miao (shrines), thereby placing monasteries in an appended third category, rather than an appended second, worried some compilers, e.g, Gaocheng xianzhi (1720), 2.9b. 40 Zhengding fuzhi (1762). A variation on this option was to append both shrines and monasteries to the chapter on schools, on the logic that schools also functioned as shrines to Confucius; this is done in Tianjin xianzhi (1739). 41 One compiler who does this labels his monasteries section explicitly as “old monasteries” (gusi); Renqiu xianzhi (1762), 3.16b. A rare variation, found in Yongqing xianzhi (1779), was to group all buildings by canton (xiang) without distinguishing which were in the sacrificial corpus. The design of this gazetteer curiously imitates Ye Chunji’s Huian handbook (see Chapter 2). 42 Jiading xianzhi (1673), fanli. 6a. 43 Leting xianzhi (1877), fanli. 5a, referring to Jiading xianzhi (1605) and Wugong xianzhi (1519). For the Siku quanshu editors’ criticism of Han Jun for relegating both “ancient sites” and “monasteries” to the closing chapter of his gazetteer, see Yongrong, Qinding siku quanshu zongmu, 74.23b. 44 Zhang Xuecheng, Wenshi tongyi, pp. 935–6, cited by Fang Chaoying in his biography of Lu Longqi in Hummel, Eminent Chinese of the Ch’ing Period, p. 547. David Nivison has noted that Zhang “was too broad-minded to ignore Buddhism or dismiss it as nonsense” and understood the best of Buddhist writings to be on a par with those of the great Confucians (The Life and Thought of Chang Hsüeh-ch’eng, pp. 76, 127). The expression “serving the great” (shi da) was more commonly used to describe the relationship of tributary states to the Chinese throne. 45 Qingxian zhi (1673), 2.2a. 46 Wuqiang xian xinzhi (1694), 2.24b. 47 Ningjin xianzhi (1679), 2.19b. 48 Qizhou zhi (1672), fanli. 1b. 49 For example, Gu’an xianzhi (1565), fanli. 3b. 50 Yongning xianzhi (1602), 47b. 51 Fengrun xianzhi (1570), 12.4b. 52 Linzhang xianzhi (1506), fanli. 1b. 53 Dingzhou zhi (1736), 2.36a. 54 Shulu xianzhi (1671), 2.20a 55 Xincheng xianzhi (1617), 12.1b. 56 Wanping xianzhi (1684), 1.47a. The term “Great Way” comes from the ancient Classic of Rites. When the Great Way was pursued, in James Legge’s translation of the passage (Li Ki, pp. 364–5), “a public and common spirit ruled all under the sky; they chose men of talent, virtue, and ability; their words were sincere and what they cultivated was harmony Thus men did not love their parents only, nor treat as children only their sons.” 57 Anping xianzhi (1687), 3.3a.

Notes

211

58 Luan zhi (1618), 6.58b. 59 Kdizhou zhi (1674), 2.33a. 60 zhengding fuzhi (1762), 9.52a. Regarding Qing elite anxieties concerning the “condominium” among the Three Teachings ideas and the possibility of a radical syncretism within popular religion, see my “Rethinking Syncretism,” pp. 25–35. 61 Fengrun xianzhi (1570), 12.4b–6a; Fengrun xianzhi (1755), 2.4a–7b. 62 Cangzhou zhi (1584) Juan 2; Cangzhou zhi (1680) Juan 3; Cangzhou zhi (1743) Juan 4. 63 Gucheng xianzhi (1594), juan 3; Gucheng xianzhi (1614), juan 2; Gucheng xianzhi (1727), juan 1. Similarly, the compiler of the second seventeenth-century gazetteer of Guangping county (Guangping xianzhi, 1676), 1.28a–b, disagreed with what the first (1608) had done by including considerable information about Buddhist monasteries. He leaves the first’s entries intact, but adds a diatribe against Buddhism and Daoism at the end of that section. The Republican-era compiler removes the commentary (Guangping xianzhi (1939), 9.7a– 10b). 64 Hejian fuzhi (1540) (see 7.6a for a negative comment on Buddhist practices); Hejian fuzhi (1615), juan 2; Hejian fuzhi (1677), Juan 6. 65 Zhending fuzhi (1549), juan 14; Zhending xianzhi (1612), juan 4; Zhengding fuzhi (1762), juan 9. The same trend may be found in the gazetteers of Guangzong: the 1598 edition omits monasteries (juan 3) and the 1693 edition lists them (juan 3). The 1624 gazetteer of Anci county lists only state temples and no monasteries (juan 4), whereas the 1677 edition includes Buddhist monasteries along with the dates they were founded in the list of all temples (juan 5); the next editor in 1749 retains the information, though he is careful to keep state temples and Buddhist monasteries in two separate sections (juan 3). 66 Xinhe xianzhi (1679), 2.21a. 67 The 1929 Xinhe gazetteer roughly replicates what appears in the 1679 edition. Unlike some of his contemporaries, the editor does not create a new “religion” (zongjiao) section in his gazetteer but shunts the monasteries into “ancient sites” (guji); Xinhe xianzhi (1929), shigu zhi, 15b–17b. 68 For example, Luan zhi (1548) lists no monasteries and Luan zhi (1618) does, whereas Luanzhou zhi (1810) packs them off to the end of the last juan. Similarly, Cizhou zhi (1526) puts siguan in the first chapter (1.19a–21a), whereas Cizhou zhi (1700) moves it to the end in juan 7. Yuanshi xianzhi (1642) has a siguan section (4.24a–b), whereas it disappears in Yuanshi xianzhi (1758) and is not reintroduced in Yuanshi xianzhi (1875). 69 Xincheng xianzhi (1617), 12.1a–b; Xincheng xianzhi (1675), juan 2; Xincheng xianzhi (1838), juan 10. 70 Daming fuzhi (1445), juan 4; Daming fuzhi (1506), 4.32b–35a; Daming fuzhi (1672), juan 15; Daming xianzhi (1790), juan 18; Daming fuzhi (1853), juan 6. 71 Wanxian zhi (1732), 2.27a–b. 72 Wanquan xianzhi (1742), 2.44a. 73 Nangong xianzhi (1559), 2.8a. 74 Qiuxian zhi (1576), 3.14b. 75 Guangping xianzhi (1676), 1.28b. 76 Nanpi xianzhi (1680), 3.14a. 77 Qizhou zhi (1756), 2.32a.

Conclusions: states of the field 1 Tilemann Grimm, “State and Power in Juxtaposition,” pp. 47, 49–50. 2 Paul Ricoeur, Time and Narrative, p. 125. 3 Aristotle, The Politics, III, ix.

Notes

212

4 E.g. William of Occam and Marsilius of Padua; see Brendan O’Leary, The Asiatic Mode of Production: Oriental Despotism, Historical Materialism, and Indian History, pp. 46–7. Niccolò Machiavelli, among others, relied on this Aristotelian distinction. 5 Thomas Hobbes, Leviathan, p. 133. 6 John Locke, Two Treatises of Government, p. 400. 7 Thomas Jeffreys, Natural and Civil History of the French Dominions in North and South America, cited in P.J.Marshall and Glyndwr Williams, The Great Map of Mankind, p. 247. 8 Montesquieu, The Spirit of the Laws, p. 63; see also Gregory Blue, “China and Western Social Thought in the Modern Period,” pp. 66–7, 87–8. 9 G.W.F.Hegel, The Philosophy of History, pp. 161, 105. 10 Ibid., p. 104. 11 Robert Fossier, The Cambridge Illustrated History of the Middle Ages, 1250–1520, p. 438, quoted in David Levine, At the Dawn of Modernity, p. 117. 12 For an analogous approach to the history of comparative political economy, and an inspiration for my own work, see R.Bin Wong, China Transformed, ch. 4. Wong builds his analysis on a comparison rooted in the eighteenth century; my procedure is to move the point of comparison back to the sixteenth. 13 Robert Fossier, Le moyen âge: le temps des crises, 1250–1520, p. 110, quoted in Immanuel Wallerstein, “The West, Capitalism, and the Modern World-System,” p. 48. 14 Levine, At the Dawn of Modernity, p. 109. Bin Wong (China Transformed, p. 87) phrases his concept of the absolutist state differently, as having developed “a measure of autonomy from the elites and common people they governed.” 15 Levine, At the Dawn of Modernity, p. 119. 16 Ibid., p. 125. 17 As Michael Szonyi has shown in his study of kinship organization in Ming Fujian (Practicing Kinship, p. 173), lineage leaders appropriated the lijia system for purposes other than those the state intended, and then used their conformity to that system as a political resource to compete with the leaders of other lineages. 18 My thoughts on these matters have been shaped by Philip Kuhn, Origins of the Modern Chinese State. 19 Gregory Blue, “China and Western Social Thought in the Modern Period,” pp. 66–7. 20 Shenzong shilu, 228.7b (1590, 10th month), 256.4a. (1593, 1st month).

Bibliography Gazetteers Note: Gazetteers are listed by province in their order in the Ming shi, juan 40–6. North Zhili Anping xianzhi (Gazetteer of Anping county). 1687. Baodi xianzhi (Gazetteer of Baodi county). 1745. Baoding fuzhi (Gazetteer of Baoding prefecture). 1886. Baoding xianzhi (Gazetteer of Baoding county). 1673. Cangzhou zhi (Gazetteer of Cangzhou subprefecture). 1584. Cangzhou zhi (Gazetteer of Cangzhou subprefecture). 1680. Cangzhou zhi (Gazetteer of Cangzhou subprefecture). 1743. Chengde fuzhi (Gazetteer of Chengde prefecture). 1887. Cizhou zhi (Gazetteer of Cizhou subprefecture). 1526. Cizhou zhi (Gazetteer of Cizhou subprefecture). 1700. Daming fuzhi (Gazetteer of Daming prefecture). 1445. Daming fuzhi (Gazetteer of Daming prefecture). 1506. Daming fuzhi (Gazetteer of Daming prefecture). 1672. Daming fuzhi (Gazetteer of Daming prefecture). 1853. Daming xianzhi (Gazetteer of Daming county). 1790. Dingzhou zhi (Gazetteer of Dingzhou subprefecture). 1736. Dingxing xianzhi (Gazetteer of Dingxing county). 1779. Dingxing xianzhi (Gazetteer of Dingxing county). 1893. Fengrun xianzhi (Gazetteer of Fengrun county). 1570. Fengrun xianzhi (Gazetteer of Fengrun county). 1755. Fuping xianzhi (Gazetteer of Fuping county). 1874. Gaocheng xianzhi (Gazetteer of Gaocheng county). 1698. Gaocheng xianzhi (Gazetteer of Gaocheng county). 1720. Gaoyang xianzhi (Gazetteer of Gaoyang county). 1730. Gu’an xianzhi (Gazetteer of Gu’an county). 1565. Guangping xianzhi (Gazetteer of Guangping county). 1676. Guangping xianzhi (Gazetteer of Guangping county). 1939. Gucheng xianzhi (Gazetteer of Gucheng county). 1594. Gucheng xianzhi (Gazetteer of Gucheng county). 1614. Gucheng xianzhi (Gazetteer of Gucheng county). 1727. Hejian fuzhi (Gazetteer of Hejian prefecture). 1540. Hejian fuzhi (Gazetteer of Hejian prefecture). 1615. Huairou xianzhi (Gazetteer of Huairou county). 1604. Jifu tongzhi (Gazetteer of the Metropolitan Region). 1735.

Bibliography

214

Jifu tongzhi (Gazetteer of the Metropolitan Region). 1910. Jingzhou zhi (Gazetteer of Jingzhou subprefecture). 1612. Jinzhou zhi (Gazetteer of Jinzhou subprefecture). 1690. jizhou zhi (Gazetteer of Jizhou subprefecture). 1524. Kaizhou zhi (Gazetteer of Kaizhou subprefecture). 1674. Leting xianzhi (Gazetteer of Leting county). 1755. Leting xianzhi (Gazetteer of Leting county). 1877. Lingshou xianzhi (Gazetteer of Lingshou county). 1685. Linyu xianzhi (Gazetteer of Linyu county). 1756. Luan zhi (Gazetteer of Luanzhou subprefecture). 1548. Luan zhi (Gazetteer of Luanzhou subprefecture). 1618. Luanzhou zhi (Gazetteer of Luanzhou subprefecture). 1810. Mancheng xianzhi (Gazetteer of Mancheng county). 1713. Nangong xianzhi (Gazetteer of Nangong county). 1559. Nangong xianzhi (Gazetteer of Nangong county). 1831. Nanhe xianzhi (Gazetteer of Nanhe county). 1749. Nanpi xianzhi (Gazetteer of Nanpi county). 1680. Ninghe xianzhi (Gazetteer of Ninghe county). 1779. Ningjin xianzhi (Gazetteer of Ningjin county). 1679. Panshan zhi (Gazetteer of Pan Mountain). 1696. Pingshan xianzhi (Gazetteer of Pingshan county). 1854. Pingxiang xianzhi (Gazetteer of Pingxiang county). 1878. Qingxian zhi (Gazetteer of Qing county). 1673. Qingyuan xianzhi (Gazetteer of Qingyuan county). 1677. Qingyuan xianzhi (Gazetteer of Qingyuan county). 1873. Qiuxian zhi (Gazetteer of Qiu county). 1576. Qizhou zhi (Gazetteer of Qizhou subprefecture). 1756. Renqiu xianzhi (Gazetteer of Renqiu county). 1762. Shahe xianzhi (Gazetteer of Shahe county). 1757. Shulu xianzhi (Gazetteer of Shulu county). 1671. Shunde fuzhi (Gazetteer of Shunde prefecture). 1750. Shuntian fuzhi (Gazetteer of Shuntian prefecture). 1886. Tangxian zhi (Gazetteer of Tang county). 1878. Tianjin xianzhi (Gazetteer of Tianjin county). 1739. Tongzhou zhilüe (Abbreviated gazetteer of Tongzhou subprefecture). 1549. Wangdu xianzhi (Gazetteer of Wangdu county). 1771. Wanping xianzhi (Gazetteer of Wanping county). 1684. Wanquan xianzhi (Gazetteer of Wanquan county). 1742. Wanxian zhi (Gazetteer of Wan county). 1732. Weizhou zhi (Gazetteer of Weizhou subprefecture). 1877. Wuan xianzhi (Gazetter of Wuan county). 1547. Wuqiang xian xinzhi (New gazetteer of Wuqiang county). 1694. Xincheng xianzhi (Gazetteer of Xincheng county). 1617. Xincheng xianzhi (Gazetteer of Xincheng county). 1675. Xincheng xianzhi (Gazetteer of Xincheng county). 1838. Xinhe xianzhi (Gazetteer of Xinhe county). 1679.

Bibliography

215

Xinhe xianzhi (Gazetteer of Xinhe county). 1929. Yizhou zhi (Gazetteer of Yizhou subprefecture). 1645. Yizhou zhi (Gazetteer of Yizhou subprefecture). 1747. Yongnian xianzhi (Gazetteer of Yongnian county). 1758. Yongnian xianzhi (Gazetteer of Yongnian county). 1877. Yongning xianzhi (Gazetteer of Yongning county). 1602. Yongping fuzhi (Gazetteer of Yongping prefecture). 1517. Yongqing xianzhi (Gazetteer of Yongqing county). 1779. Yuanshi xianzhi (Gazetteer of Yuanshi county). 1642. Yuanshi xianzhi (Gazetteer of Yuanshi county). 1758. Yuanshi xianzhi (Gazetteer of Yuanshi county). 1875. Zhaozhou zhi (Gazetteer of Zhaozhou subprefecture). 1498. Zhendingfuzhi (Gazetteer of Zhending prefecture). 1549. Zhending xianzhi (Gazetteer of Zhending prefecture). 1612. Zhengding fuzhi (Gazetteer of Zhengding prefecture). 1762. Zhuozjiou zhi (Gazetteer of Zhuozhou subprefecture). 1373. ZJiuozjiou zhi (Gazetteer of Zhuozhou subprefecture). 1875. Zunhua zhouzhi (Gazetteer of Zunhua subprefecture). 1676. South Zhili Bochi shanzhi (Gazetteer of Bochi Mountain). 1920. Changshu xianzhi (Gazetteer of Changshu county). 1537. Changzhou xianzhi (Gazetteer of Changzhou county). 1571. Dengwei sheng’en sizhi (Gazetteer of Sheng’en Monastery on Dengwei Mountain). 1644. Fengtai xianzhi (Gazetteer of Fengtai county). 1814. Gaochun xianzhi (Gazetteer of Gaochun county). 1881. Gaoshan zhi (Gazetteer of Gaoshan subprefecture). 1877. Gusu zhi (Gazetteer of Suzhou prefecture). 1506. Hanshan sizhi (Gazetteer of Hanshan Monastery). 1911. Huizhou fuzjii (Gazetteer of Huizhou prefecture). 1502. Huizhou fuzhi (Gazetteer of Huizhou prefecture). 1566. Jiading xianzhi (Gazetteer of Jiading county). 1605. Jiading xianzhi (Gazetteer of Jiading county). 1673. Jiading xian xuzhi (Supplementary gazetteer of Jiading county). 1684. Jiangdu zhi (Gazetteer of Jiangdu county). 1599. Jiangning fuzhi (Gazetteer of Jiangning prefecture). 1880. Jiangyin xianzhi (Gazetteer of Jiangyin county). 1840. Jiangyin xianzhi (Gazetteer of Jiangyin county). 1878. Jinling fancha zhi (Gazetteer of Buddhist monasteries of Nanjing). 1607. Kaiyuan sizhi (Gazetteer of Kaiyuan Monastery). 1922. Louxian zhi (Gazetteer of Lou county). 1788. Niushou shanzhi (Gazetteer of Niushou Monastery). 1597. Pingshan tang tuzhi (Illustrated gazetteer of Pingshan Hall). 1705. Qingliang shanzhi (Gazetteer of Qingliang Monastery). 1661.

Bibliography

216

Shanghai xianzhi (Gazetteer of Shanghai county). 1588. Songjiang fuzhi (Gazetteer of Songjiang prefecture). 1512. Songjiang fuzhi (Gazetteer of Songjiang prefecture). 1630. Suzhou fuzhi (Gazetteer of Suzhou prefecture). 1692. Suzhou fuzhi (Gazetteer of Suzhou prefecture). 1759. Reprint, 1883. Taibo meili zhi (Gazetteer of Taibo and Meili cantons). 1897. Taicang zhi (Gazetteer of Taicang county). 1548. Reprint, 1629. Tongxiang xianzhi (Gazetteer of Tongxiang county). 1887. Wujiang xianzhi (Gazetteer of Wujiang county). 1561. Wujiang zhi (Gazetteer of Wujiang county). 1488. Wujin xianzhi (Gazetteer of Wujin county). 1605. Wujin yanghu xianzhi (Gazetteer of Wujin and Yanghu counties). 1886. Wuxi xianzhi (Gazetteer of Wuxi county). 1753. Wuxian zhi (Gazetteer of Wu county). 1642. Xiuning xianzhi (Gazetteer of Xiuning county). 1693. Xiuning xianzhi (Gazetteer of Xiuning county). 1815. Yangzhou fuzhi (Gazetteer of Yangzhou prefecture). 1733. Shandong Anqiu xianzhi (Gazetteer of Anqiu county). 1589. Jinan fuzhi (Gazetteer of Jinan prefecture). 1840. Jining zhouzhi (Gazetteer of Jining subprefecture). 1672. Shenxian zhi (Gazetteer of Shen county). 1548. Wenshang xianzhi (Gazetteer of Wenshang county). 1717. Wucheng xianzhi (Gazetteer of Wucheng county). 1549. Yanzhou fuzhi (Gazetteer of Yanzhou prefecture). 1596. Yanzhou fuzhi (Gazetteer of Yanzhou prefecture). 1613. Zhucheng xianzhi (Gazetteer of Zhucheng county). 1764. Shanxi Changzi xianzhi (Gazetteer of Changzi county). 1513. Dingxiang xianzhi (Gazetteer of Dingxiang county). 1727. Jiaocheng xianzhi (Gazetteer of Jiaocheng county). 1882. Lucheng xianzhi (Gazetteer of Lucheng county). 1625. Henan Guangshan xianzhi (Gazetteer of Guangshan county). 1556. Guide zhi (Gazetteer of Guide prefecture). 1545. Guide fuzhi (Gazetteer of Guide prefecture). 1753. Henan fuzhi (Gazetteer of Henan prefecture). 1695. Linzhang xianzhi (Gazetteer of Linzhang county). 1506. Lushan xianzhi (Gazetteer of Lushan county). 1552. Mianyang zhi (Gazetteer of Mianyang subprefecture). 1531.

Bibliography

217

Nanyang fuzhi (Gazetteer of Nanyang prefecture). 1577. Neixiang xianzhi (Gazetteer of Neixiang county). 1485. Weihui fuzJii (Gazetteer of Weihui prefecture). 1788. Xiangcheng xianzhi (Gazetteer of Xiangcheng county). 1911. Zhangde fuzhi (Gazetteer of Zhangde prefecture). 1522. Shaanxi Yichuan xianzhi (Gazetteer of Yichuan county). 1753. Jiangxi Anren xianzhi (Gazetteer of Anren county). 1543. Dehua xianzhi (Gazetteer of Dehua county). 1872. Ji‘an fuzhi (Gazetteer of Ji’an prefecture). 1648. Jianchang fuzhi (Gazetteer of Jianchang prefecture). 1517. Jing’an xianzhi (Gazetteer of Jing’an county). 1565. Nanchang fuzhi (Gazetteer of Nanchang prefecture). 1588. Raozhou fuzhi (Gazetteer of Raozhou prefecture). 1511. Ruian xianzhi (Gazetteer of Ruian county). 1555. Ruichang xianzhi (Gazetteer of Ruichang county). 1568. Ruijin xianzhi (Gazetteer of Ruijin county). 1542. Shizhong shanzhi (Gazetteer of Shizhong Monastery). 1883. Yuanzhou fuzhi (Gazetteer of Yuanzhou prefecture). 1874. Huguang Anhua xianzhi (Gazetteer of Anhua county). 1543. Changde fuzhi (Gazetteer of Changde prefecture). 1538. Cili xianzhi (Gazetteer of Cili county). 1574. Dangyang xianzhi (Gazetteer of Dangyang county). 1866. Hengzhou fuzhi (Gazetteer of Hengzhou prefecture). 1593. Huguang tujing zhishu (Illustrated gazetteer of Huguang province). 1522. Jiayu xianzhi (Gazetteer of Jiayu county). 1449. Lianzhou fuzhi (Gazetteer of Lianzhou prefecture). 1637. Liuyang xianzhi (Gazetteer of Liuyang county). 1561. Xiangtan xianzhi (Gazetteer of Xiangtan county). 1554. Xiangyin xianzhi (Gazetteer of Xiangyin county). 1554. Xingguo zhouzhi (Gazetteer of Xingguo subprefecture). 1554. Zhejiang Ayuwang shanzhi (Gazetteer of King Asoka Mountain). 1619. Bianli yuanzhi (Gazetteer of Bianli Cloister). 1765. Reprint, 1830. Da zhaoqing lüsi zhi (Gazetteer of Great Zhaoqing Vinaya Monastery). 1882. Dinghai xianzhi (Gazetteer of Dinghai county). 1563.

Bibliography

218

Du Bai erhu quanshu (Complete book of Du and Baiyang lakes). 1805. Haiyan xian tujing (Illustrated gazetteer of Haiyan county). 1624. Haiyan xianzhi (Gazetteer of Haiyan county). 1876. Hangzhou fuzhi (Gazetteer of Hangzhou prefecture). 1475. HangzJiou fuzhi (Gazetteer of Hangzhou prefecture). 1922. Huzhou fuzhi (Gazetteer of Huzhou prefecture). 1491. Huzhou fuzhi (Gazetteer of Huzhou prefecture). 1542. Huzhou fuzhi (Gazetteer of Huzhou prefecture). 1649. Jiashan xianzhi (Gazetteer of Jiashan county). 1894. Jiaxing fuzhi (Gazetteer of Jiaxing prefecture). 1681. Jiaxing fuzhi (Gazetteer of Jiaxing prefecture). 1883. Jiaxing xianzhi (Gazetteer of Jiaxing county). 1637. Jiaxing xianzhi (Gazetteer of Jiaxing county). 1685. Jiaxing xianzhi (Gazetteer of Jiaxing county). 1909. Jingci sizhi (Gazetteer of Jingci Monastery). 1888. Jinhua xianzhi (Gazetteer of Jinhua county). 1598. Longquan xianzhi (Gazetteer of Longquan county). 1878. Longxing xiangfu jietan sizhi (Gazetteer of Longxing, Xiangfu, and Jietan Monasteries). 1894. Ningbo fuzhi (Gazetteer of Ningbo prefecture). 1560. Putuo shanzhi (Gazetteer of Putuo Monastery). 1611. Qiantang xianzhi (Gazetteer of Qiantang county). 1718. Quzhou fuzhi (Gazetteer of Quzhou prefecture). 1623. Renhe xianzhi (Gazetteer of Renhe county). 1687. Shang tianzhu jiangsi zhi (Gazetteer of Upper Tianzhu Teaching Monastery). 1646. Reprint, 1897. Taizhou fuzhi (Gazetteer of Taizhou prefecture). 1722. Tiantong sizhi (Gazetteer of Tiantong Monastery). 1811. Tianzhu shanzhi (Gazetteer of Tianzhu Monastery). 1875. Wenzhou fuzhi (Gazetteer of Wenzhou prefecture). 1605. Wucheng xianzhi (Gazetteer of Wucheng county). 1638. Wulin fanzhi (Gazetteer of the Buddhist monasteries in Hangzhou). 1615. Reprint, 1780. Xiaofeng xianzhi (Gazetteer of Xiaofeng county). 1877. Xiushui xianzhi (Gazetteer of Xiushui county). 1596. Xiushui xianzhi (Gazetteer of Xiushui county). 1685. Yinxian zhi (Gazetteer of Yin county). 1788. Yongjia xianzhi (Gazetteer of Yongjia county). 1566. Yuhang xianzhi (Gazetteer of Yuhang county). 1808. Yunju shengshui sizhi (Gazetteer of Yunju-Shengshui Monastery). 1773. Fujian Fangguang yanzhi (Gazetteer of Fangguang Monastery). 1612. Fujian tongzhi (Gazetteer of Fujian province). 1871. Funing zhouzhi (Gazetteer of Funing subprefecture). 1593.

Bibliography

219

Fuqing xianzhi (Gazetteer of Fuqing county). 1747. Fuzhou fuzhi (Gazetteer of Fuzhou prefecture). 1613. Fuzhou fuzhi (Gazetteer of Fuzhou prefecture). 1876. Gutian xianzhi (Gazetteer of Gutian county). 1606. Huian xianzhi (Gazetteer of Huian county). 1530. Huian xianzhi (Gazetteer of Huian county). 1803. Reprint, 1936. Jianning fuzhi (Gazetteer of Jianning prefecture). 1493. Longxi xianzhi (Gazetteer of Longxi county). 1762. Luochuan zhi (Gazetteer of Luochuan county). 1545. Ninghua xianzhi (Gazetteer of Ninghua county). 1684. Ninghua xianzhi (Gazetteer of Ninghua county). 1869. Quanzhou fuzhi (Gazetteer of Quanzhou prefecture). 1613. Quanzhou fuzhi (Gazetteer of Quanzhou prefecture). 1763. Shouning daizhi (Provisional gazetteer of Shouning county). 1637. Reprint, 1983. Tongan xianzhi (Gazetteer of Yongan county). 1798. Xianju xianzhi (Gazetteer of Xianju county). 1612. Yanping fuzhi (Gazetteer of Yanping prefecture). 1526. Yongfu xianzhi (Gazetteer of Yongfu county). 1612. Zhangping xianzhi (Gazetteer of Zhangping county). 1935. Zhangzhou fuzhi (Gazetteer of Zhangzhou prefecture). 1613. Zhangzhou fuzhi (Gazetteer of Zhangzhou prefecture). 1877. Guangdong Caoxi tongzhi (Comprehensive gazetteer of Cao’s stream [Nanhua Monastery]). 1672. Chaozhou fuzhi (Gazetteer of Chaozhou prefecture). 1547. Dianbai xianzhi (Gazetteer of Dianbai county). 1825. Dinghu shanzi (Gazetteer of Dinghu Mountain). 1717. Foshan zJiongyi xiangzhi (Gazetteer of Zhongyi canton, Foshan county). 1921. Gaoyao xianzhi (Gazetteer of Gaoyao county). 1826. Guangxiao sizhi (Gazetteer of Guangxiao Monastery). 1935. Guishan xianzhi (Gazetteer of Guishan county). 1783. Lingshan xianzhi (Gazetteer of Lingshan county). 1733. Panyu xianzhi (Gazetteer of Panyu county). 1871. Qiongshan xianzhi (Gazetteer of Qiongshan county). 1917. Qiongzhou fuzhi (Gazetteer of Qiongzhou prefecture). 1619. Shunde xianzhi (Gazetteer of Shunde county). 1585. Shunde xianzhi (Gazetteer of Shunde county). 1996. Xinhui xianzhi (Gazetteer of Xinhui county). 1841. Xining xianzhi (Gazetteer of Xining county). 1830. Yazhou zhi (Gazetteer of Yazhou subprefecture). 1900. Yongan xianzhi (Gazetteer of Yongan county). 1586. Zengcheng xianzhi (Gazetteer of Zengcheng county). 1820. Zhaoqing fuzhi (Gazetteer of Zhaoqing prefecture). 1588. Zhaoqing fuzhi (Gazetteer of Zhaoqing prefecture). 1876.

Bibliography

220

Guangxi Xiangshan zhi (Gazetteer of Xiang Mountain). 1682. Books and articles Ai Nanyang. Tianyongzi ji (Collected writings of heaven’s hired hand). 1836. Reprint, Taibei, 1980. An Pingqiu and Zhang Peiheng, eds. Zhongguo jinshu jianshi (Brief history of book prohibition in China). Shanghai: Shanghai wenyi chubanshe, 1992. Anderson, Benedict. Imagined Communities: Reflections on the Origin and Spread of Nationalism. Rev. ed., London: Verso, 1991. Blackstone, William. Commentaries on the Laws of England (1765–69). Reprint, Chicago: University of Chicago Press, 1979. Blue, Gregory. “China and Western Social Thought in the Modern Period.” In China and Historical Capitalism: Genealogies of Sinological Knowledge, ed. Timothy Brook and Gregory Blue, pp. 57–109. Cambridge: Cambridge University Press, 1999. Bray, Francesca. The Rice Economies: Technology and Development in Asian Societies. Berkeley, CA: University of California Press, 1986. ——. Sicience and Civilisation in China: Agriculture, vol. VI, pt. 2 of Joseph Needham, Science and Civilisation in China. Cambridge: Cambridge University Press, 1984. Brook, Timothy. “Capitalism and the Writing of Modern History in China.” In China and Historical Capitalism: Genealogies of Sinological Knowledge, co-ed. Gregory Blue, pp. 110– 57. Cambridge: Gambridge University Press, 1999. ——.“Confucianism.” In Censorship: A World Encyclopedia, ed. Derek Jones, pp. 570–2. London: Fitzroy Dearborn, 2001. ——. The Confusions of Pleasure: Commerce and Culture in Ming China. Berkeley, CA: University of California Press, 1998. ——. “Edifying Knowledge: The Building of School Libraries in Ming Chia.” Late Imperial China 17:1 (June 1996), pp. 93–119. ——“Family Continuity and Cultural Hegemony: The Gentry of Ningbo, 1368–1911.” In Chinese Local Elites and Patterns of Dominance, ed. Joseph Esherick and Mary Rankin, pp. 27–50. Stanford, CA: Stanford University Press, 1990. ——. “Funerary Ritual and the Building of Lineages in Late Imperial China.” Harvard Journal of Asiatic Studies 49:2 ( December 1989), pp. 465–99. ——. “Geografia e cartografia.” Storia della Scienza, vol. 2, pp. 493–502. Rome: Istituto della Enciclopedia Italiana, 2001. ——. Geographical Sources of Ming-Qing History. Ann Arbor, MI: Genter for Chinese Studies, University of Michigan, 1988; 2nd edn., 2002. ——. “Native Identity under Alien Rule: Local Gazetteers of the Yuan Dynasty.” In The Uses of Literacy, 1200–1330, ed. Richard Britnell, pp. 235–45. London: Boydell and Brewer, 1997. ——. Praying for Power: Buddhism and the Formation of Gentry Society in Late-Ming China. Cambridge, MA: Council on East Asian Studies, Harvard University, 1993. ——. “Profit and Righteousness in Chinese Economic Culture.” In Culture and Economy: The Shaping of Capitalism in Eastern Asia, co-ed. Hy Van Luong, pp. 27–44. Ann Arbor, MI: University of Michigan Press, 1997. ——. “Rethinking Syncretism: The Unity of the Three Teachings and their Joint Worship in Late Imperial China.” Journal of Chinese Religions 21 (Fall 1993), pp. 13–44. Brook, Timothy, and Gregory Blue, eds. China and Historical Capitalism: Genealogies of Sinological Knowledge. Cambridge: Cambridge University Press, 1999.

Bibliography

221

Brook, Timothy, and B.Michael Frolic, eds. Civil Society in China. Armonk, NY: M.E. Sharpe, 1997. Brook, Timothy, and Hy Van Luong, eds. Culture and Economy: The Shaping of Capitalism in Eastern Asia. Ann Arbor, MI: University of Michigan Press, 1997. Buck, John Lessing. Land Utilization in China. Shanghai: Commercial Press, 1937. Cai Meibiao. Yuandai baihuabei jilu (Transcripts of Yuan-period inscriptions in vernacular Chinese). Beijing: Kexue chubanshe, 1955. Cao Wanru et al., eds. Zhongguo gudai dituji: Mingdai (Ancient Chinese maps: Ming dynasty). Beijing: Wenwu chubanshe, 1994. Cartier, Michel. Une reforme locale en Chine au XVIe siecle: Hai Rui a Chun’an, 1558–1562. Paris: Mouton, 1973. Chambers, Douglas. “A Catalogue of the Library Bishop Lancelot Andrewes.” Transactions of the Cambridge Bibliographical Society 5:2 (1970). Chan, Anita, Richard Madsen, and Jonathan Unger. Chen Village. Berkeley, CA: University of California Press, 1984. Chan, Wing-tsit, trans. Reflections on Things at Hand: A Neo-Confucian Anthology. New York: Columbia University Press, 1967. Cheek, Timothy. “Contracts and Ideological Control in Village Administration: Tensions in the ‘Village Covenant’ System in Late Imperial China.” Unpublished paper. Chen Hongmou, ed. Xueshiyigui (Rules bequeathed to scholars); Xueshiyigui bu (Supplement to Rules bequeathed to scholars). Reprint, 1879. Chen Longzheng. Jiting quanshu (Complete writings of Chen Jiting). 1665. Chen Menglei, ed. Gujin tushujicheng (Compendium of texts ancient and modern). 1725. Chen Renxi. Huang Ming shifa lu (A compendium of contemporary regulations of the Ming dynasty). 1630. Chen Yuan. Mingji Dian-Qian fojiao kao (Studies of Buddhism in Guizhou and Yunnan in the Ming). 1940. Reprint, Beijing: Zhonghua shuju, 1962. Chen Zilong, ed. Ming jingshi wenbian (Essays on statecraft of the Ming dynasty). 1638. Reprint, Beijing: Zhonghua shuju, 1962. Chen Zugui. Dao (Rice). Beijing: Zhonghua shuju, 1958. Chia, Lucille. Printing for Profit: The Commercial Publishers of Jianyang, Fujian (llth-17th Centuries). Cambridge, MA: Harvard University Asia Center, 2002. Chikusa Masaaki. Chugoku bukkyo shakaishi kenkyū (Studies in the social history of Chinese Buddhism). Kyoto: Tōhōsha, 1982. ——. “Sodai funji kō kō” (Study of tomb monasteries in the Song dynasty). Tōyō gakuho 61:1–2 (1979), pp. 35–66. Chow, Kai-wing. The Rise of Confucian Ritualism in Late Imperial China. Stanford, CA: Stanford University Press, 1994. Ch’ü, T’ung-tsu. Local Government in China under the Ch’ing. Cambridge, MA: Harvard University Press, 1962. Clunas, Craig. Chinese Furniture.London: Victoria and Albert Museum, 1988. ——. “Oriental Antiquities/Far Eastern Art.” positions 2:2 (Summer 1994), pp. 318–55. Cole, Alan. Mothers and Sons in Chinese Buddhism. Stanford, CA: Stanford University Press, 1998. Da Ming huidian (Statutory precedents of the Ming dynasty). 1588. Da Ming lü (The Ming code), ed. Huai Xiao. Beijing: Falü chubanshe, 1999. Da Ming yitong zhi (Unity gazetteer of the Ming dynasty), ed. Li Xian. 1461. Da Qing huidian shili (Statutory precedents of the Qing dynasty, with cases). Reprint, Taibei: Wenhai chubanshe, 1992. Da Qing lichao shilu (Veritable records of the reigns of the Qjng dynasty). Reprint, Taibei: Hualian chubanshe, 1964.

Bibliography

222

Da Qing lüli (The Qjng code with substatutes), ed. Tian Tao and Zheng Qin. Beijing: Falü chubanshe, 1999. Daoan. Dazangjing diaoke shihua (Printing history of the Tripitaka). Beijing: Zhonghua shuju, 1978. Dardess, John. “The Cheng Communal Family: Social Organization and NeoConfucianism in Yuan and Early Ming China.” Harvard Journal of Asiatic Studies 34 (1974), pp. 7–52. de Bary, W.Theodore. “Chinese Despotism and the Confucian Ideal.” In Chinese Thought and Institutions, ed. John K.Fairbank, pp. 163–203. Chicago: University of Chicago Press, 1957. de Groot, J.J.M. The Religious System of China. Reprint, Taibei: Literature House, 1964. Dennerline, Jerry. “Fiscal Reform and Local Control: The Gentry-Bureaucratic Alliance Survives the Conquest.” In Conflict and Control in Late Imperial China, ed. Frederic Wakeman and Carolyn Grant, pp. 86–120. Berkeley, CA: University of California Press, 1975. Dong Fu. Erlou jilüe (An abbreviated record of the city of two towers). 1720. Dore, R.P. Education in Tokugawa Japan. London: Routledge & Kegan Paul, 1965. Dreyer, Edward. Early Ming China: A Political History, 1355–1435. Stanford, CA: Stanford University Press, 1982. du Halde, Jean-Baptiste. The General History of China, trans. Richard Brookes. London: Watts, 1736. Durkheim, Emile. Review of Maurice Courant, “Les associations en Chine” (Annales des sciences politiques 1899:1, p. 68ff.). L’année sociologique 3 (1899), p. 355. Ebrey, Patricia, ed. Chinese Civilization: A Sourcebook. New York: Free Press, 1993. Eisenstein, Elizabeth. The Printing Press as an Agent of Change. Cambridge: Cambridge University Press, 1980. Elman, Benjamin. A Cultural History of the Civil Examinations in Late Imperial China. Berkeley, CA: University of California Press, 2000. ——. From Philosophy to Philology: Intellectual and Social Aspects of Change in Late Imperial China. Cambridge, MA: Council on East Asian Studies, Harvard University, 1984. Elvin, Mark. “Market Towns and Waterways: The County of Shangh-hai from 1480 to 1910.” In The City in Late Imperial China, ed. G.W.Skinner, pp. 441–73. Stanford, CA: Stanford University Press, 1977. Esherick, Joseph, and Mary Rankin, eds. Chinese Local Elites and Patterns of Dominance. Berkeley, CA: University of California Press, 1990. Fang, Chaoying. “Lu Lung-ch’i.” In Eminent Chinese of the Ch’ing Period, ed. Arthur Hummel, pp. 547–8. Washington, DC: Government Printing Office, 1943. Farmer, Edward. Early Ming Government: The Evolution of Dual Capitals. Cambridge, MA: Harvard University Asian Research Center, 1976. ——. Zhu Yuanzhang and Early Ming Legislation: The Reordering of Chinese Society following the Era of Mongol Rule. Leiden: Brill, 1995. Fei, Hsiao-t’ung. Peasant Life in China: A Field Study of Country Life in the Yangtze Valley. London: Kegan Paul, Trench, Trubner, 1939. Fossier, Robert. Le moyen dge: le temps des crises, 1250–1520. Paris: Armand Colin, 1983. ——, ed. The Cambridge Illustrated History of the Middle Ages, 1250–1520. Cambridge: Cambridge University Press, 1987. Franke, Wolfgang. An Introduction to the Sources of Ming History. Kuala Lumpur: University of Malaya Press, 1968. Fu Yiling. “Lun Ming-Qing shidai Fujian tudi maimai qiyue zhong de ‘yinzhu’” (On “silverowners” in land purchase contracts from Fujian in the Ming-Qing period). Dousou 52:1 (1983). Fuzheng. Hanshan dashi nianpu shuzu (Annotated chronology of the life of Master Hanshan). Taibei: Zhenmeishan, 1967. Gallagher, Louis, trans. China in the Sixteenth Century: The Journals of Matthew Ricci: 1583– 1610. New York: Random House, 1953. Gao Jie. Houhu zhi (Black Lake records). 1621. First published 1611.

Bibliography

223

Geng Ju. Changshu xian shuili quanshu (Complete handbook of water resources in Changshu county). 1606. Goodrich, L.Carrington. The Literary Inquisition of Ch’ien-Lung. Baltimore, MD: Waverly Press, 1935. Goodrich, L.Carrington, and Chaoying Fang, eds. Dictionary of Ming Biography, 2 vols. New York: Columbia University Press, 1976. Gordon, Scott. Controlling the State: Constitutionalism from Ancient Athens to Today. Cambridge, MA: Harvard University Press, 1999. Grendler, Paul. The Roman Inquisition and the Venetian Press, 1540–1605. Princeton, NJ: Princeton University Press, 1977. Grimm, Tilemann. “State and Power in Juxtaposition: An Assessment of Ming Despotism.” In The Scope of State Power in China, ed. S.R.Schram, pp. 27–50. London: School of Oriental and African Studies, University of London, 1985. Grist, David H. Rice, 4th edn. London: Longman’s, 1965. Grove, Linda, and Christian Daniels, eds. State and Society in China: Japanese Perspectives on Ming-Qing Social and Economic History. Tokyo: University of Tokyo Press, 1984. Gu Yanwu. Rizhi lu jishi (A record of knowledge gained day by day, with commentaries). 1834. Reprint, Shanghai: Shanghai guji chubanshe, 1984. ——. Tianxia junguo libing shu(The strengths and weaknesses of the various regions of the realm). 1662. Reprint, Shanghai: Shangwu yinshuguan, 1936; Kyoto: Chumon shuppansha, 1975. Gu Yu, ed. Gui Qian chidie (Model letters of Gui Youguang and Qian Qianyi). 1699. Reprint, Zhigu tang, 1911. Guy, Kent. The Emperor’s Four Treasuries: Scholars and the State in the Late Ch’ien-lung Era. Cambridge, MA: Council on East Asian Studies, Harvard University, 1987. Hai Rui. Hai Rui ji (Collected writings of Hai Rui). 1962. Reprint, Beijing: Zhonghua shuju, 1981. Hamashima Atsutoshi. Mindai Kōnan nōson shakai no kenkyū (Studies in Jiangnan rural society during Ming dynasty). Tokyo: Tokyo daigaku shuppankai, 1982. ——.“Minmatsu Sekkō no Ka-Ko ryōfu ni okeru kinden kin’eki hō” (The land tax and service levy equalization system in Jiaxing and Huzhou prefectures, Zhejiang, during the late Ming). Tōyō bunka kenkyūjo kiyō 52 (1970), pp. 139–88. ——. “The Organization of Water Control in the Kiangnan Delta in the Ming Period.” Acta Asiatica 38 (1980), pp. 69–92. Handlin, Joanna. Action in Late Ming Thought: The Reorientation of Lii Kun and Other Scholar Officials. Berkeley, CA: University of California Press, 1983. Hansen, Valerie. Review of Chikusa Masaaki, Chugoku bukkyo shakaishi kenkyū. Bulletin of Sung Yuan Studies 20 (1988), pp. 97–108. Hartwell, Robert. “Demographic, Political, and Social Transformations of China, 750–1550.” Harvard Journal of Asiatic Studies 42:2 (1982), pp. 365–442. Hasebe Yūkei. Min-Shin bukkyo shi kenkyū jōsetsu (Introduction to the study of the history of Buddhism in the Ming and Qing). Taibei: Xinwenfeng, 1979. He Changling, ed. Huangchao jingshi wenbian (Essays on statecraft of the imperial dynasty). 1827. Reprint, 1973. Hegel, G.W.E The Philosophy of History, trans. J.Sibree. New York: Dover, 1956. Hegel, Robert. Reading Illustrated Fiction in Late Imperial China. Stanford, CA: Stanford University Press, 1998. Heilbron, J.L. The Sun in the Church: Cathedrals as Solar Observatories. Cambridge, MA: Harvard University Press, 1999. Higman, Francis. Censorship and the Sorbonne. Geneva: Droz, 1979. Ho, Ping-ti. Studies on the Population of China. Cambridge, MA: Harvard University Press, 1959. Hobbes, Thomas. Leviathan, ed. Michael Oakeshott. Oxford: Basil Blackwell, 1960. Hou Han shu (Dynastic history of the Latter Han). Beijing: Zhonghua shuju, 1965. Hsiao, Kung-ch’üan. Rural China. Seattle, WA: University of Washington Press, 1960.

Bibliography

224

Hu, Bangbo. “Maps in the Gazetteer of Tung-an County (Yung-an hsien-chih)” Gest Library Journal 6:1 (Spring 1993), pp. 85–100. Huang Changxuan. Ming’ou tangji (Writings from Shining Seagull Hall). 1628. Huang, Liuhong. Fuhui quanshu (Complete book concerning happiness and benevolence). 1694. Reprint, 1893. Huang Liu-hung. A Complete Book Concerning Happiness and Benevolence, trans. Djang Chu. Tucson, AZ: University of Arizona Press, 1984. Huang, Philip. The Peasant Family and Rural Development in the Yangzi Delta, 1350–1988. Stanford, CA: Stanford University Press, 1990. Huang, Ray. 1587: A Year of No Signiftcance. New Haven, CT: Yale University Press, 1981. ——. Taxation and Governmental Finance in Sixteenth-Century Ming China. Cambridge: Cambridge University Press, 1974. ——. “Wang Ying-chiao.”InDictionary of Ming Biography, ed. L.Carrington Goodrich and Chaoying Fang, pp. 1451–4. New York: Columbia University Press, 1976. Huang Zuo. Taiquan xiangli (The rural rituals of Taiquan). 1549. Reprint, 1821. Siku quanshu zhenben (series 4) ed., 1973. Hucker, Charles. The Censorial System of the Ming Dynasty. Stanford, CA: Stanford University Press, 1966. ——. A Dictionary of Official Titles in Imperial China. Stanford, CA: Stanford University Press, 1985. ——. “Governmental Organization of the Ming Dynasty.” Harvard Journal of Asiatic Studies 21 (1958), pp. 1–66. ——. The Ming Dynasty: Its Origins and Evolving Institutions. Ann Arbor, MI: Center for Chinese Studies, University of Michigan, 1978. ——, ed. Chinese Government in Ming Times: Seven Studies. New York: Columbia University Press, 1969. Hummel, Arthur, ed. Eminent Chinese of the Ch’ing Period. Washington, DC: Government Printing Office, 1943. Hunan shengli cheng’an (Hunan provincial statutes with case precedents). 1820. Huters, Theodore, R.Bin Wong, and Pauline Yü, eds. Culture and State in Chinese History: Conventions, Conflicts, and Accommodations. Stanford, CA: Stanford University Press, 1997. International Rice Research Institute. Rice Research and Production in China. Manila, 1978. Iwai Shigeki. “Kasei yonjuichi-nen Sekkō Genshū-fu Suian-ken juhachi-to kaichi-zu fueki kōsatsu zanbon’ kō” (A study of the surviving fragment of the 1562 Yellow Register for a ward in Township 18 of Suian county, Yanzhou prefecture, Zhejiang). In Chugoku Min-Shin chiho toan no kenkyū (Studies on local documents of Ming-Qing China), ed. Fuma Susumu, pp. 37–56. Kyoto: Kyoto daigaku daigakuin, 2000. Ji Yun, ed. Siku quanshu zongmu (General bibliography of the Complete Library of the Four Treasuries). 1781. Reprint, Taibei: Yiwen, 1969. Jiang, Jin. “Heresy and Persecution in Late Ming Society: Reinterpreting the Case of Li Zhi.” Late Imperial China 22:2 (December 2001), pp. 1–34. Jiao Hong. Guochao xianzheng lu (Biographies of men of attested virtue of the dynasty). 1616. Reprint. Taibei: Xuesheng shuju, 1965. Jingzhao Gui shi shipu (Lineage genealogy of the Gui family of Jingzhao). Ghangshu, 1913. Katayama Tsuyoshi. “Shindai Kanton-shō Shūkō deruta no zukōdsei ni tsuite: zeriyō, kōseki, dozoku” (The ward-tithing system in the Pearl River Delta in Guangdong in the Qing: taxes, household registration, and lineage co-membership). Tōyō gakuho 63:3–4 (1982), pp. 1–34. ——. “Shinmatsu Kanton-shō Shūkō deruta no zukōhyō to sore o meguru shu mondai” (Wardtithing charts from the Peal River Delta in Guangdong in the late Qing and their interpretation). Shigaku zasshi 91:4 (1982), pp. 42–81. Kawakatsu Mamoru. Chūgoku hōken kokka no shihai kōzō (The structure of control of the Chinese feudal state). Tokyo: Tōkydō daigaku shuppankai, 1980.

Bibliography

225

Kuhn, Philip. “Local Self-Government Under the Republic.” In Conflict and Control in LateImperial China, ed. Frederic Wakeman and Carolyn Grant, pp. 257–98. Berkeley, CA: University of California Press, 1975. ——. Origins of the Modern Chinese State. Stanford, CA: Stanford University Press, 2002. ——. “Political Crime and Bureaucratic Monarchy: A Chinese Case of 1768.” Late Imperial China 8:1 (June 1987), pp. 80–104. ——. Rebellion and its Enemies in Late Imperial China. Cambridge, MA: Harvard University Press, 1970. ——. Soulstealers: The Chinese Sorcery Scare of 1768. Cambridge, MA: Harvard University Press, 1990. Kuribayashi Nobuo. Rikōsei no kenkyū (Studies in the lijia system). Tokyo: Bunrui shoin, 1971. Kutcher, Norman. Mourning in Late Imperial China: Filial Piety and the State. Cambridge: Cambridge University Press, 1999. Lai Xinxia, ed. Zhongguo gudai tushi shiye shi (History of library enterprises in ancient China). Shanghai: Shanghai renmin chubanshe, 1990. Legge, James, trans. The Chinese Classics, vol. 1 (Analects, Great Learning, Doctrine of the Meari); vol. 2 (Mencius). 1867–76. ——, trans. Li Ki. The Sacred Books of the East, ed. Max Miiller, vol. 27. 1885. Lei Menglin. Dulii suoyan (Desultory reading notes on the Ming Code). Beijing: Falü chubanshe, 2000. Levine, David. At the Dawn of Modernity: Biology, Culture, and Material Life in Europe after the Year 1000. Berkeley, CA: University of California Press, 2001. Levine, Norman. “The Myth of Asiatic Restoration.” Journal of Asian Studies 37:1, (1977) pp. 73– 85. Lewis, Mark. Sanctioned Violence in Early China. Albany, NY: State University of New York Press, 1990. Li Guohao, Zhang Mengwen, and Cao Tingqin, eds. Explorations in the History of Science and Technology in China. Shanghai: Shanghai Classics Publishing House, 1982. Chinese translation: Zhongguo keji shi tansuo. Shanghai: Shanghai guji chubanshe, 1986. Li Jixian. “Mingdai tangzhang shuliie” (Brief description of embankment captains in the Ming dynasty). In Mingshi luncong (Essays on Ming history), ed. Wang Chunyu, pp. 156–74. Beijing: Zhongguo shehui kexue chubanshe, 1997. Li Le. Jianwen zqji (Miscellaneous notes on things seen and heard, 1610); Xu jianwen zaji (Continuation of miscellaneous notes on things seen and heard, 1612). Reprint of 1632 edition, Shanghai: Shanghai guji chubanshe, 1986. Li Shizhen. Bencao gangmu (Comprehensive pharmacopoeia). Shanghai: Tongwen shuju, 1904. Li Yesi. Gaotang shichao (The unedited poetry of Li Gaotang), Siming congshu ed. 1932–40. Li Zhi. Fen shu (A book to be burned); Xufen shu (Continuation of A book to be burned). Beijing: Zhonghua shuju, 1975. ——. Xufen shu (Continuation of A book to be burned). 1618. Liang, Fang-chung (Liang Fangzhong). “The Tax Captain System in the Ming Dynasty.” In Chinese Social History, ed. E-tu Zen Sun and John de Francis, pp. 249–69. Washington, DC: American Council of Learned Societies, 1956. Liang Fangzhong. “Lun Mingdai lijiafa he junyaofa de guanxi” (On the relationship between the lijia and the equalization of labor levy during the Ming). Xueshu yanjiu 4:5 (1963). ——Mingdai liangzhang zhidu (The tax-captain system in the Ming dynasty). Shanghai: Shanghai renmin chubanshe, 1957. ——Zhongguo lidai hukou, tiandi, tianfu tongji (Historical statistics concerning population, land, and taxation in China). Shanghai: Shanghai renmin chubanshe, 1980. Lin Xiyuan. Lin Ciya xiansheng wenji (The collected writings of Master Lin Ciya). 1612. Reprint, 1752. Lin Zexu. Jifu shuiliyi (Memorials on hydraulics in the capital region). Fuzhou: Lin shi, 1876.

Bibliography

226

Littrup, Leif. Subbureaucratic Government in China in Ming Times. Oslo: Universitetsforlaget, 1981. Liu Dunzhen. Beiping huguo si canji (Record of the remnants of Huguo Monastery in Beiping). Beiping: Zhongguo yingzao xueshe, 1935. Locke, John. Two Treatises of Government, ed. Peter Laslett. Cambridge: Cambridge University Press, 1963. Long Wenbin. Ming huiyao (Essential regulations of the Ming). Reprint, Beijing: Zhonghua shuju, 1956. Lu You. Ru Shu ji (Diary of a journey into Sichuan). 1170. Zhibuzu zhai congshu (Collections from the Studio of Insufficient Knowledge) ed., 1776. Luo Hongxian. Nian’an wenji (Collected writings of Master Nian’an), Siku quanshu zhenben ed. Reprint, Taibei: Shangwu yinshuguan, 1974. Luo Mengzhen. Zhongguo gudai muluxue jianbian (Concise survey of ancient Chinese studies in bibliography). Chongqing, 1983. Luo Weiguo and Hu Ping, eds. Guji banben tiji suoyin (Index to bibliographic notices of old books). Shanghai: Shanghai shudian, 1991. Lyon, Bruce. A Constitutional and Legal History of Medieval England, 2nd edn. New York: Norton, 1980. Mann, Michael. The Sources of Social Power, vol. 1. Gambridge: Cambridge University Press, 1986. Mano Senryū. “Mindai no bukkyo to Mincho” (Ming Buddhism and the Ming dynasty). Reprinted in his Mindai bunkashi kenkyū (Studies in the cultural history of the Ming period), pp. 243–323. Kyoto: Dōhōsha, 1979. Marshall, P.J., and Glyndwr Williams. The Great Map of Mankind. London: Dent, 1982. Matsumoto Yoshimi. Chugoku sonraku seido no shiteki kenkyū (Historical studies of China’s village system). Tokyo: Iwanami shoten, 1977. Matsuo Takane. “Rice Culture in China.” In Rice in Asia, ed. Association of Japanese Agricultural Scientific Societies. Tokyo: University of Tokyo Press, 1975. McKnight, Brian. Village and Bureaucracy in Southern Sung China. Chicago: University of Chicago Press, 1971. Meskill, John, trans. Ch’oe Pu’s Diary: A Record of Drifting across the Sea. Tucson, AZ: University of Arizona Press, 1965. Mihelich, Mira. “Polders and Politics of Land Reclamation in Southeast China during the Northern Sung.” Ph.D. dissertation, Cornell University, 1979. Miki Satoshi. “Minmatsu no Hokken ni okeru hokosei” (The baojia system in Fujian in the late Ming). Tōyō gakuho 61:1–2 (1979), pp. 67–107. Min, Tu-ki. National Polity and Local Power: The Transformation of Late Imperial China, ed. Philip Kuhn and Timothy Brook. Cambridge, MA: Center for East Asian Studies, Harvard University, 1988. Ming huiyao (Essential regulations of the Ming), ed. Long Wenbin. Shanghai: Zhonghua shuju, 1956. Ming shi (Standard history of the Ming dynasty), ed. Zhang Tingyu. Reprint, Beijing: Zhonghua shuju, 1972. Minglü jijie fuli (The Ming Code with explanations and appended substatutes). 1899. Montesquieu, Charles Louis de Secondat de. The Spirit of the Laws. Cambridge: Cambridge University Press, 1989. Mote, Frederick. “The Growth of Chinese Despotism.” Oriens Extremis 8 (1961), pp. 1–41. ——. “The Transformation of Nanking, 1350–1400.” In The City in Late Imperial China, ed. G.William Skinner, pp. 101–53. Stanford. CA: Stanford University Press, 1977. Mote, Frederick, and Denis Twitchett, eds. The Cambridge History of China, vol. 7: The Ming Dynasty, 1368–1644, Part I Cambridge: Cambridge University Press, 1988.

Bibliography

227

Muzong shilu (Veritable records of the Longqing reign). Nanjing: Liang Hongzhi, 1940. Ming shilu reprint, vols. 92–5. Naquin, Susan. Millenanan Rebellion in China: The Eight Trigrams Uprising of 1813. New Haven, CT: Yale University Press, 1976. Needham, Joseph. Science and Civilisation in China, vol. 3. Cambridge: Cambridge University Press, 1976. ––. Science and Civilisation in China, vol. 4, pt. 3. Cambridge: Cambridge University Press, 1971. Nishiyama Takeo. “Ajia nōdgyō no genryu” (The origins and development of Asian agriculture). Ajia keizai 6:3 (1965). Nivison, David. The Life and Thought of Chang Hsueh-ch’eng. Stanford, CA: Stanford University Press, 1966. Noguchi Tetsurd. “Mindai shoki no bukkyokai: Sazensei Keishō o megutte no nooto” (Buddhist associations in the early Ming: notes on Left Enlightener Jixiao). Chūgoku no shūkyō to shakai (Chinese religion and society). Tōyō shigaku ronshū (Essays in Far Eastern historiography), vol. 7. Tokyo: Fumaidō shoten, 1965. Noshomusho (Ministry of Agriculture and Commerce). Shina no kome ni kansuru chosa (Investigations on rice in China). Tokyo: Tōa insatsu, 1917. Ogawa Takuchi. Shina rekishi chiri kenkyū (Studies in historical Chinese geography). Tokyo: Kōbundō, 1940. Okuzaki Yūji. Chūgoku kyōshin jinushi no kenkyū (Studies of Chinese gentry landlords). Tokyo: Kyūko shoin, 1978. O’Leary, Brendan. The Asiatic Mode of Production: Oriental Despotism, Historical Materialism, and Indian History. London: Basil Blackwell, 1989. Overmyer, Daniel. Folk Buddhist Religion: Dissenting Sects in Late Traditional China. Cambridge, MA: Harvard University Press, 1976. Peng Xinwei. Zhongguo huobi shi (History of Chinese currency). Shanghai: Shanghai renmin, 1958. Perdue, Peter. Exhausting the Earth: State and Peasant in Hunan, 1500–1850. Cambridge, MA: Council on East Asian Studies, Harvard University, 1987. Perkins, Dwight. Agricultural Development in China, 1368–1968. Chicago: Aldine, 1969. Prip-Møller, Johannes. Chinese Buddhist Monasteries. Copenhagen: Gads Forlag, 1937. Qian Qizhong. Qingxiyigao (Surviving drafts from Qingxi), Siming congshu (1932–40) ed., pt. 2, vol. 1. Reprint, 1966. Qingdai wenzjyu dang (Documents concerning the Qjng literary inquisition). Beijing, 1934. Qiujim. Daxue yanyi bu (Supplement to Explications on the Great Learning). 1487. Rankin, Mary Backus. “The Origins of the Chinese Public Sphere: Local Elites and Community Affairs in the Late Imperial Period.” Etudes chinoises 9:2 (Autumn 1990), pp. 13–60. Ricoeur, Paul. Time and Narrative, vol. 1, trans. Kathleen McLaughlin and David Pellauer. Chicago: University of Chicago Press, 1984. Robinson, David. Bandits, Eunuchs, and the Son of Heaven: Rebellion and the Economy of Violence in Mid-Ming China. Honolulu: University of Hawaii Press, 2001. Saeki Yuichi. “Reforms in the Service Levy System in the Fifteenth and Sixteenth Centuries.” In State and Society in China: Japanese Perspectives on Ming-Qing Social and Economic History, ed. Linda Grove and Christian Daniels, pp. 279–310. Tokyo: University of Tokyo Press, 1984. Said, Edward. Orientalism. New York: Vintage, 1978. Schipper, Kristofer. “Vernacular and Classical Rituals in Taoism.” Journal of Asian Studies 45:1 (November 1985), pp. 21–57. Schoppa, Keith. Chinese Elites and Political Change. Cambridge, MA: Harvard University Press, 1982. Shanghai bowuguan tushu ziliaoshi (Library materials section of the Shanghai Museum). Shanghai beike shiliao xuanji (Collected epigraphic materials from Shanghai). Shanghai: Shanghai renmin chubanshe, 1980.

Bibliography

228

Shen Bang. Wanshu zaji (Miscellaneous notes from the Wanping county office). 1596. Reprint, Beijing: Zhonghua shuju, 1980. Shenzong shilu (Veritable records of the Wanli reign). Nanjing: Liang Hongzhi, 1940. Ming shilu reprint (Taibei, 1966), vols 96–123. Shiba Yoshinobu. “Ningpo and its Hinterland.” In The City in Late Impenal China, ed. G. William Skinner, pp. 391–439. Stanford, CA: Stanford University Press, 1977. Shimizu Taiji. “Mindai no jiden” (Monastic land in the Ming dynasty). Tōa keizai kenkyū 8:4 (1924), pp. 46–67. Shizong shilu (Veritable records of the Jiajing reign). Nanjing: Liang Hongzhi, 1940. Ming shilu reprint (Taibei, 1966), vols 70–91. Skinner, G.William, ed. The City in Late Imperial China. Stanford, CA: Stanford University Press, 1977. Sogabe Shizuo. Chugoku oyobi kodai Nihon ni okeru kyoson keitai no hensen (Changes in village structure in China and oldjapan). Tokyo: Yoshikawa kobunkan, 1963. Song Lian. Song xueshi quanji (The complete works of Academician Song). 1552. Song shi (Standard history of the Song dynasty). Beijing: Zhonghua shuju, 1977. Sonkeikaku bunko (Sonkeikaku library), ed. Sonkeikaku bunko kanseki bunrui mokuroku sakuin (Classified catalogue and index of Chinese books in the Sonkeikaku Library). Tokyo, 1935. Steurmer, John. “Polder Construction and the Pattern of Land Ownership in the T’ai-hu Basin during the Southern Sung Dynasty” Ph.D. dissertation, University of Pennsylvania, 1980. Su Boheng. Su Pingzhong wenji (The collected writings of Su Pingzhong). 1442. Sibu congkan reprint, 1929. Sun Dianqi. Liulichang xiaozhi (Short monograph on Liulichang). Beijing: Beijing guji chubanshe, 1982. Sun, E-tu Zen, and John de Francis, eds. Chinese Social History. Washington, DC: American Council of Learned Societies, 1956. Sun Yi. Dongting yuren ji (Collected writings of a Dongting fisherman). c. 1556. Sun Yutang and Zhang Jiqian. “Qingdai de kentian yu dingkou de jilu” (Records of cultivated land and population in the Qjng dynasty). Qingshi luncong (Essays on Qing history), no. 1 (1979). Szonyi, Michael. Practicing Kinship: Lineage and Descent in Late Imperial China. Stanford, CA: Stanford University Press, 2002. Taizu shilu (Veritable records of the Hongwu reign). Nanjing: Liang Hongzhi, 1940. Ming shilu reprint (Taibei, 1966), vols 1–8. Tanaka Masatoshi and Tsurumi Naohiro. “Ryūkōsha to nōmin” (The treadle pump and the peasants). Kindai Chugoku kenkyū iho (Tōyō bunko) no. 1 (1979). Tatsuike Kiyoshi. “Mindai no sokan” (Clerical officials in the Ming dynasty). Shina bukkyo shigaku 4:3 (1940). ——–. “Ming Ming Ming taizu de fojiao zhengce” (The Ming founder’s policies on Buddhism). Reprinted in Ming-Qing fojiao shipian (Historical essays on Ming-Qing Buddhism), ed. Zhang Mantao, pp. 1–16. Taibei, 1977. ——. “Minsho no jiin” (Monasteries in the early Ming). Shina bukkyo shigaku 2:4 (1938), pp. 9– 29. Thomas, Donald. A Long Time Burning: The History of Literary Censorship in England. London: Routledge and Kegan Paul, 1969. Tōyō bunka kenkyūjo (Institute of Oriental Culture). Tokyo daigaku toyo bunka kenkyūjo kanseki bunrui mokuroku (Classified catalogue of Chinese books in the Institute of Oriental Culture, Tokyo University). Tokyo: Kyūko shoin, 1981. Tsai, Shih-shan Henry. The Eunuchs in the Ming Dynasty. Albany, NY: State University of New York Press, 1996. Tsur, Nyok-ching (Zhou Yiqing). “Forms of Business in the City of Ningbo in China” (1909), trans. Peter Schran. Chinese Sociology and Anthropology 15:4 (1983), pp. 3–131.

Bibliography

229

Tsurumi Naohiro. “Kōki jūgo-nen joryo: Soshu-fu Chōshū-ken gyorin zusatsu no dento tōkeiteki kōsatsu” (An examination of land statistics in the fish-scale register for 1676 of Changzhou county, Suzhou), pt. 1. In Kimura Masao sensei taikan kinen Tōyōshi ronshū (Essays on East Asian history on the occasion of Professor Kimura’s retirement), pp. 311–44. Tokyo: Kyūko shoin, 1976. ——. “koKōki jūgo-nen jōryō: Soshu-fu Chōshū-ken gyorin zusatsu no dentō tokeiteki kōsatsu” (An examination of land statistics in the fish-scale register for 1676 of Changzhou county, Suzhou), pt. 2. In Nakajima sensei kōki kinen ronshū (Essays presented to Professor Nakajima on the occasion of his seventieth birthday). Tokyo, 1981. ——. “Rural Control in the Ming.” In State and Society in China, ed. Linda Grove and Christian Daniels, pp. 245–77. Tokyo: Tokyo University Press, 1984. ——. “Soshū-fu no gyorinsatsu ni kansuru ichi kōsatsu” (An examination of Fish-Scale Registers from Suzhon prefecture). Shakai keizai shigaku 34:5 (1969), pp. 1–31. Tu Long. Baiyu ji (The collected writings of Tu Baiyu). 1600; reprint, 1849. Taibei: Weiwen chubanshe, 1977. ——. Shaluo guan qingyan (Pure sayings from Śāla Hall), Congshu jicheng ed. Shanghai: Shangwu yinshuguan, 1936. Unger, Roberto. Social Theory: Its Situation and its Task, vol. 1 of Politics. Cambridge: Cambridge University Press, 1987. Wada Tamotsu. “Hoku Shina no tochi kairyo ni kansuru chōsa hokokusho yoshi” (Summary of the report on an investigation into the improvement of land in north China). Mimeograph, 1938. Wakeman, Frederic, and Carolyn Grant, eds. Conflict and Control in Late Impenal China. Berkeley, CA: University of California Press, 1975. Wallerstein, Immanuel. “The West, Capitalism, and the Modern World-System.” In China and Historical Capitalism: Genealogies of Sinological Knowledge, ed. Timothy Brook and Gregory Blue, pp. 10–56. Cambridge: Cambridge University Press, 1999. Wang Guoping and Tang Lixing, eds. Ming-Qing yilai Suzhou shehuishi beike ji (Collected epigraphy on social history from Suzhou since the Ming and Qmg). Suzhou: Suzhou daxue chubanshe, 1998. Wang Liqi. Tuan-Ming-Qing sandai jinhui xiaoshuo xiqu shiliao (Historical materials concerning the proscription of novels and operas during the Yuan, Ming and Qing dynasties). 1958. Rev. ed., Shanghai, 1981. Wang Wenlu. Bailing xueshan (The Bailing compendium). 1584. Wang Xingya. “Mingchu qian Shanxi min dao Henan kaoshu” (Study of the migration of Shanxi natives to Henan in the early Ming). Shixue yuekan, 1984:4, pp. 36–44. Wang Zhen. Nong shu (Agricultural handbook), ed. Wang Yuhu. Beijing: Nongye chubanshe, 1981. Wei Jinxi, et al. Xuezheng quanshu (Complete handbook on educational administration). 1775. Reprint, Taibei, 1968. Wei Qingyuan. Mingdai huangce zhidu (The yellow register system of the Ming dynasty). Beijing: Zhonghua shuju, 1961. Welch, Holmes. Buddhism under Mao. Cambridge, MA: Harvard University Press, 1972. -. “The Fate of Religion.” In The China Difference, ed. Ross Terrill. New York: Harper and Row, 1979. Wen Juntian. Zhongguo baojia zhidu (The baojia system in China). Shanghai: Shangwu yinshuguan, 1936. Willj Pierre-Etienne. Bureaucratie etfamine en Chine au 18e siècle. Paris, 1980. Translated as Bureaucracy and Famine in Eighteenth-Century China, trans. Elborg Forster. Stanford, CA: Stanford University Press, 1990. Williams, Raymond. Keywords: A Vocabulary of Culture and Society. London: Fontana, 1976. Witek, John. Controversial Ideas in China and Europe: A Biography of Jean-Francois Foucquet, S.J. (1665–1741). Rome: Institutum Historicum, 1982.

Bibliography

230

Wittfogel, Karl. Oriental Despotism: A Comparative Study of Total Power. New Haven, CT: Yale University Press, 1957. Wong, R.Bin. China Transformed: Historical Change and the Limits of European Experience. Ithaca, NY: Cornell University Press, 1997. ——. “The Political Economy of Agrarian Empire and its Modern Legacy.” In China and Historical Capitalism, ed. Timothy Brook and Gregory Blue, pp. 210–45. Cambridge: Cambridge University Press, 1999. Wu Bangqing. zenong yaolu. (Essential records for hydraulic agriculture). In Jifu hedao shuili congshu (Collectanea on the rivers and hydraulics of the capital region), 1824. Wu Han. Jiang zhe cangshujia shiliie (Brief history of book collectors in Jiangsu and Zhejiang). Reprint, Beijing: Zhonghua shuju, 1981. Wu Kwang Tsing. “Ming Printing and Printers.” Harvard Journal of Asiatic Studies 7:3 (1943), pp. 203–60. Wu Lien Teh. “A Striking Example of Scientific Farming in North China.” Journal of the Chinese Philosophical Society 1:1 (1911). Wu Zhefu. Qingdai jinhui shumuyanjiu (Studies of indices of banned books during the Qing dynasty). Taibei, 1969. Wu Zhenyu. Yangjizhai conglu (Collected notes from Yangji Studio). 1896. Reprint, Beijing: Beijing guji chubanshe, 1983. Xiamen daxue lishixi (Amoy University History Department), ed. Li ^hiyanjiu cankao ziliao (Reference materials for the study of Li Zhi). Fujian: Fujian renmin chubanshe, 1975. Xianzong shilu (Veritable records of the Chenghua reign). Nanjing: Liang Hongzhi, 1940. Ming shilu reprint (Taibei, 1966), vols 39–50. Xiaozong shilu (Veritable records of the Hongzhi reign). Nanjing: Liang Hongzhi, 1940. Ming shilu reprint (Taibei, 1966), vols 51–60. Xie Guozhen. Wan Ming shiji kao (Studies of historical records of the late Ming). Beiping: Guoli Beiping tushiguan, 1933. Xin Tang shu (New standard history of the Tang). Beijing: Zhonghua shuju, 1975. Xizong shilu (Veritable records of the Tianqi reign). Nanjing: Liang Hongzhi, 1940. Ming shilu reprint, vols. 124–33. Xu Fuyuan. Jinghe tangji (Writings from the Hall of Sincerity and Harmony). 1594. Xu Guangqi. Nongzheng quanshu jiaozhu (Complete handbook of agricultural management, annotated), ed. Shi Shenghan. Shanghai: Shanghai guji chubanshe, 1979. Xu Hongzu. Xu Xiakeyouji (The travel diaries of Xu Xiake). Reprint, Shanghai: Shanghai guji chubanshe, 1980. Xu Kaixi. Hanshan kao (Notes from Han Mountain). 1639. Xu Xueju. Guochao dianhui (Collection of statutory precedents of the dynasty). 1601. Xu Yikui. Shifeng gao (Drafts from Shifeng Stream). Wulin wangzhe yizhu (Surviving writings of former Hangzhou scholars). Hangzhou, 1894. Xu Zhenming. Lushui ketan (Chats of a guest at Lushui). Hangzhou, 1584. Xuanzong shilu (Veritable records of the Xuande reign). Nanjing: Liang Hongzhi, 1940. Ming shilu reprint (Taibei, 1966), vols. 16–21. Ye Chunji. Huian zhengshu (Administrative handbook of Huian). 1573. Reprinted in his Shidong wenji (Collected works of Master Shidong), 1672. Reprint, Fuzhou: Fujian renmin chubanshe, 1987. Ye Dehui. Shulin qinghua (Limpid comments on the world of books). 1911. Reprint, Taibei: Shijie shuju, 1970. Ye Mengzhu. Yueshi bian (A survey of the age). Early Qing. Reprint, Shanghai: Shanghai guji chubanshe, 1981. Yingzong shilu (Veritable records of the Zhengtong reign). Nanjing: Liang Hongzhi, 1940. Ming shilu reprint (Taibei, 1966), vols. 22–38.

Bibliography

231

Yinxiang. “Yi xianqinwang shuchao” (Memorials of Prince Yi). Appended to Wu Bangqing, zenong yaolu. Jifu hedao shuili congshu (Collectanea on the rivers and hydraulics of the capital region). Beijing, 1824. Yongrong et al. Qinding siku quanshu zongmu (Imperially endorsed general bibliography of the Complete Library of the Four Treasuries). 1782. Reprint, Taibei: Yiwen yinshuguan, 1969. Yü Chün-fang. The Renewal of Buddhism in China: Chu-hung and the Late Ming Synthesis. New York: Columbia University Press, 1981. Yu Jideng. Diangu jiwen (Casual notes on statutory precedents). 1601. Reprint, Beijing: Zhonghua shuju, 1981. Yuan Huang. Liaofan zazhu (Miscellaneous writings of Master Liaofan). 1607. Yuan shi (Standard history of the Yuan dynasty), ed. Song Lian. 1370. Reprint. Beijing: Zhonghua shuju, 1976. Zhang Huang. Tushu bian (The illustrated compendium). 1613. Zhang Lu, ed. Huang Ming zhishu (Handbook of Ming-dynasty regulations). 1597. Reprint in 2 vols, Tokyo: Koten kenkyūkai, 1967. Zhang Xuecheng. Wenshi tongyi (Comprehensive explanations of texts and histories), ed. Ye Ying. Beijing: Zhonghua shuju, 1985. Zhao Bingzhong. Jiangxi yudi tushuo (Notes accompanying the geographical atlas of Jiangxi), Wanli-era (1573–1620) Jilu huibian edition. Reprint, Taibei: Chengwen chubanshe, 1965. Zheng Kesheng. “Guanyu Mingdai Tianjin de shuitian” (On paddy fields in Tianjin in the Ming dynasty). Nankai shixue 1980:1. zhongguo difangzhi lianhe mulu (Union catalogue of Chinese local gazetteers). Beijing: Zhonghua shuju, 1983. Zhongyang tushuguan (Central Library), ed. Mingren zhuanji ziliao suoyin (Index to Ming biographical materials), 1st edn. Taibei, 1965; 2nd edn., 1978; reprint, Beijing: Zhonghua shuju, 1987. Zhou Tiandu. Tongsu bian (Compilation on popular customs). 1751. Zhu Hua. Hucheng beikao (Notes on the city of Shanghai). Reprint, 1935. Zhu Xi. Zhuzi wenji (Collected writings of Master Zhu), ed. Chen Junmin, 10 vols. Taibei: Defu wenjiao jijinhui, 2000. Zhu Yizun. Rixia jiuwen (Old tales heard under the sun). 1688. Zhu Yuanzhang. Yuzjii dagao (Imperially compiled grand pronouncements). 1385–7. Zuo Buqing. “Qianlong fenshu” (The Qianlong book-burning). Reprinted in Ming-Qing renwu lunji (Essays in Ming-Qing biography). Chengdu, 1982.

Index abbeys (conglin) 144, 150, 153 Administrative Affairs of Chun’an County (Chun’an xian zhengshi) 55–6 Administrative Records of Huian County (Huian zhengshu) 46–56, 59, 220n agricultural handbooks 94–5 Ai Nanying (1583–1643) 131–2 altars 171, 219n astronomy 120, 213n attached households (daiguanhu) 34 bagu wen (regulated-meaning essays) 124 bandits 93, 199n bantu (half-words) 196n baojia (mutual watch system) 19, 32–33, 36–41, 55, 78 Beijing 31, 81, 85, 152, 196n Blackstone, William (1723–80) 9 Blessings of Doing Good Secretly (Weishan yinzhi) 105, 107, 109 Bookfor Cultivators, Craftsmen, and Merchants (Wunongjiyi shanggu shi) 104 books 12, 122; circulation of 124, 134; difficulty acquiring 110, 113, 210n; palace editions 104–6, 110; storage 111–12; see also censorship; libraries booksellers’ guilds 135 bookshops 134–5 border affairs 7, 44 boundaries: confluence or non-confluence of in different systems 38, 66, 68–9, 75–6; new 41; obscurity of on maps 50–1; persistence of 19; political role of 44 Buck, John Lossing 82, 85 Buddhism: as competing source of authority 146, 189; as contributing to public order 173–5, 179; hostility toward 160–2, 169–71; sects 143–4; social constituency of 141; state suppression of 142–6; state toleration of 146–51

Index

233

Buddhist registrars 143, 148–9 Buddhist registries 143, 147, 149; see also Central Buddhist Registry bureaucracy 136, 140, 148 burial see grave land Cai Xian (b. 1697) 128–9 cadastral surveys see registration: of population canton see xiang cartography: grid method 47–50, 57–8; Ming methods 54–6; modern European methods 56–8; see also maps censorship 12, 118–36 census see Dazao; registration: of population Central Buddhist Registry 142, 146 Chambers’ Cyclopaedia (1728) 183 Changshu county (Suzhou) 72–4, 130 Changzhou county (Suzhou) 31, 74 Changzhou prefecture (South Zhili) 71, 74 Chen Fengwu (1475–1541) 109, 210n Chen Hongshou (fl. 1771) 95–96 Chen Hu (fl. 1645) 76–7 Chen Jinyan (fl. 1755) 170–2 Chen Longzheng (1585–1645) 204n Chen Yinming (fl. 1609) 121 Cheng Yi (1033–1107) 119 Chenghua emperor (r. 1465–87) 7, 150–2 Chikusa Masaaki 141 Cisheng, Empress Dowager (1546–1614) 154, 156 Classics Workshop (jingchang) 106 Clunas, Craig 191n Collected Rites of the Ming Dynasty (Da Ming jili) 105, 108–9 Collected Writings of Heaven’s Hired Hand (Tianyongzi ji) 131–2 commercialization 20; of publishing 128–9 communications, as state capacity 3, 188 communities 19–20, 42; polder 66, 77–9 Complete Library of the Four Treasuries see Siku quanshu Comprehensive Gazetteer of Cao’s Stream (Caoxi tongzhi) 130–2 Confucian canon 105–7; see also Great Compendium editions Confucian prophylaxis against Buddhism 162–3, 165–9, 180 Confucianism and censorship 119 constitution 9, 158–9, 181; Buddhism in 159–61, 178–81;

Index

234

Ming 7, 13, 190 contracts 75, 197n corruption 6, 8 corvée labour 31, 147; see also service levy counties, numbers of 21, 193n Covenant Recitations Halls (jiangyue suo) 38–9 dang (compact) 37, 199n Daoan(fl 1567) 211n Dazao (great register-compiling year) 34, 45, 199n, 204n Deliberations on the Great Rites Controversy (Dali jiyi) 106, 108–9 despotism 8, 182–4 Ding Yuanjian (1563–1628) 63, 68, 76, 188 Dong Rang (fl. 1499) 3–4, 6–7 du (township) 22–30, 48–9, 75, 201n du Halde, Jean-Baptiste (1674–1743) 101 Edict of Seclusion 144–6 elders (lilao) 33, 38, 45, 48, 71, 194n Elvin, Mark 69 eunuchs 6–7, 64; as monastic patrons 165; Grand Defenders (zhenshou) 3, 6–7 Europe, comparisons with 120, 126–7, 134–6, 158–9, 181, 183–6, 189–90 Exhortation to Goodness (Quanshan shu) 105, 109 family (jia) as unit of registration 36 Fan Qi (1616-after 1694) 65 illus fang (urban township) 22, 32–3 Fang Xianfu (d. 1544) 217n Farmer, Edward 159 Fei Xiaotong 66 Feng Rubi (js. 1532) 72–4 Fish-Scale Registers (yulin tuce) 29, 31, 46, 69, 72, 77, 187 Fossier, Robert 185–6 Foundation Cleric (zherji daoreri) 144 Fuben(fl. 1626) 122 Gao Ruxing (js. 1521) 94 Gazetteer of the Ming Dynasty (Da Ming zhi) 105 gazetteers 43–4, 55, 116; as complete records 172; as genre 219n; ordered submitted to the capital 2; as permanent records 111; as sources for local history 15, 21, 86, 90, 162–3, 176–8, 203–4n Gengju (js. 1601) 72–3, 75, 193n gentry 64, 98, 208n; and Buddhism 162, 179–81;

Index

235

as leaders of subcounty units 41; as patrons 113–14, 156–7; and the state 140, 161; vertical integration with state 188, 190 Goodrich, Carrington 125 Grand Pronouncements (Dagao) 32, 105, 107–9, 197n grave land 1–5, 7, 14 Great Compendium (Daquan) editions 105–11 Great Compilation on Virtue Illuminated (Minglu dadian) 106, 108–9 great household (dahu) 198n Great Unity of the Nine Palaces (fiugong taiyi) 120 Grimm, Tilemann 182 Gu Yanwu (1613–82) 37, 71, 81–2, 85–6, 93, 97, 106, 196n, 212n Gui Youguang (1506–71) 130, 132 Guo Shoujing (Yuan) 85 Guo Zaoqing (fl. 1573) 44–5, 48, 54 Guy, Kent 126 Hai Rui (1513–87) 55, 198n Hamashima Atsutoshi 203n, 205n Han jun (js. 1598) 172 Han Yu (768–824) 165–6, 168 Hanshan Deqing (1546–1623) 153 Hartwell, Robert 193n, 198n Hegel, G.W. F. (1770–1831) 5, 8, 11, 158–9, 187 heterodoxy 119–20, 124, 159–60, 162, 166 hierarchies 19–20, 42 Hippocrates 183 historiography: of Buddhism 140, 151; of Ming studies 8–9, 182–4, 192n Hobbes, Thomas (1588–1679) 183 Hongwu: dispensation 141, 156; emperor (r.1368–98) 20, 32, 64, 80, 104, 106–7; and Buddhism 139–46, 153, 158, 167–8, 174 Hongzhi emperor (r. 1488–1505) 1, 4–6, 13–14, 149–50 household (hu) as fiscal unit 36, 70, 154 Hsiao Kung-ch’üan 42 Hu Weiyong (d. 1380) 142 Huang Liuhong (fl. 1694) 43 Huang, Philip 11 Huang, Ray 71 Huang Tingkui (1691–1759) 213n Huian county (Quanzhou, Fujian) 31, 43–55 Huidian see Statutory Precedents of the Ming Dynasty hundred captains (lizhang) 33, 38, 41–2, 67, 71–2 Hundred-Day Edict 144–5 Huo Tao (1487–1540) 217n Huzhou prefecture (Zhejiang) 63, 68–70, 76, 128

Index

236

illiteracy 128–9, 213n individual and the state 9 information as imperial resource 5, 43–4, 59 irrigation 24, 29, 66–9, 75, 91–3, 95–7 Jeffreys, Thomas (fl. 1750) 183–4 Jesuits 58 illus, 101 jia (tithing) 33, 37, 197n Jiajing emperor (r. 1522–66) 105–6 Jiang, jin 124 Jiangnan 11–12, 35–6, 64–8, 74–9, 134 Jianwen emperor (r. 1399–1402) 148 Jiao Hong (1541–1620) 124 Jin Cao (fl. 1494) 71, 74 Johnson’s Dictionary of the English Language (1755) 183 Kang Gao (fl. 1572) 45 Kangxi emperor (r. 1662–1721) 83 Kawakatsu Mamoru 205n kinship, as an administrative principle 42 Korea under Japanese attack 94 Kuhn, Philip 20, 159, 200n Kuribayashi Nobuo 38 labour: costs for irrigation 96; time-extensive 208n land, magnification of arable 95, 186 land records 72–4 landlordism 77, 93, 97–8 Levine, David 185–6 li (hundred) 33, 197n; redefined by acreage 74 li (subcanton) 21–4, 50 Li Bangyan (js. 1572) 154 Li Guangdi (1642–1718) 88, 208n Li Shimian (js. 1404) 119–22, 128 Li Shizhen (1518–93) 82 Li Weizhen (1547–1626) 124 Li Zhi (1527–1602) 122–4, 154 Liang Fangzhong 193n, 200n Liangjin (fl. 1542) 149, 217n libraries: catalogues 107, 110–11, 116; loss of books from 103, 210n; monastic 102, 112; relationship to knowledge 104, 115–17, 188; inschools 12, 102–3, 111–17; see also Pavilions for Revering the Classics; Pavilions for Storing the Classics

Index

237

lijia (hundred-and-tithing system) 19, 32–6, 64, 67, 76; control of by local elites 41, 77; fiscalization of 36–8; units redefined on basis of acreage 35, 74 Lin Yingxun (js. 1571) 74 Lin Zexu (1785–1850) 81–2, 86, 91, 93, 97–8 lineage as fiscal household 36 “literary inquisition” 12, 125–6 litigation 2–5 Liu Duo (js. 1616) 122 LiuTong (d. 1637) 86 Locke, john (1632–1704) 183 Longqing emperor (r. 1567–72) 200n Lu Cheng (js. 1493) 199n Lu Guangzu (1521–97) 154 Lü Kun (1536–1618) 39 Lü Liuliang (1629–83) 130, 132 Lu Longqi (1630–93) 86, 169–73, 177 Lu Shiyi (fl. 1635) 77 Luo Hongxian (1504–64) 47–8, 50–1, 55–9, 115–17, 201n Ma Jinglun (js. 1589) 123 magistrates 31, 43, 114–15; assassination of 202n; and local society 21, 37, 41, 92, 187 Manchu legitimacy 125, 133 Mann, Michael 10 maps 187; of border regions 48; in gazetteers 24–26, 44–6, 176–7; legends 51–52; as mode of organizing knowledge 46, 59; urban 52–3; see also cartography markets, in relation to baojia 37 Marx, Karl (1818–83) 125, 184 Mei Guozhen (js. 1583) 123 Mencius (371–289 BC) 168 Miki Satoshi 199n military: agricultural colonies 87–9, 93; grain 81; system 6 militia 24, 194n Milukov, Paul 9 Ming Code (Da Ming lü) 105, 107, 109, 120–1, 159–60 Ming Statutes (Da Ming ling) 105, 109 ministry/minister: of Justice 5, 122; of Personnel 5; of Revenue 152;

Index

238

of Rites 105, 107–8, 119–20, 144, 152–3, 171, 173, 198n; of War 5, 7, 219n; of Works 71, 106, 108 Model Letters of Gui and Qian (Gui Qian chidie) 130–3 monarch/subject (Jun/chen) relationship 161, 169 monasteries 12–13; amalgation 144–6, 149–50, 164; ban on founding of 144, 159, 164; fiscal status of 159; landholding by 144, 148, 216n; legitimacy of 167; as serving the state 155, 172; state regulation of 12–13, 142–54, 215n; under the People’s Republic 145, 216n; see also patronage: of monasteries Mongols 150–1 monks: complaints about 151; in monastic communities 146; quotas on 146–8; as state emissaries 141 Montesquieu, Charles Louis de Secondat de (1689–1755) 8, 184 name plaque (e) 145, 153, 216n Nanhua Monastery (Guangdong) 130, 154 Nanjing 35, 110, 129 National Academy (Nanjing) 108, 110, 119 New Tales while Trimming the Wick (Jiandeng xinhua) 119–21 North Zhili (province) 81–98, 163–5 north-south contrasts 11–12, 21, 24, 81–4, 93, 96–7, 187, 195n, 206n, 208n nostalgia 79, 188 On the Five Relationships (Wulun shu) 105, 107–9 ordination 142, 153; certificates 151–2 “Oriental despotism” 8 paddy fields 11, 29, 82–88, 95 illus pai (placards) 37 Pak Chi-won (fl. 1780) 125–6 patronage: of libraries 113–15; of monasteries 12–13, 154–6 Pavilions of Exhibition (Shenming ting) 38 Pavilions for Revering the Classics (Zunjing ge) 102–3, 112 Pavilions for Storing the Classics (Cangjing ge) 102–3 Pei Xingjian (619–82) 85 Pei Xiu (224–71) 47 Perkins, Dwight 85 “Placard for Instructing the People” (Jiaomin banguuen) 105, 109, 191n, 198n

Index

239

plenary masses 141–2 polder captains (weizhang, yuzhang) 67–8 polders 63–80, 186 population see registered population post houses (pu) 52, 54 prices: books 107; land 88; ordination certificates 151; printing 213n; rice 97 Princes’ Book (Gongzi shu) 104 printing 49, 54, 127; see also publishing proxy for state, local officials as 6, 41–2 public authority 13, 139–42, 146, 155–7, 181 public halls (gongguan) 194n public interest 160, 215n publishing: and censorship 126–34; commercial 110–11, 130, 209n, 211n; relationship to knowledge 106; by schools 110; by the state 104–6 Putuo lsland 218n Qian Jingkai (fl. 1780) 134 Qian Qianyi (1582–1664) 130–1 Qian Qizhong (js. 1628) 124 Qianlong emperor (r. 1736–95) 125–6, 128, 133–6 Qing Code (Da Qing lü) 121 Qing dynasty contrasted with Ming 78, 87, 98, 157, 164–5, 175–6 Qiujun (1420–95) 109, 113–17 qu (sectors) as lijia units 35, 67; as polder-registration units 68–75 Quesnay, François (1694–1774) 189 Records of the Great Imprisonment Controversy (Dayu lu) 106, 108–9 Records of the Rise and Fate of the Great Dynasty (Dachao qiyun lu) 121 registered population 23, 27–8, 31; decrease of 33, 35, 198n registration: evasion of 31, 46; fraudulent 207n; of land 7, 9, 45, 63–4, 205n, 215n; of monks 142; of polders 68–70; of population 6, 20, 45, 196n rice: as military asset 93; as percentage of grain production 90–1;

Index

240

types of 82–3, 206n Ricci, Matteo (1552–1610) 47, 56–7, 127–8 Ricoeur, Paul 183 ritual corpus (lidiari) 167, 170, 180; see also sacrificial corpus rituals, community 55 riziculture 81–98; see also paddy fields roads 53–4 rural covenant system (xiangyue) 19, 38–9, 41, 55 sacrificial corpus (sidian) 167, 171, 220n; see also ritual corpus Schipper, Kristofer 218n schools: community 38, 49, 54–5, 216n; county or prefectural 1, 102–4, 106–12, 114–17 Schoppa, Keith 41 sector see qu “seditious writings” (yaoshu) 120–1, 136 service levy 70–5; see also corvée labour Shang Kexi (d. 1676) 131 Shanghai 30, 109 she (ward) 22 Shen Deqian (1673–1769) 129 Shen Yiguan (1531–1615) 123 Shiba Yoshinobu 199n Shunzhi emperor (r. 1644–61) 135 siguan (monasteries) section of gazetteer 163, 167–72, 175–7 Siku quanshu (Complete Library of the Four Treasuries) 125, 173 Sima Qian (145–86? BC) 170 Single-Whip reforms (yitiao bianfa) 45, 59, 70–6, 78, 187 Skinner, G.William 40–1 social networks 8, 10, 12–13, 157 society, capillary influences on politics of 4, 10–11, 13, 187, 190 society-making 8–10, 40 Song dynasty 22, 24, 32, 85 Song Yiwang (js. 1547) 74 Songjiang prefecture (South Zhili) 30, 71, 78, 112, 134 South Zhili (province) 24, 30, 34 state 3, 6, 13, 19–20, 41–2, 182–4; absolutist 186, 189–90; and Buddhism 139–57; and maps 44; in society 6, 8–10, 14, 77–80, 187–90; percolating influence on society of 4, 10–11, resilience of 80, 190; strong/weak 78–80; “stronger than society” 9 state-cult temples 29, 171, 174

Index

241

state-making 9–10, 40 Statutory Precedents of the Ming Dynasty (Da Ming huidian) 160–2, 168, 180 Su Kui (js. 1487) 2–3, 6–7 subcounty administration 10–11, 19–42 supernumerary households (qilinghu) 34 Supplement to “Exposition on the Great Learning” 109 surveys of land 45–8, 66–7 Suzhou prefecture (South Zhili) 38, 67, 134, 153 Szonyi, Michael 222n Taicang subprefecture (Songjiang) 25, 68–9, 76–7 Tanaka Masatoshi 208n Tang Bin (1627–87) 214n tax captains (liangzhang) 35, 67, 69 taxation 7; conflict over 78–9; effects of land improvement on 97; equitable redistribution of 69; evasion 33, 142 technology, agricultural 94–6; introduction of Western 88 Testimonies to Filiality and Obedience (Xiaoshan shishi) 105, 107, 109 Tianjin 81, 89, 94 Tianshun emperor (r. 1457–64) 147–8, 150 Tianyi ge (Number One Library under Heaven) 101 tithing see jia topography, visually indicated on maps 53–4 township see du Tripitaka: Northern 105, 154; Southern 105 Tsurumi Naohiro 193n, 208n tu (map) 46 tu (ward) 22, 30–2, 39, 67 Tu Long (1542–1605) 123 tuan (regiment) 37 tun (ward) 22 Tuotuo (Yuan) 85–6 Unger, Roberto 9–10 Unity Gazetteer of the Ming Dynasty (Da Ming yitong zhi) 105, 109 universal registers (zhouzhi wence) 142 urban administrative units 25–6, 30–2 Veritable records (Shilu) 1, 7–8, 14, 107–8, 120–2 Wang Anshi (1021–86) 24, 32 Wang Fuzhi (1619–92) 212n Wang Shouren (1472–1529) 37, 115 Wang Wenlu (1503–86) 29

Index

242

Wang Xijue (1534–1611) 190 Wang Yaliang (js.1604) 124 Wang Yangming see Wang Shouren Wang Yingjiao (d. 1628) 89, 93–4 Wang Zhen (fl. 1313) 65, 81–2, 85 Wang Zhideng (1535–1612) 211n Wang Zhiquan (fl. 1737) 96–7 Wanjuan lou (Library of Ten Thousand Fascicles) 101 Wanli emperor (r. 1573–1620) 190, 219n ward remittance (tuyun) 74 ward see tu water-lifting technology 95–6 Weber, Max 184 weisuo (military guard) 33 Wen Zhengming (1470–1559) 153 Wenchang (god of literature) 209n White Lotus 159 Wittfogel, Karl 8–11, 184 Wong, R.Bin 192n, 221n Wu Bangqing (1766–1848) 94–6 Wu Qiong (js. 1469) 2, 6, 8 Wutai Mountain 147, 150, 218n Wuxi county (Changzhou) 74–5 xiang (canton) 21–24, 68 xiangyue see rural covenant system xianq (suburban township), 22, 32–3 Xie Zhaozhe (1567–1624) 210n Xu Changzuo (1558–1609) 86 Xu Guangqi (1562–1633) 65, 81–3, 86, 93, 97–8 Xu Hongzu (1586–1641) 29 Xu Yikui (1318–c. 1400) 20 Xu Zhenming (js. 1572, d. 1590) 86, 92–3, 98 Yan Song (1480–1565) 114 Yang Yigui (fl. 1599) 92, 96 Yao Wenhao (js. 1484) 71, 74 Ye Chunji (b. 1532, jr. 1552) 11, 43–59, 187, 200n Ye Dehui (1864–1927) 127 Ye Mengzhu (fl. 1660) 112 Yellow Registers (huangce) 34, 72, 77, 187 Yinxiang (1686–1730) 88, 94 Yongle emperor (r. 1403–22) 6, 105, 146–8, 150 Yongzheng emperor (r. 1723–35) 88, 97, 130 yu (urban township) 22 Yu ji (1272–1348) 85 Yu Qian (1398–1457) 217n Yuan dynasty: origins of Ming practices 22, 32, 85–7, 95, 102, 143, 195n, 196n; transition to Ming 141, 145 Yuan Huang (1533–1606) 94

Index Yunnan province 143 Zhang Wenda (js. 1583, d. 1625) 122–4, 128 Zhang Xuecheng (1738–1801) 172–3 Zhang Yue (1492–1553) 44 Zheng He (1371–1435) 6 Zhengde era (1506–21) 41 Zhou Zhongshi (fl. 1604) 168, 170 Zhu Siben (1273–1337) 47 Zhu Xi (1130–1200) 102, 104–5, 109, 169 Zhu Yizun (1629–1709) 97 Zhu Yuanzhang (1328–98) see Hongwu: emperor Zou Yuanbiao (1551–1624) 81 Zuo Guangdou (1575–1625) 91, 94

243