Labour, Coercion, and Economic Growth in Eurasia, 17Th–20Th Centuries Studies in Global Social History

Labour, Coercion, and Economic Growth in Eurasia, 17th–20th Centuries Studies in Global Social History

Series Editor Marcel van der Linden International Institute of Social History, Amsterdam, The Netherlands

Editorial Board Sven Beckert Harvard University, Cambridge, MA, USA Philip Bonner University of the Witwatersrand, Johannesburg, South Africa Dirk Hoerder Arizona State University, Phoenix, AZ, USA Chitra Joshi Indraprastha College, Delhi University, India Amarjit Kaur University of New England, Armidale, Australia Barbara Weinstein New York University, New York, NY, USA

VOLUME 11

The titles published in this series are listed at brill.com/sgsh Labour, Coercion, and Economic Growth in Eurasia, 17th–20th Centuries

Edited by Alessandro Stanziani

LEIDEN • BOSTON 2013 Cover illustration: Miniature of Indian craftsmen, painted by Svami in the 18th century. Courtesy of the Bibliothèque nationale de France.

Library of Congress Cataloging-in-Publication Data

Labour, coercion, and economic growth in Eurasia, 17th–20th centuries / edited by Alessandro Stanziani. p. cm. — (Studies in global social history, ISSN 1874-6705 ; 11) Includes bibliographical references and index. ISBN 978-90-04-23112-2 (hbk. : alk. paper)—ISBN 978-90-04-23645-5 (e-book) 1. Forced labor—Eurasia—History. 2. Labor—Eurasia—History. 3. Economic development— Eurasia—History. I. Stanziani, Alessandro.

HD4875.E83L33 2012 331.11’730950903—dc23 2012028033

This publication has been typeset in the multilingual “Brill” typeface. With over 5,100 characters covering Latin, IPA, Greek, and Cyrillic, this typeface is especially suitable for use in the humanities. For more information, please see www.brill.com/brill-typeface.

ISSN 1874-6705 ISBN 978-90-04-23112-2 (hardback) ISBN 978-90-04-23645-5 (e-book)

Copyright 2013 by Koninklijke Brill NV, Leiden, The Netherlands. Koninklijke Brill NV incorporates the imprints Brill, Global Oriental, Hotei Publishing, IDC Publishers and Martinus Nijhoff Publishers.

All rights reserved. No part of this publication may be reproduced, translated, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission from the publisher.

Authorization to photocopy items for internal or personal use is granted by Koninklijke Brill NV provided that the appropriate fees are paid directly to The Copyright Clearance Center, 222 Rosewood Drive, Suite 910, Danvers, MA 01923, USA. Fees are subject to change.

This book is printed on acid-free paper. contents

Notes on Contributors ...... vii

Introduction: Labour, Coercion, and Economic Growth in Eurasia, Seventeenth–Early Twentieth Centuries ...... 1 . Alessandro Stanziani

Part one JURIDICAL MODELS AND LABOUR DYNAMICS

The Duty to Work: A Comparison of the Common Law and Civil Law Systems from the Eighteenth to the Twentieth Centuries ...... 29 Simon Deakin

Dutch Imperial Anxieties about Free Labour, Penal Sanctions and the Right to Strike ...... 63 Ulbe Bosma

Children and Forced Labour in the Indian Ocean World, circa 1750–1900 ...... 87 Gwyn Campbell

part two DEPENDANCE AND SERVITUDE AT WORK: LOCAL CUSTOMS AND GLOBAL DYNAMICS

Factors that Shaped the Organization of Labor and the Labor Market in Tokugawa Japan: Kyoto and Central Japan ...... 115 Mary Louise Nagata

Contractual Relations, Tariffs and Customs in the Lyon Silk Industry in the Nineteenth Century ...... 145 Pierre Vernus vi contents

The Circulation of Commercial Manpower in an Indian Worldwide Trading Network in the Early Twentieth Century ...... 175 Claude Markovits

part three DANGEROUS TIES: STATE, LANDLORDS AND LONGUE DURÉE SERVITUDES

Constrained Labour in Early-Modern Rural East-Central and Eastern Europe: Regional Variation and Its Causes ...... 191 Markus Cerman

Rights and Bondage in Russian Serfdom ...... 215 Alessandro Stanziani

Acting As Master and Bondservant: Considerations on Status, Identities and the Nature of “Bond-Servitude” in Late Ming China ...... 237 Claude Chevaleyre

Public Works and the Question of Unfree Labour ...... 273 Chitra Joshi

References ...... 289 Index ...... 323 NOTEs ON CONTRIBUTORS

Ulbe Bosma is Senior Researcher at the International Institute of Social History in Amsterdam and published several books, among which are Being ‘Dutch’ in the Indies: A History of Creolisation and Empire, 1500–1920 (Singapore University Press/Ohio University Press), with Remco Raben, and articles on colonialism and migration in International Migration Review, Journal of Southeast Asian Studies and Journal of Global History. His recent research is on labour regimes, migration and sugar production in South and Southeast Asia.

Gwyn Campbell is a Canada Research Chair, and Director of the Indian Ocean World Centre, at McGill University. He has published extensively on Africa and the Indian Ocean world, including An Economic History of Imperial Madagascar, 1750–1895 (Cambridge, 2005) and David Griffiths and the Missionary “History of Madagascar” (Brill, 2012). He is currently completing a work entitled Africa and the Indian Ocean World from early times to 1900 to be published by Cambridge University Press.

Markus Cerman is Associate Professor of Economic and Social History in the University of Vienna. His research concentrates on economic development in late medieval and early modern Europe and comparative rural history. His publications include European proto-industrialization, Cam- bridge 1996 (with Sh. Ogilvie, eds.); Wirtschaft und Gesellschaft. Europa 1000–2000, Innsbruck 2011 (with F. Eder et al., eds.); Demesne lordship and rural society in East Central and Eastern Europe, c.1500–c.1800 (Agricul- tural History Review 59, 2, 2011, special issue) (with R. Hoyle, eds.); Villag- ers and lords in Eastern Europe, 1300–1800, Houndmills, 2012.

Claude Chevaleyre pursues doctoral studies in Chinese history at the Ecole des Hautes Etudes en Sciences Sociales (Paris), under the direction of Professor Pierre-Etienne Will (Collège de France). His dissertation is dealing with the social status and identity of so-called “bondservants” in the Ming (1368–1644) and Qing (1644–1911) dynasties. His research focuses on late imperial social and legal history and pays particular attention to the Ming-Qing transition period. Among his publications are Bibliographie d’Etienne Balázs: Œuvres, projets, appréciations”, in Actualité d’Etienne viii notes on contributors

Balazs (1905–1963). Témoignages et réflexions pour un centenaire. Paris: De Boccard, 2010, pp. 111–130. “Under Pressure and out of Respect for Human Dignity: The 1910 Chi- nese Abolition”. Forthcoming in a collective book, directed by Myriam Cottias et Marie-Jeanne Rossignol, on the worldwide influence of the western abolitions of slavery.

Simon Deakin is a Professor of Law at the University of Cambridge. He specialises in labour law, company law and private law, with a focus on the empirical and historical dimensions of legal change. He is the co- author of ‘The Law of the Labour Market’ (with Frank Wilkinson, 2005) and co-editor of ‘Capacitas: Contract Law and the Institutional Precondi- tions of a Market Economy’ (with Alain Supiot, 2009).

Chitra Joshi teaches history at Indraprastha College, University of Delhi, India. She is the author of Lost Worlds. Indian Labour and its Forgotten Histories (Delhi: Permanent Black, 2003 and London: Anthem Press, 2005) and of numerous book chapters and articles in i.a. Indian Economic and Social History Review, International Review of Social History and Studies in History. She is currently working on the history of roads and labour in nineteenth-century India.

Claude Markovits is a Senior Research Fellow Emeritus at the Centre National de la Recherche Scientifique and is affiliated with the Centre of Indian and South Asian Studies at the Ecole des Hautes Etudes en Sciences Sociales, Paris. His publications include The Global World of Indian Mer- chants (Cambridge: Cambridge University Press, 2000), The Un-Gandhian Gandhi: The Life and Afterlife of the Mahatma (Delhi: Permanent Black, 2004), and Merchants, Traders, Entrepreneurs (Palgrave Macmillan, Bas- ingstoke and New York, 2008).

Marie Louise Nagata is associate professor at the department of History, Francis Marion University. She is the author of Labor Contracts and Labor Relations in Early Modern Central Japan (Routledge Curzon Press, 2005) and has published numerous articles in the field of East Asian history, with a focus on family life and labor issues in Japan.

Alessandro Stanziani is Directeur d’études (full professor) at the EHESS and Senior Researcher at the CNRS, Paris. He works on Russian, European notes on contributors ix and the Indian Ocean history of Labour, Food and Economic Institutions. He is the author of: L’économie en révolution, le cas russe, 1870–1930, Paris, Albin Michel, 1998; Histoire de la qualité alimentaire en France, XIXe–XXe siècle, Paris, Seuil, 2005; Rules of exchange. French capitalism in comparative perspective, 18th–20th centuries, Cambridge, Cambridge University Press, 2012. Bâtisseurs d’Empires. Inde, Russie, Chine à la croisée des mondes, XVe– XVIIIe siècle, Paris, Liber, 2012.

Pierre Vernus is lecturer in contemporary history at the Univer- sity Lumière Lyon 2 and deputy director of the Rhône-Alpes Historical Research Unit (modern and contemporary history). He published Art, Luxe et Industrie. Bianchini Férier, Un Siècle de Soieries Lyonnaises, 1888– 1992 (2006) and articles about silk industry and business interests associations. He recently edited with Danièle Fraboulet Genèse des Organisations Patronales en Europe (XIXe–XXe siècles) (2012).

INTRODUCTION

LABOUR, COERCION, AND ECONOMIC GROWTH IN EURASIA, SEVENTEENTH–EARLY TWENTIETH CENTURIES

Alessandro Stanziani

The Main Scope and Argument

The history of the forms of “free” labour is intimately linked to that of coerced labour. This book shows that between the seventeenth and the twentieth century, forms of labour and bondage all over Eurasia were defined and practised in reference to each other. A whole spectrum of forms of dependence, bondage and labour existed in Russia, India and Indonesia as well as in Japan, China, Britain, France, Prussia and the Indian Ocean World. This was true not only in specific areas—rooted in local institutions, values and economic relationships—but also on a global scale. Serfdom in Prussia and in Russia expressed an extreme variant of Western European notions and practices of labour as domestic service and social obligation. Similarly, in the European colonies, indentured labour was conceived as an extreme form of domesticity and servants’ subordination in Europe. Without the identification of the worker as a servant in Britain, France and the Netherlands, indentured labour in the British, French, and Dutch colonies would have not been possible. Common tendencies and local perspectives on labour relationships found their sources not only in the global circulation of models, peoples, goods and institutions, but also in market dynamics. From the seventeenth and the late nineteenth century, proto-industry, agriculture, trade and manufacturing experienced unprecedented growth throughout Eurasia. Mostly labour-intensive, this long-term growth put considerable pressure on labour resources and contributed to increased coercion and legal constraints on labour mobility in both Asia and Europe.

The Uncertain Boundary between “free” and “unfree” Labour

Anthropologists, sociologists and historians have highlighted different aspects of labour relationships, according to their disciplines, in an 2 alessandro stanziani attempt to draw the line between “free” labour and “forced” labour, particularly slavery. They have discussed social status (membership in or exclusion from the clan, the family, the local community), religion, legal status (the form of dependence, freedom of movement, the hereditary character of such constraints), socio-economic conditions (dependence, non-economic advantages, coercion, etc.), political rights and legal (and procedural) rights.1 Researchers have pinpointed several variables, but without reaching a consensus. These issues have been debated even more fiercely in the last twenty years as cultural studies and subaltern studies brought out the relativity of the notions of freedom and coercion. As a result, the question has now become whether or not a given form of dependence, bondage, etc. found in a particular society in Africa, Asia, the Indian Ocean or the Americas could be considered “slavery”. If the answer is yes, then by implication slavery existed before and independently of colonialism; conversely, if the answer is no, it means that these forms of dependence and bondage were specific to a particular place and “imperialist” and revisionist culture would like to call them “slavery” to minimise the West’s “debt” to the Third World. The debates have become all the more virulent as they are no longer confined to academia. One obvious example is the quarrels that have pit- ted international organisations against countries and firms accused of practising and legitimising hidden “slavery”.2 It is symptomatic and sometimes paradoxical to see cultural relativism espoused both by academics critical of “imperialism”, “colonialism” and now globalisation and by local managers and multinational companies that exploit child labour. The aim of this book is not to take sides in favour of one or the other “general” definition of labour and forced labour, but rather to set the boundary line between free labour and forced labour in specific historical and institutional contexts and explain why, in a given context, this line was conceived and put into practice in one way rather than another.3 By undertaking a radical re-examination of the historical forms of labour and how they were defined, we are not seeking to relativise and deconstruct categories in order to assert, for example, that “forced labour did not exist”

1 Testart, L’esclave; Meillassoux, Anthropologie; Finley, Ancient slavery; Miers, Kopytoff, Slavery in Africa; Williams, Capitalism; Bush, Serfdom and Slavery; Engerman, Terms; Klein, Breaking; Patterson, Slaver; Lovejoy, Transformations. 2 ILO, International; IPEC, Every Child; Cunningham, Viazzo, Child Labour; Miers, “Con- temporary forms”. 3 Campbell, Miers, Miller, Children. introduction: labour, coercion, and economic growth 3 or that it is an “intellectual invention”. On the contrary, by viewing these elements in their proper historical contexts, we hope to provide an original explanation of the dynamics of forms of labour. Instead of attempting to establish the moment when “free labour” and “civilisation” emerged, or conversely, stigmatising the continuation of the “guild tradition” or even of latent forms of slavery, we want to grasp the dynamics at work in certain historical forms of labour starting from the historically situated tension between freedom and constraint. The first hypothesis underlying this book is that so-called “free” forms of labour and bondage were defined and practised in reference to each other, not only within each country and region, but also on a global scale. In the Anglo-Saxon world (chapter by Deakin), until almost the end of the nineteenth century, there were fewer “wage earners” than servants, i.e. workers whose status resembled that of servants in husbandry. These workers enjoyed limited rights in relation to their employers: the absence or unjustified breach of a work contract was punishable by criminal penalties. Yet it was precisely the definition of “free” labour, based on the unequal legal statuses of employers (actually masters) and wage earners (servants) that enabled several forms of bondage to be considered a contractual “free choice” at the time.4 In the British empire, the Ameri- can colonies and later the United States, from the seventeenth century to around the middle of the nineteenth century, the indenture contract5 which historians today usually classify as a form of forced labour, was in no way opposed to free labour during this period.6 The indenture contract was seen as an expression of contractual free will; in India, this form of servitude continued until the twentieth century.7 It is therefore important to note the formal and factual link between the forms of bondage and the definition of “free” labour: if the wage earner had not been defined as a servant, it would never have been possible to consider an emigrant signing an indenture contract a “free worker”. That explains why indenture was not classified as a form of “forced” labour in the colonies until the collapse of the Masters and Servants Acts, when free labour was redefined (1875) in Britain.8

4 Steinfeld, Coercion; Deakin, Wilkinson, The Law. 5 Steinfeld, The Invention; Galenson, White Servitude; Northrup, Indentured Labor. 6 Hay, Rogers, English Society; Hay, “Masters and servants”, in Steinmetz, Private Law. 7 Hay, Craven, Masters. 8 Northrup, Indentured Labor. 4 alessandro stanziani

The connection between the forms of labour in the “centre” and in the colonies was by no means specific to the British empire. It also existed in the colonies of the Dutch empire, especially Indonesia, where colonial leaders based their conception of post-slavery labour on the constraints imposed on servant wage earners in the Netherlands (chapter by Ulbe Bosma). The conceptions and practices of labour in Europe and its main colonies influenced each other and entered into global dynamics. Inden- tured labour was not the antithesis of “free” wage labour but an extreme form of the peculiar historical meaning that wage labour acquired in modern Europe. But then if the definition of “free” labour included the indenture contract and allowed for numerous constraints and penalties placed on wage earners, what distinguished it from officially recognised forms of bondage, beginning with serfdom? As a matter of fact, what has been called the “second serfdom” in East- ern Europe needs to be redefined.9 Serfs were never legally defined as such in Prussia10 and Central and Eastern Europe (chapter by Markus Cerman), or even in Russia (chapter by Stanziani). The documents usually cited as proof of the introduction of serfdom actually refer to forms of constraint on mobility and designated those who had the right to own and transfer inhabited estates, i.e. various categories of nobles. Forms of bondage beyond those justified by “free contracts” were widespread in India (chapter by Joshi) and in Africa (chapter by Campbell). These types of bondage close to slavery were based less on European law than on local customs. Hence, a continuum of solutions appears, with a variety of forms of debt bondage, indenture contract, and other forms of bondage contracts between English masters and Indian coolies, as well as genuine slavery.11 The same complexity was found in China (chapter by Chevaleyre), where the status of servants and bonded labourers differed from Euro- pean experiences, sometimes resembling more closely the situations of Ottoman palace slaves and servants in husbandry. The complexity of statuses and real conditions and the inadequacy of “European” categories to express them testify first to the widespread presence of these situations and second, to their specific character. The same issue arises in

9 Stanziani, “Free labor-forced labor”. 10 Melton, “The decline”; Ogilvie “Communities”. 11 Patnaik, Dingwaney, Chains; Tinker, A New System; Pouchepadass, Paysans; Condom- inas, Formes extrêmes; Reid, Slavery. introduction: labour, coercion, and economic growth 5 every study of slavery as soon as we step outside of the ancient world or nineteenth-century North America, where in both cases slave status was passed on to descendants. Yet local specificities must be put into a global context. In time as well as in space, continuities and links between free and unfree labour prevail over clear-cut oppositions. Unlike the conventional thesis—celebrating the triumphal march towards freedom starting in the nineteenth century after centuries if not millennia of bondage—recent approaches in historiography underline the changes in slave-trading systems prior to official abolition and, conversely, the continuation of forms of bondage and slavery after the reforms. The conditions of African-Americans and of “free” labourers in the colonies are used to confirm this position.12 Similarly, historians of serfdom in Eastern Europe have revealed the decadence of the system well before the arrival of Napoleon and the French codes,13 just as French historians have shown how medieval serfdom evolved prior to the Enlightenment.14 Both groups have also brought to light aspects of bondage and service in husbandry that continued after the official abolition of serfdom.15 New trends have also arisen among historians of wage labour. Anglo-Saxon historians have underscored the coercive nature of the British and American norms applied to labour (Masters and Servants Acts) until roughly the end of the nineteenth century.16 In France, the break between the world of guilds and the Old Regime on the one hand, and the world of free labour on the other, has likewise been questioned.17 The rise of collective bargaining has been deemed at least as important in terms of its effects on labour relations as the abolition of guilds and of labour service (corvée).18 Our approach confirms these new contributions of historiography and emphasises the continuities rather than the breaks between the late seventeenth century and the early twentieth century in the area of labour and its institutions. Despite institutional and political shifts, important continuities in labour regulations and practices are revealed and confirmed by the cases of Britain (Deakin), France (Vernus), Austria and Prussia (Cer- man), Japan (Nagata), China (Chevaleyre), and Russia (Stanziani).

12 Miller, Slavery and Slaving; Drescher and Engerman, eds., A Historical Guide. 13 Hagen, Ordinary. 14 Bois, La crise; Duby, Les trois ordres; Bonnassie, From Slavery. 15 Hagen, Ordinary, 2004, Kolchin, Unfree Labor. 16 Steinfeld, The Invention. 17 Minard, La fortune. 18 Didry, Naissance. 6 alessandro stanziani

Continuities were even stronger in colonial empires where a significant link has been established between the conditions of European wage earners (mainly servants in husbandry and household servants) and indentured immigrants in Indonesia (Bosma) and throughout the Indian Ocean World (Campbell). However, the continuities between free and unfree labour are important not only in time but also in space. Surprisingly enough, the recent research orientations we have just mentioned have never come together: there has been little dialogue between historians of slavery and historians of wage labour, and consequently neither group has challenged the presupposition that these two worlds were and remain separate or even opposed. The aim of this book is precisely to overcome this fracture by revealing the connections between these elements underpinned by chro- nologies that are in fact too common to be unconnected or to have come about merely by chance. To answer this question, the historical dynamics of labour must be understood at once in a global dimension and in their local specificities. We will therefore examine them on these various levels first by studying the circulation of economic and legal knowledge and second, how it was applied locally.

Circulation of Knowledge and Labour Dynamics

Not all the contributions to this work necessarily adopt a global or comparative approach. On the contrary, we think specific national and even regional and local features should be taken into consideration in order to understand how the whole system worked. The strength of global history lies not in collecting second-hand banalities common to a number of different worlds, but rather in achieving a relevant representation of this multiplicity through local and national specificities. The comparative dimension applied at various levels is therefore an integral part of our method, which is why we propose to place side by side situations and historical experiences that are usually studied separately: wage labour in France and England, Russian serfdom and slavery in China, work discipline in India and Japan. This approach reveals striking similarities as well as the differences between the contexts and forms of labour. From the standpoint of global history we have adopted, two main variables are worthy of mention: the circulation of economic and legal knowledge and economic and institutional dynamics between the seventeenth and twentieth centuries. Economic knowledge and legal models introduction: labour, coercion, and economic growth 7 circulated along with people and goods. This circulation led not only to increased homogeneity among systems but also to differentiation and even hierarchies of areas and countries. For example, the identification of serfdom with an archaic world in the “East” and even the notion of “Eastern Europe” itself were pure inventions of the Enlightenment.19 Mon- tesquieu published The Spirit of the Laws in 1748, soon to be followed by the first volumes of the Encyclopédie.20 In these works, the serfdom of absolutist and medieval Europe was contrasted with the free labour of Enlightenment Europe. The advances of the Enlightenment contributed to the invention of a historiographic break between “enlightened France” and the old France on the one hand, and between Western and Eastern Europe on the other. These two “inventions” mirror each other, e.g., in the fact that serfdom could be found in Eastern Europe as well as in medieval Europe. Similarly, British and Dutch colonisers in India and Indonesia strove to translate local forms of dependence and bondage into their own categories; the disparities and similarities between the forms of slavery and bondage in the two worlds are key factors in grasping the complexity of the definitions and of the labour practices themselves (chapter by Ulbe Bosma). If labour practices had not been translated into English legal terms, the forms of post-slavery and post-colonial bondage would have developed differently.21 Conversely, the use of written contracts rather than informal arrangements reveals the need on the part of important Indian traders in Cairo to be able to bring their shop assistants and servants before English courts if necessary (chapter by Markovits). The contracts sought to translate the relationships between caste, type of labour and remuneration into terms and categories derived from British norms. Yet the circulation of ideas and practices did not necessarily indicate a relationship of dependency of the so-called “periphery” on the core; colonial discourse aside, forms of bondage indeed existed in Russia, India, Africa and the Americas well before the arrival of Europeans.22 And even during the colonial period, dependency was not simple and straightfor- ward. To return to the example of Indonesia, servitude was defined and put into practice based on indigenous notions and practices that had

19 Stanziani, “Free Labor”. 20 Duchet, Anthropologie. 21 Prakash, Bonded Histories. 22 Coquery-Vidrovitch, Moniot, L’Afrique noire; Thornton, Africa; Meillassoux, L’esclavage. 8 alessandro stanziani

continually influenced those discussed and applied in the Netherlands itself (chapter by Bosma). In like manner, in the eighteenth and nineteenth centuries, there were constant references to the case of Russia in debates over work discipline in Europe, where the assessment of the conditions of “serfs” led people to take a position regarding wage labour, which was viewed either in opposition to serfdom or, on the contrary, considered a new form of slavery.23 From this same point of view, the specificity of Japan did not consist in simply adapting the Western model. In traditional historiography, Japanese development has always been tied to Meiji reforms and West- ern influences. On the contrary, recent analyses including Marie-Louise Nagata’s contribution here, demonstrate that Japan’s economic and social dynamism dates back to the eighteenth century. To sum up, the circulation of ideas and models is important, but it cannot be taken exclusively as a synonym of dependency because 1) the “centre” and its colonies often influenced each other, 2) bondage and slavery did in fact exist, although in different forms, before and after colonisation and 3) finally, emancipation did not come about solely under pressure exerted by the “advanced West”. The role of the circulation of knowledge therefore has to be associated with economic and social trends as a whole.

Global Intensification of Labour

How can we possibly justify the increasing limitations placed on free labour throughout Eurasia between the seventeenth and the nineteenth centuries? Was this not a period of strong demographic growth and increasing capital development? Logically those processes should have led to reducing constraints on labour rather than strengthening them. This line of reasoning fails, however, to take into account the above-mentioned values and notions of labour as a social service and an obligation on the one hand, and the fact that this period witnessed unprecedented labour-intensive growth all over Eurasia. Indeed, over the past two decades at least, economic historians have moved away from the interpretation developed in the early nineteenth century by the so-called “classical” economists (Smith, Ricardo)

23 Stanziani, “Free Labor”. introduction: labour, coercion, and economic growth 9 and Marx, according to which the agrarian revolution and later the industrial revolution were accompanied by the substitution of labour by capital and the concentration of production units. In fact, they occurred mainly in the twentieth century. Before that date, in most of Eurasia, proto- industry, agriculture and even industry rather followed a labour-intensive path. First, with regard to agriculture: conventional histories of economic growth stress the relative decline of agriculture during the industrialisation process; at the same time, agriculture was supposed to provide goods to feed a growing urban population. This outcome required greater productivity and yields, which in turn were obtained through a shift in the organisation of economic units and in the relative weight of the factors (decreasing labour and land and increasing capital). These views have been increasingly challenged, starting with the role of capital in agriculture. Recent empirical analyses show that in many parts of England livestock densities were stable throughout the modern period until the mid-nineteenth century;24 conversely, there was more livestock in Asia and Eastern Europe at the time than is commonly believed.25 In other words, in Britain, the increase in yields that occurred before 1800 cannot be explained by rising livestock numbers.26 On the rest of the continent as well, the long-term trend of rising wheat prices (roughly between 1680 and 1815) led to reducing the surface area devoted to livestock and livestock feeding while increasing the acreage for wheat cultivation. The role of machines and new sources of power is also undergoing revision: after steam became the dominant form of power employed in manufacturing, the major sources of energy available to farmers continued to be men, animals, wind and water.27 Mechanisation in farming proceeded slowly because agricultural operations were more separated in time and space than industrial processes. Mechanisation was therefore a relatively unimportant component of the changes in agriculture technology up to the mid-nineteenth century28 when the appearance of commercial fertilizers and the development of mechanical harvesting equipment began significantly to affect methods of production.29 Until the machine age, i.e. after 1850, much of the rise in productivity and the growth of output depended

24 Allen, Enclosures. 25 Pomeranz, The Great. 26 Allen, “Tracking”: 226. 27 O’Brien, “Agriculture”. 28 O’Brien, “Agriculture”. 29 Grantham, “Agricultural Supply”. 10 alessandro stanziani more on the intensive use of known technology than on novel methods. So-called “new husbandry” was indeed not so new30 and it required more labour, not less.31 Only in agricultural systems like those in the United States, characterised by the high opportunity cost of labour, did economic pressure to mechanise come to resemble that experienced by some sectors of industry. On the contrary, in Russia and France as well as Prussia, from the seventeenth century through the second half of the nineteenth century, labour was not only the major input in agriculture, either directly or embodied in land improvements, but its weight even increased during this period.32 Recent analyses converge towards the same conclusion: labour and labour intensity are now identified as the main source of agriculture growth before 1850, with human and physical capital playing a secondary role.33 Labour-intensive techniques linked to the dissemination of knowledge and attractive markets (with rising agriculture prices) were widely used between the seventeenth century and the last quarter of the nineteenth century, when this trend reversed (agricultural prices went down and wages went up).34 However, increasing labour demand in agriculture had to compete with similar processes in proto-industry and manufacturing. Theories of proto- industrialisation originally associated proto-industry with demographic growth and proletarianisation on the one hand, and a decline of urban guilds and feudal institutions on the other.35 Further analyses have called these assertions into questions and nowadays it is widely accepted that on a comparative scale no single, uniform link can be established between proto-industrialisation and any of the aforementioned variables. Proto- industry developed in Western, Central and Eastern Europe from the end of the seventeenth century in response to market demand and demographic pressure.36 It retained its central position all over Europe at least until the mid-nineteenth century. After that date, and only after, some areas declined and manufactures and industries replaced the putting- out system.37 However, the shift was by no means complete, and in

30 Federico, Feeding. 31 O’Brien, “Path Dependancy”. 32 Clark, “Productivity growth”. 33 Grantham, “Agricultural Supply”; Allen, Enclosures; O’Brien, “Agriculture”. 34 Thompson, “The Second”. 35 Mendels, “Des industries rurales”; Mendels, “Proto-industrialization”; Kriedte, Medick, Schlumbhom, Industrilization before industrialization. 36 Hagen, “Capitalism in the countryside”. 37 Ogilvie, Cerman, European proto-Industrialization. introduction: labour, coercion, and economic growth 11 many European areas and districts, proto-industry continued to play a leading role during the second half of the nineteenth century and even in the twentieth century.38 This timing is even more relevant for Asia: authors such as Lee and Sugihara have maintained that the “industrious revolution” identified by De Vries in Europe39 was also visible in areas of Asia.40 As in Russia, the success of proto-industry in Japan (chapter by Nagata) and China, too, particularly rural proto-industry, was at the root of a labour-intensive path of growth.41 In all these areas, as in most Russia regions (chapter by Stanziani), agriculture did not turn into a simple supplier of produce and labour-force for industry; on the contrary, estates and peasants took part to the development of local and national markets for both wheat and proto-industrial products. In Russia (Stanziani), Japan (Nagata), and Central Europe (Cer- man), commercialization of produce was not the consequence of extra- economic coercion but responded to market dynamics.42 As in many areas of Russia and Western Europe, increases in agricultural output and income in Japan led to growth in demand for manufactured goods, which was met by an expanding rural industry using labour-intensive technology. The resulting growth in rural non-agricultural activity in turn generated increased incomes for rural households and hence greater demand for agricultural output.43 Similar results are now available on India. Contrary to the usual view, which stressed the decline of cottage industry under British rule and the growth of international markets, new research shows that the use of “traditional” labour-intensive techniques has remained widespread up to the present day. Thanks to their flexibility, these techniques have allowed a labour-intensive pattern of growth linked to family units to remain integrated in both agriculture and industrial markets.44 The persistent, global strength of agriculture and proto-industry had an unanticipated effect, however: urbanisation and the supply of labour for urban manufacturing were mostly seasonal.

38 Sabel, Zeitlin, Worlds of possibilities. 39 De Vries, “The industrial revolution”. 40 Sugihara, “Labour-Intensive”; Lee, “Trade and Economy”. 41 Frank, Rural Economic; Pomeranz, The Great. 42 Sugihara, “Labour-Intensive”. 43 Frank, Rural Economic; Saito, “The Labour Market”. 44 Roy, The Economic History; Perlin, “Proto-Industrialization”. 12 alessandro stanziani

Until the mid-nineteenth century, double employment (rural and urban) was the rule rather than the exception not only in Russia (Stan- ziani), Japan (Nagata) and France,45 but also in Britain. According to Lindert and Williamson, for the larger occupational groups such as “agriculture”, “commerce” and “manufacturing trades”, the census and statistical error margins are probably within the range of minus 40 to plus 66 per cent!!46 Seasonal needs in agriculture were the crucial variable here. In Rus- sia (Stanziani), China and Japan (Nagata) as well as Britain, France, and Central Europe (Cerman), the solution to seasonal and local shortages of manpower lay in interregional migration and, later in the nineteenth century, in the transformation of hand harvesting techniques and tools.47 In fact, the labour requirements for harvesting were especially high since labour input peaked sharply at harvest time.48 Thus, as Markus Cerman clearly shows in his chapter, landowners in Brandenburg satisfied part of their manorial labour needs by requiring the children of their peasant tenants to serve on the manor for a period of three years. However, at harvest time, the labour services provided by peasants and their children fell short and recourse to seasonal day labour was widespread. The proportion of rural dwellers that worked part time for corn growers depended on what they did outside the peak season. These activities were extremely problematic for rising manufacturing and industry which, like the other parts of the economy, mostly relied upon labour. Because the price of capital was still high, urban employers sought to cope with labour shortages by putting legal pressure on labourers, who were not allowed to leave before the end of their terms, as well as on competitors through strong penalties for unfair competition, etc. The first industrial revolution in Britain was far from the overall substitution of labour with capital. Feinstein’s estimations show that in Britain, capital and labour grew at about the same rate from 1760 through 1830, so that there was effectively no change in the capital-labour ratio during these seven decades. In the last three decades, the ratio slowly rose, as capital per worker increased at a rate of about 1/2% per year.49 The rate

45 Postel-Vinay, “The Dis-Integration”. 46 Lindert, Williamson, “Revising”; Lindert, Williamson, “English workers”. 47 For Britain: Collins, “Migrant labour”; for france: Postel-Vinay, “The Dis-integration”; For Germany: Melton, “Population structure”. 48 Grantham, “Divisions of labour”. 49 Feinstein, “Capital Formation”. introduction: labour, coercion, and economic growth 13 of capital formation in Britain was relatively slow on the whole until the mid-nineteenth century and the capital-labour ration increased sharply only afterwards.50 By 1850, relatively few workers were employed in factories: only a small proportion worked in technologically advanced industries such as cotton, iron and steel, and metalwork, and the full impact of steam power in transport and production was yet to be felt.51 This view breaks with the traditional scenario of the first industrial revolution and strongly reduces the gap between Britain and other countries, particularly France. Histori- ography long considered France “backward” in comparison with Britain, precisely because of its slow rate of concentration and capital intensification in industry.52 Recent analyses have consistently modified this view: if British capital intensification and the increasing capital-labour ratio in industry have been revised downwards, on the contrary, French dynamics have been corrected upwards. From this standpoint, the relatively slow pace of capital growth in France, and the increasing ratio of labour to capital industry are no longer seen as exceptional or “inefficient”.53 This means that not only in Russia, but also in France and most of the European countries, economic and industrial growth in the eighteenth and nineteenth centuries remained on a small scale and was labour intensive. Growth was most often achieved within the same production function whose scope expanded slightly until the mid-nineteenth century.54 There is evidence that much of the productivity increase was not associated with specific innovations, but rather with workers operating more machines.55 Christine Mac Leod reveals that the most frequently declared goal of innovation was either improving the quality of the product or saving on capital, not labour. And if inventors were not particularly intent on saving labour, those who judged their inventions were even less so. In other words, economic actors did not wish to substitute labour with capital and the final outcome for the whole economy was an increasing demand for labour.56

50 Craft, British Economy; J. Williamson, “Why was British growth”; Harley, “British industrialization”. Deane, “Capital Formation”; Feinstein, Pollard (eds.), Studies in capital formation; Craft, British Economic Growth. 51 Deakin, Wilkinson, The law: 20. 52 Crouzet, British ascendant; Lévy-Leboyer, Bourguignon, L’économie française. 53 Craft, British Economic; O’Brien, Economic Growth in Britain and France. 54 Craft, British Economic. 55 Clark, “Productivity Growth”. 56 Mac Leod, Inventing: 158–181. 14 alessandro stanziani

Overall, in British industry during the second half of the eighteenth century, labour input grew at about 1.2–1.3 per year, 2/3 of which was caused by a larger population, and the remaining third coming from longer working hours.57 Following E.P. Thompson,58 legions of historians have endea- voured to show that working time actually increased with the industrial revolution. Indeed, this process started much earlier, in the seventeenth and eighteenth centuries, with the “industrious revolution”59 and the mul- tiplication of farming and industrial activities.60 Thus, the global history of the eighteenth and nineteenth centuries in all the countries we mentioned shows a similar evolution of working time, i.e. workdays became longer as labour became more intensive.61 The innovations and technical improvements recorded in agriculture, industry and trade did not take place at the expense of labour but actually fostered employment. This mechanism is confirmed in the study of manufactures in France and Eng- land, Japan and India. From the point of view of labour institutions and the relationship of labour to other production factors, it thus became a common “wave of capitalism” (to borrow Braudel’s expression but not his chronology) that lifted Eurasian capitalism as a whole from the seventeenth to the mid-nineteenth century. Hence questions concerning working time deserve to be studied within a more complex set of dynamics: if labour was already becoming more intensive in every sector during the preindustrial era, then how could presence at work be controlled when families, landowners, traders and manufacturers were competing for it?

Child and Female Labour

In labour relations, children have always played a crucial role in every context as the focus of controversies relating not only to economics but also the social order, religion and institutions. Of course it is necessary to consider how “child” was defined, which changed according to the period and place62 (chapter by Gwyn Campbell). At the same time,

57 Voth, “Time and Work”. 58 Thompson, “Time, work-discipline”. 59 De Vries, The industrious; Voth, “Time and work”. 60 Craft, British Economic Growth; Mokyr, The Economics. 61 Cross, Worktime; Cross, A quest; Fridenson, Reynaud, La France et le temps. Schmiechen, Sweated Industries. 62 Ariès, Centuries. introduction: labour, coercion, and economic growth 15 this relativism does not rule out identifying similarities and differences as well as continuities and breaks between these definitions and child labour. The differences concern the status of children relative to that of their parents: the children of slaves were not always or necessarily slaves themselves. Conversely, pawnship, i.e. the pawning of children by debtors, was a form of non-hereditary belonging, at least from a formal standpoint. Above all, unlike slavery in the strict sense, pawnship expressed a form of integration in society rather than the alienation of a slave in another society.63 Hereditary status influenced the evolution of families and societies before, during and after Western expansion. Slavery as practised by Westerners in Africa led first to an intensification of local forms of bondage. Later on, when the Westerners themselves decided to prohibit pawnship, which was considered slavery, new forms of dependency were introduced in various African societies. This interaction with the Western world was essential: the status of children in eighteenth century England was linked to the authority of the head of the household, the employer and children’s rights. Children—as children or as apprentices—had fewer rights in relation to their fathers/employers or their masters outside the family. These unequal rights accompanied and supported the industrious revolution64 and later the industrial revolution in the strict sense, both in Europe (chapter by Deakin) and in Japan (Nagata). Child bondage was not incompatible with the growth of the industrial world and that connection was transplanted to the colonies, where slaves were immediately treated like children with fewer rights. When slavery was abolished, first the British (1832–1842), then the French (1848–1860) and towards the end of the nineteenth century the Dutch (chapters by Bosma and Campbell) assigned temporary “apprentice” status to their former slaves.65 This status was naturally used to demonstrate that former slaves were not fully civilised, but also and more precisely, the fact that, as apprentices in Europe, they were still subject to the authority of their masters/employers. The special status attributed to children and former slaves underpinned the encounter between European countries and Asian and African societies in the nineteenth century. These worlds had a reciprocal influence on each other, which led to perpetuating a special status for children within the societies concerned.

63 Lovejoy, Fayola, Pawnship. Lovejoy, Transformations. 64 De Vries, The Industrious. 65 Davis, The Problem; Engerman, Terms; Lovejoy, Transformations. 16 alessandro stanziani

Women, too, were subject to unequal rights. Keeping in mind the fundamental differences between matrilineal and patrilineal societies underlined by Laslett, Meillassoux and many others,66 it is nevertheless important to think about how these forms of hierarchy interacted and how people were placed in situations of dependency or slavery. In several African and Asian societies, even before the arrival of the Europe- ans, female slaves were sold for a specific use. Concubines and servants certainly played a role, but the rise of rural and later plantation slavery changed the context. The type of crops influenced the choice of male or female slaves, and hence, the price of women. Once again, the status of women was related to the form of economic activity in the broad sense of the term in both “free” and slave-holding societies. In Europe, women had fewer rights than men and these inequalities were intensified rather than reduced by the advent of urbanisation and industrialisation in the eighteenth and nineteenth centuries. The diminished rights of women workers, not to mention widespread workhouses in Britain as well as France, Germany and Russia, testify to these continuities.67 This trend was not reversed until the twentieth century, and even then, mainly during the second half.68 In view of these facts, it is possible to conclude that, not only in the seventeenth but also the eighteenth and much of the nineteenth centuries, the continuing importance of cottage industries and frequent migration between the city and the country created a strong link between two types of constraint: presence at work and compliance with working hours, on the one hand, and competition among employers (including heads of households), manufacturers, landowners and trader-entrepreneurs for the control and appropriation of labour, on the other. This is where presence at work and discipline encountered institutions and labour law. Along with the rules governing workshops, agricultural estates and plantations in Europe and Asia, a set of provisions was devised to control mobility and presence at work, such as the worker’s booklet, laws against poaching workers and begging, forms of bondage, etc. Constraint in the organisational sense became linked to institutional constraints; presence at work conveyed concerns about internal organisation, competition and public

66 Meillassoux, Anthropologie; Laslett, The world. 67 On France: Perrot, Les femmes; on Britain: Burnette, “An Investigation”; on the USA: Goldin, “The Quiet Revolution”; on Russia: Alpern Engel, Women in Russia; on Germany: Biernacki, The Fabrication. 68 Orloff, “Gender and the Social Rights”. introduction: labour, coercion, and economic growth 17 order. Yet if this was the case, what were the specificities of each area and how did they enter into global dynamics?

Empires of Eurasia: Global Dynamics/Centres of Development

Why Eurasia? We chose to focus on this area for a number of reasons, first of all because studies of the evolution of labour in “Western” Europe has been excessively dominated by a Eurocentric approach that views the industrial revolution and the French Revolution as the major breaks. On the contrary, we would like to show that those turning-points were merely partial, at least as far as labour institutions were concerned, and that the dynamics at work in France, England, Russia and Europe in general can be grasped only in their interaction with other parts of the world.69 Our decision to focus on Eurasia and only in one paper (Campbell) on Eastern Africa in no way indicates a lack of interest in the rest of Africa and the Americas. The choice was partly based on the existing bibliography pertaining to slavery and labour. Researchers, especially in the last few years, have written at length about the interaction between Europe and the Americas and between Europe and Africa, whereas the relationships between Europe and Asia have received relatively less attention. Examining Europe and Asia together also has the advantage of avoiding retrospective thinking about Europe; we will see that labour institutions and practices in Europe were connected to what was happening in its colonies and in Asian empires. Of course it was necessary to select particular areas. We have included a great number of regions of Asia and Europe: Russia, Great Britain, France, Prussia and Austria, India, Japan, China, Indonesia and the Indian Ocean region extending from China to Eastern Africa. These choices were made not because the regions are statistically representative, but rather because they are especially relevant to the questions we are asking. Thus, the French case is of interest not because it was the land of Colbertism opposed to liberal England, or because nineteenth century France was the country of free, codified law compared with Germany, which still lagged behind. On the contrary, France is of interest because its labour norms in the nineteenth century were actually quite well suited both to a capitalist economy and to the heritage of the Old Regime. This case is all the more

69 Pomeranz, The great. 18 alessandro stanziani likely to raise new questions as, contrary to a widespread preconcep- tion, common law in England was in fact accompanied by a considerable degree of regulation and state intervention and labour remained subject to criminal constraints until the end of the nineteenth century (chapter by Deakin). Highlighting the case of France and comparing it with Eng- land leads us to question the differences between liberalism and regula- tionism or between free labour and guilds within the capitalist world, and from there to narrow the distance separating so-called “free” labour from the varieties of bondage. We have paid special attention to India, in particular, because the time is ripe for a new analysis of the forms of dependence on England and its Empire specifically based on labour status. Unlike traditional analyses, these new studies, including the excellent examples presented here, are drawn from archival sources and provide a concrete picture of how norms and the labour market actually worked. We will show that while British norms and perceptions translated into various forms of bondage and slavery in India, and thereby helped perpetuate slavery well after its official abolition, the latter nevertheless existed prior to any British intervention. The solution adopted in India and the practices that were accepted did not result solely from British influences, but rather from interaction between those influences and local traditions. From this point of view, the chapters presented here by Joshi and Markovits concur with those of Campbell on the Indian Ocean and Bosma on Indonesia as well as with an increasingly extensive bibliography demonstrating the complexity of non- Western area “dependence” in the nineteenth century, thereby explaining the rise of proto-industry and then industry in India from the nineteenth century to the present day.70 Finally, the case of Japan (chapter by Nagata) stands out because the ste- reotyped image of its economic dynamism as a reproduction of the West- ern model adapted to its own institutions and local culture is countered here by an analysis that emphasises the specificity of Japanese capitalism within a global dimension. Proto-industrial growth in the eighteenth and nineteenth centuries and the role of the family in those dynamics were part of a more general trend of the worldwide economy. The French, Japanese and Indian examples, with the similarities and differences we have just mentioned, are put to the test by comparing them with other regions in Asia and Europe. The Russian empire is interesting

70 Roy, The Economic History. introduction: labour, coercion, and economic growth 19

(Stanziani): although serfdom and its history have been abandoned for decades, they constitute a kind of intellectual icon, apparently unshake- able in spite of new studies on slavery and on serfdom in Eastern Europe and above all despite new data in Russian economic history revealing considerable economic and demographic growth at the time of serfdom. We explain this data by showing that Russian serfdom was never actually institutionalised and was in fact more flexible and more complex than is usually assumed. These conclusions coincide with those regarding Central and Eastern Europe where the evolution of institutional relations, social relationships and economic dynamics did not wait for Napoleon to adopt a market orientation. As the chapter by Markus Cerman shows, there was sizeable proto-industrial growth at work in these regions. Once serfdom—especially the second serfdom—has recovered its place within the comparative history of forms of labour, we will take another look at the differences and similarities in relation to other possible configurations of the labour world. Within the structure of this book, China in particular seems to pose a problem.71 The chapter by Chevaleyre demon- strates the specificity and complexity of the forms of bondage and the special legal status of bonded people compared with the familiar social and legal forms in the West during the same period. Like the chapters devoted to other contexts, the one on China attacks the traditional breaks proposed by historians. In particular, it re-examines the transition from the Ming dynasty to the Qing dynasty around 1644 precisely from standpoint of the status and condition of serfs, which encompassed various situations ranging from servitude and debt bondage to hereditary slavery. The continuities exceed the so-called changes in Chinese history. Indeed, historical research72 has underlined a Chinese specificity compared with the rest of the Eurasian continent: the studies show that slavery and the most restrictive forms of bondage were weaker and eliminated sooner in China than in the West. This outcome is often explained by the demographic factor, namely, the abundance of available labour in the country.73 This is an altogether traditional, rather widespread argument in the literature on

71 I acknowledge my debt to Kenneth Pomeranz, Christine Moll-Murata, Yves Chevrier, Christian Lamouroux, Choi Wai for having helped me on this point. 72 Zurndorfer, Change; Hansson, Chinese Outcast; Campbell, Lee, “Free and unfree”; Cartier, “Travail; Schottenhammer, “Slaves”. 73 Pomeranz, The Great, 82–84. 20 alessandro stanziani the historical forms of forced labour. Following Herman Nieboer,74 count- less scholars have explained forced labour by population scarcity.75 This argument is now being challenged by historians of serfdom and slavery and of wage labour. The latter have shown that coercion in wage relations increased rather than declined between the early 1750 and the mid- nineteenth century Britain, at a moment when the population growth was already perceptible.76 On the other hand, the cases of Australia and Canada testify to the fact that the colonisation of new territories did not necessarily call upon massive imports of slaves, as in the United States.77 To be sure, in Aus- tralia, convicts initially played an important role but their number were small compared to forms of indentured and contract labour immigration.78 These last forms also prevailed in Canada in the eighteenth and early nineteenth centuries. Although “free” and “unfree” immigration can no longer be opposed as such,79 and conditions for indentured immigrants were harsh (in particular for Asian immigrants during the second half of the nineteenth century), they still cannot be assimilated to slaves because their legal status was never passed on to their children and they could always return home. However difficult their situation, they did not endure the radical exclusion of slaves.80 Thus, while Canada and Australia were not lands of freedom, they never experienced massive slave immigration like Brazil or the United States. Indeed, for indentured contracts in general, the forms of immigration to Australia and Canada seem to have been linked far more to immigrants’ motivations than to population scarcity.81 Last but not least, the case of Russia (chapter by Stanziani) also shows that the introduction and implementation of serfdom rules was never connected to population density and the availability of labour. Thus, from a reverse angle, it seems legitimate to have doubts on this subject with regard to China as well. In comparison to the rest of world, the demographic rise in China, already considerable in the sixteenth and

74 Herman Nieboer, Slavery as an Industrial System. Cambridge: Cambridge University Press, 2010 (Original: 1900). 75 Just a few example in a huge bibliography using this argument: for Russia: Blum, Lords; for the USA: Fogel, Engerman, Time. 76 Deakin, hereafter. 77 Engerman, Terms, “Introduction”. 78 Northrup, Indentured Labor. 79 Lucassen and Lucassen (eds.), Migration. Eltis, Coerced and free migration. 80 Northrup, Indentured Labor. 81 Galenson, White servitude. introduction: labour, coercion, and economic growth 21 seventeenth centuries, even accelerated in the eighteenth century.82 At the same time, there is no proven correlation between the increase of population and the abolition of bondage. There were significant regional differences in mobility throughout the nineteenth century. In particular, though restrictions on the movements of a portion of the rural population were soon abandoned—in certain regions and for certain peasants, especially those on government-owned estates—, constraints nevertheless survived for a long time, at least until the end of the nineteenth century.83 In general, a kind of caste separation between liangmin (good people) and jianmin (lowly people) continued until the nineteenth century. These different statuses were hereditary; the category of lowly people included servants, prostitutes and musicians as well as craftsmen and wage earners. These categories enjoyed fewer rights than the other social groups. They were also subject to more severe criminal punishments and greater discrimination in labour relationships.84 In this context, debt bondage, bonded servants and rural labourers were quite widespread, not to mention slave prostitutes.85 Though there are no hard figures to quantify this phenomenon, legal disputes on the subject suggests it was a considerable problem throughout the nineteenth century.86 And even until the late twentieth century, serfdom and rural slavery were widely practised by the Yi in the mountains of Sichuan and Yunnan.87 In the cities, too, the heritage of the guild system, which guaranteed fair competition between employers just as it did in Japan (chapter by Nagata), conditioned the mobility of labourers by laying down rules of apprenticeship and remuneration.88 All the same, it would be worth doing detailed studies on the forms of labour in China in the eighteenth and nineteenth centuries until the beginning of the twentieth century to determine whether forms of constraint and bondage were maintained or actually disappeared. In short, it is possible to identify similarities and differences in work constraints in Eurasia. The similarities lie mainly in the dissemination of

82 Lee, Feng, “Malthusian”. 83 Campbell, Lee, “Free and unfree”. 84 Hansson, Outcasts. 85 Schottenhammer, “Slaves”. 86 Hansson, Outcasts. 87 Hui, “Les esclaves”. 88 Moll-Murata, “Chinese Guilds”. 22 alessandro stanziani those constraints from the end of the seventeenth century to the early twentieth century, along with an intensification of labour during periods of agricultural, trade, proto-industrial and industrial growth. Labour intensification in support of growth, rather than a workforce shortage as such, accounts for the constraints. Continuity in the forms of labour regulation during this period was particularly evident in Japan and Great Britain. The differences concern primarily the strength of proto-industry, which was greater in Asia than in Europe, but also the early liquidation of the forms of bondage in China. This liquidation, which no doubt should be qualified, was in any case accompanied by a strengthening rather than the disappearance of the forms of infra- and inter-village solidarity, especially during migrations to the city. At the time, the same forms of solidarity were found in Russia, which, unlike China, nevertheless maintained “serfdom” well beyond the period of eighteenth century growth. Finally, the forms of labour in Japan stood in stark contrast to these models and were equally far removed from the solution adopted in West- ern Europe around the end of the nineteenth century, i.e. the social state. Japanese enterprises encouraged workforce loyalty and set up internal systems of social protection, both of which were among the main arrangements developed to support the transition from small to large industry. At the same time, these comparisons at the national and imperial level are valid only as a rough approximation. No doubt legal rules (civil, tax and customs laws) refer to the national dimension of these phenomena, yet those rules were only one of the components of economic action, along with symbolic, cultural and political aspects. Hence, we cannot ignore the importance of local components and the strong differences between the dynamics of different regions within a single country. Let us look at that dimension now.

Global versus Local Labour Dynamics

The forms of forced labour—the existence of bondage or even slavery alongside “free” labour—often varied from one city to the next and from one place to another. This observation is especially relevant in our case as the institutions and economic activities in the world we are studying were extremely fluid, multiple and local from the eighteenth to the early twentieth century. Several institutions coexisted at the local level and even when a process of national unification took place, institutional introduction: labour, coercion, and economic growth 23

pluralism continued.89 It was more widespread at the level of empires where legal pluralism was an important instrument of economic and political action.90 Local practices and customs played an important role and they were recognised in nineteenth century Russia with regard to property, in Islam with regard to transferring ownership of slaves, in Kyoto and Lyon with regard to work discipline. These elements account simultaneously for common phenomena (restrictions on labour mobility), the diverse ways they were expressed and their source (worker’s booklet, Russian serfdom, criminal punishment in the British Empire). They also explain the differences in the dynamics of Lancashire, the south of France and western Russia as well as those between individual English or Japanese factories and villages. Different solutions were adopted within a few miles of each other and similarities developed more frequently with factories in distant regions than with those nearby. The solutions adopted concerning working time testify to local irreducibility within a space that was nevertheless global. The chapters by Nagata and Vernus confirm this phenomenon. In Japan as in France, despite national norms and theoretically common techniques, the solutions that were adopted differed from one region to the next, often significantly. On the other hand, surprising similarities can be detected between certain districts of these two countries. These results confirm the similarities between certain Chinese regions and English districts, which have been the subject of recent research,91 just as those between proto-industrial districts in Europe.92 These resem- blances express the close link between technical solutions, economic and local institutional heritage, national discipline and international phenomena.

Book Outline

This book is divided into three parts, reflecting the coexistence of several levels of analysis. The first part reveals how the circulation of economic knowledge and legal models was related to economic behaviour and dynamics through case studies of England, Indonesia and the Indian

89 Royer, Histoire. 90 Benton, Law. 91 Pomeranz, The great. 92 Cerman, Ogilvie, European. 24 alessandro stanziani

Ocean World. Simon Deakin shows that, contrary to its reputation as the country of common law and free labour, Great Britain multiplied the statutes and constraints governing labour, unlike France where guild-based constraints proved less important than is usually acknowledged. Thus it becomes easier to understand the interface between the categories of European colonisers and those at work in the colonies. Ulbe Bosma discusses the attempts by Dutch colonial elites to establish certain labour relationships in Indonesia using the specificity of that region compared with their homeland as a lever. Similarly, the chapter by Gwyn Campbell takes a very long-term view in examining the forms of labour, bondage and slavery in the different societies of the Indian Ocean. They contributed to the proto-industrial growth experienced in other regions, which raises the question of their different trajectories over the very long term, even though the dynamics in which they took place were identical.

The second group of articles examines in greater detail the way local forms of dependence interacted with global dynamics. Marie-Louise Nagata explains how families and production units in Japan under the Tokugawa regime coped with proto-industrial growth. Taking the example of silk production in Tokyo, she brings out the interaction between family ties, the market and working time. This serves as a basis for comparing its similarities with and differences from Lyon, another silk production region. Pierre Vernus studies the solutions developed in this industrial district in the face of rapidly growing international markets in the nineteenth century and the undermining of local practices concerning labour and product quality. Claude Markovits looks at these same constraints in Indian trading companies. This case opens up at least three interesting perspectives. First, it fills a gap in the study of labour in India, which is too often focused on manufacturing. Next, it shows that Indian emigration was not necessarily limited to coolies and bonded labourers but also involved trading networks. Third, the presence of Indian trade organisations at the international level reflected international economic dynamics in which Asia, far from being a mere spectator of European development, in fact played an active role.93 In his study of silk marketing by Indian firms, Markovits provides an important supplement to the chapter by Vernus by bringing out the connection between the dynamics of silk production

93 Chaudhury, The Trading; Tracy, The Rise. introduction: labour, coercion, and economic growth 25 in nineteenth century Asia and the rise of local trade networks that were solidly integrated in the international markets. Taken together, the three cases of Kyoto, Lyon and an Indian trading firm in Cairo testify to the importance of forms of dependence rooted in family networks, which was to a certain extent common to Europe and Asia. This aspect is the core of the third and final part of the book, devoted to what Braudel described as “long-term prisons”. We will be discussing these constraints more from the standpoint of labour relations than of food. We have selected four cases: Central-Eastern Europe, Russia, China and India. The first two cases lead to questioning the “second serfdom” in Central and Eastern Europe, and from there, to re-examining the slow evolution of these societies. Instead of a shift from feudalism to capitalism, in keeping with the favoured tradition of theoreticians and historians, we will emphasise the interaction over the very long term between forms of bondage, constraints and market dynamics. The obvious differences between these situations, which we have already underlined, cannot conceal the common problem they all faced, namely economic and territorial expansion and the conflict between elites that it produced. From this point of view, the polymorphous legal status of “serfs” corresponded to the difference in status separating servants, coolies, shop assistants and agricultural labourers from their masters. The status of the different forms and degrees of service in husbandry and bondage reflected different economic conditions. Work discipline and its organisation were never a simple contractual matter but rather took place at the intersection between the private order and the public order of the markets. That is why it is important to compare the evolution of the second serfdom with that in China and later in colonial India. In the case of China, we will also learn in the chapter by Claude Chevaleyre that the existence of a multitude of forms of dependence ultimately adapted quite well to the order of civil servants, “castes” and naturally the markets. In the end, these elements will be compared with the forms of constraint in colonial India, particularly in the area of public works (article by Chitra Joshi); to meet workforce requirements, the British government resorted as readily to (Indian) criminals as to the “free” population. Underlying these measures was the notion of labour as social service. This connection was possible due not only to the hazy distinction in English law between a servant and a bonded labourer, but also to the fact that in Mogul villages in India, peasants were traditionally hired for public works projects (road construction, labour during periods of famine, etc.). It was the convergence of these different traditions that made it possible to mobilise compulsory labour to build imperial roads. 26 alessandro stanziani

The same convergence of private and public interests—of the state and the village economic community—, already noted in France and Russia, accounts for the emergence and duration of the system. We have thus come full circle: we find the link between freedom and constraint discussed in the first section in the organisation of labour and the movable boundary line between the private order (contractual) and the public order (regulatory and legal) of labour. Part one

Juridical Models and Labour Dynamics

The Duty to Work: A Comparison of the Common Law and Civil Law Systems from the Eighteenth to the Twentieth Centuries

Simon Deakin*

Introduction

Through the concept of the ‘contract of employment’, the law defines and regulates the individual employment relationship within the context of a market economy. The idea that, in a capitalist economic framework, work relations should take a ‘contractual form’ is nevertheless a surprisingly recent one. In the first stages of industrialisation in Europe and north America, work relations were not uniformly ‘free’ and the obligations of the parties were not, on the whole, defined using the language of contract. The model of employment that we are familiar with today only covered a small part of the working population, those who were ‘salaried’ managers or professionals. The story of how, from these beginnings, the model of the contract of employment came to provide the foundation for the legal account of work relations is of interest not just from the point of view of legal history and comparative law, but for the insights it provides into the evolution of law under industrial capitalism. In both the common law countries and those of the civil law, the ‘contract of employment’ that we know today is essentially an invention of the late nineteenth and early twentieth centuries. It emerged alongside the large, integrated industrial enterprise and the modern welfare state. Nevertheless, the process took different forms in various national systems. Each legal system developed its own distinctive concept of the employment relationship, based in part on its own juridical culture, as well as on the social and economic context provided by the path it took to industrialisation. The evolution of labour law both reflected and reinforced these national differences.

* University of Cambridge. This paper builds on and extends my work with Frank Wilkinson, first published in our joint work, The Law of the Labour Market (Oxford: OUP, 2005), and my chapter, ‘The Comparative Evolution of the Employment Relationship’, in Guy Davidov and Brian Langille (eds.) Boundaries and Frontiers of Labour Law (Oxford: Hart Publishing, 2006). I am grateful to the publishers and editors of these earlier works to draw on them here. 30 simon deakin

The Evolution of Concepts in Labour Law

The ‘contract of employment’ (in English law: the equivalent of ‘contrat de travail’ in French labour law) is a highly complex amalgam of many different kinds of norms and values surrounding the employment relationship. We should not expect the lawyer’s term ‘contract of employment’ to have a point-for-point correspondence with the employment relationship as it is observed and experienced in a social or economic context. The legal concept must obey certain internal rules of system-congruence which are specific to the juridical order.1 At the same time, the legal system will function well or badly according to how far it can increase the volume of external information which it can handle, while maintaining the internal mechanisms (specifically legal acts and processes) through which its reproduction is ensured. In that sense, the capacity of the legal order to process the information which it receives from the economy is in issue. If the legal concept of ‘employment’ were to bear no relationship at all to the ‘economic’ one, its functioning would, at some point, be impaired. This leads on to the observation that in the process of evolution of legal concepts, we simultaneously observe continuity and change. Adaptations are explained by a reference back to an existing conceptual maxim or principle. The rate of conceptual evolution is therefore much slower than the rate at which the content of the rules themselves changes. We can, nevertheless, identify conceptual mutations, in particular during those periods of rapid change when one paradigm displaces another in the manner of ‘punctuated equilibrium’.2 The interpretive task of ‘decoding’ of legal concepts has the effect of highlighting the contingent nature of implicit legal values; they can be seen to be the result of historically specific circumstances, which may no longer hold.3 But at the same time, the reshaping of the law which accompanies attempts at reform virtually never starts from a blank sheet. Concepts already available are put to new uses. This does not mean that the legal forms themselves to do not change; however, they change much more slowly than shifts in the substance of the law, and in ways which reflect the previous pattern of development. Legal evolution, then, is path-dependent. Conceptual adaptations are piled one on top of another, with the result that the structure of legal thought at any given point in

1 Luhmann, Law as an Autopoietic System. 2 Eldredge and Gould, “Punctuated equilibria”. 3 Atleson, Values and Assumptions. the duty to work 31

Table 1 Legal Classifications of Work Relationships from the Eighteenth Century to the Mid-Twentieth Century 1600–1800 Servant Worker (typically unmarried) engaged in service under a yearly hiring, entitled to payment in cash or in kind whether or not there was work during the period of the hiring, and with a right to a poor law settlement after the hiring ended Labourer Daily or casual manual worker in agriculture or the unregulated trades Master, journeyman, Worker in trades protected by guild regulation apprentice 1800–1875 Servant Manual worker in industry or agriculture under the disciplinary regime of the master and servant legislation, with little security or wages or employment Employee Clerical, managerial or professional worker outside the master-servant regime, with a degree of contractual income and employment security Independent contractor Independent artisan outside the scope of master and servant legislation 1875–1950 Workman Manual worker subject to the semi-disciplinary provisions of the Employers and Workmen Act 1875, increasingly protected as the period went on by collective bargaining, workmen’s compensation and social insurance legislation Employee At the beginning of the period, a non-manual worker with managerial or professional status; by the end of the period, a wage or salary-dependent worker, either manual or non-manual Self-employed Independent worker not employed under a contract of employment time incorporates forms which, although in some sense superseded, nevertheless continue to shape the path of the law. Table 1 indicates the main classifications of labour relationships in Eng- lish law at three different periods in its historical development.4 From this

4 See further, Deakin, “The contract of employment” and ‘‘The personal employment contract”. 32 simon deakin it can be seen that the modern distinction between the ‘employee’ and the ‘self-employed’ worker (which is roughly equivalent to that between ‘travail subordonné’ and ‘travail indépendant’) is a relatively recent one. It did not exist in its modern, paradigmatic form until the 1950s. Prior to that point, elements of the same idea can be found, but there were equally powerful divisions in the law, in particular between ‘servants’ and ‘employees’, a distinction drawing on notions of class and status as much as function. A ‘genealogical’ analysis helps us see that the modern employment contract has been superimposed on an older notion, the ‘master-servant’ relation, which had few of the features which compensate today’s employee for entering into a situation of personal and economic dependence upon the employer. The key to understanding this process lies in the analysis of the legislation and case-law surrounding the terms ‘servant’ and ‘employee’. At the mid-point of the twentieth century, it was believed that the prevailing legal concept for defining the employment relationship during the nineteenth century had been the ‘control’ test. Kahn-Freund influentially suggested that the control test developed in the context of the common law relating to the employer’s vicarious liability for torts of a servant acting in the course of employment.5 It is certainly true that, in the post-1945 period, this was one of the contexts in which the control test was still being applied. However, the nineteenth-century authorities which were (and are) cited for the ‘control’ test were not concerned with the issue of tortious liability. Close examination of origins of the concept of the contract of employment suggests that it was not the common law of vicarious liability which provided the context for the emergence of that concept, but, rather, social legislation dealing with taxation and national insurance. One of the most widely cited cases for the control test, Yewens v. Noakes,6 concerned the definition of a live-in servant under tax legislation. This case did not turn on the distinction between ‘employees’ and the ‘self-employed’ which is familiar to modern employment lawyers and which Kahn-Freund was writing about in the 1960s, nor even to the roughly equivalent nineteenth century distinction between ‘servants’ and ‘independent contractors’. The court’s decision was based on its refusal to believe that a salaried clerk earning a substantial salary could be a ‘servant’, since, according to the

5 Kahn-Freund, “Personal scope”. 6 (1880) 6 QBD 530. the duty to work 33 court, such a person was more clearly akin to ‘the manager of a bank, a foreman with high wages, persons in the position almost of gentlemen’. Yewens v. Noakes, then, was not concerned with the modern distinction between employment and self-employment for which it is still, even today, being cited. Rather, it was concerned with a completely different status- based divide, that between ‘servants’ and labourers in manual employment, on the one hand, and those employed in higher-level occupations and managerial and clerical work, on the other. To see why this distinction was more important at that point than it is today, a now-vanished feature of the nineteenth century legal landscape must be borne in mind. The distinction between manual and non-manual work had been central to the operation of the nineteenth century master- servant legislation; only ‘servants’ were subject to fines and imprisonment for breach of service contracts. This same distinction was carried over as a kind of ‘frozen accident’ into early social legislation concerning workmen’s compensation and social insurance. It was in this context of the law of the emerging welfare state that the ‘control’ test was established in a series of early twentieth-century decisions.7 Why did twentieth-century courts light upon the otherwise obscure decision in Yewens v. Noakes? The (re)discovery and adaptation of the control test in the 1900s and 1910s was a doctrinal innovation which was introduced at the same time as the courts were being called on to define the boundaries of what was then regulatory legislation of a wholly novel type. Nor was this judicial innovation particularly welcoming to the new legislation. The element of compulsion in social legislation went strongly against the grain of prevailing common law values. As a result, the courts regularly held that professional and managerial workers were outside the scope of these new laws.8 The control test, as applied by twentieth century courts, was also linked to disputes about employer’s liability in the context of the widespread practice of internal contracting. The contract system of hiring labour through an intermediary was still the predominant form of industrial organisation in road building, construction, shipbuilding, mining and quarrying, and iron and steel. The adoption of the

7 Simmons v. Heath Laundry [1910] 1 KB 543; Scottish Insurance Commissioners v. Edin- burgh Royal Infirmary 1913 SC 751; Hill v. Beckett [1915] 1 KB 578; Underwood v. Perry [1923] WC & I Rep. 63. The relevant statutes were the Workmen’s Compensation Acts of 1897 and 1906 and the National Insurance Act 1911. 8 See, Waites v. Franco-British Exhibition (1909) 25 TLR 441; Bagnall v. Levinstein [1907] 1 KB 531; Dow v. McNeil [1925] WC & I Rep. 32; re South Dublin Union Officers [1913] WC & I Rep. 245. 34 simon deakin control test enabled employers to avoid responsibility for the social risks of illness, injury and unemployment which it had been the aim of social legislation to impose, at least in part, upon them. In short, the rise of the control test in the early years of the twentieth century tells us much about prevalent employment disputes at that time clash between freedom of contract and the welfare state, and about shifts in the structure of the business enterprise which were also going on at this time. The ‘unitary’ model of the contract of employment which came to extend to all categories of wage-earners, including salaried and clerical workers, was only clearly adopted when further reforms were enacted to social legislation, in particular the extension of social insurance which took place in the National Insurance Act 1946, and when the internal contracting system gave way to integrated management, a development which in some industries, such as coal mining, occurred as late as the post-war nationalization process. A major aspect of the Beveridge Report of 1942 was the abolition of distinctions between different categories of employees: henceforth, all wage or salary earners, regardless of their annual income or of their professional status, would come under the same contributory classification.9 It was in the context of this new situation that the courts abandoned the old distinction between low status and high status employees when seeking to define the contract of service.10 The control test itself came to be regarded as excessively artificial, and gave way to the more recognizably modern tests of ‘integration’ and ‘business reality’. At around the same time, the term ‘servant’ mutated into the modern ‘employee’. By these means, a more inclusive notion of the employment relationship came to be established for the purposes of determining the scope of employers’ liabilities in respect of personal injuries, employment protection and social insurance. To understand this process, it is necessary to go back to the early modern roots of labour regulation in the British context. In the following section, the role of master-servant legislation in the British case is considered in some detail, with a brief account of the American case providing a point of contrast.

9 Social Insurance and Allied Services Cmd. 6404, November 1942, at para. 314. 10 See, in particular, Stevenson, Jordan & Harrison v. McDonald & Evans [1952] 1 TLR 101. the duty to work 35

The Evolution of the Contract of Employment in the Systems of the Common Law: England and the United States

England: The Emergence of the Master-Servant Model The institutional origins of the market economy in the case of England can be traced to the later middle ages. At that point, England possessed a unified legal system that operated on a national scale. The first laws on the work relationship, adopted in the wake of the Black Death of 1346, demonstrated the growing importance, at that point, of legislation as a mechanism of economic control. Their passage also indicates the presence of a system of labour contracting, operating alongside the compulsory labour of serfs and villeins.11 The Statute of Labourers of 1351 did not just formalise the regulation of wages; it triggered the development of the action of action of assumpsit, forerunner of the modern law of contract.12 By the fifteenth century, as the role of the Church as a source of social protection was disappearing, the regulatory role of the state expanded, with the passage of numerous poor law statutes and of the Statute of Artificers of 1562. This statute was to provide the basis for wage fixing, the constitution of the urban guilds and the regulation of the agricultural labour market into the nineteenth century.13 Formally, labour under the Statute of Artificers was not ‘free’: service in agriculture was obligatory for those who had no independent means of subsistence. Within the urban guilds, relations between masters, journeymen and apprentices more closely resembled those of producers subject to the common rules of the trade, than they did those of capitalist entrepreneurs and subordinated workers. The putting-out system consisted of networks of independent contractors, linked to merchants and other commercial intermediaries, rather than a class of wage-dependent labourers. Thus we have to exercise great care when considering the meaning given to the terms ‘labourer’ and ‘servant’ at this time. The century after 1750, generally associated with the period of the ‘industrial revolution’, was a time of legal transformation to match that of the changes which were taking place in the wider society and economy.

11 ‘As villeinage ceases, the poor law begins’ was Tawney’s assessment (Tawney, The Agrarian at p. 47), and the point can be extended to wage regulation more generally. 12 See Palmer, English Law in Age of Black Death. 13 See Deakin and Wilkinson, Law of the Labour Market, 44–51. 36 simon deakin

For Arnold Toynbee, writing in 1870, the essence of the industriali revolution was the displacement of medieval regulation by the forces of competition.14 Competition in the labour market was marked by the abolition of the wage-fixing clauses of the Statute of Artificers in 1813 and by the repeal of the apprenticeship laws in 1814. The Statute was by no means a dead letter at this point. The repeal of the apprenticeship laws was accompanied by a series of strikes. Prior to the final repeal, the judges had been weakening the law through a series of restrictive interpretations of its provisions.15 The willingness of the eigthteenth century courts to uphold guild rules protecting the ‘mystery’ of the trade16 gave way to judicial astonishment that the apprenticeship laws might be applied to the owners of the newly emerging industrial firms.17 By these legal means, a major institutional constraint on capitalist forms of enterprise was removed. Even this was only a step towards the ‘contractualisation’ of work relations. In certain professional and managerial occupations, a type of employment contract emerged, into which the judges began to imply a number of terms which reflected the idea of reciprocal obligations of the parties. At the start of the nineteenth century we also see the growing recognition by the courts of the common law action for damages for wrongful dismissal. This contractualisation of the employment relationship did not however apply to workers falling under the statutory régime of the Master and Servant Acts: in their case, breach of contract was a crime, under which thousands of workers were successfully prosecuted every year between the 1820s and the 1870s. The first of the Master and Servant Acts was enacted in 1747. This Act gave the justices jurisdiction to examine and rule on disputes between masters and servants in husbandry and between masters and ‘artificers, handicraftmen, miners, colliers, keelmen, pitmen, glassmen, potters and other labourers employed for any certain time, or in any other manner’ whether or not any rate or assessment of wages had been made for them in that year. They were also supplied with the power to order payment of wages due and to punish the servant or labourer for any ‘misdemeanour, miscarriage or ill behaviour’ by abating wages or committing him to the

14 Toynbee, Lectures on the Industrial Revolution, 92. 15 Prothero, Artisans. 16 Hobbs v. Young (1689) 1 Show KB 267. 17 Kent v. Dormay (1811), Kingston Assizes, August 14 reported in Chitty, A Practical Treatise, 122. the duty to work 37 house of correction for up to a month; they could also discharge the servant from the contract. An Act of 1758 extended their jurisdiction to cover servants in husbandry hired for less than one year18 and that of 1766 made it an offence for the servant to quit before the end of the agreed term.19 This last provision was an attempt to bring up to date the similar prohibition in section 15 of the Act of 1563. An Act of 1823 established new crimes of absconding from work and refusing to enter into work under a contract of hiring, and provided for imprisonment of workers for up to three months.20 Thus far from abandoning this particular part of the Elizabe- than labour code, therefore, Parliaments of the eighteenth and nineteenth centuries significantly strengthened and extended it. Complementing the Master and Servant Acts were numerous measures passed specifically to deal with theft and embezzlement by servants, labourers and outworkers. At this point, merchants and middle-men who put out goods for finishing up remained the legal owners of the material or produce throughout the process; these laws gave them powerful procedures for enforcing discipline against the rural manufacturers whom they normally employed on a nominally independent basis.21 Precedents for these measures exist from the seventeenth century;22 again, the pace of legislative change increases after the mid-eighteenth century. Under Acts of 1740, 1749, 1777, and 179223 it became an offence punishable by imprisonment for persons employed in manufacturing various goods to divert or sell materials sent to them for finishing up, or to detain them for more than twenty-one days after completing the work agreed. Justices could issue a search warrant for the inspection of premises of those con- victed or charged of these crimes. Those buying or receiving stolen goods

18 31 George II c. 11, s. 3. 19 6 George III c. 25, s. 4. This was an attempt to impose a general restraint of the kind previously enacted around this time for particular trades, including the journeyman shoemakers (9 George I c. 27, s. 4, 1722), woollen manufacturers and framework knitters (12 George I c. 34 s. 2, 1725), glove and shoe manufacturers (13 George II c. 8, 1739, s. 8). The 1766 Act was stated to apply to any ‘artificer, calico printer, handicraftsman, miner, collier, keelman, pitman, glassman, potter, labourer or other person contracting for any time or times whatsoever’. 20 4 George IV c. 34, ss. 3, 4 and 5. 21 On the so-called ‘proto-industrialisation’ model based on the putting-out to production to household forms of labour, see Berg, Age of Manufactures; the evidence for and against ‘proto-industrialisation’ as a distinct mode of production is assessed by Daunton, Progress and Poverty ch. 6. 22 See 13 & 14 Charles II c. 15, 1672–3; 20 Charles II c. 6, 1677; 8 & 9 William III c. 36, 1696–7. 23 13 Geo. II c. 8; 22 Geo. II c. 27; 17 Geo. III c. 56; 32 Geo. III c. 44. 38 simon deakin from servants also committed crimes. An Act of 1800, noting that ‘it often happens’ that colliers and miners breach their contracts ‘to the great and lasting prejudice of their employers’, made it an offence for persons contracted to raise coal or minerals to do so ‘in a different manner to [the owner’s] stipulations thereto, and contrary to the directions and against the will of the owner’.24 The use of the criminal law to enforce the contractual obligations of labourers and artisans found ready justification. ‘Imprisonment may be viewed as a mode of compelling the performance of contracts’, wrote J.E. Davis in his account of the Master and Act of 1867.25 ‘In some cases, damages might recompense a master for the breach of a contract by his servant, but the latter is seldom in a position to pay damages, and therefore, in the absence of any other remedy, he might set his employer at defiance’.26 It was expected that the jurisdiction over employment disputes would be very widely used, and that the regular courts would not be able to cope: No substitute . . . can be found for the jurisdiction of the magistrates, stipen- diary or otherwise, in the case of master and servant. The County Court does not sit sufficiently often for this purpose. In populous districts it is held once a month, so that, to say nothing of frequent adjournments of cases from one court to another, persons would be almost without the means of having their cases heard for many weeks together, instead of having the magistrates’ court to resort to every week-day in the most populous districts, and in rural districts seldom less than once in every week.27 Court procedure was streamlined and, generally, beneficial to the employer who would normally be the one to bring the complaint. In complaints of neglect of work, which took the form of proceedings for a conviction as opposed to an order, the servant could not be a witness in his or her own defence.28 The Act of 1867 remedied this and replaced imprisonment with fines as the principal remedy for breach of the Act, although imprisonment remained a possibility for, among other things, ‘aggravated’ misconduct.29 Fines also offered an advantage over discharge,

24 39 & 40 Geo. III c. 77. 25 Davis, The Master and Servant Act 1867, 6. 26 Ibid., 7. 27 Ibid., 8. 28 Ibid., 10. 29 30 & 31 Vict c. 141. The principal remedy envisaged by the Act was compensation for non-performance of the contract of service (s. 9), but there was the possibility of imprisonment for up to three months for non-payment of compensation (s. 11) and for aggravated the duty to work 39 which ‘although at first sight a desirable course to adopt, was impracticable as a punishment (and therefore impracticable as a remedy) owing to the demand for labour in many branches of manufacturers and the ability of men to get work elsewhere’.30 The courts implied into the contract of service employers’ obligations to provide work and to maintain the relationship in being through depres- sions in trade, as the necessary complements to provisions for extended notice or duration. Without such terms, a worker’s agreement to serve the employer exclusively for a period of years would be void on the grounds that it was in restraint of trade. The contract might provide for payment on the basis of piece rates or time rates. In an agreement for exclusive service for twelve months with payment on piece rates and provision for three months’ notice on either side, the court found a necessary implication that the employer shall find reasonable work and pay for the articles manufactured . . . The necessity of giving notice clearly shows that there is some obligation on the part of the employer. What is that? To find reasonable employment according to the state of the trade. That is not a unilateral agreement, but a mutual agreement with something to be done on both sides.31 However, as contract terms for the employee’s protection, these obligations were more or less notional. In the principal, reported cases of this period, the higher courts consistently rejected employee’s claims for wages based on the employers’ duty to find work.32 They found that employer had an implied right to lay off without wages, even in the case of an annual pit bond binding the workers to a year’s exclusive service.33 In this sense, long-service agreements effectively benefited only the employer; the worker was bound without having the protection of security of income or employment. The principal purpose of finding mutuality was to trigger the disciplinary provisions of the Act against the worker, or to form the misconduct (s. 14). No wages were to by payable during a period of imprisonment (s. 17). Servants and their spouses were deemed to be competent witnesses by s. 16. On the extensive use, in practice, of powers of imprisonment under this Act, see Steinfeld, Coercion and Contract, 81. 30 Davies, Master and Servant Act 1867: 10. 31 R. v. Welch, (1853) 2 E & B 356, per Lord Campbell, above; see also Pilkington v. Scott, (1846) 16 M & W 657. 32 This is in contrast to case law of the late medieval and early modern period in which claims by servants for wages due in return for being willing to serve frequently succeeded, via an action on the Statute of Labourers of 1351: see Ibbetson, Historical Introduction, 75. 33 Williamson v. Taylor (1843) 5 QB 175; see Steinfeld, Coercion, Contract, 102–124, for discussion of this case and other decisions in the same vein. 40 simon deakin basis for an action by the employer against another employer for entic- ing away the servant. Moreover, the courts’ construction of the contract of employment differed according to the statutory context which they were considering. ‘Butty’ workers—subcontractors in coal mining—were excluded from the coverage of the protective Truck Acts on the grounds that they were independent workers, while being simultaneously subject to the Master and Servant Acts as servants.34 Nor did the concept of mutuality extend to the continuation of the master’s traditional obligations of care under the service relationship. In this sense, the nineteenth century model of master and servant had little or nothing in common, beyond the use of certain terminology, with persistence of reciprocal obligations which can still be seen in parts of the eighteenth century case-law on the annual hiring. Thus at the turn of the century, the courts passed over the old authorities to find that a master had no obligation to maintain a servant or to provide him or her with medical care and expenses in the event of sickness or injury.35 These cases suggest that the courts at this time had no consistent conception of the contract of employment as a legal institution governing the reciprocal obligations of industrial workers and their employers; the classification of work relationships was determined not so much by contractual practices, as by the forms of regulatory legislation which operated on the service relationship. This juridical transformation of the service relationship needs to be set in the context of the organisational and technological transformation of production which was taking place during the late eighteenth and early nineteenth centuries. Although guild-based and independent forms of production were in decline at this time, a process which, as we have seen, was hastened by legislative change and judicial interpretations, they did not give way straight away to directly employed labour. Direct labour relations were by no means the inevitable outcome of technological change; at this time, many of the new technologies, such as those associated with steam power, were used in small workshops and in artisans’ cooperatives as well as in factories.36 Moreover, certain of the pressures favouring the adoption of subordinated labour were not technological in nature. Fac- tory owners took steps to undermine forms of profit-sharing between

34 Sleeman v. Barrett (1863) 2 H & C 934, where the court explicitly rejected an earlier suggestion (made in Bowers v. Lovekin (1856) 6 E & B 584) that the Truck Acts and the Master and Servant Acts should be interpreted in tandem. 35 Newby v. Wiltshire (1784) 2 Esp 739; Wennall v. Adney (1802) 3 B & P 247. 36 See generally Berg, Age of Manufactures. the duty to work 41 independent producers: ‘some capitalist entrepreneurs had had to fight for a system in which they could keep all the decision-making, all the control, and all the subsistence of the workers in their own hands, and had had to destroy earlier and more equable systems, such as the laws and organisations of the free miners in the Stannaries, the Mendips, Der- byshire, the Forest of Dean, Alston Moor and similar areas of Scotland’.37 Close supervision and control of the work process was a feature, by contrast, of conditions in the growing number of pauper factories and workhouses in which those dependent on poor relief were set to work. Pauper apprentices were employed on a similar basis in private industry, often as a condition of their parents continuing to receive relief from the poor law authorities.38 Thus it was not surprising that independent artisans and outworkers associated factory labour with the workhouse and, as far as they could, resisted it, drawing on a long cultural tradition of hostility to the loss of autonomous status.39 In sectors where employers did establish direct employment relations, they sometimes took a form not far removed the ‘unfree’ labour of the workhouse. This was exemplified by the institution of the pit bond in the north-eastern coal fields, by which colliers were bound by long-term contracts incorporating complex disciplinary codes. Pit-bonds were regularly enforced in the courts as a means of breaking strikes and instilling work discipline. This was a nineteenth century practice but it had roots going back to the early modern and late medieval period, when long-term employment contracts akin to a form of industrial serfdom had been widespread in the mining industry of Scotland and parts of the north east of England. The traditional pit bond was, however, exceptional, and was in any case in decline by the 1830s.40 In most industries, when the adoption of integrated factory production took place, it did not lead immediately to the establishment of direct employment relations between owners and workers. Factory sites often began as mills to which outworkers and artisans brought goods for finishing up; in practice, there was often an unclear boundary between merchants putting materials out for production, and employers of direct labour. When the transition from independent

37 Pollard, Genesis of Modern Management, 39. 38 Ibidem, 161. 39 Hill, “Pottage”. 40 See Steinfeld, Coercion, Contract, 167–182, on the evolution of contracts in coal mining away from annual bonds and monthly hirings to short-term ‘minute contracts’ between the 1830s and 1860s. 42 simon deakin

production occurred it was mediated in most industries by the institution of internal contracting, through which artisans or supervisors contracted with the works owner or entrepreneur for the performance of a job in return for piece rates, and were themselves responsible for the employment of the labourers and underhands on time rates.41 Against this background, the significance of the master and servant legislation lay in providing employers with a mechanism for imposing discipline on workers who otherwise had only a loose organisational connection to the firm, and who would often be in a position to take advantage of labour shortages to push up wages. Most of the more contentious mid-nineteenth century decisions on the coverage of the Master and Ser- vant Acts were precisely concerned with the position of skilled artisans and intermediaries such as butty workers; the central issue concerning the scope of the Acts was how far they could be used by employers to discipline skilled workers who otherwise retained a semi-independent status. This was reflected in the practice of enforcement of the Act. Pros- ecutions played an important role in relation to the business cycle, by providing employers with a disciplinary weapon to counter the effects of a tight labour market; prosecutions for absconding or refusing to work rose at times of high levels of activity, as workers sought better paid or less dangerous work. In terms of numbers of prosecutions alone, the material impact of the Acts was considerable, in particular in the mining and engineering trades and in certain regions, such as the Potteries.42 In Eng- land and Wales, numbers of prosecutions never fell below 7,000 per year between 1858 and the repeal of the Acts in 1875; a peak of over 17,000 prosecutions was reached in 1872. The historical evidence suggests that the disciplinary mechanism of the Acts were widely used as instruments of economic regulation during a period when modern managerial techniques had yet to develop, and when shifts in the business cycle could lead to considerable fluctuations in the bargaining power of employers and workers. In sum, the master-servant model was not a hang-over from the corpo- rative regime of the Statute of Artificers; on the contrary, the core disciplinary powers of employers and of the courts were enacted in legislation passed in the century from around 1750 as Parliament responded to the

41 Daunton, Progress and Poverty, ch. 7, for an account of the ‘coming of the factory’. 42 Simon, “Master and servant”; Woods, “Master and Servant Acts in the Black Country”; Hay, “Master and servant”; Steinfeld, Coercion, Contract, ch. 2. the duty to work 43 interests of the new employer class. The model of subordination embedded in the Master and Servant Acts was transferred over to the common law, with the result that long after the repeal of the Acts in 1875, the language of ‘master’ and ‘servant’ remained, a feature which was to weigh heavily on the future of labour law in Britain and the wider Empire.43

A Comparative Perspective: The American Case

During the nineteenth century the American systems followed their own route, which also eventually led to the emergence of a general model of the employment contract covering all work relations, but differing in significant respects from that which developed in England. By the start of the twentieth century, virtually all US states had adoped the rule of employment at will, according to which either party could terminate the contract without giving a reason for doing so, on minimal notice. Thus the US courts did not develop a common law action of wrongful dismissal similar to that which emerged in England. An important turning point was the 1870s. Before that, courts on both sides of the Atlantic had inserted provision for notice periods into the contract of employment, based on the type of payment (by the hour, day, week, month and so on). Employees with longer notice periods enjoyed a kind of legal job security; those with minimal notice rights were in effect employed at will. However, the US courts gradually came to regard the at-will form of employment as the default, which could only by evidence of clear intention of the contracting parties. This rule originated in Payne v. Western & Atlantic Railroad 44 and began to take shape in the context of claims arising from termination of employment such as Martin v. New York Life Insur- ance Co.,45 a case concerning a salaried worker who in England at the same time would have benefited form the presumption in favour of extended notice periods. A further reason for the embedding of at-will employment was the debate, within constitutional law, concerning the legitimacy of legislation concerning workers rights to form trade unions. The legal right of the employer to dismiss at will was a crucial weapon in resistance to unionisation. The courts struck down legislation protecting union rights, for example by outlawing the ‘yellow dog’ contract under which workers

43 Hay and Craven (eds.) Masters, Servants and Magistrates. 44 81 Tenn. 507 (1884). 45 148 NY 117 (1895). 44 simon deakin agreed not to become union members as a condition of their employment, by invoking at-will employment. In Britain, by contrast, only a segment of the wide range of work contracts was presumed, as a matter of construction, to be at will. The constitutional context which reinforced the movemtn towards at will employment of US law was missing. The Trade Disputes Act of 1906 confirmed the support of the state for collective bargaining and the freedom to take strike action in support of union membership and recognition, and the courts gradually adjusted the content of employment law to this new context.

The Emergence of the Contract of Employment in British Labour Law: The Role of Social Security Legislation

In common with other jurisdictions, Britain repealed its master and servant legislation in the second half of the nineteenth century. However, the legacy of the master and servant code was the assimilation by the common law of a hierarchical, disciplinary model of service. The persistence of this model owed much to the Employers and Workmen Act 1875, which provided statutory backing for a wide view of managerial prerogative and also for the application of quasi-criminal sanctions against certain categories of workers well into the twentieth century. The removal of criminal sanctions from the individual employment relationship in the 1870s was soon followed by the first legislative interventions of the welfare state. This wave of ‘social legislation’ began with the Employers’ Liability Act 1880, which mitigated the effects of the doctrine of common employment. This rule held that an employer could not be held vicariously liable in tort where negligence by one employee, acting in the course of employment, caused personal injury to another. The basis for the rule was the fiction that the contract of service contained an implied term under which the employee consented to the risk of being injured by a fellow worker. The 1880 Act made a limited exception, by excluding the common employment rule in a case where the plaintiff was injured as a result of the negligence of a fellow employee exercising managerial or supervisory responsibilities. From this small beginning, the principle that the employer should assume responsibility for social and economic risks arising from the employment relationship began to take shape. The first Workmen’s Compensation Act was introduced in 1897 and the first National Insurance Act in 1911. The workmen’s compensation the duty to work 45 scheme imposed liability on employer for workplace-related injuries and disease, and prompted the widespread use of employers’ liability insurance to spread the risks in question. Social insurance spread the more general risks of interruption to earnings through illness, unemployment and old age throughout the working population through the device of state-run insurance funds. National insurance contributions, levied on employers and employees, introduced a kind of fiscal regulation of the employment relationship which was later extended through the income taxation scheme. These interventions, by their purpose and nature, presupposed a certain degree of stability and regularity in the employment relationship, so reinforcing tendencies towards the growth of firms. Once the enterprise became a mechanism for the redistribution of social and economic risks, the resulting imposition of regulatory and fiscal costs on to employers created economies of scale which favoured larger firms at the expense of smaller productive units. The welfare state extended its influence in lock- step with increasing vertical integration of production and the emergence of the public sector as a significant employer. This process can be observed at the juridical level in terms of the decisions of courts called on to determine the scope of social legislation. Early social legislation did not adopt the ‘unitary’ model of the contract of employment, which emerged only after 1945; instead, it persisted in strati- fying the labour force by reference to distinctions drawn from restrictive legislation and from the common law. Hence salaried, non-manual workers were distinguished from hourly or daily-paid wage earners, who were distinguished in turn from casual workers and others with irregular wage relationships. At the core of this process of stratification was the ‘control’ test, the predominant judicial test of the time for identifying the service relationship. The Employers’ Liability Act of 1880, which made a limited exception to the doctrine of common employment, and the Workmen’s Compensa- tion Act of 1897 both adopted the term ‘workman’ rather than ‘employee’ to describe the workers within their scope. As we have seen, the term ‘workman’ had a specific meaning, related to manual labour, under the Employers and Workmen Act 1875; to this extent it was a legacy (if in somewhat attenuated form) of the status distinctions drawn under the master and servant regime. The 1880 Act explicitly adopted the meaning of the term under the Act of 1875, and the two Acts—the one protective, the other restrictive—were subsequently interpreted in tandem. 46 simon deakin

In a further reinforcement of status distinctions, supervisory employees were also expressly excluded from the protection of the 1880 Act. The term ‘workman’ was also applied to truck legislation following the Truck Amendment Act of 1887. Following the passage of these Acts, the term ‘workman’, previously given a wide reading, began to receive a more restrictive interpretation in the courts. In Morgan v. London General Omnibus Company the Court of Appeal, in a claim for personal injuries compensation under the 1880 Act, held that a bus conductor ‘who earns the wages becoming due to him through the confidence reposed in his honesty’, was not a workman: ‘labourer’ cannot in its ordinary acceptation include an omnibus conductor . . . The mere fact that a man works with his hands is not enough to constitute a workman within [the Act].46 Similarly, a locomotive guard was not within the Truck Acts,47 a sales assistant48 and a hairdresser49 were outside the Employers and Workmen Act, and a tramcar driver was outside the Employers’ Liability Act.50 As a result of these decisions, a large class of employees, who were neither supervisory employees within the meaning of the Employers’ Liability Act nor managerial employees under the Truck Acts, were excluded from the protective scope of those statutes. The first Workman’s Compensation Act of 1897 adopted a wider definition, referring to a workman as ‘[including] any person who is engaged in an employment to which this Act applies, whether by way of manual labour or otherwise’. The employments covered by this Act were restricted to the railways, mining and quarrying, and factory and laundry work.51 However, the attempt to give the Act an extended scope failed to survive the Court of Appeal judgment in Simpson v. Ebbw Vale Steel, Iron & Coal Co.52 Here, a widow of a colliery manager sued his former employers for compensation in respect of his death in an underground accident. Her claim was

46 See (1884) 13 QBD 832, 834 (CA), and (1883) LR 12 QBD 201, 207 (DC). 47 Hunt v. Great Northern Railway Co. [1891] QB 601. 48 Bound v. Lawrence [1892] 1 QB 226. 49 R. v. Louth Justices [1900] 2 Ir. R. 714. 50 Cook v. North Metropolitan Tramways Ltd. (1887) 18 QBD 683, not followed in Scot- land: Wilson v. Glasgow Tramways & Omnibus Co. (1878) 5 SC (4th. Ser.) 981; while in Smith v. Associated Omnibus Co. [1907] 1 KB 916 the court distinguished Cook by deciding that a bus driver who had to start his engine with a hand pump and used spanners and wrenches to fix it if it broke down was, for these reasons, a ‘workman’. 51 WCA 1897, s. 7. 52 [1905] 1 KB 453. the duty to work 47 rejected, the court concluding that while it was theoretically possible for a non-manual worker to be within the Act, he ‘must still be a workman’. The Simpson decision was based on the assumption that protective legislation of this kind was inappropriate for a high status (salaried) worker: It presupposes a position of dependence; it treats the class of workmen as being in a sense ‘inopes consilii’, and the Legislature does for them what they cannot do for themselves: it gives them a sort of State insurance, it being assumed that they are either not sufficiently intelligent or not sufficiently in funds to insure themselves. In no sense can such a principle extend to those who are earning good salaries.53 Despite the hostile language used here, it should not be thought that it was the courts alone which perpetuated these status distinctions. Such distinctions were deeply embedded in the law as a result of the influence of the Employers and Workmen Act, and the levels of compensation established by the Workmen’s Compensation Act were low enough to justify a belief that the Act was principally intended for the manual industrial workforce. Lawyers could be explicit about the class distinctions underlying the Acts; in Simpson counsel argued that ‘the Legislature were contemplating a class of workers who may be described as belonging to the working class in the popular sense of the term—a wage earning class’.54 The Workmen’s Compensation Act of 1906 established yet another scheme of definition, which in due course formed the basis for the classifications drawn by the legislation of the pre-war and inter-war periods. A ‘workman’ was now defined as ‘any person who enters into or works under a contract of service or apprenticeship with an employer, whether by way of manual labour, clerical work or otherwise, and whether the contract is expressed or implied, is oral or in writing’; but there were significant exceptions, including non-manual workers employed on annual remuneration greater than £250, casual workers employed ‘otherwise than for the purposes of their employer’s trade or business’, outworkers and family workers. This scheme, then, excluded high status workers at the top and casual workers at the bottom. In similar vein the National Insurance Act of 1911 applied its health insurance provisions to ‘employed’ persons, who were defined principally as those employed under a contract of service or apprenticeship, but

53 [1905] 1 KB 453, 458 (Collins MR). 54 [1905] 1 KB 453, 456. 48 simon deakin

excluding non-manual workers on an annual salary of more than £160, casual workers not dependent on their employer’s business, commission agents, and a large class of public sector workers including civil servants, military personnel and teachers, on the grounds that they were better served by state schemes already in existence.55 Male outworkers were included in the Act,56 but most female outworkers were initially excluded by the provision excepting outwork employment ‘where the person so employed is the wife of an insured person and is not wholly or mainly dependent for her livelihood on her earnings in such employment’.57 The Insurance Commissioners later exercised delegated powers58 to bring female outworkers within the Act,59 but both male and female outworkers remained outside the unemployment insurance part of the scheme. Male industrial workers were also the principal beneficiaries of the unemployment insurance scheme. The 1911 Act applied only to a specified number of industrial trades;60 women workers were only covered in practice after the extension of insurance in the Unemployment Insurance Act of 1921, and even then the legislation set differential contribution and benefit rates which had the effect of providing women workers with less extensive protection than men. The hierarchical distinction between clerical/managerial workers, industrial workers and casual workers which is to be found in this legislation filtered through into the common law in the form of the ‘control’ test for identifying the contract of service. ‘Control’ is traditionally seen in labour law as a test inherited from pre-industrial traditions of employment, and in particular the forms of domestic and artisanal production. This was the sense referred to by Kahn-Freund when he described the control test as: based upon the social conditions of an earlier age . . . It reflects a state of society in which the ownership of the means of production coincided with

55 See NIA 1911, s. 1 and Schedule 1. 56 NIA 1911, Sch. 1, Part I, para. (d). The courts held that master tailors were within the Act even if they themselves hired others: re Master Tailors as Outworkers (1913) 29 TLR 725; and outworkers were held to come under the Trade Boards Act 1909 (Street v. Williams [1914] 3 KB 537). Statutory definitions of the term ‘homeworker’ subsequently acknowledged that outworkers employing others could come under protective legislation. See Wages Act 1986, s. 26(1); National Minimum Wage Act 1998, s. 35. 57 NIA 1911, Sch. 1 , Part II, para. (j). 58 Contained in s. 1(2) of the 1911 Act. 59 SR&O 1912/921; SR&O 1914/880. 60 See NIA 1911, Sch. 6. The regulated industries were named as: building, construction, shipbuilding, mechanical engineering, ironfounding, vehicle construction, and sawmilling. the duty to work 49

the possession of technical knowledge and skill and in which that knowledge and skill were largely acquired by being handed down from generation to the next by oral tradition and not by being systematically imparted in institutions of learning from universities down to technical schools. The control test postulates a combination of managerial and technical functions in the person of the employer.61 Certainly, the control test was close to the test of ‘exclusive service’ which was the principal test of subordinated employment under the Master and Servant Acts; however, there is little to suggest that it was a remnant of pre-capitalist forms of employment. ‘Control’ was not an important test until the later nineteenth and early twentieth centuries, and when it began to be used it was in the context of the social legislation of that period and in particular the status distinctions which were being drawn at that time. An early case referring to ‘control’ in the context of the master’s vicarious liability for the torts of the servant is Sadler v. Henlock (1855),62 in which Crompton J. defined the relevant test as ‘whether the defendant retained the power of controlling the work. No distinction can be drawn from the circumstance of the man being paid at so much a day or by the job.’ In contrast to a servant, an independent contractor ‘chooses the mode in which the work is done, and the persons who do it’. This was a case concerning a labourer employed by the defendant to clear a drain on his land. It is unlikely that the control test, as such, was clearly established as a general test of status at this time; up to the repeal of the Master and Ser- vant Acts the judicial focus was on ‘exclusive service’ and after that on the definition of the term ‘workman’. Although ‘control’ may have been more important in the area of tortious liability for personal injury, the main issues there concerned the scope of the defences—common employment, consent (volenti non fit injuria) and contributory negligence—rather than the question of workers’ status. The control test was only clearly asserted later, in cases concerning not the common law of vicarious liability but the scope of social legislation. The leading cases in which the test was established were decided in the twentieth, not the nineteenth century—Simmons v. Heath Laundry63 and Underwood v. Perry64 in the field of workmen’s compensation, and

61 Kahn-Freund, “Servants and independent contractors”, 505. 62 4 E. & B. 570. 63 [1910] 1 KB 543. 64 [1923] WC & I Rep. 63. 50 simon deakin

Scottish Insurance Commissioners v. Edinburgh Royal Infirmary65 and Hill v. Beckett 66 in national insurance. The nineteenth century cases which were most frequently cited in support of the test had nothing to do either with vicarious liability or with contractual disputes between employer and employee. The 1873 case of R. v. Negus,67 in which Blackburn J. referred to control as the principal test, was concerned with the definition of ‘servant’ under the Larceny Act 1868. The subsequently much-cited case of Yewens v. Noakes68 involved a claim for exemption from inhabited house duty under the Customs and Inland Revenue Act 1869, under which premises used for trade only would be exempt from the tax if it was occupied merely by a caretaker, defined by the Act as ‘a servant or other person . . . for the protection thereof ’. The Court of Appeal held that a clerk, earning £150 per annum, did not fall into this category. In both these cases, the judges were clear that the status of ‘servant’ was distinct from that of high status ‘employee’, and that the main basis of differentiation was not the knowledge of the trade or craft but the presence of a wide-ranging power of the master to give orders to the servant, something the courts had previously associated with the idea of exclusive service. Blackburn J.’s judgment in Negus reflected this, when he said that ‘the test is very much this, whether the person is charged and bound to obey the orders of his master. He may be so without being bound to devote the whole of this time to this service; but if bound to devote his whole time to it, that may be very strong evidence of his being under control’.69 Lord Justice Bramwell evidently had the same distinction in mind when he said in Yewens v. Noakes that ‘a servant is a person subject to the command of his master as to the manner in which he shall do his work’.70 Lord Justice Thesiger thought it obvious that a salaried clerk was not a servant, any more than would be ‘the manager of a bank, a foreman with high wages, persons in the position almost of gentlemen’.71 Why did courts, several decades later, light upon these cases as author- itative guidance to the classification of employment relationships? The answer would not be problematic if the control test had become well

65 1913 SC 751. 66 [1915] 1 KB 578. 67 (1873) LR 2 CP 34. 68 (1880) 6 QBD 530; see Merritt, “Control v Economic Reality”, 113: ‘the “control” test has been misapplied almost from the moment of its emergence’. 69 (1873) LR 2 CP 34. 70 (1880) 6 QBD 530, 532–533. 71 (1880) 6 QBD 530, 538. the duty to work 51 established in tort law or in the interpretation of labour statutes, but as we have seen, this does not appear to have been the case. A more convincing explanation is that rediscovery and adaptation of the control test was a doctrinal innovation which was introduced at the same time as the courts were being called on to define the boundaries of regulatory legislation of a novel type. The element of compulsion in the legislation went strongly against the grain of prevailing common law values, as the judgments of the Court of Appeal in Simpson v. Ebbw Vale make clear. It was in the context of employers’ efforts to limit the scope of the legislation and a sympathetic judiciary that the control test was taken up. The particular effect of the control test was that it reinforced status distinctions between the ‘labouring’ and ‘professional’ classes, on the one hand, while excluding casual and seasonal workers, to whom the employer made a limited commitment, on the other. Simmons v. Heath Laundry72 concerned the employment status of a young laundry woman who gave piano lessons in her spare time. It was held that in respect of the music lessons she was not a ‘workman’ [sic], so that when she injured her hand at the laundry she could not claim for diminution of earning capacity as a pianist. Rather than saying that she was an independent contractor for the purposes of the lessons, the Court of Appeal used the control test in order to place her in the category of professional workers: ‘the question to be asked is what was the man [sic] employed to do; was he employed upon the terms that he should within the scope of his employment obey his master’s orders, or was he employed to exercise his skill and achieve an indicated result in such manner as in his judgment was most likely to achieve success?’.73 In other cases in this line, a lecturer, chemist, nurse, doctor and poor law officers were held to be outside the Act.74 This exclusion reflected the view that compulsory insurance was inappropriate for salaried or professional employees; it was also a means of reducing the potential liability of poor law guardians and hospitals, at a time when funding was more precarious than it subsequently became.75

72 [1910] 1 KB 543. 73 [1910] 1 KB 543, 553 (Buckley LJ). 74 See, respectively, Waites v. Franco-British Exhibition (1909) 25 TLR 441; Bagnall v. Levinstein [1907] 1 KB 531; Dow v. McNeil [1925] WC & I Rep. 32 (although not followed Wardell v. Kent County Council [1938] 2 KB 769; re South Dublin Union Officers [1913] WC&I Rep. 245. 75 In similar vein, the control test could even be used to hold that a nurse was an employee or ‘servant’ of a hospital for one purpose (administrative duties) but not another (work on the ward): Hillyer v. St. Bartholomew’s Hospital [1902] 2 KB 820, a tort case, not 52 simon deakin

The position of casual workers was complicated by the prevalence of sub-contracting in many industries. The contract system of hiring labour through an intermediary was still the predominant form of industrial organisation in road building, construction, shipbuilding, mining and quarrying, and iron and steel. As we have seen, this meant that there was unlikely to be a contractual nexus between workmen hired by the butty worker or foreman, and the ultimate owners of the site, plant or materials on which they worked. It was for this reason that the Factory and Workshop Acts of the period imposed obligations not upon employers in respect of their employees, but upon occupiers or owners of factories in respect of workers on their premises, regardless of whether there was a contractual relationship between them.76 The control test, as applied by twentieth century courts, often had the effect of classifying foremen as independent contractors, given their responsibility for hiring their own gangs;77 while the gang workers or labourers themselves had no claim against the ultimate users of their labour since the latter did not ‘control’ the performance of their work.78 Piecework payments were also treated by the courts as strong evidence of independent contractor status, notwithstanding clear statutory signals that this was not to be the case.79 Share fishermen—inshore trawlermen who were paid on a proportion of the profits from individual voyages— were found to be outside the social insurance legislation.80 Seasonal and casual workers who were directly employed by the ultimate employer were the subject of particular statutory provisions, speci- fying that they were to be included in the legislation where they were economically dependent on their employer’s business. This did not prevent considerable litigation arising. The courts held that a single, one-off hiring

followed in the workmen’s compensation case of Wardell v. Kent County Council [1938] 2 KB 769, and overruled for tort law purposes (but with Greer LJ dissenting) in Gold v. Essex County Council [1942] 2 KB 293. 76 See Factory Act 1844, ss. 41, 73; Factory and Workshop Act 1878, ss. 93, 94. 77 The courts reached opposing outcomes in different cases. Butty workers and foremen were found to be within the protection of the Workmen’s Compensation Acts in Evans v. Penwelt Dinas Silica Brick Co. (1901) 18 TLR 58 and Paterson v. Lockhart (1905) 42 SLR 755, but outside them in Simmons v. Faulds (1901) 17 TLR 352, Hayden v. Dick (1902) 40 SLR 95 and Vanplew v. Parkgate Iron & Steel Co. [1903] 1 KB 851. 78 Crowley v. Limerick County Council [1923] 2 Ir. R. 178; Littlejohn v. Brown & Co. Ltd. 1909 SC 87; although cf. M’Ready v. Dunlop (1900) 37 SLR 779; Doharty v. Boyd 1909 SC 87. 79 WCA 1906, s. 13; NIA 1911, Sch. 1, Part I, para. (a). 80 Scottish Insurance Commissioners v. M’Naughton 1914 SC 826. the duty to work 53 was not necessarily outside the scope of the Acts,81 and were also prepared to include seasonal workers who returned to their employer on a regular basis82 and part-time workers with long service.83 However, in other cases they excluded workers under task contracts,84 temporary workers85 and casual workers with short-term service.86 Trainees and unemployed workers receiving instruction and work experience at government training centres were also beyond the scope of the legislation.87 The adoption of the control test enabled employers to avoid responsibility for the social risks of illness, injury and unemployment which it had been the aim of social legislation to impose, at least in part, upon them. The drawing of fine distinctions to defeat the legislation in this way was not universally accepted. The Scottish courts, influenced, perhaps, by the civilian tradition of giving a purposive interpretation to social legislation, were on occasion reluctant to allow employers to contract out of the Acts by these means (just as, in the nineteenth century, they had been reluctant to accept the doctrine of common employment which held that the worker impliedly accepted the risk of a fellow servant’s negligence).88 In Paterson v. Lockhart,89 a case concerning the status of a foreman, Lord McLaren considered that it would be a serious restriction of the scope of the Act if it were possible by introducing some condition into an agreement to take it out of the category of a pure contract of service and so to avoid liability under the Act . . . It is quite in accordance with custom for a superior workman to choose his own assistants. An engineer may choose his own fireman or a mason his hodman. But that does not prevent their being servants paid by a common employer.

81 Boothby v. Patrick & Son [1918] W.C. & I. Rep. 340. 82 Smith v. Buxton [1915] W.C. & I. Rep. 126. 83 Dewhurst v. Mather [1908] 2 KB 754. 84 Alderman v. Warren [1916] WC & I Rep. 266. 85 Stoker v. Wartham [1919] 1 KB 499. 86 Knight v. Bucknill [1913] WC & I Rep. 175; Withams v. Larsen Ltd. [1928] WC & I Rep. 323. 87 Broome v. Ministry of Labour [1927] W.C. & I. Rep. 232; Watson v. Government Instruc- tional Centre [1929] W.C. & I. Rep. 265; re Leeds Corp. and Chadwick (1928) 44 TLR 797 (cf. Unemployment Insurance Act 1927, s. 15; Unemployment Insurance Act 1927, s. 15); McGeachy v. Dept. of Health for Scotland 1938 SC 282. 88 It was not until the House of Lords decided Bartonshill Coal Co. v. Reid (1856) 4 Macq. 266 that the Scottish courts accepted the doctrine, which had its origins in decisions of the English courts. 89 42 SLR 755, 757. 54 simon deakin

In Dunlop v. M’Cready,90 Lord Adam said of a group of gang workers that ‘to call them independent contractors is a mere playing with words’. These were, however, isolated examples of a more tolerant judicial attitude. More typical of the case-law was the finding of the Court of Session in Littlejohn v. Brown91 to the effect that a rivet boy, hired by the head riveters in a shipyard, had no contract of service with the main employer: ‘if the test is direct and immediate selection, payment and control and power of dismissal, the evidence is all one way, namely, that [the plaintiff ] was in the service of Gammell and Lacey, John Brown & Company being only indirectly connected with the boy’s employment’. The ‘unitary’ model of the contract of employment which, in modern labour law, extends to all categories of wage-earners, only came into being when further reforms were enacted to social legislation, in particular the extension of social insurance which took place in the National Insurance Act 1946. A major aspect of the Beveridge Report was the abolition of distinctions between different categories of employees: henceforth, all wage or salary earners, regardless of their annual income or of their professional status, would come under the same contributory classification.92 Accord- ingly, the 1946 Act established two principal classes of contributors: Class I covered ‘employed earners’, defined as ‘any persons gainfully occupied in employment . . . being employment under a contract of service’, and Class II covered those employed on their own account.93 The latter paid a lower rate of contribution and were excluded from the unemployment insurance part of the scheme. In this way the fundamental division between employees and the self-employed was established. The same distinction was adopted for the purposes of income taxation94 and, in due course, under the employment protection legislation which was introduced first

90 (1900) 37 SLR 7797, 782. 91 1909 SC 169, 174. 92 Social Insurance and Allied Services Cmd. 6404, November 1942, at para. 314. 93 NIA 1946, s. 1(2). 94 The modern division between self-employment and employment in tax law emerged gradually during the inter-war period. The principal division in the Income Tax Act 1918 was between earnings from public-sector employment, which fell under Schedule E, and earnings and profits from all other employment, which fell under Schedule D. Schedule D therefore covered both the self-employed and certain groups of private-sector employees. The Finance Act 1922, s. 18, then transferred earnings from private-sector employment into Schedule E, leaving only the self-employed in Schedule D. The Pay-As-You-Earn system of automatic deductions of tax from the earnings of employees was applied to most Sched- ule E employments by the Income Tax (Employments) Act 1943. the duty to work 55 in the early 1960s.95 The ending of the old divide between manual and non-manual workers was epitomised by the merging of the concepts of the contract of service and of employment: for statutory purposes, these were now synonymous with each other.96 Faced with this new situation, the courts abandoned the old distinction between low status and high status employees when seeking to identify the contract of service.97 The control test itself came to be regarded as excessively artificial, and gave way to the tests of ‘integration’ and ‘business reality’. These stressed economic as opposed to personal subordination as the basis of the contract of employment. The test of the worker’s ‘integration’ into an organisation was used to explain how professionals such as doctors and journalists could be classified as employees notwithstanding the high degree of autonomy they enjoyed in their work.98 ‘Economic reality’ had the effect of extending protection to casual workers and outworkers who were dependent on the business of another, as opposed to being entrepreneurs with a business and employees of their own.99 By these means, a more inclusive notion of the employment relationship came to be established for the purposes of determining the scope of employers’ liabilities in respect of personal injuries, employment protection and social insurance.100 Thus the rise of the welfare state helped to confirm the movement towards the contractualisation of employment relations. Nevertheless, we can see the preservation of elements of the master-servant model during the period of transition from the poor law to social security, a process which retarded the devleopment of British labour law. It is only in the mid-1950s that the clear division between the ‘employee’ and the ‘self- employed’ worker was recognised in labour law. There were some precedents for this approach in tax law and social security law from the 1910s onwards, but even here there was a tendency to sub-divide the class of

95 The first such statute was the Contracts of Employment Act 1963; the relevant provision is now contained in the Employment Rights Act 1996, s. 230(1). 96 See, in the context of social insurance, Vandyk v. Minister of Pensions and National Insurance [1955] 1 QB 29. 97 See, in particular, Stevenson, Jordan & Harrison v. McDonald & Evans [1952] 1 TLR 101. 98 Cassidy v. Minister of Health [1951] 2 KB 343; Roe v. Minister of Health [1954] 2 QB 66; Beloff v. Pressdram Ltd. [1973] 2 All ER 241. 99 Market Investigations Ltd. v. Minister for Social Security [1969] 2 QB 173; Lee Ting Sang v. Chung Chi-Keung [1990] ICR 409. 100 See Mitchell and Howe, “Evolution of the contract of employment in Australia” for discussion of how far a similar process of conceptual evolution to that described in the text occurred in Australian labour law during the same period. 56 simon deakin wage-dependent workers by reference to occupational status and income level. It was only with the passage of the National Insurance Act 1946, implementing key elements of Beveridge’s postwar scheme for social insurance, that the modern legal division between employment and self- employment became clear. This distinction was then carried over into the first employment protection statutes of the 1960s. The term ‘contract of employment’ which dates from this time truly is a recent development.

The Overseas Diffusion of the Master-Servant Model

The transmission of master-servant legislation to British colonies began as early as the seventeenth century (in the case of certain American states and the West Indies) and carried on into the twentieth century (in the case of some African countries), well after the repeal of the original laws in Britain. Over the course of the three hundred years or so, ‘almost 2000 statutes and ordinances made their appearance in more than 100 colonies, developing a colonial master and servant law that drew upon, elaborated, and often subverted the metropolitan models’.101 The ending of slavery in the Empire in the early nineteenth century and the formal instantiation of freedom of contract was accompanied by the enactment of pass and police laws and vagrancy legislation. Systems of plantation labour in the West Indies and Assam, and mine labour in South Africa, were underpinned by the coercive powers which master-servant law provided to employers. There was also a link to migration and racial segregation. Legislation was passed in several colonies in the 1830s and 1840s for the purpose of regulating the waves of indentured labour (millions of workers) moving around the Empire at that time. These laws created new forms of status around racial and cultural categories. Legislation of this type was being enacted in east Asia, the Carribean and Africa into the middle decades of the twentieth century. Whereas in Britain, master-servant laws were used to stabilize the labour supply, reduce the bargaining power of workers and shore up managerial prerogative in the mainly small-scale manufacturing enterprises which were characteristic of the industrial structure of the parent system, in the colonies the same types of laws were used to assist in the dispossession and separation from the land of indigenous populations and to maintain

101 Hay and Craven, “Introduction”, in Hay and Craven (eds.) Masters, Servants and Magistrates. the duty to work 57 the supply of cheap labour which was essential in plantation and mining- based economies. Plantation systems simply were not profitable without criminal enforcement of labour contracts; and the prosecution rates in some plantation societies were fully fifty times as high as those in parts of England, such as the west Midlands, where the master-servant laws were most heavily relied on by employers. In late nineteenth century Trinidad, around one fifth of all criminal convictions involved breach of contract. Enforcement took the form not simply of fines and imprisonment, but of judicially-supervised performance of the work contract, and the addition of extra periods of service as compensation for breaches by the worker.102 The end of master-servant laws did not come about through economic development. These laws lingered longest in systems without the democratic suffrage and without recognition of basic labour rights. Britain’s own experience had been similar: it was only as the franchise was extended that the political conditions for the repeal of master-servant laws were gradually established. Across the Empire, notwithstanding some liberalis- ing moves from the Colonial Office which local employers often resisted, it was the pressure of the International Labour Organisation in the 1920s and 1930s and then the decolonisation process itself which brought about the abandonment of penal laws.103 The influence of the master-servant model is a good example of a legal origin effect which occurs when transplantation is combined with colo- nialisation to produce strong path dependencies.104 This case is, however, further repudiation of the claim that common law systems were more inherently disposed to a liberal contractual model of work relations than civil law ones. Master-servant law was essentially status-based; that is to say, it preserved distinctions based on class, in the British case, and race, in the colonies. This can be seen from the evolution of the law in South Africa. The master-servant laws there were facially neutral, but with the passage from the middle decades of the twentieth century of industrial relations legislation regulating the mainly (and later exclusively) white occupations, the master-servant regime was applied the agricultural and domestic sectors which were traditionally non-white.105

102 Turner, “British Caribbean”. 103 Banton, “Colonial Office”. 104 Deakin, “Legal origin”. 105 Le Roux, “Evolution of the Contract of Employment in South Africa”. 58 simon deakin

There were prosecutions of tens of thousands of workers each year into the 1950s.106

The Civil Law Systems: France and Germany

The law followed a different path in the civil law systems. The codes of the post-revolutionary era made use of the traditional concept, derived from Roman law, of the locatio conductio for the purpose of defining the new forms of waged work. In adopting the idea of the locatio, the drafters of the civil codes were aligning work relations with other types of contract, an indication that they regarded the employment relationship as ultimately founded on an exchange. Thus labour, or labour power (Arbe- itskraft), was treated as a good or product to which the contract attached a ‘price’ (not necessarily a ‘wage’). It was classified as part of the law of things rather than the law of persons: the codes made no mention of the ‘subordination’ of the worker.107 The reality of work relations was very different : all the civil law systems recognised the power of the employer to give orders, to fix the rules of the workplace, and to to retain the worker in his service. The laws underpinning managerial prerogative were now, however, regarded as part of private law; they were seen as part of specia- lised regulatory or police powers, a feature which helps to explain their near invisibility at the level of case law and doctrine. The French Civil Code of 1804 used two versions of the locatio conductio.108 The louage d’ouvrage, loosely based on the locatio conductio operis of Roman law, applied to piece work or to the engagement of a worker to perform a specific task. The louage de services was modeled on the locatio conductio operarum and as such was based on the principle of the hire of services. In truth, any link to the original content of Roman law was tenuous; the concepts used in the Code were genuine adaptations.109 The ideological aim in the use of the locatio was to avoid any hint of the subjection of the worker; rather, the work relation was to be based on the idea of the formal equality of contracting parties. Partly for this reason, the Code did not admit the idea of a permanent or indefinite hiring.

106 Chanock, “South Africa”. 107 Simitis, “Case of the Employment Relationship”. 108 Code civil, art. 1780. 109 Veneziani, “Evolution of the Employment Relationship”. the duty to work 59

It is possible to see in the distinction between the louage d’ouvrage and the louage de services the same fundamental division as that in English law between servants and independent contractors. However, the binary divide may not have operated as straighforwardly as that. The concept of the louage de services tended to be applied to day workers and domestic workers, with the louage d’ouvrage covering other categories. In the case of the louage de services, the employer was the sole judge of whether the contracted for work had been performed.110 The division of risks between the parties was not greatly different in either case, and it is likely that in practice they both were closer to the practice of waged labour than to that of commercial relationships.111 It is therefore inappropriate to see the contract of employment as a direct descendant of either the louage de services or the louage d’ouvrage. Such a direct line of descent was maintained by nineteenth century jurists who thereby hoped to legitimise the newly emerging notion of the contrat de travail (the term began to be used in the 1880s).112 This process has more recently been termed a ‘mystification’.113 The contrat de travail served the purpose, from the employer’s point of view, of providing a basis for the worker’s duty of obedience. The concept was also used around this time in legislation governing liability for workplace accidents.114 Commis- sions of jurists charged with the task of establishing a framework for the law relating to collective bargaining and health and safety at work helped to spread its use.115 All these developments saw the idea of ‘subordination’ as lying at the core of the new concept. This implied an open-ended duty of obedience. In exchange, the employer accepted certain responsibility for social risks. The German Civil Code adopted in 1896, the BGB, used an apparently similar terminology to that of the French code at the start of the century. The Dienstvertrag (literally, ‘contract for service’) applied to wage labour while the Werkvertrag, contract for work, to subcontracting. In other respects, however, the BGB broke with the model of the locatio. The two contract forms were placed in separate parts of the Code, indicat- ing not just the fragmentation of the locatio, but, more fundamentally,

110 Code civil, art. 1781. 111 Petit and Sauze, “Une lecture historique de la relation salariale”; Cottereau, “Droit et bon droit.” 112 Ibid. 113 Cottereau, “Industrial tribunals”, 20. 114 Veneziani, “Evolution of the Employment Relationship”, 64. 115 Ibid., 68. 60 simon deakin separate conceptions of the work relation. Within the structure of the Dienstvertrag, the employer’s duty of care (Fürsorgepflicht) was balanced by the worker’s duty of loyalty (Treuepflicht). Gierke had argued that the Code should make reference to principles of solidarity which he claimed to have found in pre-modern antecedents of the Dienstvertrag. By these means, the employment relationship was aligned with the law of persons, and with an emergent communitarian notion of the enterprise. Even so, the BGB did not clearly recognise the ‘binary divide’ between subordinated and independent work.116 The Dienstvertrag covered employment contracts, but also some forms of the supply of services.117 This came later, with the appearance in the inter-war years of the concept of the Arbeits- verhältnis, a development associated with the rise of collective bargaining and the welfare state. In both countries, the form of the contract of employment was a reflection of prevailing juridical cultures.118 The French system was modeled on the idea of the power of the state to regulate basic conditions of work (ordre public social). The implicit logic in this approach was that the state, in acknowledging the formal contractual equality of the parties, came under a duty to provide protectoin to the weaker party, the employee, who was thereby placed in a position of ‘juridical subordination’. By contrast, the German approach was based on the communitarian model of the firm, reflecting the influence of Gierke who had inveighed against the ‘individualistic’ and ‘capitalistic’ orientations of the French Civil Code. In German labour law, the subordination of the worker took the form of their adhesion to the enterprise (Tatbestand), a process conferring a status akin to membership of a community.119 These two, alternative conceptions of the work relationship greatly influenced the subsequent development of the law governing the employment relationship in continental Europe, giving rise to a ‘structural ambivalence’ at the core of European labour law.120

116 Ibid., 59; Sims, Good Faith, 85–86. 117 BGB, para. 611. 118 Mückenberger and Supiot “Ordre public”. 119 Supiot, Critique du droit du travail, 18. 120 Ibid., 19. the duty to work 61

Conclusion

Without the welfare state and collective bargaining, there would be no contract of employment in the sense that we recognise today. The history of labour law teaches us that this concept serves two distinct functions: on the one side, organising, from a juridical point of view, the coordination of the enterprise and, on the other, partitioning and redistributing social risks arising from labour market participation. Solidararistic collective bargaining and social security systems are not, perhaps, inevitable features of a market economy, but they are compatible with it. While these models arguably reached their high point in western Europe and north America in the middle decades of the twentieth century, it would wrong to see them as specific to these systems or to a particular point in the development of industrial capitalism. Labour law continues to evolve alongside and in response to shifts in modes of production and societal structure.

Dutch Imperial Anxieties About Free Labour, Penal Sanctions And The Right To Strike

Ulbe Bosma1

Introduction

Thanks to massive inputs of forced labour the Netherlands-Indies emerged as a major plantation colony in the nineteenth and twentieth century. Crucial in this process was the Cultivation System (1830–1870) introduced by Count Johannes van den Bosch and designed as a policy of compulsory cash crop growing imposed on the rural population of Java. The Cultiva- tion System encroached upon the freedom of the Javanese landholding peasants who no longer could assign their land to the crops they deemed to be most profitable. This was in contrast to the tradition of free labour that had existed since the Dutch Republic and at variance with the principles of colonial rule that had been announced, shortly after the Nether- lands Indies had been returned to the Dutch by British in 1816. The history of colonial Indonesia shows that if there was any movement from unfree to free labour it was a trajectory that was far from linear but filled with twists and turns. Van den Bosch’s intention, I will argue, was to relieve the Javanese from feudal servitude and from the thralls of moneylenders, and make him a productive subject of the Dutch empire. Over time the emerging plantation economy seemed to vindicate Van den Bosch’s policies. Ever larger numbers of Javanese became coolies in the rapidly growing complex of cash crop production. In the 1860s the burgeoning sugar sector even led to tightening labour markets in East Java, where sugar factories had to offer advance payments to attract coolies. Again, the issue of forced labour resurfaced on the colonial agenda. This time to prevent coolies from running away with their earnest money without performing work. For that reason the factories demanded and got from the colonial government a penal sanction on breaches of labour contracts, in spite of the fact that it

1 I would like to thank Jan Lucassen, at the International Institute of Social History, for his suggestions with regard to the intricacies of labour law in the Dutch Republic and the Kingdom of the Netherlands. 64 ulbe bosma was a violation of Dutch labour law. A few years before when the Dutch Parliament had formalized the right to collective action it had already decided that there was no place for penal sanctions in ordinary labour contracts. It could not come as surprise therefore that it did not accept that the government of the Netherlands Indies had introduced the penal sanction as a measure to enforce labour discipline.2 And yet, only one year later the penal sanction was introduced for the Outlying Provinces, and for East Sumatra in particular. In this case, however, it did not originate from a labour regulation but as an addendum to a decree ordering civil servants to check whether coolies had been engaged on a voluntary basis. The first version of a regulation on recruiting practices dated from 1868. In 1880 penal sanctions were added to these regulations and the whole set was given the name of Coolie Ordinance, which had to safeguard coolies against deceit and the employers against their labourers running away before they had worked off their recruitment costs. The penal sanction in practice enforced the already existing immense inequality of power between employers and labourers, the former being European citizens, the latter colonial subjects.3 In the early twentieth century the continued existence of these coerced labour conditions were defended by employers who insisted on the crucial role of the Sumatra plantations for the colonial and imperial economy. The Sumatra employers compared the position of their coolies to Dutch sailors, whose freedom of movement was also severely restricted by law. This was an argument that began to backfire in a rather peculiar way once the ILO (International Labour Organization) was established in 1919. As part of its campaign against slavery and forced labour, the ILO also began to critically review the internationally customary penal sanctions in the sailors’ labour contracts. There is some irony in the fact that while international legislation could not enforce the abolition of the penal sanction in the Netherlands Indies, it did compel the Dutch metropolitan government to attune its own maritime labour legislation to more liberal standards of international law. This paper consists of four parts each covering a particular phase in the development of labour relations on Java from the early nineteenth century to the end of Dutch colonial rule. I will first describe how Van den Bosch developed his ideas about forced labour as a bridge between vagrancy and

2 Kalma, Het arbeidscontract, 11. 3 Bosma, Karel Zaalberg, 100. dutch imperial anxieties 65 civilisation in the Dutch metropolitan context. The second part focuses on the implementation of the Cultivation System on Java and details how the existing systems of traditional tributes were reshaped into a colonial system of land rent and forced labour. As a result a fast growing army of marginal and landless peasants found employment in the emerging plantation economy on a wage basis. The third section addresses the question how to discipline this labour force of rural workmen and women on the plantations, which was no longer tied by links of bondage or dependency to Javanese (village) elites. The final stage is after World War I, when the Indonesian nationalist movement emerges. In this phase the right to collective action becomes a contested one and is eventually practically suppressed by the colonial government in 1923. This was done with reference to the paramount interest of preserving the colonial order and economy in particular. Likewise, ‘paramount economic interests’ and ‘extraordinary circumstances’ were used as arguments by colonial enterprise and government to preserve the penal sanction to regulate labour relations on the Sumatra plantations.

From Free Wage Labour to Work As Fulfilment

The intricate connection between personal freedom and the rise of free labour (in this context labour free from non-pecuniary compulsion) has been amply discussed, for example by Steinfeld in his book on free and unfree labour in the United States and Great Britain.4 Contrary to the picture Steinfeld is giving us about the Anglo-Saxon world, where the concept of free labour only emerged in the nineteenth century, in the Netherlands labour migration was free, labour contracts were free and there were no penal sanctions on breach of contract during the days of the Dutch Republic. Particularly its western, maritime and highly urbanised provinces was one of the few parts of early modern Europe where unfree labour was absent. Many artisans’ and a limited number of journeymen’s guilds existed. Moreover, labour conflicts, including strikes have been documented.5

4 See Steinfeld, Coercion, Contract. 5 According to Lucassen ‘the Dutch labour market was one of the few, or maybe the only labour market of the Ancien Régime which was completely free’. See Lucassen, “Labour and early modern economic development”, 395. See also Lucassen “Mobilization of labour” and Lucassen, De Moor and Van Zanden, “Introduction”. 66 ulbe bosma

Under the Dutch Republic two types of labour contracts existed. One regulated the position of domestic servants, who were part of the household and the other the subcontracting of all other wage earners. Although local in character, because of the federal structure of the Republic, and therefore showing substantial differences, some features were common. Subcontracting labour was seen as an integral part of merchant law, i.e. a contract between equal parties and was sanctioned by civil law. Urban servants were free to engage in wage labour and both parties had access to the local courts for enforcement or dissolution of the contract. Sanctions were of a pecuniary nature.6 The free wage labour conditions that were ubiquitous in the highly urbanised western part of the Dutch Republic were more or less extended to the urban enclaves in the colonies.7 Sailors of course being the exception to the rule, an exception that was based upon labour regulations that were adhered to by a number of European seafaring nations and dated from the medieval times. Interestingly enough, however, the Dutch East Indies Company (voc) did not differentiate between Asian and European sailors as far as their labour conditions were concerned.8 In the Asian port cities of the voc freedom was a privilege for city-dwellers who were exempted from corvée service and for persons who were in the position to exact services from other people. By the late seventeenth century half of Batavia’s (contemporary Jakarta) population consisted of slaves.9 The Dutch East Indies Company itself was the most important slave trader in the archipelago if not in the entire Indian Ocean. The total volume of the Dutch Indian Ocean slave trade has been estimated by Vink as being as large as 15–30 per cent of the Atlantic slave trade.10 In the urban environment they were usually part of the household, worked as artisans or clerks, and could easily become part of a grey zone between slavery and freedom, where he or she could earn money to pay for his or her manumission.11 The slave trade on Java was abolished in 1811 by the British during their interregnum over the island. Slavery in the European realm

6 In the nineteenth century the Dutch versions of the Code Napoléon featured some novelties like the abolishment of the guilds and in fact the impossibility of any sort of ‘combination’: trade unions and strikes were officially allowed only from 1855, respectively 1872, although de facto much earlier, see Brugmans, Arbeidende klasse, 184–9, 249–52. 7 Lucassen, “Labour and early modern economic development”, 383. 8 See Lucassen, “A Multinational and its Labor Force”, 20–1. 9 Vink, “ ‘The World’s Oldest Trade’,” 148. 10 Ibidem, 168. 11 Bosma and Raben, Being “Dutch” in the Indies, 47–9; Blussé, Strange Company. dutch imperial anxieties 67 of Java continued to exist until 1 January 1860, as a matter of fact, but was a marginal phenomenon and largely confined to domestic services in coastal towns.12 To the institutes of slavery and servitude under the Javanese rulers I will return in the next section. The first Constitution of the Netherlands Indies of 1818 was still written in a liberal mode and inspired by the ideas of free labour and free labour contracting.13 But this liberal spirit did not last long, primarily not because of developments in the Netherlands itself. Widespread unemployment and indigency reigned over the newly created Kingdom of the Netherlands. In this situation the issues of freedom and (unfree) work became redefined by a peculiar blending of Enlightenment ideas with Ancien Regime and Calvinist notions. The result was the idea of work as fulfilment, which served as the motivating force behind Benevolent Societies. These associations wanted to relieve people from poverty, which they considered to be a dehumanizing condition, and to educate them to an arbeidzaam (industrious) existence. This was done, for example, by establishing beggar colonies in the Netherlands. From 1818 onwards indigent from Amster- dam, Rotterdam and other large Dutch cities found themselves relocated to endless moors in the northern provinces of the Netherlands in so-called ‘peat colonies’. These peat colonies would become directly linked to the Dutch colonial policies, as they were created by Johannes van den Bosch (1780–1844). In 1830 he became Governor-General of the Netherlands Indies, where he inaugurated the forced Cultivation System. Van den Bosch was inspired by Philipp Emanuel von Fellenberg, the Swiss ‘Volkserzieher’ whom he even visited at his model farm near Bern. The idea of coercion was alien to these experiments to edify the paupers, but Van den Bosch made a distinction in this respect between paupers, beggars and vagrants. The latter category being defined as indigent without permanent address. Begging was prohibited in the Netherlands since the Napoleonic times. Under the influence of the Enlightenment the idea took root in Dutch society that pauperism had to be banished, begging criminalised, and that indolence was a cause of poverty that needed to be cured. This very idea that the government had the right to put people to work for their own good and

12 An exception is the employment of slaves at the nutmeg plantations of the Banda islands, at the time of abolition there were still some 1200 plantations slaves here. See for example Bosma and Raben, Being “Dutch” in the Indies, 155. 13 See GB of 1819, Staatsblad van Nederlandsch-Indië, 5 January 1819, no. 10. 68 ulbe bosma the commonwealth was transplanted to the Netherlands Indies by Van den Bosch. To put the record straight: Van den Bosch was in favour of the abolition of slavery. Moreover, he was criticized by conservative contemporaries who condemned his attempts to emancipate the poor by giving them work, which in their vision was a subversion of the social order created by God, which was best served by charitable institutions. On the other hand, more liberal minded contemporaries considered Van den Bosch to be a conservative who refused to distinguish between coerced labour and free wage labour. As most of his upper class contemporaries, Van den Bosch saw free labour only existing under conditions of economic independence, which in fact made every form of wage labour qualify as unfree.14 This in fact was a deviation from the notions of labour contract and free wage labour that had existed since the Dutch Republic. Again, the Dutch situation is at variance with the widely accepted notion that from the early nineteenth century citizenship and free labour contracting became linked to each other in a progressive way. It is precisely at this assumed linearity that Prakash directs his well-known critique about the European binary of free and unfree, which in his view had been written on Indian history like it had been on African or Javanese history for that matter.15 Prakash perceives a Eurocentric linear perspective that projects slavery onto antiquity and freedom onto the future, a perspective that was taken for fact by civil servants in different European colonies.16 The Dutch trajectory, however, shows that Europe was far from homogeneous on this account. The emergence of new social ideas and economic concepts or the occurrence of unprecedented poverty, as was the case in the Kingdom of the Netherlands, could drastically shift perceptions on labour issues and even legitimize new types of forced labour. In the Dutch case the early nineteenth century saw a shift in emphasis in thinking from labour that can be contracted to labour as a constraint. Van den Bosch’s involvement in Benevolent Societies and the Cultivation Sys- tem bears witness to the fact that he was familiar with the idea of labour as a constraint, a necessary condition to reach an acceptable level of civilisation both at the individual as well as the collective level. As Schrauwers

14 The notion that all wage labour was in a way coerced labour was also shared by many working class people in the United States and Great Britain according to Steinfeld. See Steinfeld, Coercion, Contract, 13. 15 Prakash, Bonded Histories 8–9. 16 Banaji, Theory as History, 150. dutch imperial anxieties 69 has noticed: ‘Viewed in this light, Van den Bosch’s colonial regime established an alternative trajectory toward the creation of a free labour market than that documented by the English poor law and anti-slavery agitation’.17 The complexities of Javanese society would make Van den Bosch’ trajectory towards free labour a rather contorted one, however.

The Pedagogy of the Forced Cultivation System on Java

Van den Bosch’s Cultivation System was imposed under global market conditions and went along with the massive claims on Javanese land. On top of that came the unequal protection under law. The European employers were the citizens and guardians of the colonial order making the island profitable for the Dutch exchequer, promoting the welfare of the colonial subjects and protecting them from abuse. Whereas Dutch citizens in the Dutch East Indies fell under Dutch law, the Javanese were merely colonial subjects, which in the nineteenth century meant amongst other things that they did not enjoy the same legal entitlements such as recourse to an independent judiciary, fell under different tax regimes and hardly got any education. Attempts at creating legal protection for the peasants and other workers by well-meaning colonial civil servants therefore faltered on the unequal legal status between colonial subject and European citizen and the economic power of the latter. The Cultivation System basically subsumed existing systems of corvée and dependence. Systems of slavery were practically irrelevant in this respect. Beyond the cities and domains of the voc on Java slavery was rather marginal. Victorious rulers and war lords within the disintegrating Empire of Mataram could carry away populations to put them work at the construction of palaces or fortifications.18 Pawnship (cacah-kawula) for food was a rather common way of enslavement. One could also be sold to the Sultan’s palace to be put to work there as a servant or concubine. Slavery was however not an economic factor in Javanese society, as Moertono argues, because the rulers could rely on corvée labour that was abundantly available.19 These ‘traditional’ burdens on the Javanese peasantry were quite significant both in political as well as economic

17 Schrauwers, ‘The “Benevolent” Colonies’, 299. 18 De Graaf and Pigeaud, De eerste moslimse vorstendommen, 219. 19 Moertono, State and Statecraft in old Java, 147. 70 ulbe bosma terms.20 In pre-colonial times the Javanese polities were always in need of sources to feed themselves and in need of soldiers to expand or defend their polities. Until the eighteenth century Java was sparsely populated and the state apparatus of the Mataram Empire was organised in a such way that each department and thus each high official had its own district, or better a group of people, from whom he exacted services. According to custom the peasantry not only had to deliver 2/5 of their harvest to the Sultan and his officials and of the remaining 60 percent another 1/5 for the desa (village) notables, but also to maintain roads and partake in the retinue of the Sultan’s army. This left the cultivator theoretically with about half the produce of his lands. The actual situation was quite different however. Only sawah (irrigated land) holding peasants were due to these taxes, which were in fact based upon a cacah—originally the smallest unit that could supply military later one a unit producing agricultural crops. The cacah- holding peasant, the sikep, had usually a number of dependents to whom he could shift part of the burden. What made taxes less onerous than paper, finally, was that the Javanese rulers simply lacked the apparatus to lay claim to all lands, and particularly not to newly developed lands. The peasants were therefore able to conceal substantial parts of their income.21 Things might have slightly changed for the worse, however, when by the late seventeenth century forced deliveries to the Sultan were transformed into a monetized land tax that were quite often collected by Chinese tax farmers.22 From the seventeenth century onwards the voc had used the rights it had obtained via treaties or as the new sovereign to conscript labour to build and maintain infrastructural works. Each year 400 to 450 Javanese labourers were conscripted to clean the canals in Batavia, for example. Gradually a system of contingents of deliveries of cash crops was extended over West-Java that tapped into the existing patrimonial administrative

20 In my view unfree labour before the arrival of the colonial state, is not as Prakash and Miers & Kopytoff would have it, just a social and cultural phenomenon within the confines of a personal relationship. The argument of Jan Breman that the patron-client relation between landowners and landless in pre-colonial and a colonial India had the extra dimension of a permanent claim on labour of the latter, seems to me quite convincing and also applicable to the situation on Java. Likewise, the slave trade and slavery in Africa was institutional and part of economic life. Prakash, Gyan, Bonded Histories 7; Miers and Kopytoff, Slavery in Africa; Breman, Labour Bondage in West India; Lovejoy, Transfor- mations in Slavery, 24. 21 Ibidem, 73. 22 Hoadly, Towards a feudal mode, 142. dutch imperial anxieties 71 tax system.23 The bupati and mantri, the high regional officials of the waning Mataram Empire, were ordered to deliver quantities of coffee and European horticultural experts were attached to these functionaries. The whole system of forced cultivations and contingents that had been introduced under the voc was at variance with the process of monetization of taxes that had already taken place under the Mataram Empire. A monetization, however, that had also demonstrated its dark sides. For the smaller peasants it was difficult to pay theses taxes or land rents in cash, and here the Chinese and other moneylenders came in to provide advances on the crops against high interest rates of course.24 It was precisely for this reason that Van den Bosch, unlike most of his contemporaries, did not denounce the voc system of contingents. Conversely, he severely criticized the system of land rent imposed during the British interregnum under Lieutenant-Governor-General Thomas Stamford Raffles (1811–1816). The peasantry had great difficulty to produce these taxes, which made them vulnerable to moneylenders. Moreover, the corvée for local officials had all but disappeared. The hodge-podge of taxes, land rent, bondedness and different corvées that existed in early nineteenth century Java was a disaster according to Van den Bosch. But as a remedy he did not expect anything from the introduction of sound liberal economic principles, because in his view these would not work in a Javanese society that was steeped in despotism. His own proposal to impose a system of forced labour and forced cultivation—he literally used these words—at first sight could hardly be considered as less despotic, but he made it palatable by presenting the Cultivation System as an apprenticeship of some sorts, a social remedy against feudalism.25 He allegedly had declared in 1830 that forced cultivations should have been applied just for ten years, after which ‘free labour’ should have become the norm.26 I have no evidence to corroborate this, but what we know for sure is that a rapid introduction of wage relations in the forced sugar cultivation took place from the early 1830s onwards when Van den Bosch was still in office as Governor-General. Moreover, wages and monetization were a basic feature of the Cultivation System.

23 After the Priangan in West-Java had been officially handed over to the voc by the Sultan of Mataram, the Susuhunan, in 1705, forced cultivation of coffee was introduced. Hoadley, Towards a feudal mode of production 48, 106. See also Breman, Koloniaal profijt. 24 Carey, “Waiting for the ‘Just King’,” 99. 25 Van den Bosch, Nederlandsche bezittingen ni Azia, Amerika en Afrika i, 222. 26 Vitalis, Opmerking omtrent den loop der suiker-industrie, 62–3. 72 ulbe bosma

The Javanese cultivators had to pay their land rents from their crop payments. In practice the former was often deducted from the latter, in principle however the cultivator would be enabled by the government to pay for his land rent.27 This aspect seems to me an underrated element of Van den Bosch’s ‘pedagogy of forced labour’ as it was crucial to introduce the Javanese peasantry into the workings of the wage economy. When Van den Bosch presented his plan for the Cultivation System to the Dutch King, he was aware of the fact that it would be against the existing communis opinio in the Netherlands. But for most of the Dutch proletarians, living under constant fear of starvation, the whole concept of free labour did not mean anything, he argued.28 In his analysis most of the labourers in densely populated countries like England and the Netherlands had practically no access to means of production, which in practice reduced them to the same social status as slaves.29 This point of view had been foundational for the Dutch peat colonies and now became extended to Java. What Van den Bosch believed to see was that the Java- nese peasantry only needed a few hours work per day for subsistence. The Cultivation System was an exercise in teaching the Javanese population industriousness, and at the same time a mechanism to make the Nether- lands Indies profitable for the Dutch exchequer rather than a permanent loss. What Van den Bosch was aiming for was to unchain an industrious revolution in the tropics, so to speak, and to replace local circuits of moneylenders and tax farmers by an economic circuit between colony and metropolis. The semi-governmental Dutch Trading Society played a central role in connecting Dutch and Javanese markets and workers: supplying the Javanese cultivators with advances, bringing commodities to Europe and calicos back to Java and thus encouraging the cotton industry in the Eastern part of the Netherlands.30 At any rate the colonial requisitioning of labour did not right away replace ‘traditional’ dependency relations. In practice, the rural population of Java suffered under the double burden of cultivation conscription and traditional corvée well into the nineteenth century. The Javanese local power brokers did not give up their rights easily, and there was no urge for them to do so, because the Dutch colonial government had to rely on them. The resistance and also massive desertions from West and

27 See Fasseur, The politics of colonial exploitation. 28 [J. van den Bosch] “Nota”, 316–317. 29 See Berends et al., Arbeid ter Disciplinering. 30 See Schackmann, De proefkolonie. dutch imperial anxieties 73

Central Java to the more sparsely populated areas of East Java amply demonstrated that the Javanese peasantry did not consider the introduction of the Cultivation System as a first step out of despotism into the realm of freedom.31 In fact, it turned out to be far more burdensome than past arrangements. More land and more persons were included in cultivation conscription (cultuurdiensten) than in the corvée of Javanese rulers or the landrent schemes that had been imposed by Raffles. Only the elderly, sick and priests were exempted, whereas peasants who did not hold sawahs were also subject to cultivation conscription, although to a lesser degree. The Cultivation System became thus far more encompassing than past taxation schemes, and though cheating the authorities by underreport- ing the number of villagers and the acreage of land did continue, control over the rural areas became tightened up in the course of the nineteenth century. Moreover, the burden of growing cash crops often turned out to be underestimated, as for example the soil became depleted from growing indigo or much more cane had to be brought to the factories than initially had been reckoned with. And although the Javanese officials, the lurah, mantri and bupati, got their share of the so-called cultuurprocenten (cultivation bonuses) they still levied their tributes from the people in their resort. It took time to transform the patrimonial Javanese state into a modern bureaucracy.32 Another aspect that might seem to contradict Van den Bosch’s intention to familiarize the Javanese with the pursuance of his own economic interests is the fact that the Cultivation System both strengthened communal tendencies and inequalities in village life. The village elite had been put in a more powerful position towards their fellow desa men and women than ever before. In 1838 European agricultural and other enterprises were allowed to negotiate contracts with the ‘eldesten en most prominent’ members of the Javanse village (the desa) for the delivery of labour. This was a fundamental departure from the principles of the Constitution of Netherlands Indies of 1818, from which the regulation about individual labour contracting had been derived.33 Burden sharing was also strongly promoted by the Cultivation System, without becoming egalitarian however. J.H.F. Sollewijn Gelpke, a civil servant who had made extensive study of Javanese village

31 Bosma, “The discourse on free labour and the forced Cultivation System,” 400. 32 See Sutherland, The making of a bureaucratic elite. 33 GB Stbld. 25 December 1838, no. 50, Van Geuns, De suikeronderneming Djatiroto, 28, 35. Although the possibility of hiring coolies on an individual basis continued to exist, it meant a firm deviation from the principles of the GB of 1819, Stbld. 5 January 1819, no. 10. 74 ulbe bosma in the late nineteenth century, mentions that sugar planting desa’s were quite willing to accept additional villagers to become sawah holders, but only after they had first been working on the village cane fields.34 The Cultivation System indubitably featured communalisation of landholding, but it should not detract us from a far more fundamental development, namely the proliferation of wage labour, which became particularly visible in the sugar industry.35 Under the Cultivation System sugar manufactures were contracted to process the cane that was delivered from the villages assigned by the colonial government to grow cane. The factories had to sell the sugar back to the government, which via the Dutch Trading Society took care of the shipment to the Netherlands. In the sugar sector Van den Bosch’s vision of creating a wage economy was making headway in the 1830s, as we will detail at the following pages. Whereas in the initial years of the Cultivation System practically the entire payment by sugar factories had been in crop payments plant (a combination of wage and lease), in 1886 the sugar industry paid 3/4 in coolie wages and only 1/4 in land leases. The share of wage payments would still grow over the decades to come.36 The intense debates by colonial experts on the issue of ‘free labour’ should be considered against the backdrop of the rapid advance of wage labour relations.

Partial Proletarianisation and the Penal Sanction on Java

Wage labour is not necessarily synonymous to free labour and vice versa. In the eyes of the Javanese country man or woman the free labourer was not the proletarian but the gogol, the peasant who owned his or her land and did not need to rent out his labour to other employers.37 The free labourer was someone who owned sufficient land for his subsistence, and he or she was in declining order followed by the tenant and the share cropper, whereas the landless dependent worker was tailing this ranking. The proletarian, the quintessential ‘free labourer’ in the Marxist sense, hardly existed in Javanese rural society, since most labour was absorbed in some kind of dependent relationship. Apart from a small but growing

34 See Sollewijn Gelpke, Naar aanleiding van Staatsblad. 35 About the advance of wage labour in the Cultivation System see Elson, Village Java, 179–226. 36 Suermondt and Van den Berg, Nota over de suikerindustrie, 11. 37 See Sollewijn Gelpke, Naar aanleiding van Staatsblad, 1878 No. 110. dutch imperial anxieties 75 stratum of itinerant workers and urban proletarians landless rural workers were usually dependents on the desa elites.38 In the course of the nineteenth century the class of dependent sharecroppers and entirely landless Javanese would increase. The strategy of the Dutch colonial government and private enterprise was to tap on this labour reservoir and reduce the claims of the Javanese officials and notables on tribute labour. Over time, as Knight has pointed out, a reduction of their rights to corvée labour had taken place, which was very much in the interest of the factories who complained that it were these rights of local chiefs that stood in the way of the advance of wage labour, which they assimilated with ‘free labour’.39 It all fitted the transformation process of the Javanese from a landowning taxpayer, which he knew he was for centuries, to a coolie. Sugar factories were in favour of wage labour but not at all wary if this had been obtained under coerced conditions. For them free labour just meant that the labouring masses of Java were liberated as much as possible from corvée, tribute and extractions they considered to be parasitical. For the same reason, the increasingly powerful sugar factories developed a preference for individual labour contracting rather than recruitment and payment via the desa elites. It is no coincidence that the regulation about collective labour recruitment in the desa was withdrawn in 1863, by a Minister of Colonies who had been administrator of a sugar factory himself and who had been keen to pay directly to factory coolies rather than to the desa elites. Otherwise the coolie would get hardly anything, he had observed.40 On the other hand, the lurah could not be missed by the sugar factories. And thus he continued his job of recruiting dependent labour in the village for the sugar factories who provided him with a premium for every worker he could deliver.41 What sugar factories understood by free labour came down to a transformation of the lurahs from fiscal officials into a kind of mandurs (recruiter/foreman) who had to deliver a certain a number of workers. The agenda of the factories was to reshape the village into a labour reservoir for the factory.42 Meanwhile, the sugar factories were busy to wrestle the sawahs (irrigated rice fields), the means of production of a Javanese cultivator, out of his hands against sub-economic prices. The Javanese peasant usually

38 See Carey, The power of prophecy. 39 Knight, “Peasant Labour and Capitalist Production,” 259. 40 Fransen van de Putte, De regeling en uitbesteding, 30 ff. 41 Knight, “Peasant Labour and Capitalist Production,” 255–6. 42 Margana, “Hybridity,” 101. 76 ulbe bosma leased his land to the sugar factory not because he expected to gain from it, but under the compulsion of indebtedness, the pending payment of his land rent, or outright coercion by the village head. These facts were well- known and stated again by a government investigation in 1894.43 More- over, the price the Javanese peasant could ask for his land was lower at the time of puasa (ramadhan) and at the time the landrent had to be paid. This gave the factory an enormous advantage, because with a good timing it could negotiate with people who were in dire need of money.44 This was not changed by the emergence of credit cooperatives (volkskredietbanken or desa lumbungs), which had been actively promoted by the Indische government as part of its welfare policies that had been inaugurated in 1901 under the name of Ethical policies. Around 1920 it was still primarily indebtedness that made a peasant lease out his land to a sugar factory.45 It also still happened that impoverished Javanese rented their land out against a pittance to rich countrymen, who in turn gave this land in lease to sugar factories.46 This partly undid the protection which the colonial government had given to Javanese peasants by not allowing Europeans or ethnic Chinese to have land in property. Even worse had been the effects of the Agrarian Law of 1870 that allowed for long term leases of 75 years to establish plantations in the middle of Java on so-called waste land but that often enclosed villages. Meanwhile, the government of the Netherlands Indies did its part to facilitate the transition from tribute to tax. After 1854 it embarked upon a scaling down of the corvée labour for the maintenance of public works, a process that was completed only in 1916. Those conscriptions services were however replaced by monetized taxes. This idea was wholeheart- edly endorsed by colonial business circles as yet another step towards free labour for plantation estates, and compelled Javanese to do even more wage work in order to pay their taxes. It was difficult in the 1860s to get sufficient coolie labour particularly for the sugar industry, because often the planting of paddy, the preparation of the fields for the new cane and the grinding season coincided. Again, the colonial government came to assist the plantations and the sugar factories in particular. In the early 1860s one could get remission from the corvée imposed by the Indies government for public works, if one would perform coolie services. Wealthier

43 Huender, Overzicht, 90–92. 44 Kohlbrugge, Is grondverhuur, 64. 45 Huender, Overzicht, 91. 46 Verslag van de suiker-enquête commissie, 167. dutch imperial anxieties 77 farmers pressured their dependents, often landless, into coolie work to exempt themselves from these duties. This was the preferred way, because if factory mandurs had to recruit labour they had to advance considerable sums, and it was not always easy to make the coolies honour their commitment.47 That labourers who had been attracted with advances— traditionally sealing a verbal labour contract—did not always honour their commitment was a reason for employers to appeal for penal sanctions to be enclosed in labour contracts to ensure that coolies would not simply walk away with the money. The advance was also known under metropolitan Dutch labour law as handgeld (earnest money) or godspen- ning (queen’s shilling) sealing the labour contract. In Dutch law there was no such thing as a ‘masters and servants’ act, and although the servant or hired worker was not equal to his master before the law, penal sanctions did not exist.48 As said before, in the days of the Dutch Republic a distinction existed between regulations on servants who were part of the household and subcontracting, a distinction which survived Napoleonic times and continued into the nineteenth century. Gradually a more paternalistic relationship emerged in which the subcontracted workers came in more or less the same position as the domestic servants. Interestingly enough, and quite similar to the situation in Great Britain, in the Netherlands the legal definition of a servant was extended from the quotidian ‘domestic servant’ to hired worker in the course of the nineteenth century.49 In the nineteenth century regulations between masters and domestic servants could still be subject to provincial or even municipal regulations within the Dutch Kingdom and its colonies. Such a local regulation was the Surabaya police ordinance of 1829 that determined that every domestic servant on Java who quitted from his temporary contract could be sentenced to time in prison or fined to 25 guilders, and if he was a native or foreign oriental with 8 days block or 30 beatings with a rattan.50 The

47 Elson, “Sugar factory workers,” 153, 166–7. 48 Heerma van Voss e.a., Asser’s handleiding, Vol. 2, no. 2. As a matter of fact, only from 1 February 1909 did the Netherlands have a full-fledged law on labour contracts. Lourens and Lucassen, Lipsker, 38–41. For the Law on the Labour Contract (Wet op het arbeidscontract, ter vervanging van BW art 1637–1639) see Diepenhorst, De Nederlandsche arbeidswetgeving II, chapter I and see Smissaert, De arbeidsovereenkomst. 49 See for the multiple meanings of servant in England Steinfeld, “The Invention of Free Labor”, 17. For the Netherlands see Lourens and Lucassen, “Lipsker op de Groninger tichelwerken,” 40; Smissaert, “De arbeidsovereenkomst der dienstboden”. 50 Indisch Staatsblad 1851, no. 26. Van Geuns, De suikeronderneming Djatiroto, 37, 39. 78 ulbe bosma

Surabaya ordinance was adopted for the entire island in 1851, which in itself was not a problem as long as it had been confined to domestic servants. In 1872 it was however extended to Javanese workers at all kinds of enterprises, which led to an indiscriminate application of the penal sanction to all indigenous wage workers of Java. This went way too far for the Dutch Parliament, which in this very year 1872, had just abolished the prohibition of collective action (e.g. unions and strikes) that had been part of the Code Napoléon. This abolition was attended by the explicit statement in the draft of the Bill that there was no reason to attach penal sanctions on either labourers’s or employers’s breach of contract.51 In the same vein Dutch Parliament decided in 1879 that labour regulations concerning the position of indigenous (inlandsche) servants and labourers should not be part of the Penal Code for the indigenous people, but that the position of indigenous servants and workers, as well as sailors, should be arranged in accordance with the Dutch Civil Code.52 In 1879 The Hague overruled Batavia by determining that only walking away with an advance while not having any intention to perform labour was punishable.53 This was hard to prove of course. On Java the penal sanction did not resurface, but in the Outer Provinces it would as we will see in the next section. The issue of the penal sanction would become less pressing for the sugar factories, because after 1880 labour shortages became less acute. Since then coolie wages on Java hardly rose nominally, and even began to decline in real terms after 1900, to slide down to the level of hunger wages in the Depression years.54 In the early twentieth century the sugar industry came to rely on a growing contingent of (migrant) workers who did probably most of the cutting, haulage from the field and coolie work at the sugar factory. But an important share of the work in the field, the planting and tending of the cane, was done by female labourers, most of whom came from smallholder households.55 In the early twentieth

51 Van Houten, Regtstoestand, 7; Brugmans, Arbeidende klasse, 184–9, 249–2. 52 The Penal Code for indigenous was based upon the European Penal Code, but traditional and religious right could be appointed as advisers to the judge. Kleintjes, Het Staatsrecht II, 51, 57. 53 Van Geuns, De suikeronderneming Djatiroto, 37, 39; Levert, Inheemsche arbeid in de Java-suikerindustrie, 92; Kalma, Het arbeidscontract met den inlander 11, 73; Wertheim, “Planters tegen koelies.”; See also Staatsblad van Nederlandsch-Indië 1872, no. 111, art. 2, no. 27. The “Soerabaja Politiereglement”, Staatsblad van Nederlandsch Indië 1829, no. 8 and Staatsblad van Nederlandsch-Indië 1851, no. 26. 54 Coolie Budget Commission, Living conditions of plantation workers.; Scheltema, The food production. 55 Knight, Colonial production, 33–4. dutch imperial anxieties 79

century probably more than half of the cane planters were landless peasants (including female) employed by the ‘big peasants’ who held an increasing share in the ‘communal’ sawahs or increased their individually owned sawahs.56 Nonetheless, as a concomitant of emerging nationalism and Indonesian labour unions after 1910, labour issues got a new twist. Coolie wages that were clearly below subsistence and sub-economic leasing of land became sources of political contention and collective action. While the coolies of Java had escaped from the penal sanction because of a lack of basis in Dutch law, the situation was less positive as far as freedom of collective action was concerned. Freedom of association was only tentatively established on Java, and collective action had been drastically curtailed in the Netherlands for crucial economic sectors. After the big railway strike of 1903 the right to strike had been withheld from civil servants workers in economically crucial sectors in the Netherlands. This would set a prec- edent for the colony. In order to ensure the orderly conduct of the sugar campaign of 1923 the Indies government outlawed the Union of Indone- sian Sugar Workers (Personeel Fabriek Bond) and practically annihilated the Railway Union (VSTP). It was done with a regulation prohibiting ‘all strikes that could lead to disorganisation of public life or severe economic consequence’.57 This time objections in Dutch Parliament were not, and since 1903 could not be, strong enough to repeal such a clause that seriously violated the principle of free labour. The suppression of Indonesian labour unions was part of a general repressive course on which the government in The Hague had embarked after it had disavowed the more forthcoming attitude of the internationally oriented Governor-General Count Johan Paul van Limburg Stirum (1916–1921). As a diplomat by profession Van Limburg Stirum had been susceptible to a the new international atmosphere with regard to colonial self rule and colonial labour issues, epitomized by the League of Nations. He had announced in November 1918 that the penal sanction for the Out- lying Provinces would be abolished by 1 January 1926.58 In 1919 he installed a committee to investigate the labour unrest in the sugar industry and another one to investigate the desirability to establish a minimum wage for coolies on Java.59 Again, with regard to political emancipation and

56 Ibidem, 22. 57 See Ingleson, “ ‘Bound Hand and Foot’.”; Bosma Karel Zaalberg, 369. 58 Middendorp, Twee achterlijke arbeidssystemen, 35. 59 See Verslag van de arbeidscommissie. 80 ulbe bosma labour issues of the people of the Netherlands-Indies his administration had steered a more progressive course than the metropolitan government. When this Governor-General left office the announced abolition of the penal sanction was in danger.

Penal Sanctions and Coolie Scandals on Sumatra

Since the mid-nineteenth century a growing army of Javanese migrant labourers found their way to the sugar plantations, the harbour towns and finally to the plantations of Sumatra. Breman’s and Stoler’s well known publications on the coolie scandals on Sumatra and conditions on the East Sumatra plantation belt have attracted attention far beyond the circle of Indonesianists.60 Colonial enterprise defended the penal sanction by referring to the high costs for recruiting labour and the need to guarantee a steady supply to this crucial economic sector, in fact elevating their specific interests to a national cause. They also claimed that there was room for exceptions to the principle of free labour, since in the Netherlands civil servants in crucial economic sectors were not allowed to strike and sailors were not permitted to leave their employer during their engagement. Like other colonial governments the government of the Netherlands Indies considered labour contracts with penal sanctions the lesser of two evils. In order to combat recruitment malpractices, they had taken on the task of regulating a phenomenon which was at variance with the principle of free wage labour. In order to ‘regulate’ migratory labour recruitment, the colonial state had now given legal status to labour contracts issued by official recruiters and put a penal sanction on leaving work without the employer’s consent.61 But the Dutch East Indies Coolie Ordinance did not protect the coolie against mistreatment and continuing indebtedness. Moreover, recruitment excesses were widespread. Lurahs and other local Javanese officials were paid for every coolie whom they could deliver to the recruiters.62 The penal sanction in practice gave the plantation management the authority of the police; if coolies fled the dreadful circumstances at the plantations plantation guards went after them and brought them back without any interference of the police. If coolies were sent to

60 See Breman, Koelies, planters en koloniale politiek; Stoler, Capitalism and confrontation. 61 See for example Jan Breman, Koelies, planters en koloniale politiek; Baak, “Enslaved Ex-Slaves”; McKeown, Melancholy Order; Robb, “Labour in India”, 48; Roy, “Sardars”, 972. 62 Tideman, “De koelieordonnantie,” 43. dutch imperial anxieties 81 the police because of alleged breached of their contract, however, they were routinely imprisoned otherwise it would be a loss of face for the plantation management en therefore detrimental to ‘colonial prestige’.63 In its actual working the Coolie Ordinance was heavily if not almost completely biased against the coolies. For every conceivable transgression on their part it was explicitly mentioned that severe punishments could follow, but plantation staff that was known for beating up the coolies regularly had little to fear, as any real measures against them allegedly tainted ‘colonial prestige’. And thus the Coolie Ordinance had become a travesty of earlier regulations aiming at preventing coolie recruitment from falling back into systems of slavery. Since 1868, just eight years after the abolition of slavery in the Netherlands Indies, the first regulations aiming to protect Foreign Orientals from slavery and deception were issued. In 1872 these were extended to native Indonesians and particularly targeted labour conditions in Sumatra. The Coolie Ordinance of 1880 was plainly a system of indenture, as it stipulated penal sanctions on leaving the plantation within the fixed term of employment of three years.64 In spite of the announcement made in 1918 by Governor-General Van Limburg Stirum about the out-phasing of the penal sanction it remained in place until 1931 and when it was abolished it was not because of humanitarian concerns or respect for international conventions. No, the reason were purely economic. By that time the Depression had set in and the employers were massively laying off their contract coolies. Indeed, as Stoler observes, the abolition of the penal sanction helped the employers to get rid of their labour without paying them anything for their remaining contractual time.65 More- over, the US Senate had adopted the Blaine Amendment prohibiting the import of goods produced under forced conditions, which would be put into effect from 1 January 1932 onwards. The initiative for the Amendment was not rooted in humanity but a protective measure for American Vir- ginia tobacco that suffered from competition by the Sumatra leaves.66 This as far as the actual facts of the matter are concerned. What concerns us here is how the penal sanction could be reconciled with the fact that under Dutch Civil law a labour contract was a contractual arrangement between two equal partners. A fact that had become explicitly part

63 Ibidem, 58–9. 64 Koelie-ordonnantie van 13 juli 1880, Staatsblad van Nederlandsch-Indië 1880–113. See also Bosma, Karel Zaalberg, 100. 65 See Houben and Lindblad, Coolie Labour; Stoler, Capitalism and Confrontation, 43, 88. 66 See E. Kupers, Het vraagstuk van de poenale sanctie. 82 ulbe bosma of the Dutch law only in 1909, but which was nonetheless deeply rooted in Dutch history. Moreover, when the Dutch Parliament had adopted the law that legalised trade unions and strikes in 1872, it had also decided that penal sanctions should have no place in an arrangement where the government was not a party. The penal sanction for the Sumatra plantations had come in through the backdoor of the regulation of recruitment practices and was maintained on the grounds of exceptional circumstances and paramount colonial interests. The fact that it was exceptional made it already a legal anomaly, which became even more pronounced when the ‘paramount colonial interests’ were themselves were redefined after World War I. ‘Home Rule’ as adopted as the formal objective not only of British but also of Dutch colonial policies in Asia. It is no coincidence that the revisiting of the penal sanction was announced in the context of the general reform of status of Netherlands Indies inspired by the Chelmsford Montague reforms for India. The announcement of the Governor-General van Limburg Stirum in 1918 about the abolition of the penal sanction was part of a broader constitutional reform that would bring ‘Home Rule’ for the Netherlands Indies, the so-called November Promise. The East Suma- tra planters understood that they had to compromise and they came with their own proposal for a new ordinance. They were ready to acknowledge that the ‘coolie’ was a free subject and contracting partner, they also admitted that Coolie Ordinance should be better calibrated, but their proposal fell short of parting from the penal sanction.67 In principle, their position might have been accepted 15 years earlier, as the Dutch Social Democrat and specialist on colonial issues J.E. Stokvis wrote in a pamphlet against the continuation of the Coolie Ordinance. In the early years of the twentieth century there had been Labour members of Parliament who had been in favour of penal sanctions in Dutch labour law, as long as these were applicable to employers too.68 But, Stokvis went on, in colonial society where the (European) employers and (native) workers were never equals, a penal sanction only engendered underpayment and racialised violence. In fact, the labour migrants from Java to the Out- lying Provinces were recruited to work against a payment just sufficient to reproduce him or herself, which in Stokvis’s view came down to slavery.69

67 See Ontwerp eener Arbeidswetgeving. 68 Stokvis, Koloniale Schande, 4. 69 Ibidem, 5. dutch imperial anxieties 83

E. Kupers, Chair of the Social Democratic Labour Union and Member of Parliament70 as well as one of the Dutch representatives at the meetings of the ILO pointed out that it was the combination of the penal sanction, the racial inequality and the fact that the recruited Javanese labourer had no countervailing force, no labour union to stand for his rights, which made the conditions in East Sumatra unacceptable. Reading the Coolie Ordinance superficially, one might even have the impression that it would protect the coolie, but the situation at the plantations was violent if not murderous. It has been noted that around 1918 about 8 percent of the 200,000 contract coolies ended up in prison each year. This percentage may have just slightly decreased to 5 per cent in 1926, thus in the final years of the penal sanction. The Coolie Ordinance created a labour regime based upon fear, punishment and violence.71 For this reason Albert Thomas, the first director of the ILO condemned the penal sanction as a remnant of servitude. Moreover, another critic pointed out, contracting labour for the Sumatra plantations under the penal sanction was utterly uneconomical. The plantation had to pay 100 to 150 guilders per coolie, while the coolie wage per day was 55 cents. So, why not encourage free migration against higher wages?72 The question has never received a serious answer and deserves further research. I would suggest as a hypothesis that the very fact that labour recruitment was a highly profitable business may have played a role in the continuation of the penal sanction. Indentured conditions to work of the fares were not at all uncommon in the history of labour migration and was not confined to Asian labour. Over the course of the nineteenth century white labour continued to work under various degrees of bondedness or coercion in the so-called settler colonies.73 Constraints and non-pecuniary sanctions were not absent in European labour contracts either, particularly not in the maritime sector. The legal position of the sailors, as a matter of fact, had been subject of international legislation since the 12th century Rolls of Oléron came into existence as the germinal of a corpus of international maritime law. But in the new context of international labour legislation by the ILO the legal position of seafarers became connected to issues of forced plantation labour and coerced conditions. Labour conditions of the seafarer became

70 See “E. Kupers.”. 71 Tideman, “De koelieordonnantie en hare toepassing,” 55; Middendorp, Twee achterlijke arbeidssystemen. 72 Middendorp, Twee achterlijke arbeidssystemen, 49. 73 Bosma, “European colonial soldiers,” 318. 84 ulbe bosma a focal point of the ILO legislation, not only because these issues were almost by definition global in character but also because of the existence of penal sanctions in maritime law, which were considered to be detrimental to the fight against indentured conditions on plantations. Indeed, advocates of the continuation of the penal sanction on the Sumatra plantations used the fact that desertion of sailors and some categories of civil servants at railways were still under the penal sanction to shore up their cause. But the argument lost its final bits of credibility in the 1920s, when the ILO developed international legislation to ensure honest recruiting practices in the maritime sector and in 1926 to regulate the indentured condition of the seafarer.74

Conclusion

In the nineteenth century labour contracts that comprised a penal sanction were not necessarily conceived as in contradiction with free labour, as we would today, but rather as constrained freedom which in fact every type of wage labour was. The Cultivation System in the Dutch East Indies was implemented as an upshot of ideas about the edifying role of work that had first been tested in the Netherlands. Van den Bosch had been a vision- ary whose intention had been to give the Javanese peasant the fruits of his work and land. That he wanted to make the Javanese rural population work harder sounds colonial, but it is not necessarily so, and in his case it was derived from the Netherlands context, where he had established beggar colonies. It could hardly have come as a surprise that the introduction of the Cultivation System in social conditions that Van den Bosch himself had qualified as despotic, in practice aggravated the already precarious position of the majority of Java’s rural population. The colonial government not for nothing had to scale down the Cultivation System in the 1840s, which did however not diminish the inequalities at the village level. Van den Bosch’s Cultivation System was not able to break through despotism for the simple fact that it had to rely on it. The coalition between the village elites and the sugar factories, which I focused on in this article, continued even into postcolonial times. Until this very day village patriarchy

74 ILO Convention No. 9, Placing of Seaman Convention 1920 and Convention No. 22, Seaman’s Articles of Agreement Convention, 1926. dutch imperial anxieties 85 supply labour and moneylenders siphon off the resistance of the sawah owners to lease their land out against sub-economic prices.75 It has been noted that the step from bondage to contract with penal sanctions may constitute a feeble step towards freedom. For women for example it could mean an escape from patriarchy.76 In Java large segments of the population had to enter the wage labour market, simply because their access to land was encroached upon by plantation enterprise. Free labour for the Java sugar industry meant free availability of labour. As soon as the sugar manufacturers were confronted with a tightening labour market they asked for a penal sanction and they got one, although it was repealed after an intervention by Dutch Parliament. At that time it was felt that within the Dutch empire a penal sanction had no role in ordinary labour relations. The penal sanction for Sumatra came via the backdoor of regulating recruitment practices. The abolition of the penal sanction that was in force outside Java was stalled on the basis of extraordinary circumstances at the Sumatra plantation belt and its paramount economic importance for the Dutch empire. The repression of collective action in the Netherlands Indies in 1923 was motivated on behalf of the same paramount interest of preserving the colonial order. The way in which the colonial government of the Netherlands Indies invoked ‘extraordinary circumstances’ to legitimize penal sanctions caramboled however via the international arena to the Netherlands. The indentured conditions of sailors on Dutch ships were softened in 1937, when with some delay the relevant ILO Convention of 1926 came into force in the Netherlands.77

75 See Ingleson, “Life and Work”; Chandra, “The Role of Female Labor”; Mather, “Industrialization”. 76 Sen, “ ‘Without His Consent?’ ”, 77–104. 77 Schuman, Tussen vlag en voorschip, 203.

Children and Forced Labour in the Indian Ocean World, Circa 1750–1900

Gwyn Campbell

This paper examines the structure and significance of servile child labour in the Indian Ocean world (IOW) in the period c.1750–1900. To do so, it needs first to define the terms “IOW” and “child,” and delineate the main features of traditional child servitude in the region defined.

The Indian Ocean World (IOW)

The IOW refers to a vast region running from eastern Africa (from the Cape to Cairo) through the Middle East, South and Southeast Asia to the Far East. This region was witness to the first “global economy,” defined as a sophisticated, durable structure of long-distance exchange of commodities, ideas, technology and people―as distinct from the modern “international economy” which began to take shape in the nineteenth century. The economic, social and political foundations of the IOW date back two millennia and are related to the monsoons, a system of regularly alternat- ing winds and currents unique to the Indian Ocean, and Indonesian, and South and East China Seas. The monsoons exerted a huge influence over the lands and societies around these inter-connected bodies of water: Monsoon rains underpinned agricultural production, while monsoon winds created the possibility of direct trans-oceanic sail which in turn facilitated the rise of a structure of long-distance maritime exchange. The IOW global economy, which linked the three major production centres of China, South Asia and the Middle East to each other, and to other IOW regions such as Indonesia and eastern Africa, started to develop from about 300 BCE, had become a fully fledged exchange system by at least 1000 CE, and was only seriously challenged by Europeans from the late eighteenth century. Traditional IOW patterns of exchange continued to be vibrant up to the immediate post-1945 era, since when they have become steadily more attenuated. 88 gwyn campbell

Definitions of Children

The current definition of a child is generally someone under the age of eighteen years.1 In the IOW, traditional definitions of children varied. In much of the IOW, as in pre-modern Europe, children and females were considered inferior to, and under the authority of, adult males. In China, where a female could never carry on the ancestral line, an unmarried girl was submissive to her father; a married female to her husband; and a widow to her son. Females were never accorded full adult status.2 A male was similarly submissive to his father up to the time of the latter’s death. Only then did he assume full adult status.3 In China, as in nineteenth century Thailand and Madagascar, a husband could sell his wife, or children.4 On the French Mascarenes in the eighteenth and early nineteenth century, child slaves were deemed to be those under the age of fourteen.5 However, European enclaves in the IOW were rare before the late eighteenth centuries, and as, historically, most societies indigenous to the IOW did not issue birth certificates, it was often difficult to accurately divine the age of a child. Thus in Egypt in 1834, a European doctor put the age of one young boy slave at three, while two colleagues considered him to be twice that age.6 Traditionally, servile adults were differentiated from servile children by the onset of puberty for females and a combination of age and height for males. For example, in Imperial Madagascar in the early 1830s, slave- raiding armies executed male captives over a certain height (variously estimated at between 107 cm and 122 cm) while retaining those under that height―generally considered to have been ten years old and younger― as “boy” slaves.7 A similar definition seems to have applied in the 1830s by Egyptian slave-raiding armies who only enslaved captive females, and boys.8 The reason older boys and adult males were killed was because captors feared that they would resist, the costs of surveillance were much

1 Campbell, Miers and Miller, “Introduction” to idem (eds), Children in Slavery Through the Ages, 3. 2 Hirata. “Free, Indentured, Enslaved;” Dull. Ed. Han Social Structure, 4. 3 Dull. Ed. Han Social Structure, 21–2. 4 Bowring, “Historical Forms of Bondage in Siam,” 286; Dull. Ed. Han Social Structure, 110; Campbell, “Slavery and Fanompoana,” 476. 5 Allen, “Children and European Slave Trading,” 38. 6 La Rue, “Brief Life of ‘Ali,” 73. 7 Freeman and Johns, A Narrative, 39–40; Griffiths, Hanes Madagascar, 62; Callet, His- toire des Rois, 322; Ellis, History of Madagascar. vol. I, 138; Campbell, “Role of the London Missionary Society,” 119; Hastie, “Journal” (1817–18), 184, 250–1. 8 La Rue, “Brief Life of ‘Ali,” 77. children and forced labour in the indian ocean world 89 higher, and they were less adaptable. Indeed, there is growing consensus that in traditional IOW slave systems, the majority of humans trafficked were young females and children.9

The Rise of the International Economy and Bonded Labour in the IOW, circa 1750–1900

The period 1750 to 1900 witnessed the inter-related development of the international economy, abolitionist movement, and modern colonial order―events that shaped the changing nature of forced child labour in the IOW. The international economy emanated from the Industrial Revo- lution, associated technological innovations, and an international financial order centred on the City of London.10 It expanded rapidly, and by the first decade of the twentieth century had drawn all bar the most remote societies into its orbit. In the capitalist heartland of Western Europe and North America, these forces, occurring in a context of rapid demographic expansion, undermined forms of bonded labour and promoted the rise of contract wage labour. This process took longer than is traditionally thought―slave labour persisted in the United States until the 1860s―but by the late nineteenth century wage labour predominated. However, the structure of the international economy remained highly unbalanced, with on the one side an industrialising core of largely western countries, and on the other the extra- European world the economies of which remained largely pre-modern. In the IOW, population growth rates were comparatively lower and uneven, capital was comparatively more expensive, and more readily available from Indian and other indigenous creditors than from Western sources,11 and attempts at industrialisation―with the exception of Japan―failed (as in Madagascar)12 or remained small-scale with limited linkages.13 Hence, the economy of the IOW remained overwhelmingly agricultural and artisanal. Capital investment was directed predominantly into the cash crop sector, mineral extraction, and exploitation of forest resources, as well as

9 See e.g. Campbell, Structure of Slavery. 10 Cain and Hopkins, British Imperialism: Innovation and Expansion, 1688–1914. 11 See e.g. Campbell, “Indians and Commerce in Madagascar.” 12 Campbell, “An Industrial Experiment.” 13 See e.g. Pomeranz, The Great Divergence; Tomlinson, Economy of Modern India; Fahmy, All the Pasha’s men, 13–18, 42; Toledano, State and society in mid-nineteenth-century Egypt. 90 gwyn campbell into the transport and communications infrastructure required to facilitate exports to regional and Western markets.14 Given the limited nature of capital investment, such developments greatly enlarged demand for labour, not only in extractive and production processes, but also in transport (e.g. human porters; dockers; sailors). However, major troop movements and armed conflicts (notably in Euro- pean and indigenous “secondary” imperialism), increased trans-IOW communications, and natural disasters, resulted in increased mortality.15 For example, cholera, endemic in Bengal, was from 1820–22 carried by Brit- ish troops and ships throughout the IOW.16 Cholera killed approximately 13 per cent of Cairo’s population in the 1831 outbreak,17 and almost 30 million Indians in epidemics that swept Indian in the second half of the century.18 The evidence from China illustrates what could happen when man-made disasters coincided with natural catastrophes. One of the few IOW regions initially characterised by very fast demographic growth, China’s population leapt from approximately 300 million to 420 million between 1800 and 1850. However, due a combination of exceptionally cold weather in South China from 1876 to 1895,19 epidemics experienced in 36 of the 61 years from 1811 to 1872,20 almost constant warfare from 1850 to 1878, and famine, population growth stagnated and by 1900 the Chinese population stood at only 450 million.21

The Abolitionist Movement in the IOW

As most IOW labour was tied in traditional servile relations to indigenous courts and elite households, there existed a very restricted free

14 Heydenrych, “Railway Development in Natal”; Maestri, “Naissance et premiers dével- oppements d’un outil économique: le chemin de fer de la Réunion”; Porter, Victorian Ship- ping; Cain and Hopkins, British Imperialism, 333; Kenwood and Lougheed, Growth of the International Economy, 30. 15 For Southern and eastern Africa see Chrétien, “Demography and ecology in East Africa”; Paillard, “Les recherches démographiques sur Madagascar”; See also Austen, Afri- can Economic History, 67; Campbell, “State and Pre-colonial Demographic History,” 415–45; Kjekshus, Ecology Control, esp. ch. 1; Campbell, “Disease, Cattle and Slaves.” 16 McNeill, Plagues and Peoples, 231–4. 17 Ibid., 231–4. 18 Chandra, “Colonial Legacy,” 11. 19 Landsberg, “Past Climates from Unexploited Written Sources,” 61–2. 20 McNeill, Plagues and Peoples, 268–9. 21 Ponting, Green History of the World, 241; Barraclough, Times Atlas of World History, 174, 232. children and forced labour in the indian ocean world 91 wage labour sector. The commercial boom associated with the expanding international economy thus greatly stimulated the regional demand for servile labour. In this context, abolition was viewed as a means of releas- ing labour resources onto the free labour market where the forces of supply and demand would operate. However, the abolitionist movement was spearheaded by western powers, notably Britain. It had an initially muted influence in the IOW except in the Cape where, well before the 1834 abolition of slavery, the 1826 removal of measures protecting wine had diverted investment into the far less labour intensive wool-producing sector―a measure that effectively undermined slavery.22 Elsewhere, the labour intensive nature of local economies and lack of free wage labour rendered largely academic any arguments that slave labour might be inefficient.23 Until the abolition of slavery in European territory (in British colonies in 1833; French colonies in 1848; the Dutch East Indies in 1860), European- held territory in the IOW initially depended largely on slaves to meet their labour demands. For instance, Robert Farquhar, the first British governor of Mauritius, delayed anti-slave import measures in acknowledgement of the cheap labour requirements of local sugar planters.24 The British, in a variant adopted by other colonising powers, declared newly-conquered territories to be “protectorates” and thus avoided enforcing some abolitionist measures compulsory in “colonies.” India, under East India Com- pany rule until 1859, was considered a separate case.25 Overall, abolitionist measures failed to significantly augment the free labour market in the IOW where both European and indigenous authorities resisted complete emancipation of slaves who, when genuinely freed, as on Mauritius from 1839, sought livelihoods independent of their former slave owners.26 In the late-nineteenth-century imperialist surge in the IOW, abolition formed a central justification for the imposition of European colonial rule. Moreover, colonial regimes, governed by precepts of self-financing, viewed “liberated” slaves as a vital source of both taxation and manpower. How- ever, another priority of colonial authorities was to retain the goodwill of local slave-owning elites whose collaboration was required to administer the colony. Thus, while moving quickly to hinder slave trading, colonial authorities were reluctant to enforce rapid abolition lest it spark revolt, as

22 Worden, “Indian Ocean slavery and its demise.” 23 Clarence-Smith, Economics of the Indian Ocean Slave Trade, 4–5. 24 Anderson, “Bel Ombre Rebellion”; see also Oliver, “Sir Robert Townsend Farquhar.” 25 Chatterjee, “Abolition by Denial.” 26 Allen, Slaves, Freedmen, and Indentured Laborers, 105–35. 92 gwyn campbell occurred in the Muslim province of the Southern Philippines after abolition was enforced in 1904.27 In Somalia, the colonial regime initially permitted European settler farmers access to slave labour, and even returned fugitive slaves to their owners,28 while in German East Africa, European planters were permitted to “ransom” slaves who were obliged to work for their “liberators” until the ransom had been paid off.29 A widespread variant of this was the practice of Christian missions in Africa of redeeming slave children whom they subsequently employed working for them.30 Anti-slavery measures in European controlled territories occurred fit- fully well into the twentieth century. In Africa, the internal slave traffic remained buoyant for some 50 years after the banning of the external slave trade. In the Sudan, effective measures to curtail slavery were taken only in the late 1920s.31 On the eastern side of the IOW, the French first seriously applied anti-slavery measures in Indochina in 1897, while the British abolished slavery in Hulsawng valley in eastern Burma only in 1926. Slavery was outlawed in the Netherlands Indies in 1860, but the Dutch then possessed only one quarter of the Indonesian territory that was to pass under their control by 1910―in much of which they tolerated slavery. Slavery endured in remoter regions of French Indochina and the Dutch Indies into the 1940s.32 In the Middle East, drawn into the British informal empire after the First World War, abolitionist pressure remained muted until the post-1945 era.33

Types of Servile Labour in the Period 1750–1900

The failure to abolish traditional forms of servitude, and the search for new forms of coerced labour, resulted in a particularly wide range of types

27 Clarence-Smith, “Islam and the abolition of the slave trade”; Salman, “meaning of slavery.” 28 Eno, “abolition of slavery and the aftermath stigma,” 83–9. 29 Miers and Klein, Slavery in Colonial Africa, 6; see also Kopytoff and Miers, “African ‘Slavery’,” 73–4. 30 See e.g. Clarence-Smith, “Redemption of Child Slaves by Christian Missionaries.” 31 Miers, “Slavery and the Slave Trade in Saudi Arabia”; see also Kopytoff and Miers, “African ‘Slavery,’ ” 72; Miers and Klein, Slavery in Colonial Africa, 1–2, 4–5. 32 Klein, “Emancipation of Slaves in the Indian Ocean”; Boomgaard, “Human Capital”; Delaye, “Slavery and Colonial Representations in Indochina”; see also Reid, Slavery, Bond- age and Dependency, 34; Klein, “Introduction: Modern European Expansion and Tradi- tional Servitude in Africa and Asia,” 24. 33 Miers, “Slavery and the Slave Trade”; Mirzai, “The 1848 abolitionist farmān”; Klein, “Emancipation of Slaves in the Indian Ocean”; Clarence-Smith, “Islam and the abolition of the slave trade.” children and forced labour in the indian ocean world 93 of servile labour from 1750 to 1900. Traditional IOW systems of servile or unfree labour that were perpetuated into the post 1750 period varied from chattel slavery to the client slavery that characterised the prazos of Mozambique, to pawnship, to agrestic servitude in India, to debt bondage.34 Forms of enslavement varied, but the five types dominated: enslavement through warfare (captives); justice (criminal penalties); raids/kidnapping; sale of dependants; and indebtedness.35 Traditional slavery continued to be important in the IOW. Moreover, children were in greater demand than adults as slaves because they could be more easily moulded, learn the host language, and imbibe local religious values and traditions. They also more easily forgot their own language, customs and memory of homeland, and forged relationships of total dependence upon their master.36 For this reason, children born into slavery were highly valued,37 but demand was not met locally, partly because of a generally low rate of slave reproduction, partly because of a relatively elevated rate of manumission. Certainly a steady trickle of slaves was assimilated into many African communities, depleting local slave stocks and encouraging further slave imports. A similar process was evident in Muslim communities, where the sharia extolled manumission as meritorious, stipulated that children born to her owner by a slave women would be free, and that a concubine who bore a child to a free Muslim would, upon his death, be manumitted.38 The rate of manumission could theoretically be high; whereas a rich Muslim was legally restricted to four wives, the number of concubines he might possess was unlimited.39 These factors ensured a high demand for imported slave children. Such was demand that the IOW slave trade peaked in the nineteenth century despite growing abolitionist scrutiny that induced slavers to adopt indirect routes and pass slaves off as non-slave porters, sailors, domestics, and even as children or other kin.40 Large numbers of East Africans were

34 For Africa see Isaacman, Mozambique: The Africanization of a European Institution, The Zambesi Prazos; Kopytoff and Miers (eds.), Slavery in Africa; Lovejoy, Transformations in Slavery; Meillassoux, Anthropology of Slavery; Manning, Slavery and African Life. For Asia and the IOW see Campbell, Structure of Slavery; Reid, Slavery, Bondage and Dependency; Watson, Asian and African Systems of Slavery; Prakash, Bonded Histories; Klein, Breaking the Chains; Clarence-Smith, Economics of the Indian Ocean Slave Trade. 35 Lasker, Human Bondage in Southeast Asia, 16–17. 36 Law, “Slavery and Debt Bondage in Sarawak,” 291. 37 Manning, Slavery and African Life, 114. 38 Sheriff, “Slave trade and its fallout”; Clarence-Smith, “Islam and the abolition of the slave trade.” 39 Miers, “Slavery and the Slave Trade.” 40 Campbell, “Introduction: abolition and its aftermath in the Indian Ocean world.” 94 gwyn campbell shipped to Zanzibar, Pemba, Somalia, Madagascar, the Mascarenes, and some to Cape Town. They were also exported to India and the Ameri- cas. Malagasy slaves were sent in considerable numbers to Réunion and Mauritius. Indian slaves were shipped to Indonesia, Mauritius, Cape Town and the Middle East. Most slaves to the Middle East initially originated from the Caucasus, Eastern Europe and Africa but were joined in the nineteenth and early twentieth century by slaves from the Makran coast of Iran, some from Western India and a few from Indonesia and China.41 Indonesians were dispatched mainly to Southeast Asia (but also to markets such as Mauritius), Indochinese and Koreans to China, and Chinese to Singapore and San Francisco. In all of these trades, sources, markets, routes, and slave functions, varied considerably.42 It is currently impossible to estimate with any precision the number of slaves traded in the IOW given the duration of the slave trade there, and the limited nature of extant records. Even in European-dominated enclaves, statistics are at the very best, patchy. Research is only just beginning into slave-ship journals and records in the region,43 but, in contrast to the Atlantic system, IOW slaves rarely constituted a specialist cargo. This was especially the case on indigenous ships, but also on many Euro- pean-captained slaving vessels.44 Thus scholars’ attempts to divine quantities of slaves traded in the IOW are at best “guesstimates” that inevitably have considerable range. Moreover, they rarely take into account variance in the number of slaves embarked and those reaching their destination: Mortality rates aboard ship decreased significantly over the centuries as slave traders paid increasing attention to improving slave hygiene, nutri- tion and security aboard ship,45 so that historians have emphasized that by the nineteenth century distance was the main factor influencing mortality rates: Thus there appears to be consensus that the slave mortality

41 Miers, “Slavery and the slave trade in Saudi Arabia”; Sheriff, “slave trade and its fallout”; Klein, “emancipation of slaves in the Indian Ocean,” 198–218. 42 Allen, “Mascarene Slave-Trade”; Warren, “Structure of Slavery in the Sulu Zone”; Delaye, “Slavery and Colonial Representations”; Alpers, “Flight to Freedom”; Schotten- hammer, “Slaves and Forms of Slavery in Late Imperial China”; Machado, “Forgotten Cor- ner of the Indian Ocean”; Boomgaard, “Human Capital”; Worden, “Indian Ocean slavery and its demise,” 29–49; Campbell, “Unfree labour,” 66–82; Sheriff, “The slave trade and its fallout.” 43 See e.g., Ross, “Dutch on the Swahili Coast”; Allen, “Constant Demand of the French”; Westra and Armstrong, Slave Trade with Madagascar. 44 Campbell, Economic History of Imperial Madagascar, 228; Allen, “Constant Demand,” 64–5. 45 Westra and Armstrong, Slave Trade with Madagascar, 13, 15, 25, 33, 35. children and forced labour in the indian ocean world 95 rate aboard ships to the Mascarenes from other regions was 20 to 25 per cent (from India), 25 to 30 per cent (West Africa) and, in the late eighteenth century, 12 per cent (Madagascar) and 21 per cent (East Africa).46 However, this takes little account of disease, which played a major role in the nineteenth century, or mounting pressure by anti-slave trade patrols in the closing decades of the nineteenth century which forced the cap- tains of slave ships to take more risks, overcrowd their ships, and jettison slaves should their ships be threatened with capture. Because of the conventional focus on Black slaves, East Africa and Mad- agascar are the only IOW regions with concentrated slave export guesstimates―respectively ranging from between 800,000 to over two million,47 and between about 72,000 to double that figure.48 In total terms, factoring in other regions and non-African slaves, it is probable that the cumulative number of slaves traded across the maritime space of the IOW over the centuries well exceeded the 10 to 12 million landed in the Americas. However, it is highly probable that the greatest IOW slave traffic was over- land, notably within Africa, Hindu India and the Confucian Far East: in 1841, Bartle Frere estimated that there existed eight million to nine million indigenous slaves in India alone49―double the number of black slaves in the United States in 1865. This again speaks to the major difficulty faced by Europeans, but largely ignored in the literature―the tying up of the IOW workforce in indigenous systems of bonded labour.50 Again, while adult males were the most valued in some markets, such as Cape Town in the late eighteenth century,51 and the plantations on Zanzibar and the French Mascarenes,52 children continued to form a

46 Allen, “Mascarene Slave-Trade,” 39. 47 Austen, “The 19th Century Islamic Slave Trade from East Africa”; Sheriff, Slaves, Spices & Ivory, 226; Campbell, Economic History of Imperial Madagascar, 238; Alpers, Ivory and Slaves, 15 1, 185–7; Allen, “Mascarene Slave-Trade,” 38–9; Capela and Medeiros, O Tráfico de Escravos, 24–5, 41; Liesegang, “First Look at the Import and Export Trade,” 463; Kjekshus, Ecology Control, 14–16. 48 Filliot, La traite des esclaves, 157–9; Campbell, Economic History of Imperial Madagas- car, 55–6, 238; Allen, “Mascarene Slave-Trade,” 35, 38. 49 Balfour, Cyclopædia of India, 674. 50 Watson, “Transactions in People,” 235. 51 In 1777 at Zanzibar, a Dutch slaving ship from Cape Town took aboard 328 slaves, 70 percent of whom were males, mostly adults, and 27 percent of whom were adult females― Sheriff, “Localisation and Social Composition of the east African Slave Trade,” 138. See also Ross, “The Last Years of the Slave Trade.” 52 This is reflected in the values placed on slaves in Mozambique coast markets where slaves were split into three categories: the pesca, comprising healthy young male and female adults aged between 18 and 25 years, followed by the pote d’agua—healthy slaves 96 gwyn campbell majority of slaves exported from East Africa. For example, in 1856, of the 69 African slaves freed by the British from ships travelling to the Persian Gulf, 29 percent were aged less than 10 years old, and 36 percent aged between 10 and 19 years.53 None of the males was older than 25 years (38 percent were aged under 10 years); whereas the average age for females was 19 years—17 percent being over the age of 25.54 In Southeast Asia, children formed the major target of slave raiders and kidnappers. Thus Hugh Low commented of Sarawak in the early nineteenth century: The slaves (ulun-ulun) in Borneo are generally Dyaks and their descendants who have been captured by the rulers of the country . . . it was no uncommon thing for these tyrannical chiefs . . . to send up parties to the Dyaks to bring down all the young girls and boys they could catch. I have been told by Dyaks, and by dependants of the Sereib, that 300 girls and boys have frequently been brought down at one time; such of these as the chief selected as likely to suit his purposes were reserved, the remainder were sold to whoever would buy them, the chief taking a considerable share, though his servants who had kidnapped the poor children generally contrived to cheat him of a considerable portion of the produce.55 “Primitive” hill peoples were particular targets of kidnappers who seized and sold children, notably girls.56 Sometimes such kidnappings were so extensive―as by the Rawa of Sumatra on the Mantra of Malacca―that it forced the communities subject to the raids to abandon their ancestral lands.57 Children who entered slavery were seldom sent directly to their final market. For example, in interior East Africa, the passage to the coast of those younger slaves destined for coastal and export markets was frequently delayed, or made in stages, possibly in order that they learned to be doc- ile, but also to ensure that remained healthy and thus retained their value. During the often lengthy interval between enslavement and shipment,

aged between 14 and 18, and finally a group comprising less desirable slaves―Campbell, “Madagascar and Mozambique in the Slave Trade of the Western Indian Ocean 1800–1861,” 178. See also Géraud, “Les esclaves à l’épreuve de l’industrie,” 297–8. 53 Sheriff, “Localisation and Social Composition of the east African Slave Trade,” 139. 54 Ibid., 139. 55 Law, “Slavery and Debt Bondage in Sarawak,” 289. 56 See e.g. the enslavement of Negrito children by urbanized Christian Filipinos; and of hill girls captured by the Muslim Sulu―Lasker, Human Bondage in Southeast Asia, 37, 40–1. 57 Lasker, Human Bondage in Southeast Asia, 48. children and forced labour in the indian ocean world 97 they often changed hands a number of times, and were employed by their owners in a variety of ways―as domestic servants; tenders of animals, gardens, or even of children younger than themselves; and fetchers of fire- wood and water.58 Similarly, the children kidnapped and sold by the Red Karen in nineteenth-century Burma were frequently sold several times before finally being retained by a master.59 Child slaves also constituted a commodity money, “small capital suited for negotiating deals, settling debts, or establishing credit”60 in East Africa; and in Central Africa were commonly used in payment for ivory and transport.61 In much the same way, merchants visiting Mecca on the Haj, frequently took slaves who they used as a kind of “travellers cheque,” selling upon arrival in order to finance their return trip.62 The traffic in children reflected both supply and demand factors. In times of dearth, and notably famine, families throughout the IOW commonly sold their children in order to give them, and the remaining members of the household, a better chance of survival.63 However, slavers did not take very young children, who they considered a burden, likely to slow down a caravan and prone to high mortality.64 On the demand side, elite males purchased slaves, notably children and young women, chiefly as items of conspicuous consumption to reflect their power, prestige and wealth.65 Indeed, in some cases, the costs of maintaining a slave exceeded the benefits accruing from his/her services which, in exceptional circumstances, could lead to the financial ruin of the owner.66 Neverthe- less, a false dichotomy has often been assumed between “productive” and “unproductive” slave activities67 many of which had considerable economic significance. Children and wives acquired for kinship groups were both status symbols and important additions to the group’s productive

58 Morton, “Small Change,” 60–1. 59 Lasker, Human Bondage in Southeast Asia, 44. 60 Morton, “Small Change,” 55. 61 Renault, “Structures of the Slave Trade in Central Africa,” 155–6. 62 Ahmad, “Ethiopian Slave Exports at Matamma, Massawa and Tajura,” 99. 63 Lasker, Human Bondage in Southeast Asia, 53; Morton, “Small Change,” 65; Ferny- hough, “Slavery and the Slave Trade in Southern Ethiopia,” 107; Deutsch, “Notes on the Rise of Slavery,” 84. 64 Manning, Slavery and African Life, 114. 65 Ahmad, “Ethiopian Slave Exports,” 93; Boomgaard, “Human Capital”; Lasker, Human Bondage in Southeast Asia, 49–50; see also Patnaik and Dingwaney, Chains of Servitude,” 2–4, 26; Reid, Slavery, Bondage and Dependency, 13; Goody, “Slavery in Time and Space,” 36–7; Klein, “Introduction,” 8–13. 66 Boomgaard, “Human Capital.” 67 See Kopytoff and Miers, “African ‘Slavery,’ ” 55–7, 64–6. 98 gwyn campbell and reproductive capacity, while throughout the IOW slaves acquired by elite households were often encouraged to engage part-time in profit bearing activities.68 In the eighteenth and early nineteenth centuries, slaves comprised between 20 and 30 percent of the population of many IOW societies, rising to 50 percent and over in parts of Africa and in Indonesian ports.69 Boys were traditionally valued chiefly as household servants, guards, and soldiers. For example, one of the main destinies of Serb, Greek and Alba- nian Christian boys aged 10 to 18 years absorbed into the devşirme system that functioned in the Ottoman Empire until 1826 was the janissary corps of the army.70 Slave soldiers enforced law and order―a pre-requisite for economic growth―and safeguarded vital trade routes, supply centres and markets. Most were “remunerated” with battle spoils or expected to engage part-time in economic activities. Indeed, the capture and exchange of slaves (chiefly women and children) was a principal objective of most pre-colonial armies and navies. Slave armies thus often generated slaves, sometimes―as in the Sulu case―on their own initiative, and for their own as well as their owner’s material interests.71 Like slave girls, slave boys were also valued for sexual purposes by masters, but also―clandestinely―by the master’s wife.72 Sometimes, sexual exploitation occurred haphazardly, at other times sexual services formed an integral part of a boy slave’s functions. For example, in societies with an established tradition of homosexuality, as in parts of the Ottoman Empire, Pakistan, Muslim India, Afghanistan and Turkestan, there was considerable demand for slave boys (including eunuchs) used as sexual companions,73 prostitutes,74 and as bacchá―boys who, chosen for their height, size and beauty, were trained to sing and dance for elite male audiences amongst whom they were traded for sexual favours.75 In Persia,

68 See e.g. Reid, “Introduction,” 14. 69 Kim, “Nobi: A Korean System of Slavery”; Boomgaard, “Human Capital”; Kopytoff and Miers, “African ‘Slavery,’ ” 60–1; Reid, “Introduction,” 12, 29; Campbell, “Slavery and Fanompoana,” 474–5. 70 Yilmaz, “Becoming a Devşirme,” 121–8. 71 Warren, “Structure of Slavery in the Sulu Zone,” 111–28; Goody, “Slavery in Time and Space,” 26–7. 72 Morton, “Small Change,” 65. 73 Murray, “Homosexuality among Slave Elites in Ottoman Turkey.” 74 Neill, Origins and Role of Same-Sex Relations, 169. 75 Mathee, Pursuit Of Pleasure, 49, 285. children and forced labour in the indian ocean world 99 most slave boy prostitutes and dancers were aged between 10 and 16 years, and were of Circassian, Georgian and Armenian origin.76 The highest valued child slave was the eunuch, commonly sought by sovereigns throughout the IOW from the Ottoman Empire to China, notably as personal bodyguards, and keepers of the royal harem. There were different types of castration: In Korea, these were listed as “involuntary,” “accidental,” “congenital” and “self castration”―meaning “voluntary.”77 In the case of slaves, it was mostly “involuntary,” enforced on enslaved boys; but some submitted voluntarily in an attempt to achieve the higher status and corresponding privileges accorded to eunuchs.78 However, the risks were high: It has been estimated, despite the care taken because of the value of a eunuch, that in Korea between 20 to 50 percent 79 and in Africa up to 90 percent80 of boys died whilst undergoing surgical castration. Of those who survived, some of the “self-castrated” had only their testicles removed, which meant that for some, sexual intercourse remained possible―there are, for example, cases of love affairs developing between court eunuchs and ladies, and of highly-placed Korean eunuchs taking wives and concubines.81 The commercial boom from the late eighteenth century led to rising demand for male slaves for activities associated with the production, pack- ing and transport of export commodities, such as pearls, dates, cotton, wool and opium in the Persian Gulf,82 cloves, coconuts and grain on the Swahili coast,83 sugar on Mauritius, and hides in imperial Madagascar,84 and the return traffic comprising staples such as cloth, iron goods, firearms, gunpowder, and alcohol.85 Most of this demand was for adult slaves, but there was also significant demand for boys. Thus Richard Allen considers that even in the case of Mauritius, a plantation economy, the trade was as much in boys as in young adult males.86

76 There is also reference to Russians―Mathee, Pursuit of Pleasure, 169, incl. fn. 139, 170. 77 Kim, “The Third Gender. Palace Eunuchs,” 138. 78 Ibid., 138. 79 Ibid., 147. 80 Lovejoy, Transformations in Slavery (Cambridge: Cambridge University Press, 2000), 35. 81 Kim, “Third gender,” 130. 82 Ricks, “Slaves and Slave Traders in the Persian Gulf,” 60, 65; Clarence-Smith, Islam and the Abolition of Slavery, 10. 83 Cooper, Plantation Slavery, 3. 84 Campbell, Economic History of Imperial Madagascar, 166–88. 85 See e.g., for Madagascar―ibid., 249–59. 86 Allen, “Children and European Slave Trading,” 36. 100 gwyn campbell

Nevertheless, most of the children traded as slaves were girls who consistently fetched higher prices than non-eunuch boy slaves. For example, in the mid nineteenth century, the average price for a male slave at Gondar, in Ethiopia, was from 15 to 18 thalers, and for females from 20 to 25 thalers (the price for an exceptionally beautiful girl could rise to 80 thalers); while the price of a eunuch was 80 to 120 thalers.87 The price paid on the coast was invariably higher than at source—15 times more in the case of female slaves; and on the Arabian coast doubled again (while prices paid for beautiful Ethiopian girls rose even more dramatically in Persian Gulf markets).88 Thus the kidnapping of young girls was common across Ethiopia―although traffickers often used captive girls as concubines during the trip to market, despite the fact that the loss of virginity lowered a girl’s sale price.89 Traditionally, at destination, some girl slaves were employed as water carriers, and in agriculture, textile production and mining,90 but most were absorbed by wealthy households that employed them predominantly in personal domestic and sexual services (for masters and their wives)91 and entertainment.92 Female slaves who were secondary wives, concubines, entertainers and domestic servants of the wealthy, enjoyed a lifestyle and―as Anthony Reid argues for Southeast Asia―a respect, superior to that of female peasants.93 There are instances of concubines in the Middle East sending for family members to join them, albeit as non- slaves.94 Female slaves were also less likely to be sold.95

Indebtedness

Possibly the majority of people entering slavery in the IOW did so through indebteedness. Impoverishment frequently forced families into debt which was normally expressed in monetary terms, although it was often incurred in non-cash forms such as food or tools. Commonly, household heads attempted to meet debt payments through the sale or mortgage

87 Ahmad, “Ethiopian Slave Exports,” 96, 98. 88 Fernyhough, “Slavery and the Slave Trade in Southern Ethiopia,” 113. 89 Ibid., 107, 112. 90 Goody, “Slavery in Time and Space,” 21, 32. 91 See the extraordinary case cited by Anwar, “Attitudes towards Homosexuality,” 404. 92 See Miller, “A Theme in Variations,” 169–94. 93 Reid, “Introduction,” 25–6. 94 Miers, “Slavery and the slave trade in Saudi Arabia,” 120–36. 95 Reid, “Introduction,” 25–6. children and forced labour in the indian ocean world 101 of a household member, notably a girl, but if in dire straits, also boys and wives. This was a common pattern across the IOW, from China96 to East Africa.97 Moreover, in many IOW regions, enslavement was legally enforced for debtors and their relatives. In Imperial Madagascar, for instance, creditors could, through recourse to justice, enslave a debtor, his wife and children.98 This was also the customary practice in Thailand and Malaya.99 In addition, the punishment for certain crimes was exacted in fines, which often led to indebtedness and subsequent enslavement.100 If the debt was paid off, an enslaved debtor could regain non-slave status. Here, slavery needs to be distinguished from debt bondage with which however, it could overlap. Enslavement for indebtedness was involuntary, whereas most people entered debt bondage voluntarily, as a credit secur- ing strategy. Mortgaging a child, or wife, to raise a loan was common practice in the IOW from early times.101 Certainly by the nineteenth century, debt bondage embraced a vast range of people in the IOW, from farmers mortgaging future harvests and potential grooms borrowing a bride price, to small traders living off credit from larger merchants, the ubiquitous rural gambler of Southeast and East Asia and opium addicts in nineteenth-century China.102 During catastrophes, people often entered debt bondage or slavery in return for subsistence as a survival strategy, either voluntarily, as was the case of many dvija caste members in India, or pro- pelled by their kin group.103 They did so whether they lived in regions of relatively low population density, such as Cambodia, Laos and Indonesia, the Middle East and Africa, or of relatively high population density, such

96 Hirata, “Free, Indentured, Enslaved,” 4; Dull. Ed. Han Social Structure, 110. 97 Morton, “Small Change,” 59. 98 Campbell, Economic History of Imperial Madagascar, 295–6. 99 Lasker, Human Bondage in Southeast Asia, 147, 150. 100 Reid, “Introduction,” 10. 101 Thus the Code of Hammurabi (1795–1750 BC) in Ancient Mesopotamia stated: ‘If an obligation is outstanding against a man and he sells or gives into debt-service his wife, his son or his daughter, they shall perform service in the house of their buyer or of the one who holds them in debt-service for three years; their release shall be secured in the fourth year”―cited in Harris, “Did Solon Abolish Debt-Bondage?” 418. Different provisions applied to slaves: ‘If he should give a male or female slave into debt-service, the merchant may extend the term (beyond three years), he may sell him; there are no grounds for a claim.’—ibid. 102 Boomgaard, “Human Capital”; Delaye, “Slavery and Colonial Representations”; Schottenhammer, “Slaves and Forms of Slavery in Late Imperial China”; see also Watson, “Transactions in People,” 228–36. 103 Klein, “Introduction,” 11; Patnaik and Dingwaney, “Chains of Servitude,” 25–6. 102 gwyn campbell as Bengal, Vietnam, Korea and South China.104 Moreover, most victims appear to have been pushed into debt bondage as children―in early twentieth century Thailand they were rarely aged over ten years.105 Those subject to debt bondage could outnumber slaves. For example, they were possibly the most numerous social category in Majapahit, in Java, while in central Thailand in the eighteenth and nineteenth centuries, they formed up to 50 percent of the total population. Again, in Burma in the mid nineteenth-century, when debt bondage was far more widespread than hereditary slavery, the male head of an impoverished household frequently sold his wife and children to meet tax impositions.106 The servitude to which those in debt bondage were subject was generally taken as paying off interest on the loan they had contracted, to which was added the cost of lodging, feeding and clothing the debtor. Consequently the debt in most cases increased and servitude could become permanent, or even hereditary―at which point there was little to distinguish debt bondage from slavery.107 However, in nineteenth-century Thailand, if the interest owed by a debt slave grew to a level exceeding the original loan, the master deemed the slave a bad investment and customarily sold the slave―albeit on unfavourable terms.108 Pawnship, common in Eastern Africa, was a variant of debt bondage wherein a pawn, frequently a girl, was given as security for a loan. The creditor did not have the right to sell the pawn, whose labour he could use until the moment that the family paid off the debt. However, in cases where an inordinate amount of time passed with little or no sign or repay- ment, the creditor’s ownership of the pawn became established.109 In some groups, the sale of children became a traditional, accepted form of generating revenue. This was the case amongst many kasama (tenant farmers) in nineteenth century Philippines of whom John Fore- man (1899) commented: Under the pretext of guaranteeing a loan, parents readily sell their children (male or female) into bondage; the child is handed over to work until the loan is repaid, but as the day of restitution of the advance never arrives,

104 Campbell, “Introduction,” 15. 105 Lasker, Human Bondage in Southeast Asia, 151. 106 Ibid., 138. 107 Kim, “Nobi”; Reid, “Introduction,” 12; for debt bondage from another angle, see Miller, “A Theme in Variations.” 108 Lasker, Human Bondage in Southeast Asia, 151. 109 Lovejoy, Transformations in Slavery (2000), 13–14; Deutsch, “Notes on the Rise of Slavery & Social Change in Unyamwezi,” 89. children and forced labour in the indian ocean world 103

neither does the liberty of the youthful victim. Among themselves it is the law, and is still a practiced custom, for the debts of the parents to pass on to the children, and . . . debts are never repudiated by them.110 This appears to have also been the case amongst the Aetas, a non-Christian Negrito group, traditionally hunter-gatherers, from the isolated mountains of Luzon, who in the nineteenth century and into the early twentieth century were “adopted” by Christian Filipinos in return for a small payment to their parents or guardians. Filipino masters justified such purchases as being a means of Christianizing pagans.111 This rationale was similar to that expressed by Christian missionaries in Africa who redeemed slaves,112 and stemmed from the widespread belief in Enlightenment Europe that children from “primitive” pagan societies required “domestication,” and that slavery was the protected status best suited to that end. From the Enlightenment era, religious leaders propounded parallel concerns for the moral progress of “children” that was translated into the provision of Christian ministry to “childlike” slaves. Slave-owners needed to discipline their slave wards into accepting their “public” duties, but also had a responsibility to save their souls and teach them the moral virtues of adulthood in the personal realm of “marriage” and “family.” By this line of reasoning, bondage could be viewed as a state wherein undeveloped humans might receive the physical and moral discipline and training necessary for the attainment of civilised virtues―which alone could justify “freedom” in the sense of political participation or economic autonomy.113 European powers that established colonies in the IOW facilitated a growth in indebtedness, through imposing monetary taxes, promoting commercialisation, and enforcing credit contracts. At the same time, colonial authorities both maintained tight budgetary regimes that avoided funding public welfare programmes, and distinguished debt bondspeople from “true” slaves, whose condition they attributed solely to violent capture. As a result, debt bondage and enslavement through debt expanded considerably across the IOW in the 1800s.114 India is a prime example. In a century characterised by rising taxation and years of famine, “freedom” for members of the former slave outcastes, who had deliberately been kept

110 John Foreman (1899), cited in Lasker, Human Bondage in Southeast Asia, 131. 111 Lasker, Human Bondage in Southeast Asia, 38. 112 Clarence-Smith, “Redemption of Child Slaves,” 252–79. 113 Campbell, “Children and slavery in the new world,” 261–85; Ariès, Centuries of Child- hood, 128–33. 114 Reid, “Introduction,” 11. 104 gwyn campbell destitute and debarred from land ownership, translated into the liberty to starve. Some adopted sharecropping, but with two-thirds of the crop paid to the landlord, the risk of failure was high. In order to survive, many entered debt bondage that was from 1859 reinforced by the Breach of Con- tract Act.115 In some areas of India, members of the most depressed castes formed the overwhelming bulk of those in debt bondage. The situation closely resembled slavery in that bondage could be inherited, and the vast majority of bonded people had their geographical mobility restricted.116 Nevertheless, colonial powers across the IOW considered debt bondage to be a benign form of private welfare, and generally condoned its continuation well into the twentieth century.117 In some areas, such as Thailand, Burma and Indochina, this encouraged a revival of covert slave raiding.118 In Africa where debt bondage was represented by the “pawnship” of a person, usually a young girl, to a creditor in return for a loan, the system weakened only during the post Second World War boom.119 In all cases, the debtor had a clear market value, expressed in more monetized Asian economies in terms of cash, and in less monetized economies, as in most of Africa, in terms of “human” capital.

Apprenticeship

In the hope of avoiding manpower shortages on plantations after abolition, the British introduced an apprenticeship scheme whereby emancipated slaves would be forced to continue work for their ex-masters for a limited period. For example, when on 1 February 1835, slaves were accorded non- slave status on Mauritius, only those aged under six years were granted total freedom. The 66,613 former slaves above that age were obliged to serve their former masters as “apprentices”—for four years for non- praedial and six years for praedial former slaves (in the event non-praedial apprentices were liberated on 1 February 1839, and praedial apprentices

115 Patnaik and Dingwaney, Chains of Servitude,” 29–30. 116 Ibid., 30–1. 117 Chatterjee, “Abolition by Denial”; Klein, “emancipation of slaves in the Indian Ocean”; Salman, “Meaning of Slavery.” 118 Delaye, “Slavery and Colonial Representations”; Turton, “Violent Capture of People,” 69–82; Klein, “Introduction: Modern European Expansion and Traditional Servitude in Africa and Asia,” 23. 119 Lovejoy, Transformations in Slavery (2000), 13–14; Miers and Klein, Slavery in Colonial Africa, 12. children and forced labour in the indian ocean world 105 two months later).120 Apprentices were obliged to work 45 hours a week (Sundays and national holidays excepted), and could, in addition, take on paid work for a maximum of three additional hours a day. They could not be sold individually, but could be transferred along with the property on which they worked.121 Many female child apprentices—some as young as eleven—were hired out by their “owners” as prostitutes in Port Louis.122 Apprentices did all in their power to escape the system. Many apprenticed mothers resorted to theft and prostitution to avoid the necessity of contracting their children as apprentices,123 while many others ran away: an average of 4,300 refugee apprentices a year being captured by police.124

Indentureship

Another temporary solution to labour shortages was the placement of “prize negroes,” “liberated” from slave ships, under contract to Euro- pean settlers; those granted to Mauritian planters were “leased” to the government for four days a year to perform public works.125 However, both the Prize Negro and Apprenticeship systems were overshadowed by the indenture system. Indentured labour was an old institution used in European settlements, either formalised by contracts, or ad hoc, as with Khoi and San boys captured by Dutch farmers in the Cape interior during the eighteenth century. The captives were forced to work until the age of 25, by which time they were often married, with sons who were subject to similar obligations. Many parents refused to abandon their children, and so remained tied to the farm for life. Following abolition of the slave trade in 1807, the Caledon Code of 1809 formalised indenture in the Cape through a “pass” system, which restricted San and Khoi to farms in a system of “virtual slavery.” Thus, Nigel Worden argues, the first effective formal abolition in the Cape was not official Abolition in 1834 but the 1828 repeal of the Caledon Code.126 Upon abolition ex-slaves in the Cape (mostly Khoikhoi and San)127 and Mauritius (chiefly African, Malagasy,

120 Peerthum, “Le système d’apprentissage à L’île Maurice 1835–1839,” 285, 293. 121 Ibid., 285–6. 122 Ibid., 290. 123 Ibid., 289–90. 124 Ibid., 291. 125 Anderson, “Bel Ombre Rebellion.” 126 Worden, “Indian Ocean slavery and its demise.” 127 Lovejoy, Transformations in Slavery (2000), 239. 106 gwyn campbell

Malay, Indian and Creole)128 who received no financial assistance, were declared “apprentices” and obliged to continue working for a fixed period for their old masters. In the Cape, the system ended in 1838, but on Mauri- tius it continued until mid-century. Many apprentices fell into debt bondage to their old employer or chose to continue working for him in order to remain with their children.129 A new indenture system arose due to the failure of abolitionist measures to transform ex-slaves into pliant wage labourers. Generally involving five-year contracts, it channeled manpower resources to enterprises both within and external to the IOW―such as the sugar plantations of the Fiji Islands. The geographical range of indentureship within the IOW was enormous, running from the rubber plantations of Sumatra and Malaya from the 1860s,130 the tin mines and rubber plantations of Malaya, 131 the tobacco and rubber plantations of the Dutch East Indies,132 to the Assam tea gardens from 1859–77,133 to Burma (Myanmar), Ceylon (Sri Lanka), Queensland, Australia, where over 60,000 Pacific Islanders served as contract labour on the sugar plantations from 1863 to 1904;134 New Guinea;135 and the South African sugar plantations from 1860 and goldfields from 1903–8.136 The nominal duration of indenture contracts varied. At the outset in Natal, contracts were of 10 years duration; 137 the Assam tea gardens (from 1859) of between 3 and 5 years.138 In Queensland in the 1890s, contracts were mostly of less than a year’s duration,139 and by 1904 in New Guinea, contracts were of only 12 months.140 The historiography has concentrated on recruitment for European enterprises, particularly of Indian indentured labourers, some one million of whom were employed in India by the close of the nineteenth century; two million others were shipped to overseas plantations between 1834 and

128 Allen, Slaves, Freedmen, and Indentured Laborers, 43. 129 Worden, “Indian Ocean slavery and its demise”; see also Carter, Servants, Sidars and Settlers. 130 Gordon, “Contract Labour in Rubber Plantations.” 131 Jain, “Tamilian Labour and Malayan Plantations,” 2364. 132 Das Gupta, “Structure of the Labour Market in Colonial India,” 1801. 133 Ibid., 1784–6. 134 Fitzpatrick, “Indentured Labour in Australia.” 135 Shlomowitz, “Mortality and Indentured Labour.” 136 Richardson, “Recruiting of Chinese Indentured Labour.” 137 Brain, “Indentured Indians.” 138 Das Gupta, “Structure of the Labour Market,” 1784. 139 Shlomowitz, “Markets for Indentured and Time-Expired Melanesian Labour,” 88. 140 Shlomowitz, “Mortality and Indentured Labour,” 73. children and forced labour in the indian ocean world 107

1920.141 However, labour from many other sources were involved, including the Pacific Islands and China. Indeed, Chinese sources suggest that Chi- nese emigrant indentured labour was far greater than previously thought. Some 7.7 million people emigrated from South China in the period 1851 to 1901, of whom 86.8 percent travelled to Southeast Asia, and most of this serviced indigenous IOW rather than European demands for labour.142 Recruitment, transport, and living and working conditions of indentured labourers were often similar to those of slaves.143 However, recruitment was overwhelmingly of young adult males. Thus in Queensland from 1884–1901, females comprised only 200 of the 3,867 Pacific Island indentured workforce;144 while the last shipment of Chinese indentured labour for the Witwatersrand goldfields in November 1906 comprise 2,129 adult males, three women, and four children.145 Although regulations forbade the recruitment of children, it is clear that both recruitment agents and the recruits commonly practiced deception as to age, so that probably a significant proportion of the predominantly young males who entered indentureship were under the age of eighteen.146 The same applied to the females who in some instances, as on Mauritius,147 formed a more significant percentage of indentured recruits. In the case of the Assam tea plantations from the late 1850s, professional recruiters (arkattis) working for licensed contractors based in Calcutta, recruited and abducted many women and children from rural villages to serve as indentured workers.148 Moreover, the ranks of the children were added to between the time the destination was reached and the end of the period of indenture. As children of indentured parents grew up, they too came to form part of the Indian indentured labour force. Thus on Reynolds Broth- ers’ sugar plantations, alongside Tongaat the largest in Natal in the 1890s, a

141 Chandra, “Colonial Legacy,” 11; see also Allen, “Mascarene Slave-Trade”; Patnaik and Dingwaney, “Chains of Servitude,” 5–7, 27; Klein, “Introduction: Modern European Expan- sion and Traditional Servitude in Africa and Asia,” 20–1. 142 My thanks to Jesse Sayles for this calculation based on information from McKeown, “Global Chinese Migration, 1850–1940.” For the traditional view, see e.g. Blue, “Chinese Emigration and the Deck Passenger Trade”; Campbell, Chinese Coolie Emigration. 143 See e.g. on Indian indentured labour, Pineo, Lured Away; Mookherrji, Indenture System in Mauritius; Tinker, New System of Slavery; Carter, Servants, Sirdars and Settlers; Gerbeau, “Engagees and coolies on Réunion Island”; Carter and Gerbeau, “Covert Slaves”; Dasgupta, “Plantation labour.” 144 Shlomowitz, “Markets for Indentured and Time-Expired Melanesian Labour,” 88. 145 Richardson, “Recruiting of Chinese Indentured Labour,” 86. 146 Northrup, Indentured Labor, 57–8, 70. 147 Allen, Slaves, Freedmen, and Indentured Laborers, 58. 148 Das Gupta, “Structure of the Labour Market,” 1785. 108 gwyn campbell highly exploitative labour regime was established in which work lasted 17 to 18 hours a day for which adult males were weekly given extra one-and- a-half pounds of rice, but women and children only if they worked―thus contributing to high infant mortality rates.149 In Natal, upon the completion of their indentures, Indian workers were encouraged to re-indenture to the same or another employer, Indeed, in 1895, a law was passed, effective from 1901, imposing a tax of £3 on those time-expired workers who did not sign up for a further period of at least two years; and in 1903 this measure was extended to cover the children (boys over 16 and girls over 13 years of age) of formerly indentured men.150

Forced Labour Regimes

Penal Labour Europeans also adopted penal labour. Tens of thousands of indigenous convicts, predominantly male, were shipped to labour in European settlements, as for example to most British colonies,151 from Goa to Mozam- bique, and from Batavia to the Cape.152 This practice increased in the transition years of abolition as slave labour progressively dwindled, penal labour sometimes being leased out for use by European individuals, as on Mauritius until 1851.153 Penal labour was also widely used by indigenous authorities for the harshest types of work. For instance, the Merina regime in Madagascar used convicts in road construction, mining and foundries.154 However, it is unknown whether, as in indentured labour schemes, significant numbers of children were involved.

Corvée Labour Western abolitionist pressure prompted indigenous regimes as far apart as the Ottoman Empire, Thailand, Zanzibar, Imerina and Ethiopia to make at least official proclamations against the slave trade and slavery.155

149 Halpern, “Solving the ‘Labour Problem’,” 33. 150 Brain, “Indentured Indians.” 151 Anderson, Legible Bodies. 152 Worden, “Indian Ocean slavery and its demise”; Shirodkar, “India and Mozam- bique,” 51. 153 Anderson, “Bel Ombre Rebellion.” 154 Campbell, “Unfree labour.” 155 Delaye, “Slavery and Colonial Representations”; Klein, “Introduction,” 25; Miers, “Britain and the Suppression of Slavery in Ethiopia.” children and forced labour in the indian ocean world 109

In Thailand such measures largely ended slavery by 1900.156 In general, however, indigenous anti-slavery measures were rarely effective, or were manipulated to divert slaves from private ownership into state labour pools, as occurred in late eighteenth-century Korea,157 and in nineteenth- century Thailand158 and Madagascar.159 In these countries and in Java160 most productive labour was performed not by slave but by the nominally “free” population subject to state corvées, while in many states, as in Egypt under Muhammad Ali,161 and in Imperial Madagascar,162 indigenous regimes used the forced labour of nominally “free” subjects to drive programmes of economic modernization. Often inextricably linked with such programmes were attempts, from Korea, Thailand and Burma to Iran, Zanzibar and Imerina, to create and economically exploit “secondary empires.” Critical to their success were armies comprising mostly forced labour. Indeed, in order to escape state corvée, which in Korea, Burma and Thailand often claimed up to 50 per cent and in Imperial Madagascar up to 100 per cent of the labour of “free” draftees, some slaves rejected opportunities to gain “free” status, while some non-slaves voluntarily entered slavery.163 In some cases, a significant number of children were drafted into forced labour units. Because of their small size, children were particularly valued in mines. Thus bonded children were used in Chinese mines from early times until at least the mid-twentieth century;164 and in the mica mines of India from the late nineteenth century.165 Again, with the opening up of goldfields (mines and alluvial) in imperial Madagascar from 1883, Merina authorities summoned child labour from mission chapels and schools― where attendance was made compulsory and registers formed the basis for fanompoana (unremunerated forced labour) quotas. Missionaries and

156 Klein, “emancipation of slaves in the Indian Ocean.” 157 Kim, “Nobi.” 158 Terwiel, “Bondage and Slavery”; Turton, “Thai institutions of slavery”; Feeny, “Decline of Property Rights”; Feeny, “Demise of Corvée.” 159 Campbell, Economic History of Imperial Madagascar, 112–19, 131, 213–8. 160 Reid, “Decline of Slavery in Nineteenth-Century Indonesia”; Boomgaard, “Human Capital.” 161 Toledano, State and society, 6–7; Batou, “Attitudes of State and Society towards Industrialization in the Nineteenth-Century Third World,” 14; Fahmy, All the Pasha’s men, 10, 11, 86–93, 96. 162 Campbell, “Slavery and Fanompoana.” 163 Campbell, “Unfree labour”; Kim, “Nobi”; Reid, “Introduction,” 18–19. 164 Lasker, Human Bondage in Southeast Asia, 337 fn. 19. 165 Ibid., 357 fn. 5. 110 gwyn campbell other state-church agents played their role in drawing up the quotas and, increasingly, in overseeing the fanompoana units.166 For example, as early as 1887, Norwegian Missionary Society churches in the northern Betsileo regions of Mananadona and Fisakana were interrupted by the recruitment of 400 gold fanompoana workers comprising both sexes. Child labour in the gold fields became so widespread that in December 1888 the Mada- gascar Times declared: “by forced labor, into which even evangelists and school children are pressed, the Prime Minister is washing for gold on his own account. It is said that more than a thousand laborers are daily employed on this work.”167 Child exploitation in the gold fields intensified from the late 1880s as the threat of French intervention grew. In general, the predominantly Protestant Merina officials obliged the pupils of Roman Catholic schools to undertake the heaviest gold fanompoana, but in 1889 the same London Missionary Society (LMS) schools in the Ambo- hibeloma district, where military drill was instituted by the local missionary, were being called on to supply “gold” labour, and that same year the entire “free” population of Ambositra region, stretching from Behenjy to Ambositra and excepting only Antankaratra, were summoned.168 Resistance to fanompoana was met by force. For example, in 1889 members of the church and school at Ambanimaso were seized and militar- ily escorted to the goldfields.169 From 1890, all Malagasy entering Indian or European shops in Mahajanga were searched lest they be carriers of stolen gold, and it was commented of the neighbouring “gold” town of Maevatanana: People working steal the gold, sell it, and, if caught, their heads are cut off and stuck on poles. It is reported that there are dozens of heads of gold- stealers and so-called robbers stuck up like this at Maevatanarivo [Maeva- tanana] . . . Lately I have heard that a young girl of twelve or thirteen years, along with a young man, were caught with a small quantity of gold on them at Maeratanarivo [Maevatanana] and their heads were sawn off by a spear- head, taking hours over it.170 Finally, colonial regimes, burdened by metropolitan exigencies that insisted on self sufficiency, and local realities wherein there were shortages of labour, also fell back on forced labour regimes. This was, for exam-

166 Campbell, “Gold Mining,” 113. 167 The Madagascar Times (8 December 1888). 168 Campbell, “Gold Mining,” 113–4. 169 Ibid., 116. 170 Anon, “Gold in Madagascar.” children and forced labour in the indian ocean world 111 ple, evident from at least the 1880s in India where European collieries established zamindary rights over villages, and thus over the forced labour of villagers.171 Again, the French in Madagascar after 1895 quickly resorted to a forced labour regime that resembled, and was justified in reference to, the pre-colonial fanompoana practiced by the Merina.172

Sexual Slavery

In the nineteenth century, demand grew for girls and young women due to the rise of modern form of sexual slavery in the IOW. This, in turn, was due to rising gender imbalance in key locations as a result of the commercial boom associated with the international economy, and military action associated mainly with European imperialism (indigenous IOW army camp followers included large numbers of wives and concubines). This was reflected, for example, in the increase in the number of traditional female dancer-singers in Egypt turning also to prostitution by the mid-nineteenth century.173 On the other side of the Indian Ocean, the migration of millions of Chinese male labourers to IOW centres such as Singapore led to a huge demand for females for sexual purposes that was met by a traffic in mostly involuntary prostitutes, many of whom were girls engaged through deceit, or sold by their parents.174 A variant system was that of the mui tsai, in which in exchange for payment a girl was transferred from one family to another, as an adoptee, nominally to become a domestic servant. The mui tsai system, which continued well into the twentieth century, involved very young girls―in Hong Kong in 1921, 69 per cent of mui tsai were under the age of 14175―many of whom were by the late nineteenth century used as prostitutes by their new owners.176

171 Das Gupta, “Structure of the Labour Market,” 1788. 172 Jennings, “Forced labour in Madagascar.” 173 Nieuwkerk, ‘A Trade like Any Other,’ 34–5. 174 Jaschok and Miers, “Women in the Chinese Patriarchal System,” 19–20; Warren, “Chinese Prostitution in Singapore.” Between 1860 and 1880, most girls recruited for prostitution in America varied were aged between 16 and 25 years old: in 1870, almost 46 percent were aged under 20. Many of the female children of such prostitutes were similarly raised to be prostitutes by the brothel owners, raising the number of child prostitutes: of the total number of Chinese prostitutes in San Francisco between 1860 and 1880, 5 percent were aged 15 and under―Hirata, “Free, Indentured, Enslaved,” 21–2. 175 Poon, “The Well-Being of Purchased Female Domestic Servants (Mui Tsai),” 153. 176 Lasker, Human Bondage in Southeast Asia, 53–5. 112 gwyn campbell

Large concentrations of soldiers led to a similar demand for females to provide sexual services,177 that in European colonies was often facilitated by colonial officials who feared unrest should the sexual needs of soldiers not be met. Indrani Chatterjee notes that British garrisons in India openly employed “slave” prostitutes who in the 1860s were officially taxed―a measure which, alongside fines, only served to increase the debt bondage that more often than not had “enslaved” such girls.178 Again, in Natal in the late nineteenth-century some white planters controlled male workers’ access to women―sometimes their wives and at other times prostitutes― to reward or punish indentured Indian males.179

Conclusion

There is growing consensus that, unlike the trans-Atlantic slave trade, most victims of human trafficking in the IOW were children. The traditional systems of bondage remained vital during the nineteenth-century when the growth of a truly international economy created an unprecedented commercial boom in the IOW that established a large and growing demand for menial labour. However, there was difficulty meeting that demand, for two main reasons: comparatively low and uneven demographic growth, and the tying up of the potential workforce in largely indigenous systems of bonded labour. Abolitionist pressure in the IOW was muted as the focus of abolitionists was on the Atlantic slave system, and European officials in the IOW were reluctant to attack a ubiquitous system that most regional elites held dear. They therefore perpetuated some traditional systems of bondage, and introduced new ones. It is clear that, while the historiography has concentrated on the role of adults, children continued to play a highly significant part in these new forms of bondage.

177 See e.g. the impact of French troops in the 1798–1801 French occupation of Egypt― van Nieuwkerk, ‘A Trade like Any Other,’ 30–1. 178 Chatterjee, “Abolition by denial,” 150–68; see also Patnaik and Dingwaney, “Chains of Servitude,” 32. 179 Halpern, “Solving the ‘Labour Problem’,” 34. Part two

Dependance and Servitude at work: Local Customs and Global Dynamics

Factors that Shaped the Organization of Labor and the Labor Market in Tokugawa Japan: Kyoto and Central Japan

Mary Louise Nagata

Introduction

The organization of labor and the labor market in Tokugawa Japan, 1600– 1868, was shaped in major ways by the political policies and laws designed to establish and maintain political stability and social order, consequently changing the social, political and economic framework of society. One consequence of these policies was urbanization and the expansion of urban markets with a population dependent upon the market to supply their daily needs. At the same time, the land under cultivation was also expanding, demanding expansion of the agricultural labor force. Labor migration from village to town and the expansion of the urban labor market was one result. Rural handicraft production for urban and other markets was another. Labor and the labor market in Tokugawa Japan was largely shaped by these competing needs. The first part of this article discusses these constraints and how they shaped labor in Tokugawa Japan. The Japanese family and demographic system also interacted with these needs to shape and be shaped by labor and the labor market. Japa- nese family businesses tended to follow the stem, branch and sub branch framework of the stem family which could link villages and cities in a complex business network. Japanese family practice allowed for the establishment of non-kin as well as kin branches and used adoption, marriage, and divorce to recruit individuals into or expel them from non-kin, distant kin or close kin relations. These practices also provided important opportunities for the ambitious young person to prosper through his or her ability and work. The second section discusses various labor relations and the organization of labor in the family business as well as the role of the family system that shaped and was shaped by the labor market. Then the focus will narrow to the traditional capital city Kyoto and the organization of labor and the urban labor market. Kyoto suffered from a severe economic crisis partly due to the opening of treaty ports to international trade in 1857 with sharp price inflation in grain prices and raw 116 mary louise nagata silk 1858–1863. The political crisis that ultimately overthrew the Tokugawa regime also included random murders and violence in Kyoto and the city was under martial law 1863–8. The final part of this article uses population registers compiled by Kyoto neighborhoods to investigate the effect of these political and economic crises upon the Kyoto labor market. The results show that people reacted to the economic crisis by leaving the city if they could or by not coming to the city in the first place. However, after these initial actions, once the political crisis became severe, residents stayed in place and quietly continued their lives seeming almost to ignore the crisis. The final sections will address this analysis before the concluding discussion.

The Pax Tokugawa and the Organization of Labor

The Tokugawa regime gained official recognition from the Japanese imperial court in 1603 culminating a process of reunification begun in the middle of the sixteenth century by the warlord Oda Nobunaga, continued by Toyotomi Hideyoshi and then stabilized by Tokugawa Ieyasu. This process included several policies and practices, also begun under Oda Nobunaga around 1581 and more or less completed by Ieyasu’s grandson Tokugawa Iemitsu in 1635, that established the framework for the new political, social and economic order and shaped or constrained the organization of labor and the labor market. Although any major reorganization of society is bound to effect the organization of labor in innumerable ways, I limit the discussion here to several major components of the peace that were essential to shaping the labor market. Before the Tokugawa regime, the Muromachi government, a hegemony supported by a majority of regional warlord magnates, established control of much, but not all, of the Japanese archipelago in the late fourteenth century. By the middle of the fifteenth century this hegemony was falling apart and losing what nominal support or control of religious institutions and other political rivals it had been able to claim. One part of the decline of the Muromachi government included vassals and lieutenants of individual warlords overthrowing them to take control of the territory the warlords had claimed. Of course, if a lieutenant could overthrow his overlord, he could also be overthrown by his subordinates or others. The social order and public security broke down as local strong men or even village bullies fought to claim as much territory as they could take and control. factors that shaped the organization of labor 117

At the same time, villages, towns and some provinces organized for mutual self defense and local autonomy in egalitarian, or semi-egalitarian leagues called ikki. The process of reunification begun by Oda Nobunaga entailed bringing all of the warlords under the authority of one overlord— Nobunaga, later Hideyoshi and ultimately Tokugawa Ieyasu—by conquest or alliance. Policies designed to guarantee warlords security in their status were critical for establishing the framework of the new social order.1 Two policies originally begun by Oda Nobunaga in his own domains and later extended nationally by Toyotomi Hideyoshi were cadastral surveys and the separation of warriors from the land. With these two linked policies, men who considered themselves warriors and vassals of some daimyo warlord lost their claim to land and moved into castle towns or cities. At the same time, all agricultural land was registered with the cultivators of that land and attached to rural villages. Villages gained confirmation of local autonomy and independence from warrior interference in exchange for paying a harvest tax on rice based on the annual estimated potential for producing rice on the land attached to the village. Daimyo were granted domains defined by the amount of harvest tax they could collect from the territory under their administrative control and were to use this income to pay stipends to their vassals, who were prohibited from most commercial activity, and fund any other duties or services necessary for the administration of their domains.2 Separation of warriors from the land they had controlled and forcing them to live on stipends from their overlords in town separated hope-to-become warlords from the resources to overthrow their overlords, providing political stability and security to the overlords. Separating warriors from the land also resulted in a rapid urbanization. Since warriors, clergy and the nobility were also restricted from most commercial activity, this sudden urbanization also demanded a rapid expansion of production for urban markets and increasing demands for labor There were 260 domains during the Tokugawa period at any given time, each with a castle town. By 1650 there were 250 castle towns of various sizes with 25 of them cities with populations of 30,000 or more.3 In addition to the castle towns were three metropolises. Edo was the Tokugawa political capital. All of the daimyo were required to maintain residences in Edo

1 See for example Souyri, The World, 121–135. 2 Wakita Osamu, “The social”. 3 Nakai Nobuhiko, “Commercial change”. 118 mary louise nagata where their wives and children were required to live as hostages while the daimyo were required to alternate attendance in the shogun’s court with time back in their domains.4 Thus Edo grew rapidly to a major metropolis just from the residences, family, vassals and servants of the daimyo, not to mention the merchants, artisans and entertainers necessary to supply their needs. The population of Edo reached one million by 1700. Osaka was the financial center where a portion of the tax rice from each domain was stored or converted to cash. As the financial center, Osaka was also a major commercial center with many large merchant houses and a center for the cotton textile industry. Osaka’s population was stable at around 600,000 by the eighteenth century. The third metropolis of Tokugawa society was Kyoto, the traditional capital and home to the emperor as well as a center of handicraft industry and silk textile production with a population of around 200,000. The warrior class alone comprised about eight percent of the population at that time. When the populations of the castle towns are added to the three metropolises—Edo, Osaka, and Kyoto—and other commercial cities in Central Japan, the urban population in 1650 was around fifteen percent of the total population, estimated at 12.3 million and approached twenty percent of the population in the eighteenth century.5 Before the early seventeenth century labor was usually supplied from within the household, defined broadly to include individuals or family units that were dependent upon the main household, or supplied from casual labor. The dependents included several groups. Some were working to pay off a debt. Others had been sold or even sold themselves into service to gain the basic support of room, board and clothing. In these cases, they frequently were rewarded with parcels of land from the main household after ten to fifteen years in service, but continued to barter labor for other needs. Yet others joined the main, or master, household during the period of political disorder and war in the fifteenth and sixteenth centuries. While many of these dependent individuals and households were not blood kin of the master household, others were kin or became kin through intermarriage with the master’s family or through adoption. In the case of marriage, girls inherited their mother’s status and boys inherited their father’s status. These dependent relations could continue across generations or be limited in time. Those who remained as dependents across

4 Hall, “The bakuhan system,” 158. 5 Shinbo Hiroshi and Hasegawa Akira, “Shōhin seisan”. factors that shaped the organization of labor 119 generations, called nago, hikan, or fudai in various regions of Japan, could be regarded as hereditary servants or slaves.6 The ability to sell oneself or a family member into service was considered an important survival mechanism and potential opportunity for the poor or others of low status that could also provide training as well as protection. For example, the new daimyo warlords created by the Toku- gawa regime rewarded with domains from the territory held by warlords defeated in the Battle of Sekigahara in 1600 were fudai from within the Tokugawa house. At the same time, this system could certainly be harsh and was open to abuse. One long time problem was bandit gangs who kidnapped people off the road and sold them into service. Political and legal authority was so fragmented and limited during the fourteenth to sixteenth centuries, even during periods of relative strength and stability, that there seemed little hope for controlling the practice.7 The cadastral surveys and political stability changed these relations as dependent households gained independent claim to the land they farmed. This began the process of emancipation as well as breaking up overly large holdings and denying absentee landlords—warriors, clergy or religious institutions, or nobility—ownership or control of the land. The new regime was also much stronger and claimed far more legitimate central authority than the Muromachi regime. So, even though the political order remained relatively decentralized, it was far stronger and more central- ized than previously. As political order was established, new laws in 1616 and 1619 prohibited the permanent sale of persons. Although individuals selling themselves or members of their households into service were not considered a major problem, bandits had kidnapped members of the nobility and others of high status to sell into service as well. According to the new laws, both the seller and the buyer would be imprisoned and fined the amount of the sale thus removing the profit of the sale from the seller and doubly fining the buyer, whereas the putative slave went free. In some cases the seller was executed. These laws were the last of a series of steps taken to end the practice of bandits kidnapping people and selling them, and ultimately the market for persons.8 The sale of persons devolved into two types of service

6 Nelson, “Slavery in Medieval Japan”. See also Farris, Japan’s Medieval Population, 79–83, 235–45, 247–52. 7 Nelson, “Slavery”. See also Karen Brazell (trans.), The Confessions and Ishii Ryōsuke, Shinpen Edo, 106–7. 8 Ishii Ryōsuke, Shinpen Edo, 106–7. 120 mary louise nagata agreements. One was called “pawn service”9 in which a head of household borrowed money on the collateral of the labor of a member of his household. Sometimes the labor of the servant was considered the interest for the debt, but the servant often received a wage for his service after the debt was repaid. Since the 1619 law prohibited the permanent sale of persons, there were also time limits set for how long the servant could remain in service with maximum periods of up to ten years depending on the region. Ten years was the maximum allowed in the domains of the Toku- gawa House and most other domains followed suit.10 When a family that still owned some slaves went into debt, they typically pawned their slaves, who then usually went free at the end of the service contract. These and other opportunities for wage labor and commerce led to the near total disappearance of “hereditary servants” as slaves by the eighteenth century in Central Japan, where two of the metropolises were located that provided many economic opportunities, and in all but the most isolated regions by the early nineteenth century. The second type of service agreement was understood as a temporary sale and this developed into contract wage employment. As a temporary sale wages were originally paid in advance to the family or owner of the servant. Labor had originally been supplied from family and slave labor within the household and a network of personal connections such as the children of kin, friends and neighbors, but rapid urbanization and expansion of the market demanded more labor than available through personal connections. This increased the demand for contracts beyond those offered for debt service. Since this labor force was also supported within the household of the employer, the new conditions left the system open to abuse. Employers in need of workers would pay employment agents a commission to introduce prospective employees for a five day trial period and then negotiate a contract. The agent would receive the wages, paid in advance, and remit them to the family of the worker less another commission. Problems occurred as the agents persuaded the worker to abscond, stealing whatever s/he could, and introduced the worker to a new employer to go through the process again. With the growing demand for labor, this scam could be quite lucrative for the agents. Efforts to control these abuses resulted in the development of written labor contracts and wages paid, or credited, regularly to the worker rather than paid in advance to his or

9 質入れ奉公. 10 Ishii, Shinpen Edo 108–9. factors that shaped the organization of labor 121 her family by the employment agent. The state also organized the agents into self-policing and licensing organizations and set a flat rate paid once by the worker, so the agent could not profit so easily from this scam.11 Although wages came to be paid directly to the worker over the course of employment, in reality they were credited to the worker’s account with the balance paid in cash when the worker left service.12 The political, legal and economic relations between village and domain shaped labor in many other ways as well. With 260–70 semiautonomous domains under the Tokugawa regime, and because local autonomy meant that each village was free to develop its own accommodation to the new environment of political stability and growing commerce, there is not room here to address all of the variations in detail, nor has any study been able to accomplish this enormous task. Here I will simply note several important trends and highlight some of the ways villages developed new organizations of labor for market production. In general terms, daimyo income and domain administration depended upon the harvest tax on the estimated rice crop. This originally resulted in domain demands and restrictions to maximize the land cultivating rice and the agricultural labor force. On the other hand, the tax was commonly paid in some combination of cash and kind, usually rice which also served as another form of cash. While the rice was expected to come from the harvest, the cash could be earned from other sources. When weather or natural disaster caused the harvest to fall short of the estimated amounts, cash from other sources became even more important so peasants could buy the rice necessary for that part of the tax.13 The need for cash income from other sources encouraged the development of commercial ties to the market. For many, this meant labor migration to other villages, the castle town, or opportunities outside the domain. The distance and direction could be constrained by domain policies trying to keep agricultural labor local, but long lists of people leaving villages without reporting where they were going where domain restrictions were strong demonstrate that domain policies could constrain, but not prevent labor migration outside the domain.14 Who tended to leave the village

11 Ishii, Shinpen Edo, 91–105. 12 Nagata, Labor Contracts, 85–93. 13 The rice tax and adjustments to the system have been studied and discussed in detail by other scholars. Our focus is on how the system constrained the organization of labor and the labor market. For more information on the rice tax kokudaka system, see Chie Nakane, “Tokugawa Society,” or Smith, Native Sources, 50–70. 14 Nagata, “Leaving the Village”. 122 mary louise nagata for labor migration depended upon the family system and regional differences in access to national commercial networks. Analysis of population registers in Central Japan has shown a life course pattern of labor migration during a period of life-cycle service during the ages 11–25 followed by marriage. Population registers in Northeastern Japan have shown a longer period of labor migration after marriage during the ages 20–60.15 One reason for these differences was the reliance upon rice agriculture for the economic base. Northeastern Japan did not have a climate suitable to the strain of rice common at the time, so there was more trouble and poverty from harvest failures. Southwestern Japan, on the other hand, did so well with rice that there was less incentive for investment in other market production. The region was also less urbanized than central Japan resulting in yet other patterns of labor organization and migration, but little has been published on this topic as yet. Domain authorities sometimes encouraged local industries and commercial production to provide wage labor and economic opportunities within the domain to encourage the agricultural labor force to remain closer to home. This was often with the understanding that labor migration would take place during the agricultural off season of the winter months, as was often the case for sake and soy sauce breweries.16 Breweries also employed a large casual labor force that worked by the day, or even the half day, providing more flexibility for labor amid various constraints as commerce, industry and agriculture competed for labor force.17 Yet another solution was for villages or wealthy peasants within villages to invest in production for the market. This could include cash crops, including industrial crops such as paper mulberry, cotton or lacquer trees, or producing primary products such as raw silk from raising silk worms or dye cakes from safflower or indigo. Some villages also invested in textile manufacture with younger girls spinning or winding the cotton thread or silk floss, older women weaving the cloth and men dyeing the cloth or the thread in proto-industrial organization of household labor by-employments. In other villages, one or more of the wealthier households invested in the production of paper, sake, soy sauce or other product employing men and women from the village and surrounding villages. The organization of production labor including the use of casual labor in a

15 Nagata, “Labor Migration”, 60–84. 16 Nagata, Labor Contracts, 81–82. See also Roberts, Mercantilism. 17 Nakayama Shōtarō, “Shōyu jōzō”, 175–98. See also Yunoki Manabu, Sake zukuri and Nagata, Labor Contracts, 48–50. factors that shaped the organization of labor 123 manufactory, or using multiple households as household by-employment depended upon the production needs of the product.18 Ironically, rural investment in manufacturing as by-employments or using casual labor was so successful that its success delayed or prevented the formation of an urban proletariat and wage labor, whether contract or casual, was closely tied to agricultural income.19 Larger and more successful of such businesses, some of which are still in operation today, expanded in a network following the stem and branch organization of the Japanese stem family to have retail and manufacturing branches of the business in the major cities of Kyoto, Osaka and Edo as well as shipping their products to other markets through the national commercial networks. The next section addresses the family system and lineage network as constraint and factor shaping the organization of labor and the labor market beyond the local community.

Family As Business and the Organization of Labor

Labor in Tokugawa Japan fell into five broad categories:

1. casual labor—also known as daily wage labor hired without contract by the half day, the day or longer periods. For example, the sake brewing industry employed 100-day laborers. Casual labor was employed by all sectors of the economy as it made use of the free time from agricultural labor, either to work as a farm hand for someone else, providing transport or in market production. 2. Live-in contract labor—servants, farm hands or business employees living in the employer’s household. The age period and character of the labor force for this type of arrangement varied between northeastern and central Japan. As business employees they were employed as apprentices, clerks and managers. 3. commuting contract labor—management level employees in commercial enterprises, semi-independent artisans without independent business licenses, some rural village labor. Live-in business employees moved out and continued working as commuters when they reached

18 Nagata, Labor Contracts. This book is a study of labor contracts and labor relations from rural businesses taking each of these forms as well as urban businesses from Kyoto. See also Pratt, Japan’s Protoindustrial Elite. 19 Osamu Saito, “Wages”. 124 mary louise nagata

management or were ready to marry and establish their own households. From the late eighteenth century, villagers who lived within commuting distance of employment opportunities tended to commute instead of living in. This became more common as more labor opportunities, both agricultural and manufacturing, became available in the countryside as well as in the city. 4. Pawn service—domestic service and farm labor. These contracts rarely appear where commercial opportunities were common. Most of the pawn service agreements have been found in agricultural villages. 5. hereditary servants/(former?) slaves—could no longer be sold and often had opportunities for freedom or independence. These could become integrated as kin or almost kin members of the master’s household in the business context, but were likely to continue in dependent relations in isolated villages. Where the employer/master owned a family business, they became part of the elite management of the business because they were considered more reliable.

Labor in Japan was not the purview of any guild-like organization controlling apprenticeship or the technical standards of any industry. Individual employers trained their own apprentices and relied on the loyalty of the live-in contract employees trained in the business to guard their technical production secrets as best they could. Businesses competed with each other on quality and innovation. There were no particular rules for the above labor relations, but employers were generally expected to provide room, food, clothing and basic medical care to the live-in employees and food to the daily wage laborers and the commuters. Study of labor contracts reveals no obvious difference between proto-industrial handicraft, manufactures, or agriculture in the contract conditions, likely because the purpose of the contracts was to prevent conflict and provide information on who would settle any conflict between employer and employee.20 Pawn service contracts tended to disappear where there were other commercial opportunities. For example, records from the Sugiyama family document collection of Tondabayashi village south of Osaka include some pawn service contracts from the early eighteenth century, the most recent dated 1738.21 Other records, however, demonstrate how a poor

20 Nagata, Labor Contracts. 21 Osakaya Sanzaemon and Suzuya Shōbei, [to Sugiyama Zenzaemon], “Uke tori shōmon no koto,” Document of receipt, 11/1738, Sugiyama collection no. 62B, Kyoto Uni- versity Museum, Kyoto. Nagata, Labor Contracts, 86. factors that shaped the organization of labor 125

tenant farmer could borrow money to open a shop in town and pay off his debt to his landlord and employer by working in the landlord’s sake brewery.22 Business records rarely include pawn service contracts and most of the pawn service contracts have been for domestic service or agricultural labor, so I exclude this form from consideration in this part of the study. Evidence from northeastern Japan where pawn service remained relatively common into the nineteenth century suggests that a farmer was more likely to pawn his own labor rather than that of another member of his household. One possible reason was that the servant received a wage independent of the debt transaction. A farmer knew he would receive that wage if he pawned his own labor and left his family to figure out how to repay the loan, whereas another family member might take the wage and leave.23 In addition, although children do occasionally appear in the data as servants at ages younger than ten, the occurrence was rare. The probability that a servant would be less than ten years of age was only five per thousand in late Tokugawa Kyoto.24 This is not to say that child labor was not used, but children were not a substantial portion of the labor force. Adoption was the main means of transferring guardianship and any potential exploitation of a child from one family or individual to another. Yet, adopted children under the age of ten were also rare, or at least rarely identified as such.25 Tokugawa enterprises formed on the stem and branch framework of the Japanese stem family. In this framework, one married child remained home with the parents and, in the case of a business, apprenticed to the parents and succeeded to management as head of the family business. The other children all left to manage another branch of the business if it were large enough, marry or otherwise join another family, or establish their own alternative occupations. A small business was only one household. A large business was formed by multiple households built on this framework and the organization of labor in these businesses was constrained by this family framework. Even smaller businesses might have

22 Haya et al., [to Sugiyama Chōzaemon], “Issatsu no koto,” Letter of request, 8/1776, Sugiyama collection, no. S97, Kyoto University Museum Archives, Kyoto. Nagata, Labor Contracts, 66–7. 23 Nagata, “Leaving the Village”. 24 Analysis of the Kyoto population registers explained and used in the next two sections of this article reveal twenty-eight servants listed under the age of ten: 1 age two, 1 age five, 2 age six, 2 age seven, 4 age eight, and 18 age nine out of a total 4,794 observations for servants for a probability of 0.0058. 25 Nagata, “L’adoption à Kyôto” . 126 mary louise nagata one collateral branch managed by a younger son, the retired head or some other member of the family. I begin with the organization of labor in a small business and expand the explanation to larger and more complex businesses with multiple branches. Most business households were comprised of family members, primarily nuclear or stem kin, and live-in employees of both sexes. These employees appear in four categories: apprentices, skilled or management workers, maids and (rarely) other domestic servants. I use the term apprenticeship loosely. There were no official rules for apprenticeship-journeymen as a system. A boy or girl entered employment at ages 10–14 with the understanding that they would learn to read, write, calculate, and the basic skills of that business or industry. Their training at this point from early teens until around 17–20 was under the supervision of the wife of the head of household and business and assisting her was part of their duties. Labor turnover tended to be quite high during this period, although the details are still under investigation. Apprentices who finished their apprenticeship without being let go generally moved on to more intense technical training becoming skilled workers, who would move out after a few years to become commuters. These workers also tended to change their personal names at this stage of employment, sometimes receiving a name from the employer, and were socially identified as members of the employer’s house. Many young women also received new names while in service.26 Although the majority of apprentices following this career path were young men beginning their apprenticeship at around the age of twelve, records from the Kyoto silk textile industry show young women following this career path too, at least to skilled artisan if not management. Clerks, management employees and skilled male workers were called tedai. Although women were listed as student/apprentices, they do not appear listed as tedai, suggesting they could become skilled weavers or artisans, but not management employees. If the business involved production that could use unskilled or semi- skilled casual labor, as with brewing or dyeing, then there would be meals provided with room on the floor for the day laborers to sleep, often in the manufactory or the attic. This casual labor was employed in addition to the live-in management and domestic labor. Nearly everyone married in the family and demographic system of Tokugawa Japan, usually before the age of 30 for men and around age 23

26 Nagata, “Names”. factors that shaped the organization of labor 127 for women in central Japan. Live-in servants were expected to be single, so the live-in period of employment usually ended for men and women around the time they married. The transition from a live-in skilled or management employee to a commuter who lived out often accompanied marriage. In the extant population registers of Kyoto, for instance, adult women born outside of the city were either in service or married. Women only appear living independently and single if they were natives to the city.27 The reasons for this are not clear, but the analyses in the final section of this paper suggest that the economic and political climate of the final decades of the Tokugawa period in Kyoto were not conducive to young women living alone. Unfortunately, there is no usable population register data for the other cities or for Kyoto before 1843 that recorded age and birth province, so I cannot say if this was a general trend or the result of that particular economic and political context. Larger businesses expanded by establishing a branch household of family and business, a regular mechanism of the Japanese stem family system. Kyoto businesses often had a collateral branch where younger siblings of the heir to the business gained training or where the former head of a business could retire to. Skilled or management employees would be sent from the stem household of the business and apprentices might circulate between the branches. Larger businesses also often subsidized skilled or management employees to establish their branches of the business in the same way, and the business could continue expanding with both kin and non-kin branches and sub-branches. For example, the Konishi Breweries had their main or stem business and brewery in Itami, a rural village outside of Osaka, from the mid sixteenth century. They established a kin branch in Osaka in the seventeenth century, a non-kin branch in Edo in the early eighteenth century and a non-kin sub-branch of the Edo branch also in Edo in the nineteenth century. Apprentices were recruited by all branches for training in the stem business household and were sent to the various branches from the stem household after completing their apprenticeships.28 These larger businesses were typically controlled by a senior management council made up of the head of each branch, the head couple of the stem business and the retired head of the stem business. Clearly this council included non-kin employees. At least one larger business that

27 Nagata and Hamano, “Marriage Market”. 28 Nagata, Labor Contracts, 28–30. Yunoki Manabu, “Kinsei Itami”, 1–23. 128 mary louise nagata retained hereditary servants employed them in these top positions. These were the most trusted employees often serving as the agent or representative of the head of the stem business as well as managing branch shops in Edo or major transportation hubs.29 Since the large family businesses were controlled by such councils, disaggregating certain management decisions can be difficult. Certainly the head of the business made most everyday decisions, but the council could and did overrule him. Moreover, the family could, and sometimes did, disown and divorce the head from the family and business.30 The practices of adoption, divorce and disowning in the Japanese family system were also important for shaping the opportunities for labor as well as privileging ability over heredity in the management of family businesses, including disowning family members. Skilled employees could, of course, marry into the employer’s family. In addition, they could be adopted without marriage to a daughter of the employer’s family, although they usually married some member of the family, such as a cousin or a niece, after adoption. Moreover, members of the family, including the head of the family and business, could be forced out of the position and disowned from the family, even in favor of an adoptive son from outside the family such as an employee. The overall head of the Mitsui conglomerate was disowned from the family in the mid eighteenth century.31 Because of such actions, the family as business appeared to act as a corporate unit. Certainly different members of the family were more involved in particular decisions than others, but there has been less research and there is less information on the internal politics of family and business management. Tokugawa family practice in central Japan shaped the labor market in other ways as well. Nearly everyone married, but there often was little room for establishing new households in village society. Moreover, some time in service was a common way to gain training and experience in agricultural techniques as well as in proto-industrial skills. Pre-marital service was a useful way for young people to gain the training and skills, bring extra income into the village and send extra youths out of the village to

29 Kobata Atsushi, Okamoto, 350–1. Nagata, Labor Contracts, 87. 30 Nagata, Labor Contracts, 27. 31 Kagawa Takayuki, “Kinsei shōnin”, 173–80. For discussion of an adoptive heir who disowned his adoptive father and head of the business see also Mary Louise Nagata, “Mis- tress or Wife?” factors that shaped the organization of labor 129 find other places to settle thus circulating labor and population to where there was need for it.32 So the Japanese family provided the framework for business and therefore the organization of labor. In addition, Japanese family practice also constrained and shaped the labor market and the opportunities provided for labor. Certainly this is true everywhere. In Tokugawa Japan, however, the family was a large corporate unit that could be limited to one household and nuclear unit, but often included many branches and sub- branches and many non-kin employees and managers as well. Individu- als were rotated in and out of various levels of kin relations through the mechanisms of adoption, marriage, and divorce for the convenience and longevity of the family business. Maintaining the profitability and longevity of the business was important because the business provided the economic base for many people, both kin and non-kin, and the welfare of the group was more important than the individual wishes or success of any individual member including the head or the heir. The previous section showed how the political system and political, legal and economic policies were important constraints and factors shaping the organization of labor and the labor market. The next sections examine the interaction and effect of economic and political crisis upon urban labor in the final decades of the Tokugawa period through analysis of the population registers of Kyoto neighborhoods.

Labor in Late Tokugawa Kyoto—Data, Background and Methodology

In the 1630s Tokugawa Iemitsu, the third shogun of the regime, banned Christianity and took strict monopoly control of foreign trade, limiting trade to China, Korea, the Netherlands and later Siam. Foreign ships from these countries were allowed to come only to Nagasaki, although Japan also maintained a closer trade relationship with Korea through the daimyo of Tsushima Island.33 These events also marked the beginning of an annual compilation of religious investigation registers designed to ensure that no Japanese residents were Christian. These registers also proved useful for keeping track of the population and vital events such as marriage, adoption and divorce. Many domains also used the registers to keep track of the agricultural labor force and many villages and domains further used

32 Nagata, “Leaving the Village”. 33 For a detailed discussion of Tokugawa foreign policy, see Toby, State and Diplomacy. 130 mary louise nagata them to keep track of how much land, or agricultural production, each peasant household was responsible for.34 Kyoto was the smallest of the three metropolises of Tokugawa society after the political capital Edo and the commercial center Osaka. Kyoto was the official capital as the home of the emperor and the imperial court nobility as well as the capital of traditional culture. It was also a manufacturing center for silk and cotton textiles and other handicraft industries such as ceramics, paper and lacquerware. The best data for understanding the general nature of labor and the labor market is the religion and population investigation registers. Although Edo and Osaka were larger cities and Osaka was the larger commercial center, few of the population registers of Edo are extant and those of Osaka did not record age until the final year of the registration. In the city of Kyoto these religious and population investigation registers (shūmon ninbetsu aratame chō)35 were compiled annually in the ninth month of the lunar calendar by each neighborhood community, the urban unit of local autonomy. The registers list each resident of the neighborhood by religious sect (all Buddhist), and household of residence by name and relation to the head of household. From 1843 the listings also recorded the age and birth province of each individual and made greater effort to keep track of movement into or out of the community through notations on slips of paper pasted on the registers. In addition, the ninth month listing each year was followed by an update listing of newcomers to the neighborhood in the second lunar month of the following year. These annual listings continued until 1868, the year the Tokugawa regime fell, although some neighborhoods also compiled a final listing in 1869. Since age and birth province are important information for indentifying and linking individuals from one listing to the next and age is essential for any demographic analysis, this study uses only the listings compiled 1843–1869.36 The recording of individual ages in the Kyoto registers also means that they are more useful than those of Edo or Osaka where this critical information was not recorded until later. A neighborhood consisted of all households on both sides of one block of a city street from one corner to the next. As autonomous units, each neighborhood had its own headman and administrative council, collected

34 Laurel Cornell and Akira Hayami, “The Shumon Aratame”. 35 宗門人別改帳. 36 One neighborhood, Shimizu, included this information from 1842, so that listing is also included in the data. factors that shaped the organization of labor 131 its own dues and had its own ordinances. There were 1500 such neighborhoods in Kyoto during these final decades of the Tokugawa period, but only twenty-eight neighborhoods have extant listings available for analysis at this time. The availability of the listings is also spotty. Ten neighborhoods only have one extant listing from the data period while others may have multiple extant listings, but few or no consecutive listings. Only two neighborhood collections have more than twenty of the twenty-six possible listings of the population registers providing a nearly unbroken series. The data period of 1843–1868 is very good for investigating the effects of economic and political crisis upon labor in the urban context. The data series begins in the final year of the Tempō reforms (1841–1843) enacted by the Grand Councilor Mizuno Tadakuni to address the economic crisis after the Tempō famine of 1833–1837. Two reforms had negative effect for Kyoto and particularly the Nishijin silk textile industry. The first was a number of edicts that dissolved trade and business associations or any organization that looked like a guild or a cartel on the assumption that these associations were keeping prices high. However, these associations functioned to regulate the market, manage credit, manage distribution and enforce contracts. Moreover, many of these associations had been formed by order of the state in the eighteenth century with the purpose to control prices and keep them down. The effect of the edicts dissolv- ing and prohibiting these associations was an upset of the distribution system, lack of credit in a market that operated primarily on credit, and price inflation.37 In addition to the negative effect upon the market and the resulting price inflation, the trade associations also had insurance functions that provided assistance to members. When they were dissolved, this safety net for businesses in times of economic trouble disappeared. This was particularly a problem for the Nishijin textile industry because another part of the Tempō reforms restricted production and sale of luxury items such as silks and introduced new sumptuary laws. So the Nishijin silk textile industry took a direct hit with these reforms, which also removed one of the safety nets for businesses in the industry.38 Even so, the population trends in Kyoto neighborhoods outside of the Nishijin district show some evidence of economic growth or stability. Nevertheless there was a

37 Ishii Ryōsuke, Shōnin, 78–85. Miyamoto Mataji, Kabu nakama, 330, 337–343. See also Nagata, “Brotherhoods”. 38 Yagi Tetsuo, “Kinsei”, 169–172. Hamano Kiyoshi, “Kinsei toshi”, 211. 132 mary louise nagata further crisis in 1853 when the price of rice jumped fourteen percent due to another poor harvest. The effect of this crisis was particularly heavy in Nishijin because it came at a time when weaving workshops were shut- ting down leaving weavers unemployed due to the effect of the reforms. Ironically, records show family members of these weavers returning to live-in service positions during this time of hardship. The city of Kyoto also provided emergency welfare distributions of rice to households in distress in that year.39 1853 was also the year that the American Commodore Matthew Perry forced Japan to open wider to international trade. By 1857 Japan had treaties with five Western nations and had opened several treaty ports. The opening of the treaty ports contributed to the economic growth of some parts of Japan, especially regions in the northeast that produced raw silk. However, raw silk that was exported abroad was also raw silk that did not reach the Kyoto market. This caused the price of raw silk to rise in Japan bringing a supply crisis to the silk textile industry that now found itself in competition to buy raw silk. At the same time, brokers who dealt in raw silk took advantage of the rising prices and stockpiled the silk. In 1863 one such broker’s warehouse was attacked revealing a stock of 500,000 ryō of raw silk at a time when finding raw silk for sale in the market was difficult.40 Two years earlier, in the middle of this crisis, another neighborhood in Nishijin compiled a listing of households in need of assistance. The price of rice also began a continuous rapid rise at that time. Whereas the price of rice had been hovering at around eighty monme (300 grams) of silver per koku (180 liters) of rice, it more than doubled during the Bunkyū era 1861–64 to around 200 monme (750 grams) of silver per koku of rice. Poor though the economy was in 1861, the price of rice had only begun to rise. During the Keio era (1865–8) the price of one koku (180 liters) of rice rose to 400–1,000 monme (15–37 kilograms) of silver. The magistrate’s office could not ignore this rise and solicited money for rice and porridge directly from the large businesses, business associations and neighborhoods, establishing various centers for rice and porridge to be distributed.41 When rice became too expensive, they gave out money. The economic situation for the city of Kyoto did not improve until after the

39 “Nanjū sha kazoku tori shirabe sho shita gaki,” [Draft of an investigation of families in distress], 1853, Hanaguruma-chō collection, Kyoto City Library for Historical Documents. 40 Hamano Kiyoshi, “Kinsei toshi”. 41 Kobayashi Takehiro, “Bakumatsu ishin”, 3–15. factors that shaped the organization of labor 133 regime fell in 1868. During the 1860s, various outside political groups such as leaders who opposed the regime, daimyo who came to Kyoto to assist the Kyoto warden and, of course, the central government all competed to provide assistance to the people of Kyoto. The impossibly high and rising price of rice, however, suggests that little could be done until economic controls were re-established under the new regime. The treaties and the opening of the ports after 1857 also precipitated a domestic political crisis. As the official capital and home of the emperor and the imperial court, Kyoto was also at the center of the political crisis. Political legitimacy in Japan has traditionally depended upon recognition or support by the imperial court, usually in exchange for economic support of the imperial court. Young warriors opposing the Tokugawa regime came to Kyoto in secret to make contacts and alliances with the imperial court. Tokugawa representatives were also sent to Kyoto to prevent this and strengthen and affirm regime ties to the court. Each grouped attacked the other with assassinations, counter-assassinations and random attacks on people suspected of supporting one side or the other in the city. The Tokugawa regime sent a group of young loyalists to defend their position and maintain order placing the city under martial law from 1863 until the regime fell in 1868.42 So, the data period 1843–1868 begins with the final year of the Tempō reforms and a brief period of growth followed by another major economic crisis ultimately leading to a political crisis that finally brought the end of the Tokugawa regime. The problems were aggravated even further when much of the city burned down in the seventh month of 1864. Did these crises reduce immigration to the city or encourage increased emigration? Did they limit the labor market or did live-in employment provide refuge in time of need for people who otherwise would have been finished with the life-cycle service portion of their lives? The next section will use analysis of the population registers of Kyoto to address these questions in several ways. First I examine the probability of being a servant as it changed over time and the probability that a servant would be a native of Kyoto or an immigrant from another province, as well as the average ages of servants that appear in the registers. Second I use consecutive year data to investigate the probability of entering or leaving the neighborhood communities, comparing again natives and immigrants. More pertinently, I then limit this analysis to servants.

42 Totman, The Collapse. 134 mary louise nagata

Finally, I examine the results according to the background of economic and political crises to address the above questions.

Labor in Political and Economic Crisis

The above story of economic and political crises during the data period 1843–1868 suggests that the Nishijin silk textile district was the hardest hit. Although there are records of hardship distributions in two Nishijin neighborhoods, the population trend does not entirely support this. Fig- ure 1 shows the annual population trend for the two neighborhoods that have near continuous series of extant population registers during the data period. Seido is a neighborhood near the commercial center of the city containing a number of family businesses. Kankiji is a neighborhood in Nishijin also containing a number of family businesses including weaving workshops that employed young women as apprentices. The population trend for Kankiji neighborhood shows increase 1852–6, followed by a decline until 1861, and then rapid growth until 1865. Thus, the population was growing at the time of the 1853 hardship distribution, but had declined to the lowest point at the time of the 1861 hardship distribution. On the other hand, the period of population decline 1857–61 is the period when the opening of the ports began to affect the silk trade. The population trend for Seido, less dependent upon the silk trade, is relatively flat. The dip in population in 1864 reflects the effect of the fire that burned down much of the city that year including houses in Seido neighborhood (see figure 1). Figure 2 shows relations to head of household by age for the full data set graphically demonstrating the life-cycle nature of live-in employment (hereafter called service). The age period of life-cycle service was 11–28 for young men and 15–25 for young women. Figure 3 shows the probabilities that natives or immigrants would be in service at ages 11–30 over time. This figure shows that natives and immigrants show the same employment trend, but immigrants at these ages were far more likely to be in service than natives. In other words, most non-natives who came to the city came for labor migration. Moreover, 1853 was a year when ninety percent of non-natives and fifty percent of natives ages 11–30 were in service, an increase of roughly ten percent from previous levels. In addition, 1861 shows a dip in the proportion of both natives and immigrants in service at these ages, reflecting the economic hardship for employers, but 1864 shows a sudden peak in the year of the great fire. This suggests that people factors that shaped the organization of labor 135

300

250

200

150 Kankiji Seido

100

0 1851 186 1 1853 1855 1857 1852 184 3 1850 1845 1847 1854 185 8 1859 1863 1856 1865 1867 1862 1844 1848 1849 1860 1846 1864 1868 1866

Figure 1 Neighborhood population trends. not in service left for safer refuge elsewhere if they had other places to go, and this could also be the reason for the peak in 1853 as well. While both the native and immigrant trends show a general decline, this is less obvious among natives as periods of crisis seem to have either sent immigrants not in service outside the city or sent them back into service. Figures 4–5 provide more explanation for these trends. Figure 4 merely shows immigrant and native servants as proportions of the total populations of immigrants or natives. In the crisis years 1853, 1861 and 1864 the probability that an immigrant of any age would be in service rose dramatically, especially in 1853 and 1864 when sixty percent of the entire immigrant population was in service, suggesting both that immigrants not in service left or re-entered service. Figure 5 shows the proportion of immigrants among all servants and in the total population, confirming the explanation that times of crisis sent people looking for live-in service positions, likely for the provisions of room board and clothing that came with 136 mary louise nagata

900

800 sibling 700 stem spouse 600 sib servant 500 servants parent spouse 400 other

Observations nonstem 300 nonkin head 200 parent child child 100 head

0 14710131619222528313437404346495255586164677073767982858891 Age

Figure 2 Observations by age and relation to head.

0.9

0.8

0.7

0.6

0.5 Immigrants

0.4 Natives

0.3

0.2

0.1

1843 1845 1847 1849 1851 1853 1855 1857 1859 1861 1863 1865 1867

Figure 3 Probability that natives and immigrants ages 11–30 were servants. factors that shaped the organization of labor 137

0.7

0.6

0.5

0.4 Immigrants Natives 0.3 Proportion 0.2 Immigrant

0.1

0 185 1 1861 185 3 185 5 185 7 1843 1845 1847 185 9 1863 1865 1867 1849

Figure 4 Probability that natives or immigrants were servants.

0.4

0.35

0.3

0.25

0.2 Servants

0.15 Total

0.1

0.05

0 1 3 185 1861 185 1855 1857 1843 1845 1847 1859 1863 1865 1867 1849

Figure 5 Proportion of immigrants in the population. 138 mary louise nagata the live-in position.43 The proportion of immigrants in the total population declined from fifteen to ten percent after 1853, but the proportion of immigrants in the live-in service population increased 1860–64. After 1864, during the political crisis and martial law, the proportion of the population in service fell dramatically for both natives and immigrants as the people of Kyoto stayed in place. Possibly, those who had places to go, whether leaving the city or entering service, went during the economic crisis and then settled in to wait out the political crisis. One sign of people re-entering service for refuge could be a rise in the average age of live-in employees. Figure 6 compares the average ages of native and immigrant young men in service over time. These young men followed the career path of apprentice to skilled worker and management employee with the successful ones becoming branch managers or moving out to establish their own branches or businesses. The young average age of these employees represents the high turnover as many left service before finishing the apprenticeship period or soon after. The average age for completing apprenticeship was seventeen.44 At the same time, some firms also employed older young men in their twenties in domestic positions. Figure 6 shows the average age of immigrant men in service rising higher than native men in times of economic and political crisis. The average ages of native men in service fell during the economic crisis, possibly reflecting difficulty in advancing as skilled worker, or reflecting young men who took their skills home or outside of the city. Life-cycle service for young women was generally understood as a way to gain skills in managing a business household in preparation for marriage, sometimes including the craft skills to assist their future marital households in the textile industry. The age that young women left service was closely linked to the age at marriage. I have already noted that immigrant women in Kyoto appear either in service, married or widowed.45 Thus the average age of young women in service should be around nine- teen or twenty since they entered service at around sixteen and left at around twenty-three. Figure 7 shows that the average ages for immigrant women were higher until 1853, but the age of native women rose close to that of immigrant women in service in 1853. Then the average age of women in service, both immigrant and native, fell until 1860. The average

43 Nagata, Labor Contracts. 44 Ibid. 45 See for example, Nagata and Hamano, “Marriage Market”. factors that shaped the organization of labor 139

25 24 23 22 21

Age 20 Immigrants 19 Natives 18 17 16 15 185 1 186 1 1853 1855 1857 1843 184 5 184 7 1859 1863 1865 1867 184 9 1869

Figure 6 Average ages of immigrants and native male servants.

Ag e 21 Immigrants

19 Natives

15 186 1 185 5 1852 184 3 1858 1867 1849 1846 1864

Figure 7 Average ages of immigrant and native female servants. 140 mary louise nagata age for native women remained lower, declining to around eighteen until 1867 when it returned to the original level. The average age for immigrant women, however, rose dramatically to twenty-five and generally remained rather high until the end of the period. This analysis supports the explanation of service as a refuge in time of political and economic crisis, especially in time of political crisis.

Concluding Discussion

This discussion of the constraints and factors that shaped the organization of labor and the labor market in Tokugawa Japan has taken us from the sixteenth century and the founding of the Tokugawa regime in 1603 to the fall of the regime in 1868. The constraints and factors that have played major roles in shaping labor during this period were political, legal, economic, social and cultural. In other words, they involved all aspects of Tokugawa society. Political policies during reunification and the beginning of the Toku- gawa period changed the social and economic framework of society and the labor market. The separation of warriors from the land, the cadastral registration, laws prohibiting the sale of persons and other policies designed to establish peace and security formed the framework of Toku- gawa society creating an urban market and a population dependent upon supplying its needs from the market as well as changing the source of labor for that market production. On the other hand, political policies often had very little effect upon labor even when they were intended to address the labor market. For example, attempts to prevent or limit labor out-migration from northeastern domains were ineffective as shown by long lists of people who left without reporting where they went. Perhaps because of local autonomy, political authorities did not play a major role in shaping village industries or the rural-urban labor market. Instead, the timing and direction of labor migration depended much upon the convenience of village society as well as networks of kin, friends, region and business. The main concern of political authorities was maximiz- ing the stability of the rice harvest and therefore their tax income. Agri- cultural production and the land under cultivation was also expanding, and there was no large population forced off the land or a large landless population. Village communities and their member families balanced the need for extra income from remittances, useful for paying taxes or other investment, with the need to retain sufficient agricultural labor force. If factors that shaped the organization of labor 141

anything, political authorities and village communities in some parts of Japan worried about land left abandoned or underutilized and looked for ways to retain rural agricultural labor force in the face of demand for labor in other places. Encouraging investment in manufacturing in the villages could keep labor available for agriculture. Village economic needs and the labor market in turn shaped the demographic system. Everyone married in the Japanese family system, but only one married child inherited management of the family holdings. If the land under cultivation or the village economy expanded so there was room for more households, the village would try to retain more people and encourage those who left for labor migration to later return and settle down. If not, then labor migration was a way to send young people out to find other places to settle and other people to marry as well as circulate labor to where it was most needed. Labor within villages, both agricultural and proto-industrial or handicraft labor, and labor migration to and from villages depended upon local solutions balancing these needs as well as whether and how many families in the village or nearby villages had invested in market production. This complex mixture of concerns also shaped regional differences in the labor market and the demographic system. Villagers in Northeastern Japan where agriculture was not so productive needed outside income for survival, but also needed agricultural labor force. Marriage in this region was early and universal and labor migration took place after marriage. Villages in central Japan were close to many economic opportunities and villages could need to shed population. Labor migration took a life-cycle form and marriage was delayed, although still universal, especially for younger sons and young people of less affluent families in the village community. Through business investment and expansion following the stem, branch and sub-branch framework of the Japanese stem family, corporate family and business organizations formed a network of households that functioned as shops, manufactories and offices, often with manufacturing in the home and neighboring villages, retail shops in cities and towns with some branches also serving as warehouses, and transportation hubs that all served as residences where family members, managers and apprentices lived. Traditional techniques and the training of labor in Europe were often under the control of guilds and the apprenticeship system. Japa- nese trade associations or other guild like organizations were not concerned with training labor or controlling apprenticeship. Each employer trained its own employees in its own methods “in house”. Thus family, business and regional networks became very important as each employer 142 mary louise nagata developed training methods and tried to control their techniques without formal constraints. This was still apprenticeship in that young employees worked with or for a skilled worker and, in the process of following directions and helping out, learned the techniques developed and used by the employer’s business. But this was not subject to outside testing or valida- tion beyond success in the market. Labor contracts and the various categories of relations between employer and employee were about economic relations, prevention of conflict and mechanisms to settle conflict privately, so there was little difference between industries or sectors of the economy. The different categories reflect the degree the worker was integrated into the employer’s business. Casual labor was not integrated at all. Pawn servants were not much integrated into family or business. Live-in workers, however, became non-kin members of the household, family and business. The higher they reached in the business hierarchy, the stronger the social identification with the employer. Since one of the major issues of the Japanese labor market was trust, the hereditary servants one would normally consider at the bottom of the labor hierarchy were often at the top and considered almost family members in Japanese businesses. Japanese family practices of adoption, filiacentric marriage where an heiress is used to recruit a capable son-in-law, divorce and the integration of non-kin into the family also offered many opportunities for the capable and ambitious employee to gain a sponsor to establish his own business or a branch of the employer’s business, rise to top management or even inherit management of the employer’s business. They also provided the family business with methods to ensure capable management and leader- ship while remaining a family business. Analysis at the more micro level of the labor response to political and economic crisis in Kyoto, still the official capital and home of the emperor at the end of the Tokugawa period, revealed that live-in service continued to provide refuge from hardship. Unemployment was certainly a problem during the crisis. At the same time, people affected by the crisis tried to find more secure situations with some, particularly immigrants to the city, leaving the city and others returning to service. Especially in the crisis of the final years of the regime, the majority of immigrants were in service suggesting that others left the city or avoided coming. In spite of these effects, political and economic crises did not much change the system. In the end, labor in Tokugawa Japan was shaped and constrained by a complex package of factors. Political policies, but also the political system formed the social, legal and political framework for economic factors that shaped the organization of labor 143

development and labor. The economic needs of village as well as families and individuals were also important for the timing, direction and volume of labor migration and labor opportunities. Urbanization also played a role, but village production and the need to retain agricultural labor were also important factors. The family system also provided a business framework that could connect rural and urban manufacturing and markets as well as labor and migration opportunities through the family and business network.

Contractual Relations, Tariffs and Customs in the Lyon Silk Industry in the Nineteenth Century

Pierre Vernus

In the nineteenth century, industrialization in France combined the dynamics of several types of industrial systems. Among them, alternatives to more or less standardized mass production in factories, in which the Lyon silk industry can be included, played a preponderant role for a long time.1 The Lyon silk industry, a unique community of this kind, called the ‘Fabrique’ contributed greatly to making Lyon an important industrial and commercial center on a global scale, and for a long time it employed the majority of Lyon’s working population. Throughout the nineteenth century it increased its production and made a great contribution to French exports. The Lyon silk industry is also one of the most representative examples of a collective industry;2 a system of localized production made up of specialized productive entities of various shapes and sizes, linked by the process of manufacturing goods, in this case silk fabric. The elements of this economic ‘meso-system’3 were linked by a network of various relationships which could stretch from cooperation to competition and included the pay system. The stability of the organizational set up of this system has often been emphasized. It was marked by the domination of mercantile capital over the workshop and from the eighteenth century the vast majority of workshop heads had lost their access to the market which was controlled by merchants who organized production. But behind the organizational continuity of the industry, (at least until the early 1880s), the first decades of the nineteenth century were marked by great changes; institutional, administrative, legal and political, by a shift in the markets and consequently in production, as well as changes in personnel. We will briefly recall the main structural characteristics of the Lyon silk industry (the ‘Fabrique’) and explain the new institutional and legal framework in which the contractual relations between the merchants and the

1 Dewerpe, Le monde; Sabel and Zeitlin. “Historical”; Woronoff, Histoire. 2 This notion borrowed from Le Play was revived by Cottereau, “The Distinctiveness”. 3 On the notion of economic ‘meso-system’, see De Bandt, “La filière.” In Traité. 146 pierre vernus workshop heads from the beginning of the nineteenth century took place. Then our aim will be to try to understand the effect of these changes on the relations between merchants and workshop heads and in particular the contractual relations between them. We will show that the legal equality in the contractual relations between merchants and workshop heads established by the French Revolution did not put an end to the latters’ economic dependence on the formers that had grown gradually during the seventeenth and eighteenth centuries when the old corporations polarized increasingly. This dependence placed mainly on the workshops heads’ shoulders the consequences of the incompleteness of the contacts due to the uncertainties characteristic of the manufacture of silk fabrics but also to changes in the personnel and the markets. The workshop heads tried to reduce the social imbalance between them and merchants in negotiations which could generate major constraints on their working conditions in a competitive framework such as intensification of work and threat to their freedom to organize autonomy, by demanding the setting up of a tariff, but also by trying to improve the organization of information between weavers, and more widely inside the industry. Moreover the traditional practices of the industry, one of whose functions was to reduce the uncertainty over certain aspects of the contractual relationship often worried the weavers. The changes the industry had to make encouraged attempts to make the old customs obsolete or to set up new ones. So it will be useful to explain the strategies aimed at maintaining the practices that the different parties judged fair, or modifying them in a way that seemed more favourable.

The Apparent Permanence of the Organizational Pattern

At the beginning of the Restoration the productive structures of the ‘Fab- rique’ seemed little different from those it used at the end of the eighteenth century. It was a luxury industry depending on the extremely random fluctuations of a demand which was sensitive to the capricious nature of fashion and to political and diplomatic crises. The organization seemed to be flexible, allowing rapid adaptation of the volume of its production and the range of products. The system was run by merchants, also called merchant-manufacturers, or even ‘merchants ordering products’, but who did not produce the goods contractual relations, tariffs and customs 147 themselves. They prepared their collections of samples (several hundred for the largest houses) before the two main seasons, then, after booking in their orders, bought the raw silk fiber which they paid weavers to transform into fabric and dye, before handing them over to the workshop heads who wove the prepared silk, following the instructions they were given. The weaving was done by the ‘silk workers’ (the ‘canuts’). But this body, far from being homogeneous, was made up of two large groups, the workshop heads and the journeymen. The workshop head wove in his own shop and owned his own looms. He was in a middle ground between the merchant, a mercantile capitalist, and the journeymen or women and the apprentices, who he employed. This ternary structure was the result of a process which, inside the community during the old regime (before 1789), had led to growing social differences among the group of ‘masters’ which ended up as a split between the master-merchants and the master- workers (usually workshop heads). This development was accompanied by the adoption of various regulations which led to the traders usurp- ing the responsibilities and management of the community. Right up to the beginning of the Revolution the master workers challenged this development, at first violently, then more peacefully. Towards the end of the eighteenth century they finally saw it as inevitable. From then on the major social division was between the merchants and the large group made up of the master-workers and the journeyman workers. There was a long standing convergence of their interests and conditions. Most of the journeymen ended up by setting up as master workers, so finally, the only clear difference was that of age. Moreover the rule, maintained throughout the nineteenth century, which attributed to the journeymen a part (usually 50%), of the wages given to the master-worker, linked their pay to that of the workshop heads.4

A New Institutional Framework

The stability of this global scheme of organization of the ‘Fabrique’ in the first decades of the nineteenth century contrasted with the changes in the legislative and institutional frameworks which appeared during the Revolution and the Empire. The relations of provision of work and pay

4 Garden, Lyon et les Lyonnais; Godart, L’ouvrier en soie. 148 pierre vernus between the merchants and the workshop heads had to fit into a new order which constituted a break with the old corporate order and was based on new principles, in particular civil equality and the freedom of work and trade. Merchants and workshop heads agreed to recognize that the work relations which they had established were of a contractual nature and resulted from interpersonal negotiations, even if they disagreed on the content they gave to this notion.5 In this respect, in the legal framework defined by the Civil Code, these relations were dependent on contract law. The workshop heads were considered as work contractors and the contract made with the trader was seen as a contract for work. However, the intermediate position of the Lyonnais workshop heads did not fit into the categories lawmakers had used in the texts concerning the administration of work. The law of 12 Germinal year 11 (12 April 1803) and in particular its third section dealing with the workers’ accounts booklets, completed by the decree of 19 Frimaire, year 12, which applied “only to workers employed as journeymen or simple workers” did not apply to the Lyon- nais shop heads because they had benefitted from the status of ‘masters’ in the former corporate structure and were the employers of journeymen and workers. It was the third section of the law of 18 March 1806 which created the ‘conseil de prud’hommes’, (originally a conciliation board, but with wider legal prerogatives), which dealt specifically with their relations with the traders. It obliged the workshop heads to keep a double accounts booklet, delivered by the tribunal for each of the looms they possessed, to keep a copy and hand it over to the merchant-manufacturer for whom he worked on his loom. When a workshop head stopped working a loom for a trader, the merchant manufacturer had to settle his accounts and, when necessary, give the details of any debts he had incurred towards him. The other merchants for whom the workshop head worked also had to inspect the booklet and indicate the state of their accounts with the weaver. In the case of debt, the new merchant for whom the workshop foreman set up the loom committed himself to keeping back an eighth of the price of transformed silk to pay back his creditor colleagues, in chronological order of the creditors. The law maintained the practice of accounts booklets which had existed before the Revolution It also revived a measure installed in 1737, but abandoned in 1744, which linked the debt not to the workshop head but to the loom. Besides, this system constituted a credit

5 Cottereau, “La disincorporation”. contractual relations, tariffs and customs 149 mechanism for the workshop heads and offered a collective guarantee of their debts. In this way the shop heads had fully gained their freedom, often challenged, to change their clients and use their looms for several merchants, but also to stop working for a merchant as soon as the job he had been given was finished, even if he was in debt to him.6 By getting rid of the corporate structures the Revolution had opened up a period of uncertainty because the institutions in charge of organizing and regulating production disappeared with them. Even if the institutions had disappeared, this was not the case for demands for regulation.7 From now on what kind of regulation should be put in place? In what institutional framework? And what should be done with the old regulations? The statute of 1744 which had been modified several times and which was still in force just before the Revolution, was often cited as a reference. Thus, in 1807, the chairman of the ‘conseil de prud’hommes’, referring to article 10 of section 2 of the law of 1806 which told the board to note any infractions of the new laws and statutes or those which had been revived, considered that all the articles of this statute (of 1744) which conformed to the law of 18 March 1806 should “be carried out, until replaced by new ones”.8 There was a more general call for reflection on the ways and means of reestablishing regulation for looms; several projects were produced, but were never carried out. Nevertheless, the ‘Fabrique’ benefitted from the institutional reconstruction carried out under the First Empire9 which was marked by the propensity of the minister of the interior to consult the notables rather than the workers, as can be seen by the organization of the industrial tribunal itself. The law of 18 March 1806, even if it lowered the proportions anticipated in the first Lyonnais project10 gave five of the nine posts as councilor to merchant manufacturers, giving them a

6 The workshop head and the merchant “are linked to each other simply by mutual convenience; and when this link ceases to exist, both parties settle their accounts and separate without discussion or reproach” (Montfalcon, Histoire des insurrections); “neither one nor the other [the workshop head] work exclusively for this merchant or that, but successively and for several at the same time. The merchants’ agent usually overlooks the production.” (Villermé, Tableau), 355. On this point see also Pariset, Histoire de la Fabrique and especially Cottereau, “The fate”. 7 Minard, “Les corporations”. 8 Lyon municipal Archives (AM Lyon) 783 Wp 1, Letter from the President of the Lyon ‘conseil des prud’hommes’ to the mayor, 17 June 1807. 9 Moullier, Le ministère, chapter 14. 10 The report by Regny senior to the Lyon chamber of commerce 7 thermidor year XII, anticipated giving two thirds of the seats to the merchants (Archives of the Lyon chamber of commerce (ACC Lyon), minutes). 150 pierre vernus majority on paper.11 Beyond this, the only workshop foremen who could vote were those who were duly licensed by the tribunal, which meant the tiny minority who produced independently, (105 in 1807, only 50–60 in 1831).12 Moreover, even if the relations between the heads of workshops and journeymen were the responsibility of the tribunal, the journeymen were not represented. Nevertheless, the rapid increase in the activities of the tribunal witnesses to the fact that it answered a need on the part of all the people involved.13

Contractual Inadequacies and Social Imbalance

Contract law left both parties, merchant and workshop head, great leeway in negotiations about the content of the contract, which gave the opportunity to really negotiate, but the hazards Inherent in the production of silk goods gave rise to many uncertainties as to the carrying out of the contract. On the other hand, despite the affirmation of equality between the two negotiators in theory, their respective positions were not equal. Even though long term relations over several years between a merchant and the workshop head could arise,14 each order from the first to the second required its own contract which produced real negotiations about the price for transforming silk into fabric, delivery dates and the quality of the raw material, and the fabric itself. Two high points influenced the negotiation, firstly the reception of the raw materials and the manufacturing instructions given to the weaver, and secondly the delivery of the fabric to the merchant. The contract was written—or should have been—in the accounts booklet on which were generally written the conditions (the nature of the article, the price, delivery dates, sharing of the overheads for the setting up of the loom, etc.) As contract law stipulated, any failure by either of the parties implied paying damages (article 1142 CC). The pay was the equivalent of a salary (the price of transforming silk). So there were neither hiring for a specific period nor, unlike for industrial workers,

11 Real equality in numbers was only established by the law of 7 February 1880. 12 Rude, L’insurrection; L’Écho de la Fabrique (from now on EF), 12 (1832): 3. 13 The number of cases arbitrated tripled between 1807 and 1810, rising from 407 to 1218 (National Archives (AN) F12 934). 14 Worries about the future were very often felt both by the merchant who gave his orders, when he wished to be sure of keeping a weaver whose ability he recognised, and by the weaver who needed new orders to keep his/her looms working. contractual relations, tariffs and customs 151 a fixed time for starting and finishing the day: it was up to the workshop foreman to manage his work time to respect the delivery dates. The contract did not lead to hierarchical relations; the workshop head often suggested improvements in the designs the merchants gave them15 and were free to organize and share out the work inside their workshop. What is more, the quality control of the work was not just a single stamp after the work was finished; contacts between the two parties or their assistants, were maintained throughout the period of production of the fabric in visits to the workshops, and the fabric had to undergo two tests, when it arrived at the merchant’s, and finally when it was delivered to the customer who had given the initial order to the merchant.16 As well as this, unlike what could be seen in a factory situation, the contract concerning quality was not in the context of a company, but in this particular intermediate collective called the ‘Fabrique’, with its own rules and customs. All these practices were aimed at clearing up, at least partly, some of the uncertainties that the contract had left vague, uncertainties which were largely determined by the nature of the products, which were central to the work obligations. The silk goods which came from the looms of the Lyonnais workshops were luxury or semi-luxury items. Their production was not homogeneous. The big families of woven materials (plain, figured, novelties, velvets) were themselves subdivided into numerous categories defined according to their weaves and designs, the characteristics of their woof and weft frames and their chains. And these categories in turn included a large number of textures. For example in taffetas there were plain, glossy, shot silk, changeable taffeta, Marceline, sarsenet, ‘gros de Naples’ and poult-de- soie, etc. However, even the materials which were easier to produce, like plain taffeta demanded “extensive and rational experience”. The know-how of the weavers, like that of the other professions involved in the production of materials contributed to their final appearance, by their ability to get over the difficulties arising from imperfections in the material, difficulties in adjusting the loom or reproducing the designs. This know-how was an essential quality in production, inside which new types of material were constantly being created.

15 Montfalcon, Histoire, 17. Confirmed by Bezon who, among the people his dictionary of silks is addressed to, mentions “the workshop heads who throw themselves into serious and deep study of production, which they saw as perfectible”, Bezon, Dictionnaire. 16 Cottereau, Distinctiveness, 91. 152 pierre vernus

According to the goods to be produced the merchants thus had to go to the weaver he thought was most capable of producing the desired result. These skills were reinforced by the polyvalent nature of the workshops’ production. Unlike the Spitalfields weavers in London, who were specialized in one type of material, their Lyonnais counterparts were skilled in several types.17 Overall, in the framework of traditional silk manufacture before the depression of the 1880s, silk goods were made up of goods which could be qualified as ‘specially ordered’ or ‘interpersonal’, to use the categories suggested by Robert Salais and Michael Storper.18 From then on, several gradations around this pole could be seen, on the one hand towards ‘standard-specially ordered’ quality for the most common goods, and on the other, towards ‘immaterial quality’, for example when creating a new kind of material or a new article using combinations of thread, new patterns or new finishes. The inherent uncertainties involved in producing such a variety of articles took different forms for the merchants and the workshop heads. The first one concerned the mutual understanding of the needs expressed by the merchant and the suggestions put forward by the workshop head. It was partly eased by the fact that, because of its longevity, Lyonnais silk weaving formed a community which benefitted from particular advantages, similar to the localized industries mentioned by Marshall, and in particular a ‘shared language’ and specialized and diffuse skills, transmitted by the publication of technical journals and by being passed down from generation to generation.19 A second uncertainty came from the high price of silk. The merchant who gave it to the weaver suspected that the weaver might steal some for his own use. Weaving inevitably produced waste so they had to find out what proportion was acceptable. Another uncertainty concerned the finished quality of the goods. When the roll of material was delivered, how could a mutual agreement on whether the product met the requirements of the person giving the order be arrived at? The management of time was also important. For the merchant, if the production deadlines were not met it meant the risk of losses on the selling price or the can- cellation of the order by the client, especially for the seasonal novelties the price of which were high at the beginning of the season but fell later.

17 Cottereau, Distinctiveness, 91–93. 18 Salais and Storper, Les mondes. 19 Marshall, Principles; L’Écho de la Fabrique de 1841, 64 (1844), 1. contractual relations, tariffs and customs 153

For the weaver, if the raw materials were delivered late, some time after the loom had been set up, there was a slack period which meant a loss of income to be made up. Moreover the setting up of a loom, especially for the production of novelties and fashioned goods, could take several days or even weeks which implied high expenses for the workshop head. But if it was not followed up by an order allowing him to cover the setting up expenses, who would bear the cost? The workshop foreman? The merchant for whom the loom had been set up? Both, but in what proportion? Since there was no standardization in the process of preparing the silk threads and because of the diversity of their origins, the quality of the silks was an element which could greatly slow down the making of fabric and damage the final quality of the material. On his side the merchant had to have guarantees that he would get the necessary quality of work, and flexibility, especially if he aimed to have ‘tasteful’ goods or novelties woven for him. From this standpoint he was interested in the maintenance and renewal of the skills needed for his productions. The variations in the economic situation were a widespread risk especially since luxury goods were the first to be hit by a reduction in spending. The ebb and flow of fashion were added to these economic swings, without being totally in synch with them. After being a sudden hit an article could just as quickly be discarded. The organization of production limited the risks for the merchants whose investments in production were limited. On the other hand, they were borne more heavily by the weavers whose work and wages varied greatly with time, even though the combination of special- ties mentioned earlier partly compensated for their effects.

Shifts in Markets, Types of Production and Personnel

By unsettling the old practices, the shifts in the markets and the trade cycles increased the sources of uncertainty. In fact the strong growth in the British and American markets to the detriment of the traditional markets of continental Europe brought with them an acceleration of the rhythm of trade and great changes in the rhythm of work in the silk production centers. During the last few years the way of doing business has completely changed. In the past an order was given six months in advance. We used to work for Germany, France and Spain and we knew nothing of the great deals with America and England. The merchant used exactly the same number of looms throughout the year. He never varied more than ten or twelve: his 154 pierre vernus

clientele was stable. Today it is the opposite; the price of silks, the designs, the colors, etc. change twice per season, the orders have to be delivered at thirty days, six weeks, two months at the most. Among all these constant changes the producer has to adapt; the number of looms may go up to 300, or go down to about 150 according to the extent of his capital, but always in this sort of proportion. What effect do these ups and downs produce? He can never guarantee a worker a fixed number of days of work. The worker sometimes works for six to eight producers so that they all spend their lives running after each other. On the one hand, the worker often delivers his loom, set up in a certain way, to the first producer he finds, which forces him to totally change the setup of his loom and finds himself faced with ever renewed overheads. On the other hand, to finish on time an order needing forty looms, the manufacturer is obliged to look around for a fortnight, to knock on two hundred doors and hire workers left, right and center, who often have never worked on the article they are offered.20 The expansion of the new Anglo-Saxon markets also meant a change of clientele. The dynamism of the demand from the middle classes overtook that of the old ecclesiastic hierarchy or aristocrats. This encouraged the renewal and diversification of products which contributed to shaking up the methods and practices applied to the older types of production21 and also led some traders to try to change the rules governing their relations with the weavers. For example, the rapid changes of materials on the looms brought the question of the setting up of the looms to the fore.22 Finally, these developments tended towards the disappearance, among the main merchants, of the desire for regulations which had been expressed up to the first half of the 1820s. The new generation of merchant-producers seems to have been much more amenable to the principles of ‘laissez faire’ capitalism than the previous one, both because of the spread of the ideas of economic liberalism and because the old rules hampered the adaptation to changes in demand. What is more, the sharp increase in the numbers of merchants (which had doubled between the Empire and 1827, before settling to between 400 and 450), intensified local competition. Even though the biggest merchants were in a comfortable financial position which helped them to weather the crises, the financial situation of a good number of the new arrivals, much tighter, forced them in slack periods to sell off their stock to

20 EF, 46 (1833), 5. 21 Cayez, Métiers. 22 Benoît, Confession, 66. contractual relations, tariffs and customs 155 honor their debts and lower their prices to be in favour with the buyers.23 This movement forced itself eventually on their competitors and affected the price of transforming silk goods. Besides this, the renewal among the ranks of the merchants undermined the traditional processes of self- regulation and the internal social constraints in the ‘Fabrique’. The unease brought about by these changes soured the relations between the merchant-manufacturers and the workshop heads and this grew with the crisis at the end of the Restoration. The workshop heads complained of abuses perpetrated by the merchant-manufacturers and even though they were in a privileged position compared with other groups of workers, their position remained precarious and suffered the consequences of periodic unemployment. They were deeply affected by the difficult economic situation, especially since the drop in wages also affected the price of figured products.

Imbalance of Positions

The weavers were all the more sensitive to the mounting uncertainties concerning their relations with the merchants that the theoretical equality of contract set by contract law or by economic liberalism became in fact an imbalance of positions. In fact, in the eyes of the workshop heads, the argument upheld by many merchant-manufacturers, that the workshop heads were free to refuse a salary that was too low, was simply a trap. From the last decades of the eighteenth century the master weavers argued that: We only have the appearance of this freedom; no rule obliges us to accept work for which we would be underpaid, but the most implacable law, that of need, imposes on us the cruel reality; our workshops are not fortresses which it is easy to bring down by starving us out; we have to live every day and work every day.24 Beyond this, the prices paid for transforming silk greatly influenced the length and intensity of the weavers’ work because in order to counteract the fall in prices they had to work longer. Besides, the merchant had at his disposal various means which allowed him to transfer part of the production costs to the workshop head. He

23 Montfalcon, Histoire, 35, 48–49. 24 Godart, Ouvrier, 262. 156 pierre vernus could, for example, provide substandard threads, make the weaver pay certain costs which he had paid himself before, or extend the payment date. On the other hand during slack periods the merchant-manufacturer found himself in a better position to demand earlier delivery, which forced the weaver to lengthen his workdays if he wanted to avoid late- ness penalties. But weaving at night increases the risk of bad work due to fatigue, and it is more difficult to distinguish shades of color or see traces of smoke or grease. The workshop heads were affected in various different ways by this inequality in contractual relations. In fact as a group they were riven with splits on several levels. The first division was based on the hierarchy of profits and wealth. At the end of the 1850s L. Reybaud distinguished a tiny minority of the privileged who owned their house or who had a capital of 20–30,000 francs. Below them were the ‘majority of good workshop heads’ who managed to earn 3–4,000 francs net when work was continuous. The third level consisted of ‘the less rich heads’ whose takings did not exceed 2 000 to 2 500 francs. The base of this pyramid was made up of ‘the majority of heads who,’ because they owned fewer looms or their equipment was not up to scratch’, who earned less than 2 000 francs, which could go down to 1 200 or even 1 000 francs. Furthermore, although there was a certain diversification in the production of the workshops, there was a clear difference between the weaving of plain articles and fashioned goods. The amount paid for the former was sometimes three or four times lower than that for the latter. However the production of fashioned goods was very dependent on fluctuations in fashion. They stopped before the plain fabrics and they often started again afterwards, and during the setting up, the loom was idle, to the extent that many foremen who were capable of doing fashioned goods preferred to produce plain fabrics. Besides, a gender difference in tasks appeared; the weaving of light plain cloth, the lowest paid, was gradually given over to the women workers, often daughters or wives of the foremen, and whose pay was often half that of the men.25 As one of them pointed out, for the weavers this imbalance also fluctu- ated with the general economic situation: If there are only a few looms available [the merchant-manufacturer] pays a salary for the most expensive articles, but on the other hand, if there were a lot of idle looms he pays less because he can choose from a large number of

25 Reybaud, Etudes, 374–384; Villermé, Tableau, 168–169. contractual relations, tariffs and customs 157

looms in different workshops, so the workshops compete to hire out looms at ever lower rates. It is the total application of supply and demand.26 This situation was very unpopular with the weavers, even more so as they considered it an attack on their dignity. Their demands for equality stemmed from the principles of the Revolution, reactivated by the 1830 and 1848 revolutions. But it was also nourished by the memory of the conflicts between master workers and merchants in previous centuries and the fact that in the old community the workshop head had benefitted, on an equal footing with the merchants, from the title of ‘master’.27 It was a memory which brought with it nostalgia for a mythical golden age of the ‘Fabrique’, of a community of small independent producers crushed by the deceit of the merchants. To give a framework to contractual relations and reduce the social imbalance they suffered from, the weavers formulated two major demands: the setting up of a ‘minimum tariff’ which would limit the fall in the payment for fashioned goods during slack periods, and the establishment of regulations to put an end to the injustices they claimed to be victims of.

Reducing the Social Imbalance: Tariffs and Information

The weavers justified the setting up of tariffs by the social inequality they claimed they suffered from. It was true that this imbalance did not guarantee freedom of negotiation between the parties. The weavers put forward the idea that conventions between two parties are legal only on condition that they are free. So it was up to the authorities to step in to restore this freedom by restoring social balance. The tariff was to stop salaries from going below the workers’ real needs.28 Through collective organization, better communication of information, often difficult to access for weavers

26 Benoît, Mémoires, 44. 27 The booklet of Joseph Masson-Sibut, a workshop head and member of the founding group of the cooperative movement at the origin of the mobilisation of the silk workers at the beginning of the 1830s, bears witness to the transmission of this memory (Mois- sonnier, Maurice. La révolte des canuts. Paris: Editions sociales, 1975, 175). See also EF, 40 (1832), 2–3. 28 EF, 39 (1832): 2. This thesis, which had already been developed before the Revolution, was taken up again in 1860: “The freedom of convention [. . .] only exists under conditions of equality; it does not exist and will never exist between the worker, waiting for his salary in order to live, and the merchant, speculating on his capital”. (ACC Lyon, documents of the ‘Condition des Soies’, petition submitted to the Emperor during his visit to Lyon, 1860). 158 pierre vernus cut off in their workshops, would also contribute to correcting the imbalance between workshop heads and merchants.29 The demand for tariffs by the Lyon workshop heads since the last quarter of the eighteenth century remained central throughout the following century. It was closely linked to the payment for transforming silk, which remained in force both for the urban looms and those in the country and later in the weaving factories which flourished in the 1880s and 1890s. In Lyon itself the demands for tariffs were the source of many work stoppages in other areas of work, particularly in the hat industry and the building trade.30 We also know that they were the source of numerous conflicts throughout France and continued with the large number of wage demands in the strikes during the last 30 years of the nineteenth century.31 In the Lyon ‘Fabrique’ the tariffs which were demanded and sometimes granted were for a ‘minimum wage’ for the difficult periods and aimed at avoiding the drop in wages for transforming raw silk, particularly in plain fabrics where price competition was rife, plain goods and some popular novelties. These tariffs were a point of reference. They were negotiated more or less collectively. Interpersonal bargaining between the workshop head and the merchant could also use it as a reference.32 The tariffs were generally a list of variable length of fabrics grouped in families (plain, figured, velvets, handkerchiefs, shawls etc.) subdivided into types (for the plains: (taffetas, satins etc.) and subtypes. Each fabric was defined by a certain number of technical characteristics. For example, for the plains, the width of the ells, the number of threads which made up this width, the type of chain and the quality of the weft were taken into account. For fashioned goods it was the number of repetitions of the design in the width of the fabric or the number of wefts which followed each other regularly (the lats), the nature of the threads, an estimation of the number of passages of the woof threads (the coups) per ell. The amount paid per length or per article (e.g. for handkerchiefs). The aim of these tariffs then was to work out a minimum payment for the workshop head. From a ‘conventionalist’ point of view, we would speak of a ‘formal investment’33 aiming at standardising the criteria for

29 Godart, Ouvrier, 247–271. 30 Rude, Insurrection; Festy, Le mouvement. 31 Bourgin and Bourgin, Le Régime; Perrot, Les ouvriers. 32 Cottereau, “Désincorporation”. 33 Thévenot, “Les investissements de forme”, 21–71. contractual relations, tariffs and customs 159 paying the weaver and reducing the uncertainties about the definition of the product. In fact the wage wasn’t paid for a certain length of time but for a certain length of fabric. Before agreeing on the amount, it was necessary to agree on the nature of the object. But because Lyon’s textile production was so varied, and that this meant many degrees of difficulty in the production of goods, it was essential to fix representative indica- tors of these difficulties, or at least part of them (Length, width, number of prints etc.). The question of tariffs shows how the positions of the groups involved changed, and indicates that the variety of interpretations of the new post- revolutionary legal corpus, which could be seen in the central administration itself, opened up possibilities which were only closed at the beginning of the 1830s. Prolonging a series which opened with the Revolution, several tariffs were applied, at least momentarily, under the Empire and the Restora- tion.34 However, contrary to the process of wide consultation and collective formulation carried out by the workers in 1789, the tariffs of the first 20 years of the nineteenth century were rather the fruit of deliberations by the notables. For example, in 1811, the “tariff additional to the minimum payment for fashioning of fashioned silk goods”, which followed a tariff established in 1802 concerning plain fabrics,35 was passed during a meeting of the Chamber of Commerce, sitting as the consultative chamber of the chamber of arts and manufacture, in the presence of the municipal authorities, members of the ‘conseil de prud’hommes’, three important merchants and the ‘préfet’ who chaired the meeting. When the tariff was passed it immediately became a prefectural decree. In December 1817 in a petition, rich merchants denounced their colleagues who “took advantage of the difficult situation of the workshop heads” and demanded the reinstatement of the 1811 tariff.36 A municipal

34 Cottereau, “Désincorporation”; Cottereau, “Fate”. 35 It had been adopted on 10 frimaire year XI (1st December 1802) during a session of the ‘conseil des prud’hommes’ in the presence of ‘préfet’. A petition by merchants of plain fabrics, undated but probably written at the end of 1806 or the beginning of 1807, asked the council to “have the bill renewed and publicised so that no one could plead igno- rance” (AM Lyon, 784, Wp6, tariffs for transformation of rolls of silk). It was the ‘conseil des prud’hommes’ which had asked that the tariff in use for plain fabrics should be sup- plemented. (Archives départementales (AD) du Rhône, U 134, conseil des prud’hommes Lyon. Soieries. Président, Vice-président nomination ou élection 1822–1858, letter from the chamber of commerce to the ‘préfet’ of, 22 June 1811). 36 A similar approach had been used at the end of 1806 or the beginning of 1807 concerning the 1802 tariff. 160 pierre vernus bye-law dated 6 February 1818 “modifying the tariff for fashioning silk goods” of 18 June 1811 “satisfied their wishes.”37 So the influential merchant-manufacturers accepted tariffs. It is true that the 1817 petitioners underlined that the price for fashioned goods fixed by the 1811 tariff gave the worker just enough to renew his capacity to work.38 Moreover these prices could be lowered.39 So it was not a question of stopping the variations in price due to the law of supply and demand, but giving it a framework. However, beyond any charitable considerations, these measures were also concerned with regulating competition among the merchant-manufacturers themselves. According to the signatories of the petition, the payment of salaries below the tariff represented unfair competition towards those who respected the tariff and, they wrote, “If these differences in price were not curbed it would have fatal results for our town, since the prices fixed by these tariffs, which have not been abrogated, leave the worker below the poverty line, and by lowering them we would be forcing them to leave the straight and narrow, to beg or to go abroad.”40 For them, the functioning, the cohesion even the survival of the ‘Fab- rique’ were at stake. Faced with misery, the workers could be tempted to steal part of the raw materials they had received to bolster the parallel economy (which was called ‘stealing an ounce’). This would affect the cost price of materials negatively, or worse; if they set up shop abroad, they would transfer their know-how and reinforce the foreign silk centers whose competition was a major source of worry for the local professionals and the local authorities. On their side these authorities and the ‘préfet’ guaranteed and ratified these documents while occasionally wondering about their legality. Thus, in 1811 the chamber of commerce recognized that no law explicitly authorized tariffs, only to underline immediately that a ruling from the Royal Council on 5 September 1789 had authorized:

37 AM Lyon, 1 I 11, arrêtés, ordonnances, règlements. 38 AD Rhône U 134, ‘conseil des prud’hommes’ Lyon. Silks, president, vice-president nomination or election 1822–1858, letter from the chamber of commerce to the ‘préfet’, 22 June 1811. 39 In 1818, at the request of merchant-manufacturers in gauzes, crêpes and plain cut velvets, a cut in prices for the transformation of these articles was proposed by the ‘conseil des prud’hommes’ “given the fact that they [the merchants] found it impossible to with- stand the competition from the large number of ‘factories’ of this type which have been set up in the neighbouring districts where manpower is cheaper” (AM Lyon, 1 I 11, municipal ordinance, 10 february 1810). 40 AM Lyon, 784 Wp 6, petition from merchant-manufacturers to the president of the Lyon ‘conseil des prud’hommes’, 11 December 1817. contractual relations, tariffs and customs 161

“the setting up of a mixed commission of master workers and master merchants with a brief to provisionally fix the prices paid for fashioned goods”, and that “as soon as order had been restored in the Lyon silk industry by the ‘conseil des prud’hommes’, the Council wisely renewed its use, to which all the honest merchants quickly agreed. Today these tariffs can be considered as family law which constrains, by the pressure of opinion, those who would be tempted to speculate on the distress of the worker. They can also be used as a guideline for the arbitrators in ruling on any disputes which may arise between the master and the worker over the price for fashioned goods.”41 In 1819 the question of the legal status of these tariffs was raised once more following a conflict not in the silk ‘fabrique’ but in the hat industry. It seemed at that time that the doctrine of the Ministry of Home Affairs, which had been consulted several times, was ambiguous and that it could, under certain conditions, tolerate the proclamation of a tariff.42 However, the legality of tariffs appeared less and less clear to the administrative authorities and during the 1820s no new tariff was ratified or modernized in the silk trade even if, in 1822, the municipality of Lyon passed a statute “concerning the fixing of a minimum wage in the hat industry.”43 The events of November 1831 clearly marked the end of the authorities’ tolerance of tariffs. Even if the uprising resulted from the indignation provoked by the attitude of a majority of merchants who rejected the signed agreement, it is useful to recall that the government encouraged them in their refusal. The 1830 revolution and the change in the régime reinforced the demand for equality, freedom and dignity on the part of the workshop heads, who had taken an active part in the revolt in Lyon and encouraged their collective mobilization.44 Resuming the process of 1789, the workshop heads engaged in a process of collectively drawing up a project for tariffs which finally served as a basis for bipartite negotiations under the aegis of the ‘préfet’. A large number of merchants refused to agree to the project and balked at applying the tariff which was finally signed. Some of them laid the matter before the government which overruled the ‘préfet’ and forced him to point out that the tariff was only a “declaration on honour” with no legal status.45 Instead of tariffs the new ‘préfet’ proposed that the

41 AD Rhône, U 134, letter from the chamber of commerce to the ‘préfet’, 22 juin 1811. 42 Rude, Insurection, 78–79, 82–85, 99–101, 167. 43 AM Lyon, 1 I 8, municipal ordinance, 24 october 1822. 44 “Those who, from pride or sensitivity of character [the worker] considers with arro- gant disdain, will go back on their error and admit that the class they have so long despised is worthy of their respect and deserves to see its fate improved” (EF, 12 (1832), 2. 45 Rude, Insurrection. 162 pierre vernus

‘conseil des prud’hommes’ should establish a market price list of prices, in other words a statement giving the average prices for fashioned goods at a given moment for the different categories of fabric. When it became clear that to the workers’ minds this statement was the same as a tariff, the merchants on the ‘conseil des prud’hommes’ manoeuvred to block it and erase it from memory.46 From then on, in the name of the freedom of legally equal individuals to establish agreements, the authorities considered any tariff as a rule which at best had no legal value and at worst was illegal. What is more, in 1840 during the drawing up of Lyon’s municipal code, all the bye-laws and ordinances which had allowed tariffs were declared null and void and were discarded as “contrary to free trade.”47 This meant that in the case of collectively negotiated agreements between weavers and merchants, the local and national authorities refused to guarantee them and did not force recalcitrant merchants to apply them. The ability of the workshop heads to have this type of agreement respected depended on their degree of mobilization and organization. But from 1832, and more so from 1834 until the last years of the Second Empire, (with the exception of a gap under the second republic), the workers’ movement suffered from repressive policies of varying intensity but which overall reduced the silk workers’ ability to organise and protest collectively. It led some of them to become resigned and led others to adapt their forms of resistance and the types of their demands. After 1843, police pressure on associations relaxed a little, so the weavers formed groups corresponding to the specialized goods. It was soon clear that these groups served as cover for fighting funds when, the following year, four of them revived the practice of blacklisting the merchants who paid below a certain price for fashioned articles. But this kind of action was limited and was virtually repressed.48 The liberalization of legislation on the forming of groups during the 1860s (law of 25 May 1864), then administrative recognition of professional associations from 1868, prolonged under the third republic (law of 21 March 1884) allowed the weavers to organize collectively once more. The groupings were once more according to the categories of the articles. In 1869, when the raising of the prices for transforming the silk lagged

46 EF, 39 (1832), 2–3 and 50 (1832), 3. 47 AM Lyon, 1 I 5 to 9. 48 McDougall, Lynn. The Artisan Republic, 14. contractual relations, tariffs and customs 163 behind the renewal of industrial activity there was a vast mobilization in favour of a tariff. Groups of weavers drew up tariffs by category of article which were then submitted to the merchants concerned. These prices were accepted but, as one merchant-manufacturer wrote, it still had to be put into practice. Concerning the tariff, I haven’t changed my mind since you gave it to me. It is in your interest to keep it because if it didn’t exist, competition would turn to reducing salaries, and in periods of trade crisis we don’t know how far the reductions could go. The tariff puts all the manufacturers (i.e. the merchants) on the same footing and no one can do deals which harm wages. But to put it into practice, you have to be capable of getting it done. In order to do so the “Provident and Information Society for the work of the weavers in the Lyon ‘Fabrique’” was founded. It was made up of 14000 members in several hundred cells, grouped by specialty. The work of the cells was to get work for its members and compensate them if they moved. In fact, in their big meetings the weavers had refused the strategy of the general strike in favour of that of ‘displacement’ which consisted in refusing the orders of merchants who did not respect the tariff. If that happened, the Provident society gave the weavers the names of merchants who were looking for looms to be worked, and who respected the tariff, or it paid them compensation. In order to do this more effi- ciently the Providence society opened an information bureau. At the same time, in April and May 1870, delegations were sent to contact the weavers in the countryside, who had stopped work, to get a rise in the price of transforming. After the gap of the Franco-German war this system worked until the Society was dissolved by the authorities in 1874 during the ‘Moral Order’ period.49 This strategy could only be efficient if it was set up by an effective and representative organization with the go-ahead, or at least the benevolent neutrality, of the authorities and in favourable economic conditions. But although the organization reopened in 1876 under the name ‘The Weavers Federation’ the situation changed profoundly from the following year. The crisis of 1877 heralded a period of great change. The shift in demand towards light fabrics to the detriment of the beautiful rich fabrics which had made the name of the urban silk industry, and the acceleration of mechanization, spelt the inevitable and rapid decline of Lyon’s urban weaving. In 1889 there were only about 16000 looms, half the number in the

49 Mathé, aîné. Les tisseurs: 190; Maritch, Histoire, 226–228. 164 pierre vernus

1855–1875 period.50 The change in the economic situation brought about a fall in the prices of fashioned goods and the conditions of the weavers compared with that of the workers in the other local industries; the crisis of recruitment which had already been apparent, got worse. The average number of apprentices dropped and virtually disappeared in the 1890s. The workshop was reduced to the ageing household of the weaver.51 From then on, as the last great movement in 1885 showed, the workshop heads were no longer in a position to impose respect for the prices which had been negotiated, even if, although divided into two rival organizations, they managed to bring the merchants to discuss a rise in prices (in fact an arrangement below the tariff of 1869). But their negotiating partners manoeuvred to stop the list of prices, which had been settled, from looking like a commitment signed from organization to organization which could later be considered as ‘local practice’.52 The only thing that remained of the agreement was a list of ‘recommended prices’. The organization of a huge demonstration to try to make the forty or so recalcitrant merchants to give in, then the attempt to have the prices guaranteed by the ‘conseil des prud’hommes’, and finally, the demand for legislative sanctions, all failed. These movements revived the memories of the revolts of 1831 and 1834 and accelerated the putting-out of production outside Lyon, and with this accelerated the inexorable decline of urban workshop weaving.53

‘Mutualism’ and ‘Regeneration’ of the ‘Fabrique’

‘Mutual duty’ or ‘mutualism’ started in 182854 and aimed at being a tool for moral regeneration, aiming at opening the way for reforming abuses. At the outset this moral regeneration targeted the workshop heads group: “Walk with heads high against other abuses.”55 It was a secret society, divided into lodges of twenty members, with mutual links and was marked by the restrictions imposed by article 291 of the penal code. It was inspired by freemasonry and trade guilds, only recruited workshop heads, and had three aims; ‘information, help and assistance’, which meant encouraging

50 Lequin, Les ouvriers, 76–77, 82–91. 51 Cayez, Crises, 55–62. 52 The Association of the Lyon ‘Fabrique’, an employers’ association of merchant- manufacturers of silk goods had been founded in 1868. 53 Vernus, “ Pour une histoire”. 54 On the twists and turns of its beginnings, see Rude, Insurrection. 55 EF 1841 49 (1843), 2–3. contractual relations, tariffs and customs 165 the spread of information useful to the members, favouring mutual aid by the loan of materials and in difficult moments, helping members financially or morally, particularly for funerals.56 It started with just a few dozen members, but numbers grew rapidly from 1831. Mutualism was one of the numerous initiatives which the idea of association spawned from the 1820s.57 Its adherents saw in it a response to the atomization of the group of workshop heads, and more broadly to the forces of dissolution, in the front line of which was the selfish pursuit of personal gain, which threatened the industry as a community whose main victims, it seemed to them, were the silk workers. The more or less implicit ideal of this movement of workshop foremen was of a ‘Fabrique’ inside which the varied interest of the constituent groups could find a balance thanks to the action of joint commissions. Besides, the notion of ‘information’ which was fundamental for the founders of mutualism, was a real part of their project for regenerating the ‘Fabrique’.58 For them, the association should favour the circulation of information in the workshop heads’ milieu.59 It was one of the functions which the ‘Echo de la Fabrique’ gave itself. The journals which followed it carried on the same line. This is clearly visible in the offices of the ‘Echo de la Fabrique’ which set up a ‘special information bureau for the silk materials industry’ in Lyon. [. . .] The artisans and those who provide them with work are completely isolated from each other. The social position of one group and often the destitution of the other, impedes the contact without which the industry cannot prosper. The workshop heads and the workers experience the same isolation. From this innumerable problems arise when it comes to mounting new articles, and the enormous overheads which often overwhelm the workshop heads. They can be avoided or at least partly diminished, if everyone could turn to a common centre. The benefits that the ‘special information bureau for the production of silk fabrics’ would provide are innumerable; there, the workshop heads will find either workers and apprentices, or all the tools, harnesses and

56 Blanc, Histoire, 516–523. 57 Rosanvallon, Le modèle, 84–189; Sewell. Gens de métier, 274–281. 58 28 June 1828, the date of the durablefounding of Mutualism, was celebrated as the beginning of “Year 1 of regeneration” (Rude, Insurrection, 141). 59 The preamble of the rules for mutualism declared: “The aim of mutualism is thus between the founders and all those who will be accepted as brothers: 1° to inform each other, frankly and loyally, mutually and generally, of everything that may be useful and necessary for them in their professions; 2° to help each other by the loan of equipment as much as is in their power (. . .)” (Blanc, Histoire, 516). 166 pierre vernus

accessories for setting up looms according to the various articles; he will also find the means to get work without going all over the place knocking on the doors of the different establishments. The workers will be less afraid of being laid off because the special information bureau being the centre where all the requests for workers end up, they will know where to go and will not vegetate, going from one workshop to another looking for an available loom. But it is in the bad times that the information bureau will be most useful, because that is when there is less work and as a result, the worker is more likely to change workshops. The merchants will be able to get the number of workers they need for an order more easily. What we suggest has been tested by some commercial houses for whom we supplied the number of workers they needed in two days. So, in the interests of commerce and industry, a special information bureau for the production of silk goods will be installed from today in the offices of the ‘Echo de la Fabrique’. We will take care of I° the merchants’ requests for looms; 2°, the placing of workers in the different workshops according to the articles; 3° requests for placing apprentices; 4° the sale of looms, harnesses and accessories for all kinds of production, and finally all requests concerning the ‘Fabrique’.60 By reducing the cost of access to information and by guaranteeing better adjustment of the internal needs and resources in the ‘Fabrique’, the information bureau was supposed to reduce waste of time, to ensure the quality of production and in fine, to improve the global functioning of the ‘Fabrique’ which would thus be in a position to face up to foreign competition. The improvement in the circulation of information should also reduce the imbalance penalizing the workshop head in relation to the merchant in the contractual relationship, and partly contribute to balancing the positions of the two parties. This argument was clearly expressed in a petition launched in 1840 by the workshop foremen and the workers. “The merchant-manufacturer is informed of everything that happens among his colleagues; he knows the sales price of the fabrics, the time he spends gathering information is well used and produces its share in the profits”. On the contrary, for the worker, “every hour spent outside the workshop is time lost. Badly informed and driven by need, he must accept the conditions he is offered”.61

60 EF 18 (1832), 7–8. 61 Pétition. contractual relations, tariffs and customs 167

Fighting Abuses

Even if the workers saw tariffs as a way to compensate for their weak position by framing their contractual relations with the merchant-manufacturers, they could not reduce the abuses the merchant-manufacturers, or at least some of them, were accused of. In fact, fixing the price for transforming silk, covered by a tariff or not, only covered one aspect of the contractual relations linking the workshop head and the merchant. They included other parameters (amount of waste, overheads for setting up the looms, quality of the material, etc.) which affected the workshop head’s pay and conditions. For Pierre Charnier,62 these abuses stemmed from free competition and were the major cause of the Fabrique’s difficulties.63 Following Alain Cottereau, we should underline that although some of the decisions taken by the ‘conseil des prud’hommes’ in its general assemblies, or some of its judgments, set up a local quasi-legislation which sometimes revived certain elements of statutes from the eighteenth century, these practices were also continually submitted to criticism, more or less intensely depending on the period.64 When these attacks came from the merchant-manufacturers who could act collectively, they were considered by the weavers or their spokespersons to be abusive. At the end of the 1820s this question was particularly acute when, during the difficulties at the end of the 1820s, some merchants tried to free themselves from the traditional practices, especially as they could appear to be unsuited to a form of production which was constantly being renewed. In a series of articles in 1832, Joachim Falconnet65 denounced the most glaring of these abuses. The first one dealt with surcharges or modifications put on the workers’ account booklets by certain merchants or their assistants, taking advantage of the fact that the rule stipulating that the noting of all elements should be written by both parties on the accounts booklet of the other, was not respected. In fact it was quite common for the workers to leave their booklets with the merchant and pick them up

62 Workshop head born in Lyon in 1795, son of a grocer; founded the ‘Mutual society for supervision and information’ in September 1827. Member of the ‘conseil des prud’hommes’ 1832–1852. 63 Rude, “L’insurrection”; La Révolution de 1848 164(1938), 18–49, 165 (1938), 65–117, 166 (1938), 140–179. 64 Cottereau, Fate, 51–55. 65 Born in Lyon September 1798, son and grandson of workshop heads and himself a workshop head, founder and manager of L’Echo de la Fabrique. He was a member of the ‘conseil des prud’hommes’ 1832–1858. 168 pierre vernus only a week or two later, sometimes more. This period could be used to modify the written conventions in the booklet. Added to this the workshop heads sometimes failed, deliberately or not, to state the balance of accounts, which left him in a doubtful position towards the merchant. Moreover, some merchants only paid their workers on a fixed day and not on the day the roll of material was delivered. This practice favoured the merchant’s bank balance but weighed on the weaver’s who often had to down tools and go once more to the merchant’s office and leave his work for that time. It was particularly unpopular because the merchants on their side, if there was a delay in delivering the rolls of material, would charge a penalty on the price of transforming. Another big cause for recrimina- tion by the weaver was over the percentage of wastage accepted by the merchant during the making of fabrics. In fact some merchants, mainly for new articles, tried to question the practices which had been in place even before the Revolution, which gave a level of wastage of 3%, and recognised that the selvage depended on the weaver.66 The setting up of the looms was also a recurring source of conflict. For fashioned goods the preparation work for weaving took a long time, several days or even several weeks, and this imposed heavy costs on the workshop head. On the other hand, the manufacturer generally agreed to order a minimum number of rolls to be woven. But this practice was also questioned. Some merchants who had cancelled part of their orders, refused to compensate the worker for his overheads. Others, sometimes the same, attempted to make the weavers pay the cost of repairing the lacing of the cards used with Jacquard mechanisms, or even that of correcting mistakes in the piercing of holes, although the cards belonged to the merchants.67 The workshop heads demanded the establishment of stable regulations to counteract the abuses. There were two parts to this demand. First of all they wanted to keep the old customs against the attempts by certain merchants to question them, or more precisely against their practices which, by creating precedents, threatened to nullify the practices in place. So then it was a question of defining new practices for the new fabrics. At the beginning of the 1830s the ‘conseil des prud’hommes’ found itself at the heart of the weavers’ efforts to fight against the abuses. The

66 Selvage, or edge, the weaver had to make to adapt the fabric to the roller of the loom. 67 EF 15 (1832): 2, 16 (1832), 3, 17 (1832), 2–3, 18 (1832), 2–3, 19 (1832), 2–3, 22 (1832), 3–4, 23 (1832), 2–3. contractual relations, tariffs and customs 169 tribunal seemed, to the weavers, to be their main means of redress, especially since they had made progress on two points they had criticized. In February 1831 a petition criticized the fact that the general bureau’s meetings were held in camera. They called it ‘an occult bureau [. . .] favourable to the merchant-manufacturers’.68 The following month the meetings were publicized and, despite some attempts to question it, this lasted till the end of the second Republic. The weekly sessions of the general bureau attracted many people. The weekly minutes were published in the ‘Echo de la Fabrique’ and in the similar journals which followed it. Another criticism was of the numeri- cal majority of the merchant-manufacturers on the board. Following the revolt in 1831, the government’s only concession was on the organization of the ‘conseil des prud’hommes’. A statute of 1832 enlarged the electoral college of workshop heads and increased the number of ‘conseil’ members from the silk working industry.69 So, after the rejection of the 1831 tariff, the workshop heads put their hopes in the new reformed ‘conseil’. They were very active in the April 1832 elections, hoping that the institution would define regulations outlawing the worst of the abuses. But, although during the first few months of the new ‘conseil’s activity the path of collaboration and conciliation of the interests of the workshop heads and the merchants seemed possible and allowed the adoption of decisions favorable to the weavers, tensions appeared as early as the last few months of 1832. It was particularly the case in the 1833 elections, where merchant-manufacturers were elected who were not disposed to grant concessions and who sought to block the working of the ‘conseil des prud’hommes’.70 The relations between the two major elements of the institution finally calmed down and the ‘conseil des prud’hommes’ was able to continue its work. Its activity reached its high point in the middle of the 1840s with more than 6000 cases heard in privy session. This number dropped later because of the fall in the number of urban workshops due to the development of rural weaving. In the first half of the 1840s about 40% of cases were between merchants and workshop heads. It is worth noting the stability

68 Rude, Insurrection, 231–241. 69 It still did not establish equality of members since the merchant-manufacturers went up from 4 to 9 and the workshop heads from 3 to 8, which meant, for the council as a whole, 14 merchant-manufacturers against 11 workshop heads instead of 9 and 6. More- over, only the heads of workshops who owned more than 4 looms had voting rights, only 20–25% of the total. 70 Bezucha, “Aspects”. 170 pierre vernus of a certain number of ‘conseil’ members, especially among workshop heads where some members from the 1832–1834 period continued to sit during the 1840s and even till 1852 for J. Falconnet and 1858 for Pierre Charnier. The questioning of certain customs continued to bring about moments of tension, like in 1843 when a group of merchants contested once more the assigning of selvage, or in 1860 about ‘peignes’ (cards),71 the percentage of waste and the setting up of looms. As early as 1831 the weavers asked the ‘conseil des prud’homes’ to set up a ‘fixed jurisprudence’, which, being written down, would stabilize their acquired positions or confirm the older customs, and would play the role of regulations for the ‘Fabrique’ It was from this standpoint that Marius Chastaing72 inaugurated the publishing of the minutes of the meetings of the bureau of the ‘conseil’ in the first edition of the ‘Echo de la Fabrique’ in October 1831. One year later he presented them in the form of ‘notices of jurisprudence’, in which precedents were presented before the final decision. Then, in the first four months of 1833, he managed to persuade the ‘conseil’ to have its decisions systematically noted as minutes by the clerk, whereas this was not the case before (at least, it would seem, before Octo- ber 1832) unless it was requested immediately. At the same time Chastaing started to collect the ‘conseil’s decisions “so that the accused can use them if necessary and mention them as precedents. That would partly make up for the lack of a fixed jurisprudence, which we would have liked, and which we continue to demand and which the ‘conseil des prud’hommes’ has a duty to set up”’.73 He took up this work again, with a short break in July 1835, in ‘La Tribune Populaire’. His ‘notices of jurisprudence’ were collected in 4 then 5 series74 and were aimed at serving as a basis for a ‘code of the council’ or of ‘the silk workers’.

71 A certain length of chain which could not be woven at the end of each roll. 72 A journalist, Chastaing was the son of a bailiff who had himself followed law studies (Popkin, Jeremy D. Press, revolution, and social identities in France, 1830–1835. University Park: Pennsylvania state university press, 2002, 143). 73 EF 2 (1833), 12; 4 (1833), 1–2 and 9 (1833), 69–70. 74 When, from October 1832 Chastaing started to collect the council’s decisions and to publish them in ‘L’Echo de la Fabrique’, he grouped them in 4 categories (which he called series): first series; the workshop heads in their relation with the merchants; second series; the workshop heads in their relation with the workers or journeymen; third series; the workshop foremen in their relation with pupils and apprentices; fourth series; the workshop heads with various other persons. Later Chastaing stopped working for ‘L’Echo de la Fabrique’ in 1833. He took up publishing his ‘notes of jurisprudence’ of the Lyon ‘conseil des prud’hommes’ in a new paper, the Tribune prolétaire in July 1835. The fourth series was then called: Workshop heads in their relations among themselves or with various people and different questions concerning skills and others. He created a fifth series contractual relations, tariffs and customs 171

In 1843 in his ‘Information directory for the Lyon Silk Industry’ J. Falcon- net carried on Chastaing’s work, re-publishing his ‘notices’ and adding his personal notes. For him it was also the first draft of what should be a fixed jurisprudence for the industry, like those which already existed for civil and commercial affairs Despite moments of tension around the question of wastage it seems that there was little change on it during the 1840s.In a new edition of his work in 1849–1850, Falconnet, who, as a councilor in the ‘conseil des prud’hommes’ was well placed to judge, thought it unneces- sary to reprint the ‘notices’ since the ‘jurisprudence of the council has not changed much despite some changes’.75 However, the idea of an industrial code continued to progress. In 1860 a petition submitted to Napoleon III with 561 signatories, of whom the first were seven newly-elected worker counselors, took up the slogans both of an ‘industrial code’ (for the silk industry), which would stabilize practices, and also of fixing the correspondence with the commercial code. This petition was not heard, but from the end of the 1860s the merchant- manufacturers, worried by the rise in collective action by the workshop heads and keen to put a brake on what seemed to them to be ‘abuses [which] infiltrate[d] the’ conseil’s decisions were attracted by this desire for codification,76 and put it into practice by publishing ‘Practices of the industrial tribunal of the town of Lyon for the silk industries’ which was widely read by the merchant manufacturers. This collection of articles dealt successively with the relations between the merchant-manufacturers and the workshop heads, between the latter and their workers and apprentices, this group making up the largest part of the book, and then the relations of the manufacturers and workshop heads with the other professions which took part in the transformation of silk and the production of fabrics (reelers, warpers, throwers, dyers). So the structure of the work was similar to that of Chastaing then Falconnet. Afterwards, the modifications of the customs were negotiated between the merchant-manufacturers employers’ federation and the workshop foremen, like in 1885. A new modified version of these works was published in 1891.77

entitled: Merchants and their relations with each other or with people other than workshop heads. 75 Falconnet, Indicateur, 1843 and 1849. 76 Pariset, Histoire, 328; Enquête, 371. 77 Usages du conseil; Recueil des usages. 172 pierre vernus

Conclusion

Behind the apparent stability of the organizational framework of the Lyon silk ‘Fabrique’, the relations between merchant-manufacturers and workshop heads as well as their relative positions within the framework of these relations, varied greatly. Between 1744 and 1786, the master weavers had noted the domination of the merchant-manufacturers over the community imposed by the 1744 statute. The abandoned their claim for a return to equality in the management of the community and reoriented their demands, calling for the setting up of a compulsory general minimum tariff designed to put a frame round the merchants’ economic power.78 The Revolution disturbed the projects for restructuring the ‘Fabrique’, which, like the one presented by the intendant Terray after the 1786 uprising, aimed to cancel the intermediary status of the master workers and reduce their status to that of the journeymen. In fact, despite the disappearance of the community of the Ancien Régime, on the legal plane the special position of the workshop heads was confirmed by the law of 1806 which created the Lyon ‘conseil des prud’hommes’. Moreover this institution revived, although only partially, the demands for joint management of the ‘Fabrique’ by the merchants and workshop heads. Added to this, the political situation, the new principles for organizing society, in particular equality and the prevalent lack of manpower during the first years of the nineteenth century reinforced the position of the weavers towards the merchants. However, at first, contract law which framed the contractual relations between merchants and workshop heads left open the debate about the meaning of ‘interpersonal negotiation’. The interpretation given by the weavers’ spokespeople, according to which contractual relations could only be freely established if the positional imbalance between the two parties was balanced by barriers limiting the power of the dominant party, could still be heard at the local level up until the 1820s. It was even supported by the most influential, if not the majority, of the representatives of the merchant-manufacturers, who saw it as a way of regulating competition inside their group and of avoiding that behavior based on purely individual interest should lead in the long term to the disintegration of the ‘Fabrique’ itself. As well as that it was tolerated if not supported by local and government authorities. But the progress of liberal economic

78 Garden, Lyon, 572–582. contractual relations, tariffs and customs 173 ideology led the weavers’ idea of mutual agreement to become less and less acceptable both by the merchants and the authorities. The authorities henceforth refused to approve any tariff or even nullified any such agreements. This was the cause of the 1831 events. Having said that, the confrontation between merchants and weavers was not only around tariffs; it also concerned practices aimed at clearing up certain aspects of contractual relations of work which caused uncertainty and sometimes greatly influenced the wages earned. Several parameters contributed to fashioning the strategies and limiting the scope of means available to the different parties to push forward their demands. First of all the institutional, legal and also political frameworks defined the formal space and the types of action the workshop heads could use. The variations in this framework opened or closed room for mediation or negotiation, helped or hindered compromises and alliances and influenced their form and content. They allowed or prohibited certain strategies of the parties and conditioned their ability to attract political support. Moreover, the state of the job market and the direction of demand constituted parameters which weighed on the workers’ ability to impose a certain amount of control over the rules governing their pay as well as the intensity of their work, and to balance their position in the process of collective or individual negotiation which led to a contractual relationship. The use by the merchant-manufacturers of rural manpower, then from the 1880s, the acceleration of the mechanization of weaving encouraged by the direction of the markets, in the long term weakened the urban weavers position by reducing their ability to create a balance of power which could improve their position relative to the merchants.

The Circulation of Commercial Manpower in an Indian Worldwide Trading Network in the Early Twentieth Century

Claude Markovits

The existence of a significant worldwide circulation of commercial manpower from India has been little noticed in the existing literature. In India itself, those employed in shops and other commercial establishments are largely absent from standard accounts of labour history, although they probably represented and still represent altogether one of the largest segments of the overall labour force. Labour historians have tended to focus their attention almost exclusively on agricultural and industrial labour, to the detriment of a category that suffers from an ambiguous socio- economic position, straddling as it does the boundary between lower middle-class and working class, and has attracted little attention from colonial rulers, resulting in a paucity of source materials. As regards the circulation of Indian labour outside India, the existence of a sizeable movement of commercial manpower is also generally overlooked by historians of the South Asian diaspora. Their focus has been mostly on the movement of indentured and kangani labour (two forms of contract labour) that flourished between 1834 and the late 1930s, resulting in the emigration of millions of Indian workers towards the sugar-producing countries of the Caribbean (Guyana, Trinidad, Surinam, Martinique, Guadeloupe) and of the Indian Ocean (Mauritius, Reunion, South Africa), as well as the tea plantations of Ceylon and the rubber plantations of Malaya. These migrations are particularly well documented in the colonial archives, and have given rise to a vast secondary literature.1 However, in addition to this migration of workers strongly regulated by the Indian colonial government in coordination with the imperial government (inasmuch as 90% of the indentured workers and 100% of the kangani migrants made their way to British colonies, less than 10% going to French and Dutch territories), there has existed another movement, much smaller in volume but nevertheless significant, of traders and commercial employees. That particular

1 On this migration, see, amongst a vast and growing literature, two classics: Tinker A New System and Northrup, Indentured Labour. 176 claude markovits stream has origins that go back to the medieval and early modern period, but was considerably amplified in the nineteenth century when British India was integrated into the world capitalist economy and new trading circuits were created linking India with the Far East (China in particular), while already existing circuits connecting India with the Middle East and East Africa were reorganized and significantly expanded. It remaining largely unrecognized in the literature has to do with the fact that the colonial government of India did not organize it directly or even monitor it closely, with the result that there are few official sources available. In a previously published article,2 I used some little-known published material to evaluate the importance of that movement, that belongs to the universe of circulation rather than of permanent migration, and I came up with a global figure of some 1.5 million men (it was an almost exclusively male stream) for the period 1830–1950. I found that it was characterized by a great diversity in the geographical origins of its participants, most of whom however hailed from coastal regions of the subcontinent, in particular Gujarat on the West Coast and Tamilnadu on the South-eastern coast. Surprisingly, I came across evidence that some inland areas and towns also sent commercial migrants abroad on a significant scale. This was the case in particular of the town of Hyderabad in the province of Sindh (not to be confused with its better-known homonym, Hyderabad- Deccan), that belonged to British India from 1843 to 1947 and was then made a part of Pakistan. Although situated inland on the lower valley of the Indus at a distance of 150kms from the seaport of Karachi, Hyderabad was the cradle of a particularly dynamic trading network that, between 1860 and 1914, extended its operations practically to the entire world.

The Rise of a Worldwide Trading Network

I have given in a book3 a detailed historical survey of the rise and growth of that network, which I shall here briefly summarize. The origins of the network go back to around 1860, when some Hindu merchants from that middle-sized town with a population of 40 000 (a majority of them being Hindus, while Sindh as a whole was a Muslim-majority province) started going from Bombay to Egypt on annual voyages. In Egypt, they sold a certain kind of craft goods produced by the (Muslim) artisans of Hyderabad

2 Markovits, “Indian Merchant”. 3 Markovits, The Global World. the circulation of commercial manpower 177 and some neighbouring Sindh towns, which had become known in Bom- bay in the 1850s as “Sindwork”. Hence the appellation of “Sindworkies” under which these traders were generally known. They benefited from the fact that their arrival coincided with the beginnings of modern tour- ism in Egypt in the form of the first tours organized by Thomas Cook. Their customers were thus mostly affluent European and North American tourists. In the following five or six decades, that network transformed itself into a worldwide phenomenon. From Egypt, the Hyderabad merchants gradually extended their trading operations to the entire Mediter- ranean basin, and, from Gibraltar, penetrated into the Atlantic, reaching West Africa, the Canary Islands, Central and Southern America, and even making forays into North America. Other merchants of the same town, departing from Calcutta and Madras, reached Southeast Asia, and Singa- pore emerged as an important hub for them. They soon reached China and Japan, where, from the 1870s onwards, they started buying silk goods and various artisanal products that gradually replaced the productions of the small workshops of their native province. By 1907, it was estimated by a British official4 that there were approximately 5 000 so-called Sindwor- kies dispersed across the world in locations as widely scattered as Kobe in Japan and Panama in Central America. They specialized in the sale of silk goods and curios to a mostly European and North American clientele of affluent travellers and tourists, and had establishments at the various ports where the great transoceanic steamers alighted. These Sindworkies however were far from being a homogeneous group of merchants, as the standard literature on “trading diasporas”5 would have us believe. They actually formed a highly diversified community in which capitalists, small and big, who were proprietors of family firms of varied size, salaried managers, and different kinds of employees were represented. The core of the network consisted of a certain number of family firms based in Hydera- bad-Sindh, a hundred in all, of which five or six were really big (the largest being the firm of Pohoomull Bros that, circa 1911, had more than twenty branches outside India), the rest medium or small-scale. There were also small independent merchants, who peddled goods supplied by firms for one season, or purchased in India directly from artisans. Firms however accounted for the bulk of the turnover of the network, and the vast

4 Aitken, Gazetteer. 5 The term “trading diaspora” was first used by the Africanist scholar Abner Cohen in his seminal article “Cultural Strategies in the Organization of Trading Diasporas”, in Meil- lassoux, The Development, 266–278, and further popularized in Curtin, Cross-Cultural. 178 claude markovits majority of the circulating men were in the employ of a firm. The principals of the firms themselves generally resided in Hyderabad and only con- ducted inspection tours of their branches abroad at various intervals. The circulating Sindworkies belonged to three main categories: 1) the managers, some of whom were working partners of the principals, others salaried employees (and the proportion of salaried managers increased as the network expanded to new areas); 2) the shop assistants, by far the largest category; 3) the so-called “servants”. There was actually a fourth category, of pedlars, but its importance diminished overtime. For the purpose of this paper, I shall ignore the managers, since they uncomfortably straddle the boundary between capital and labour, as well as the pedlars, about which too little is known, and concentrate on the latter two groups, who were clearly workers, although of very different origins and status. My information about the commercial workers comes mostly from documents of a judicial nature, files concerning disputes between Hyderabadi proprietors of commercial firms and their employees that came before the Cairo consular court and other British consular courts in Egypt during the period 1900–1925, which I traced in the Foreign Office Records at the Pub- lic Records Office (now the National Archives of the UK). The reason why disputes between Indian employers and employees in Egypt were referred to the consular courts is simply that Indians in Egypt, being British subjects, were subject to the jurisdiction of the Consular Courts, in the matter of disputes with other British or foreign subjects, in accordance with the stipulations of the “Capitulations” under which Egypt, still nominally a part of the Ottoman Empire (till 1914) fell. The files contain in particular a certain number of contracts of employment signed in Hyderabad between employers and employees, which figured as exhibits in the court cases. These contracts define with a certain degree of precision the conditions under which employees of commercial firms travelled between Hydera- bad and Egypt, and their terms of employment in Egypt. My sample of cases is small, but I have no reason to think that it is not representative. Apart from the fact that Egypt was a most important destination for Sind- workies prior to the First World War, being really one of the major hubs of that worldwide network (the other ones being Gibraltar, Singapore and Hong Kong-Canton),6 the specific location of the cases has probably no

6 In 1915–16, out of a total of 572 certificates of identity delivered in Hyderabad to Sind- workies to allow them to travel abroad, 69 were given for Egypt, then the fourth destination for these men, after the Dutch East Indies (108), Gibraltar (104) and the Straits Settlements (79). See Markovits, The Global World, Table 4.1, 127. the circulation of commercial manpower 179 direct bearing on the form or content of the contracts, which I assume to have been relatively standardized. The reason why employees of Sindwork merchants were provided, at least after 1900 (no clear evidence has come to light regarding the earlier period, c. 1860–1900, in the history of the network) with written employment contracts, a practice rarely found in other similar commercial networks, either those of other Indians merchants, or those of Chinese or Lebanese traders, can only be guessed at. Growing tensions between employers and employees in Egypt were noticed from 1899 onwards by the British authorities in Egypt,7 and they led to a spate of court cases, which may have prompted an institutionalization of previously more informal arrangements. It must have become apparent in Egypt to the employers that the “normal” mechanisms for solving disputes between employers and employees in India through caste panchayats or other “customary” institutions were not available, and that the existence of written contracts, which were clearly skewed in favour of the employers, would come in handy in case of disputes in the courts. Another possible factor is that the big firms employed an increasing proportion of non- family labour; lacking the means of pressure normally available to heads of firms to squeeze family labour, which consisted mostly in applying a moral and financial pressure, they tended to seek redress in the courts in case of theft, absconding, etc. by their employees. The fact that private contracts signed in British India were deemed to have legal value in British-occupied Egypt (that was officially annexed only in 1914) and could be used as corroborating evidence in court cases tried before British consular courts reflects the existence of what I would call an imperial legal space, in which norms were largely standardized. The homogenization of legal norms across an imperial space does not mean however a unified labour regime. While employment contracts signed in Hyderabad conformed to a certain general pattern, which was directly derived from aspects of British labour law, in particular the whole body of legislation known as the Master-Servant Act,8 they actually reflect the existence of different labour regimes, in their turn linked to different

7 See the correspondence between the British Consulate in Cairo and the Foreign Department of the Government of India in May 1907, reproduced in Bombay General (Mis- cellaneous) Proceedings Emigration, November 1907, Serial no A 112, India Office Records, Asian Pacific and African Collections of the British Library, London. 8 On this question, see the following magisterial survey: Hay and Craven, (eds), Masters. 180 claude markovits occupational and social profiles. To illustrate the point, I shall present a comparative analysis of two contracts.

Two Labour Contracts

Contract no 1, dated 1901 has been partly reproduced in my book under the heading “employment contract of a shop assistant.”9 While nothing in the contract itself actually identifies the employee specifically as a shop assistant, other evidence at the trial makes it clear that it was the case. Besides, the salary scale is clearly that of a shop assistant, as known from other sources. There are altogether eleven clauses to the contract. Clause I concerns duration of service (two and a half years), location (“any place they may be willing to send me to”) and salary (45 Rs per month to be increased to 50 after 15 months), which was high by the standards of the time (the average wages of an agricultural worker in Sindh were 12 rupees per month, a figure that was higher than the Indian average). Clause II concerns travelling expenses between Cairo and Hyderabad, to be paid for by the employer. Clause III defines exactly the period of payment of the wages. In Clause IV, that is probably lifted word for word from the Master and Servant Act, the employee pledges honesty and obedience to the employer. The tasks however are not described in detail (“I shall do all the works they impose on me”). Clause V concerns the place where accounts will be settled, which can be either Hyderabad or Bombay, a clear indication that accounting took place in the city of residence of the principals of the firms and not in Cairo, where the employees worked. Clause VI mentions English calendar months as the unit of accounting, a trait that sets that kind of contract apart from more widespread Indian contracts using Indian calendar time. Clause VII concerns sanctions in case of loss of or damage to goods. Clause VIII concerns dismissal in case of breach of contract by the employee (nothing is written about breach of contract by the employer) in which one can find an echo of a 1859 Indian Act. Clause IX prescribes the large fine of 200 rupees (a sum equivalent to four or five months of wages) in case of the employee not joining the firm at the appointed time. Clause X is repetitive of Clause IX, concerning any

9 “Pohoomull Brothers, Plaintiff and Fatehchand Kayomull, Defendant”, Foreign Office Records, Embassy and Consular Archives: Egypt, Cairo Consular Court Records, Dossier no 25 of 1902, FO 841/72, Kew, The National Archives of the United Kingdom (TNA), partly reproduced in Markovits, The Global World, 298. the circulation of commercial manpower 181 breach by the employee of the clause concerning the date of expiry of his contract. Clause XI concerns an advance of 25 Rs to be given by the firm to the family members of the employee for their “maintain”, which hints at the fact that the salaries earned in Sindwork were a crucial resource for many lower middle-class families in Hyderabad. Contract no 2,10 dated 1905, makes clearer use of a lexicon of master– servant relationship. The employee is thus explicitly called a servant and the employer a master, which was not the case in contract no 1. Clause 1 defines the work of the employee as that of a “general servant” and mentions Cairo explicitly as his place of work while not excluding “any other place . . . wherever the said master may be pleased to send him”. Clause 2 fixes the duration of the contract as three rather than two and a half years. Clause 3 gives the amount of salary to be paid, which is much lower than in contract no 1: 10 Rs to be increased to 12 and 15 after one and two years. Of this 6 Rs are to be paid to the parents of the servant, and the rest kept by the master and given only on completion of the period of service. Clause 4 is about conveyance to Cairo at the expense of the employer. Clause 5 gives a detailed account of the allowances to be received by the servant: board, lodging, washing, shaving, to be provided at the employer’s expense, and a yearly allowance of 18 Rs for the purchase of clothing. Clause 6 is the usual one about obedience and honesty. Clause 7 goes into a detailed evocation of possible forms of misconduct (“gambling, hard drinking, debauchery” and “any such act not sanctioned by the society”) which it then defines as reasons for immediate dismissal. Clause 8 is concerned with the keeping of trade secrets and prohibits the employee from disclosing any information to competitors. Clause 9 is about breach of contract, and, although it does not mention a fine, it includes the threat of non-repatriation to India. Clause 10 is about the advance given to the family of the servant. A rapid comparison of the two contracts brings out firstly the large gap in salaries, with a ratio of 4,5 to one in favour of the shop assistant’s. Another striking feature is how much more paternalistic contract no 2 is: it purports to be prescriptive at the level of conduct, but also provides more in terms of living expenses. Thirdly, regarding the payment of wages, it appears that, contrary to the shop assistant, the servant is in fact not

10 “Kesumal Sobhraj, plaintiff, vs Chellaram Vasanmal, defendant”, Foreign Office Records, Embassy and Consular Archives: Egypt, Cairo Consular Court Records, Dossier no 7 of 1907, FO 841/91, TNA, partly reproduced in Markovits, The Global World, 299–300. 182 claude markovits going to receive any money: part of his salary will be paid to his parents, and the rest kept by the employer. This is at least in part linked to the fact that there was a significant difference in age between the two employees; while the shop assistant was 33, and therefore a mature adult, the servant was much younger, probably under 18, although already past adolescence, judging from the fact that shaving expenses were one of his perquisites. This comparative exercise thus brings into focus the existence of a two- tier labour market in Hyderabad for Sindwork. Although no detailed job description is offered in the contracts, other evidence from court cases shows that servants generally did not serve clients in the shops, contrary to shop assistants, but were entrusted with menial tasks, such as cleaning the premises, shelling pearls, or attending personally to the managers. It seems also that, after a certain point in time, they were expected to do the cooking, a task which, in an earlier period, was accomplished by high- caste cooks specifically recruited for that purpose. As to what the principles of this segmentation of the mercantile labour market were, we are somewhat left guessing, because of a lack of docu- mentation. Apart from age, which may or may not have been an important differentiating factor (we do not know for instance whether it was common for youths who had started as servants to later graduate to the position of shop assistants), two obvious criteria that come to mind are skills and family origin. Regarding the former, shop assistants were required to be literate and numerate, to have some basic knowledge of goods and markets as well as some rudiments of accounting, so as to be able to manage the relatively complex tasks involved in selling goods to an international clientele, and were also expected to have a smattering of English, and even of other foreign languages, so as to be able to talk to customers who were not likely to know Indian languages. Such qualifications could be found only amongst youths having frequented school for several years. The existence of adequate educational facilities up to the secondary level in Hyderabad, whose Hindu population had one of the highest male literacy rates in the whole of British India, explains that the merchants were able to find locally a readymade supply of shop assistants amongst the educated youth. The latter later acquired on the job the kind of practical knowledge necessary to the performance of their tasks. On the other hand, servants, given the kind of simple manual tasks they were expected to perform, did not need any literate skills, and certainly had no need of English. It is probable that they were illiterate like the vast majority of Indian workers at the time. Access to education being dependent on socio-economic position, one can infer that shop assistants came mostly the circulation of commercial manpower 183 from merchant families with some resources, probably in part families already engaged in Sindwork, while servants belonged to a lower social strata. The exact sociological profile of Hyderabad town is unfortunately not known to us in detail, in the absence of reliable studies. The only quantitative data available come from the decennial Censuses, but the latter are of little use as far as the socio-professional aspect is concerned. What can be said with a measure of confidence is that, in Hyderabad, contrary to most regions in India, caste cannot be treated as a proxy for class, since most of the Hindu population of Hyderabad belonged to one so-called caste, the Lohana. The latter was divided between two segments, an upper strata of clerks and professionals known as Amils, and a lower strata known as the Bhaibands, which encompassed the entire spectrum of those engaged in the trading professions, from humble shopkeepers to affluent bankers. Although practically all the Sindworkies belonged to the Bhaiband group, that in itself was no indication of the exact position occupied by an individual on the socio-economic scale. The most remarkable feature of this system of labour circulation, especially when compared to the largely contemporary system of indentured labour, is that it was not directly supported or even underwritten by the colonial state. It was purely a product of Indian private enterprise, and actually the colonial state appears to have become conscious of its existence only in the late 1890s when, as already mentioned, an influx of Sind- workies in Egypt resulted in a sudden spate of court cases which led the British authorities in Egypt to take up the matter in correspondence with the Government of India. The authorities in Calcutta seem to have given some passing attention to the question, but not to the point of actively interfering in the organization of the circulation of the employees. The latter did not fall under the purview of the vast corpus of emigration legislation, developed from the 1830s onwards to regulate the movement of indentured labour from India to the colonies. It is only after the abolition of indenture in 1917 and the promulgation of a new Indian Emigration Act in 1922, that all movement of working personnel from India became an object of state regulation. From that date onwards, contracts which had been signed in Hyderabad before a notary public had to be registered in Karachi (the capital of the province of Sindh, and an important port of embarkation for Sindworkies, although many of them left India through Bombay or Calcutta) with the Protector of Emigrants, a magistrate specifically in charge of overseeing migration and of checking abuses. The shift from a purely private system to a state-regulated one did not however entail significant modifications, as shown by a detailed comparison I have 184 claude markovits done of the 1901 contract already mentioned with a contract signed in 1929 at Karachi.11 Regarding the question of legal constraints on labour mobility, a specific dimension that has to be taken into account is that of cross-border travel, crucial in the case of an international network. Between 1870 and 1920, there occurred a gradual transition from a regime of relatively few restrictions on travel for commercial people to one in which restrictions were many. While, in the 1870s, Indian merchants and their employees could move practically everywhere, without any documents proving their identity, and simply trust in the protection afforded by their status as British subjects (following Queen Victoria’s proclamation of 1858, all residents of British India were considered subjects of the British Crown), the growth, from the 1880s onwards, of “anti-Asiatic legislation” in the United States, and in the white settler colonies of the British Empire, (Natal, the Australian colonies), while meant mostly to stem the migration of cheap Chinese (and Indian) labour, inevitably affected commercial manpower. Starting in the 1880s, travelling Sindworkies had to carry some legal proof of identity, mostly in the form of certificates of identity, which district magistrates in Sindh were empowered to issue under a regulation dated 1863. The First World War really sounded the death knell of the free travel regime, and, after that date, Sindworkies had to have proper passports to travel abroad, which did not however always allow them to enter their country of destination, including British territories. In the interwar period, growing restrictions on travel led to calls from the Sindworkies for intervention by the British government to ease them, and therefore left traces in official correspondence.12 The organization by the merchants of a small inland town in India of a system of worldwide circulation of employees, without any intervention or support from the colonial state prior to 1917, is undoubtedly a puzzling fact, and raises different kinds of questions. The only comparable example that comes to mind is that of the commercial network originating from the town of Wenzhou, in the Chinese province of Zhejiang, specialized in the sale of leather goods, which was more or less contemporary. A comparison between Hyderabad and Wenzhou could be highly instructive but unfortunately it has not been possible to trace studies of the “Wenzhou”

11 Markovits, The Global World, 172–73. 12 For a few detailed case studies, concerning in particular the Philippines and Panama, see Markovits, The Global World, 230–240. the circulation of commercial manpower 185 in Western languages. It is possible however to speculate that, in the case of Wenzhou, family and clan must have shaped the network in more direct ways than in Hyderabad. Regarding the Sindworkies, it is striking to note how the sphere of labour relations acquired a degree of autonomy vis-à-vis the sphere of family and caste relationships. Their case is clearly different from that of another well-known network originating from India, that of the Nattukottai Chettiar bankers, who dominated the field of rural credit in Burma and Ceylon between 1870 and 1930, and also played an important role in Malaya, Thailand and French Indochina. In his study of this community, American anthropologist David Rudner emphasized the central role played by caste in its success (Rudner 1994). In the case of the Chettiars, it was their caste institutions, based in Saivaite temples, that ensured the regulation of financial transactions within the community. Nothing comparable can be found in the much more amorphous social and religious universe of the Sindhi Hindus, characterized by a certain degree of syncretism with Sikhism and Sufi islam. Such a “soft” structure appears to have created a space in which it was possible for class antag- onisms to emerge within an apparently homogeneous community. The degree of polarization that existed should not be exaggerated: the shop assistants, the group most capable of manifesting an antagonism vis-à- vis the employers, not excluding forms of violent behaviour, were often themselves potential small capitalists. It was common for Sindworkies, after having served an employer for years, firstly as shop assistant, then as manager, to create their own firms and to become in their turn employers of labour. In spite of the existence of forms of mobility internal to the community, there remains the fact that a confrontational stance was frequently present in the relationship between employers and employees. Regarding the question of the degree of freedom these workers enjoyed, a comparison with indentured labour is in order. At first glance, shop assistants would appear to constitute a clear case of free labour, in striking contrast to the indenture regime, while servants could be seen as being closer to indentured labourers. Yet both groups of employees had proper contracts, signed in front of witnesses, while it is known that generally indentured labourers did not actually sign and take cognizance of their contracts before their departure from India.13 While there is still a debate about whether indentured labourers were voluntary or forced

13 Thus, in the 1840s, indentured labourers going to Mauritius would sign contracts only after their arrival in the colony. Mentioned in Northrup, Indentured Labor, 63. 186 claude markovits emigrants, there seems no doubt that Sindworkies, be they shop assistants or servants, were not coerced into entering into agreements with employers. Lack of open coercion is not however sufficient to conclude to the “free” character of a labour contract. Firstly it is possible that some of the so-called servants were actually employed as chattel to repay the debts incurred by their families to their employers. There is nothing explicit in the sources about it, but the specific payment arrangements in contract no 2 leave at least that possibility open, although it remains necessarily speculative. Regarding the actual terms of employment, a fundamental question is whether these workers were free to terminate their employment before the expiry of the period for which they had signed, without risking penal sanctions. It is not clear from the contracts themselves, but other evidence from court cases and other sources shows that it was not uncommon for a shop assistant or a servant to change employer or to start business on his own when in Egypt, where there seems to have been a steady demand for labour in Sindhi shops as well as for pedlars. The degree of freedom of the workers would actually have depended more upon the state of the labour market than upon the legal nature of contractual arrangements. One has to take into account the fact that labour in commercial establishments in a big city like Cairo could not be as tightly controlled as labour on a sugar plantation in Mauritius or Guyana, where mobility was actually very restricted for Indian indentured workers. There remained nevertheless a significant difference between the cases of the shop assistants and of the servants. The former actually pocketed their wages regularly and, given that their ordinary expenses were much lower than their wages, could accumulate some savings. Their effective saving rate however depended on two variables: on the one hand the amount they remitted to their families, and on the other hand their lifestyle; if they drank, gambled and frequented prostitutes, three temptations which loomed large in a city like Cairo, (well known for its large number of bars, brothels and casinos frequented by a cosmopolitan clientele), they could easily run into large debts. Provided however they were prudent enough, they could easily survive a period of unemployment between two jobs. On the other hand, the servants did not actually pocket their wages, which were partly paid directly to their families, partly retained by their employers, and had no way of surviving dismissal in Cairo, except if another Sind- work merchant took them immediately in his service. The shop assistants can be seen as forming a sort of labour aristocracy, their salaries being much higher not only than those of most commercial employees in India (about which little reliable information is nevertheless the circulation of commercial manpower 187 available) but also than those of many such employees in Britain. On the basis of the data provided in a book published in Manchester in 1910 by British trade-unionists dealing with the situation of commercial employees in Britain14 it appears that they formed one of the most exploited categories of British workers. Their wages were low, they worked very long hours (between 60 and 85 per week, depending on branches and categories of workers) and had precarious life conditions. Some lived on the premises of the commercial establishments, a trait that is found also in Cairo amongst employees of Sindworkies. The latter also worked long hours (although no detailed information is provided in the judicial documents), but their salaries were relatively high, a consequence firstly of the prosperity of the trade in semi-luxury articles in which the Sindhis were specialized, and secondly of the fact that they had qualifications that were more than average (in particular linguistic competence that was essential for dealing with an international clientele). Such an “aristocracy” however cohabitated in the Sindhi shops of Cairo with a group that had more resemblance to a kind of sub-proletariate, with employment conditions, as known through the contracts, that were close to debt servitude, a system that was and is still widespread in India. The cohabitation of these two very different kinds of manpower did not fail to generate certain tensions, about which we know however too little.

Conclusion

To conclude this presentation, I shall address the question of the relationship of this system of labour circulation to colonialism. Was it a direct product of the colonial situation, or the outcome of a mixture of “traditional” features and of innovations due to the colonial regime? The legal framework as already mentioned, was decidedly British colonial: the contracts, especially contract no 2, are direct derivations of contracts under the Master and Servant Act. They were apparently written in Eng- lish (although there may also have been a version in Sindhi), and signed before a notary public with three witnesses. They followed a so-called “English” calendar, different from the various Indian calendars, whether Hindu or Muslim, that regulated most of commercial activity in India. They included financial clauses which seem to belong more to a British

14 Hallsworth and Davies, The Working Life. 188 claude markovits than to an Indian universe. Payment of wages on a monthly basis in particular, as far as can be ascertained, was not a very widespread practice in the commercial world of India. On the other hand, the “moral” clauses included only in contract no 2 could be reflecting either Victorian Puri- tanism or traditional Indian merchant ethics, which were actually not that different. The system of advances on wages does not seem either to be specifically British or Indian, but could probably be found operating in both settings. The legal framework under which the employees of the Sindwork merchants circulated and laboured between India and Egypt in the early twentieth century appears thus to be the product of a pragmatic adjustment between British legal norms and Indian practices, a sort of “hybrid”, which reflects the capacity of Indian merchants to adapt so as to be able to function in an international setting. On the other hand, the actual working of the system in Egypt was the product of the interplay between market forces, which seem to have somewhat favoured the employees, at least the shop assistants, because of the existence of a shortage of qualified manpower, and a legal system which was clearly skewed in favour of the employers, producing a contentious work environment in which disputes were rife, and violence not unknown. And of course it is because of the disputes that there are archival traces of this story, which would otherwise have remained hidden. Whether any kind of broader conclusion as to the problem of labour mobility in the Eurasiatic space can be derived from such a limited case study is a point I shall leave to the appreciation of the readers. Part three

Dangerous Ties: State, Landlords and Longue Durée Servitudes

Constrained Labour in Early-Modern Rural East-central and Eastern Europe: Regional Variation and its Causes

Markus Cerman

Introduction

Several generations of rural and agrarian historians broadly agree that the establishment of a commercial demesne economy and a deterioration of the legal and social situation of the rural population were central elements in the characterization of early modern East-central and Eastern European agrarian structure. This conception lies at the heart of the construct of an ‘agrarian dualism’ between early modern Eastern and Western Europe and formed an important pillar in assessing an alleged long-term East- ern European economic backwardness in conventional but also in world- system analyses. By contrast, there is no agreement about the exact timing of the establishment of demesne lordship and a demesne economy nor about the main causes for these developments. Likewise, approaches to this alleged special agrarian system differ widely. While traditional schools of thought place most weight on legal factors, such as the rights of demesne lords and the weak position of villagers, other approaches, including Marxist ones, primarily focus on economic aspects. Among the latter, the attempts of historians of the former GDR, former Czecho- slovakia and Hungary focussed on the prevalence and burden of labour rents (forced labour services, corvée) (see below), which villagers had to perform for demesne farms and the demesne economy.1 The actual burden of labour rents has been used in research to differentiate between ‘mild’ forms of demesne lordship and ‘harsh’ ones which could include the establishment of serfdom for instance in Schleswig-Holstein, Meck- lenburg or Pomerania. The threshold has been set at two to three days of labour rent per week for a full tenant farm.2

1 Harnisch, “Gutsherrschaft in Brandenburg”, 138, 145–46; Pach, “Bauernaufstand”; Válka, “Druhé nevolnictví”. See also Bush, “Serfdom”, 211, 213; Knittler, “Between east and west”. 2 Harnisch, “Gutsherrschaft in Brandenburg”, 146–47; Harnisch, “Probleme”, 252, 272–73; Henning, Dienste, 46–70, 116–19; Bush, “Serfdom”, 213. 192 markus cerman

It remains questionable whether regimes of serfdom are accurately defined by this one variable alone, but this is not the concern of the pre sent paper. Instead, it mainly refers to the organization of work in the demesne economy and whether or not forced labour was really the dominant source of labour in early modern East-central and Eastern Europe. In order to analyze this it is necessary to discuss, in a first step after an initial discussion of the terminology, the causes for the rise of forced labour services after 1450/1500 and the ways, lords succeeded in raising them. Especially in the light of studies published in recent years it becomes clear that landlords had to negotiate higher forced labour services and make concessions to villagers. In some circumstances, higher labour services may have been more favourable for the economy of tenant households than an increase of rents in cash and kind. Secondly, the paper will highlight the fact that it is sometimes very difficult to distinguish between ‘forced’ and ‘free’ labour in the context of the early modern demesne economy and will discuss areas where characteristics overlap. Hence, studies which claim the prevalence of one or the other form need to be evalu- ated carefully as to whether this problem in distinguishing these forms is sufficiently considered. Thirdly, the main argument developed here is that the demesne economy in general and demesne farming in particular could not rely on forced labour services alone, but that a mix of various forms of labour—with considerable swings in the importance of one or the other form—prevailed. With the exception of Russia, however, there seems to have been a tendency during the eighteenth century to increasingly convert a part of the abundant labour rents into cash payments and to use wage labour for demesne farming to a higher degree. Labour rents or corvée represent (usually unpaid) compulsory work bur‑ dens of tenants which were mainly used for demesne enterprises. They were part of the ordinary rent obligations of tenant farmers, smallholders and cottagers in the areas surveyed here. It is significant that cash rents and rents in kind were typically small and the major burden for tenant holdings was represented by the obligation to render labour rents, especially in regions in which they were actually demanded and were higher than one or two days a week per holding. In the area surveyed here, labour rents are referred to in the literature as Arbeitsdienste or Scharwerk, robota, pańszczyzna or barshchina. These forced labour services or labour rents have to be differentiated from two other forms of compulsory work used by demesne lords, i.e. compulsory agricultural life-cycle service (Gesindezwang in German) and compulsory wage labour (see below). In certain terri‑ tories, it was part of the rights of demesne lords to force children of tenants regional variation and its causes 193 into employment as life-cycle farmhands and maids (for which, however, the individuals received wages) for a certain number of years. Households of smallholders and cottagers could be asked to supply compulsory wage labour, e.g. during agricultural peaks, as part of their tenancy regulations. As areas of different languages are covered in the following, the paper will use the approximate English translations introduced before.3

Why and How Did Demesne Lords Increase Labour Service Demands?

With the exception of Russia, the rural population of East-central and Eastern Europe was confronted with seigniorial demands of rising forced labour services (corvée) at some point between the second half of the fifteenth and the early seventeenth century. As with regard to actual levels of forced labour services, there was also considerable regional variation in the occurrence of these demands (in the timing or in the strength in which these claims were presented) and in the reasons for this trend. As ‘classic’ major cause, the literature identifies the growth of the commercial demesne economy in East-central and Eastern Europe from the fifteenth century onwards. Confronted with the erosion of conventional rent income during the crisis of the Later Middle Ages and the strong devalu ation of (fixed) customary rents during the sixteenth century, landlords all over Europe were desperate to generate additional revenue. While the flexibility of tenures and the development of property rights in the West allowed for systematic rent increases (such as for short- or medium term leases) or the introduction and increase of entry fines, which soon became important sources of additional income also for lords in Central Europe, various theories of demesne lordship suggest that due to the abundance of land demesne lords in East-central and Eastern Europe systematically turned to direct management and the expansion of commercial demesne farming geared at international markets where grain demand grew from the late fifteenth century onwards. The fact that increases in forced labour services were reported systematically in the sixteenth century would suggest that this period did indeed represent a watershed in this respect. Typically, the literature refers to estate registers or to negotiations on territorial Diets. These documents may reflect a shift in norms or in general views on the matter, but they

3 For an excellent discussion of the relevant German and Polish terminology see Hagen, “Village life”, and Hagen, Ordinary Prussians. 194 markus cerman cannot be regarded as sufficient evidence that higher services were ac‑ tually generally imposed. The impression remains that demesne lords may have needed labour of their rural tenants desperately, due to the lack of servants or wage labourers and high wage levels. Complaints on Diets and princely regulation of the terms of agricultural life-cycle service in this period reflect this concern.4 Contrary to often repeated images in the literature, the existence of forced labour services was not unusual in late medieval East-central and Eastern Europe, either. Yet, traditional accounts were right to suggest that apart from special social groups directly attached to demesne farms average service obligations usually were limited to a few days per year. With the tendency to abandon demesne activities during the fourteenth and early fifteenth century, they were further reduced or converted into payments. When the tide turned, lords succeeded in raising labour rents beyond the medieval levels by different steps whose timing and mechan isms could vary. Recent studies investigating the rise of forced labour services after 1450/1500 do not ignore the use of force by lords in this process, but also stress the element of negotiation with villagers throughout the early modern period. According to this argument, it seems clear that in an environment of high levels of village desertions and a strong competition for tenants—of which measures of the Estates in various territories, which were aimed at forcing lords to return fugitives to their original lords give evidence (and which, in a re-interpretation that took a number of steps, would become the basis of legal mobility restrictions in the sixteenth century and later)—the use of violence would get lords only so far, before, e.g., tenants left elsewhere. The bargaining position of villagers had definitely increased (although legislative changes were again intended to strengthen lords, see below). The gradual change therefore should be interpreted as a complex process of negotiations in which lords had to offer something in exchange for the services they wanted to receive. It has also been suggested that the original demands of forced labour services and simultaneous pressure for the increase of other rents occurred at a time of growing income for tenant farmers during the sixteenth century. They were thus able to absorb some of this pressure in a period of rising living standards among farmers and smallholders without any significant loss of income.5

4 See, e.g., for Prussia Aubin, Geschichte, 120–22, 124, 130–31. 5 Kula, Theory, 125. regional variation and its causes 195

Initially, tenant farmers and smallholders may have consented to rendering forced labour services occasionally, which were requested rather than commanded or were introduced by means of formal contracts between demesne lords and villages in return for the reduction of other payments and dues or for additional land.6 During this period, when inflation had eroded much of the original value of cash rents, the rural population may have agreed to rendering corvée, especially when levels were still low (a few days per year). Based on this, they could keep most of their production for commercial sales, which was more favourable for their income in an environment of rising agricultural prices. A recent study suggested a different argument for late medieval Poland. According to this, villagers themselves preferred the change towards forced labour services during the fifteenth century, because they were unable to sell grain at sufficient prices and hence raise the necessary cash to service their rents. By giving up some of their land as well, they succeeded to reduce overall rent levels without jeopardizing their subsistence livelihood. The earliest legislation of maximum labour services on Diets within parts of the Polish Common- wealth in 1520 is thus seen as a means to prevent that all villagers could exchange their rents in cash and kind against labour rents rather than as a push of lords to increase overall labour rent levels.7 The empirical literature reports a large number of cases which prove that increases of forced labour services in the fifteenth and sixteenth century were based on negotiations between lords and villagers. For instance, in the Brandenburg estate of Stavenow farmers only agreed to provide more labour services, because their rents in cash and kind were held constant or reduced.8 For the needs of a new demesne farm in Hindenberg in Mecklenburg in 1517, the Duke established a new general service of one day in every third week in two villages. In return, villagers’ payments in lieu of services were reduced by half and they no longer had to perform transport services which had existed before.9 In a formal contract between the lord and the villagers of Roxförde and Wannefeld in the Brandenburg Altmark, both parties held the right to choose between labour services and payments.10 As late as in 1770, a village in one of the Esterházy estates

6 Enders, Uckermark, 179, 193–200; Enders, Prignitz, 400nn; Enders, “Frondienste”; May- baum, Entstehung, 130; North, “Entstehung”, 57. 7 Mikulski/Wroniszewski, “Das Vorwerk”, 122. 8 Hagen, “Peasant rents”; Hagen, “Village life,” 168; Hagen, Ordinary Prussians, 32nn, 65nn; Scott, Society, 190. 9 Maybaum, Entstehung, 134. 10 Enders, “Frondienste”, 90. 196 markus cerman in Hungary chose to offer labour rents to avoid the introduction of a rent in kind.11 In other cases, villagers may have accepted forced labour services when lords offered certain ‘help’. A village of the Mecklenburg monastery of Dargun agreed to being ‘protected’ by a neighbouring noble lord and villagers were ready to perform some work for him under the condition that they be served meals on the days of work. When the village wanted to end this arrangement some time later, the lord turned against them and he insisted that these services had to be rendered on a regular basis. In 1573, the conflict was resolved and the village consented to paying a considerable sum in exchange for the labour rents.12 Some of the examples show that the development was more often ac‑ companied by conflicts than not and demesne lords exerted various degrees of pressure, including outright violence or imprisonment of those who resisted their claims.13 It was the rising seigniorial bargaining power based also on state concessions to the Estates during the sixteenth century which slowly and persistently weakened the position of villagers. Constant resistance against increasing demands by demesne lords, who tried to undermine earlier agreements, became the rule.14 Often, it was the rural population that insisted that a specified level of forced labour services be written down once and for all. When previous agreements stated that they were to help at harvest or at haymaking, it left them vulnerable, when demesne lords extended their demesnes or built new ones. By setting an explicit number of days or clearly specified tasks, further demands for increases could be prevented and it may have been because of this that the well-known general regulations of a certain number of service days per week first entered contemporary documents or legislation.15 Brandenburg tenant farmers and smallholders fought fierce legal battles with their demesne lords over binding arrangements for forced labour services from the 1540s on. In many cases, they turned to the territorial prince that a solution be decreed or existing contracts be obeyed.16 Some- times, outright injustice prevailed. Thus, despite the fact that the Meck- lenburg Duke agreed to relieve his tenants on the island of Poel from their

11 Domanovszky, “Gutsherrschaft”, 459. 12 Mager, Geschichte, 100. 13 Enders, Uckermark, 179, 195, 198–99; Harnisch, “Gutsherrschaft in Brandenburg”, 135–39; Maybaum, Entstehung, 135–36, 149. 14 See e.g. Enders, Uckermark, 193–201; Enders, “Frondienst”, 145. 15 Enders, “Frondienst”, 92, 94. 16 Enders, “Frondienst”, 97–9. regional variation and its causes 197 labour obligations of four days a week in exchange for an annual payment of a certain quantity of barley in 1639, his administrators continued to demand services. When the villagers complained to the Duke, he insisted to uphold the original burdens despite the agreement.17 Such moves often were accompanied by the use of violence to intimidate villagers or crush their resistance.18 Apart from the ruling of Diets and some princely concessions, traditional accounts often refer to the prior acquisition of jurisdictional lordship as an important basis to claim and receive new or increased services.19 With these powers, lords could police arrangements and gained access to forced labour services which were originally reserved for the prince (by his power of jurisdictional lordship). With respect to the medieval foundations of early modern developments, this seems an interesting point for future analysis. A careful study for Mecklenburg, however, describes this as a relatively complex process which also displayed important discontinuities and was far from being a uniform pattern even within this country.20 The evidence presented reveals that the increase of forced labour services frequently came in return for freezing or reducing other rents. Vil lagers were often prepared to accept this, because they could minimize cash and commodity rents in periods of low agricultural prices or reap maximum commercial profits during agricultural booms, because their market quota was not affected by seigniorial demands. The significant general erosion of rent value in cash and kind in real terms between the fifteenth and early seventeenth century was recently emphasized in surveys on Brandenburg and Poland21 and is evident in numerous case studies for regions in East-central and Eastern Europe.22

Identifying Forms of Forced and Wage Labour

One of the difficulties of research regarding a clear picture of the organ ization of work within demesne enterprises is that it is often impossible to differentiate between forced labour services, compulsory wage labour

17 Mager, Geschichte, 101– 102. 18 Enders, Uckermark, 195, 198–99; Göttsch, Leibeigene. 19 Blum, “Rise”, 822–26, 833–35; Schmidt, Leibeigenschaft, 127–30. 20 Maybaum, Entstehung, 42–3, 49–86, 87–95, 97–107, 126–30. 21 Guzowski, “Changing economy”; Hagen, “Village life”; Hagen, Ordinary Prussians. 22 Muszyńska, Gospodarstwo; Wajs, Powinnośći. 198 markus cerman and wage labour.23 For instance, life-cycle servants on demesne farms are often regarded as being in waged employment, which, in the strictest sense, they were. Demesne lords, however, often had the right to recruit servants from the children of their tenants’ households and force them into paid service for a certain number of years.24 The border line is equally difficult to draw in the case of day labourers. Demesne accounts often show expenses for the wages of day labourers. Yet, in many cases, these may have been smallholders or cottagers, who had to supply compulsory wage labour when they were summoned as part of their feudal obligations. To complicate matters further, custom often stipulated that villagers had to be compensated even for their tenurial forced labour service obligations. As a consequence, people often received food or beer on the days they worked on demesne farms (and these expenses would be registered as well).25 Compulsory wage labour was also systematically used to supplement labour services, such as in Lithuania when demesne farms were further extended during the eighteenth century.26 When studies emphasize a widespread use of wage labour in the demesne economy for instance in certain Central European territories prior to 1650, the existence of overlaps between forced and wage labour does not always seem to be considered in a sufficient manner.27 It thus seems problematic that the exact nature of ‘wage labour’ is not characterized in necessary detail in these examples. For instance, the employment of servants on demesne farms would usually include villagers’ children in forced life-cycle service as well as waged day labourers; hence, distinguishing between compulsory wage labour and wage labour in general terms seems difficult.28 Sixteenth- and seventeenth-century laws enabled demesne lords to force villagers’ children to enter agricultural life-cycle service on demesne farms (Gesindezwang). The term, during which they were paid wages

23 Pach, Agrarentwicklung, 30; Pach, “Problematik”; Żytkowicz, “Trends”, 71. 24 Hagen, “Village life”, 168. 25 See e.g. Pach, “Problematik”; Soom, “Gutswirtschaft”, 46–7; Stutzer, “Ertrags- und Lohnverhältnisse”, 20; Wächter, Domänenvorwerke, 63. 26 Kahk/Tarvel, Economic history, 71. 27 See Čechura, “Gutswirtschaft”; Čechura,“Dominium Smiřických”; Čechura, Grund- herrn, 56–62, 114–17, 150–54. 166nn. 28 Válka, Hospodářská politika; Válka, “Druhé nevolnictví”. regional variation and its causes 199 as servants, was usually limited to three to five years.29 For instance in Brandenburg, it was first established in 1518 and confirmed in various new ordinances during the sixteenth and seventeenth centuries.30 In Western Pomerania, a universal forced life-cycle service was usual from the beginning of the sixteenth century. The East Prussian Land Ordinance of 1577 introduced forced life-cycle service for children of subject tenants which were not needed on their families’ farms.31 In Upper Lusatia, forced ser vice with prescribed wages was first established in 1539 and regulations were regularly repeated and extended in land ordinances until 1767.32 After 1550, regulations on Bohemian crown estates repeatedly confirmed the possibility to hire subject children as servants. As examples and pro- tests show, forced life-cycle service was usual in noble estates at this time, too.33 Resolutions of the Silesian Diet governing mobility restrictions, ordinances for servants, and compulsory life-cycle service were passed in 1502, 1512, 1528, and 1565. The provisions of these were standardized in the 1652 Ordinance for Servants and Subjects.34 Yet, compulsory life-cycle service was far from universal in East-central and Eastern Europe. In Poland, Slo- vakia or Hungary, for instance, comparable regulations did not exist.35 In Mecklenburg, compulsory service was usual, but does not seem to have been formally introduced by law.36 Due to the existence of compulsory life-cycle service on demesne farms and of compulsory wage labour it is therefore difficult to conclude from wage expenses recorded in seigniorial accounts that these were only incurred for ‘free’ wage labour.37 A study on estate management in Upper

29 Boelcke, Bauer, 113–15; Henning, Herrschaft, 142–46, 157–58; Kaak, Gutsherrschaft, 20, 40, 103–106; 141, 156, 399; Maur, “Poddanská otázka”; Maur, “Das Gesinde”, 80–1; North, Amtswirtschaften, 83; Wächter, Domänenvorwerke, 53–5. 30 Enders, “Entwicklungsetappen”, 124; Harnisch, Boitzenburg, 69; Kaak, Gutsherrschaft, 103–6; Kaak, “Erbzinsrecht”, 76. 31 Aubin, Geschichte, 134–35; Henning, Herrschaft, 142–46, 157–58; North, Amtswirt- schaften, 83; Wächter, Domänenvorwerke, 53–5. 32 Boelcke, Bauer, 113–15. 33 Maur, “K demografickým aspektům”, esp. 31–2; Maur, “Gesinde”, 80–1; Urfus, “Právní postavení”. 34 Feigl, “Entwicklung”; Weber, Polizei- und Landesordnungen, 150nn, 197nn. 35 Horváth, “Poddanská otázka”, 11–2; Lubinski, “Everyday work”, 148; Špiesz, “Vývoj”. 36 Kaak, Gutsherrschaft, 20, 40, 141, 156, 399; Mager, Geschichte, 87, 163–65; Rudert, “Grenzüberschreitungen”, 361. 37 Míka, “Problém”, 233, 241–42; Válka, “Nevolnictví”, 320–21; Válka, Politika, 100. See also Jirásek, “Venkov”, 98nn; Maur, “Gesinde”, 80–1; Válka, “Stavu”, esp. 200. For compulsory wage labour on Rožmberk estates see Longfellow, “Bohemia”, 7. Also North, “Lohnarbeit”, argues for East Prussia that many wage labourers from sub-peasant strata were forced into employment and were paid in bread and beer rather than money. 200 markus cerman

Silesia at the end of the eighteenth century, for instance, shows that tenant farmers were compensated in cash and kind for the statutory forced labour services (corvée) they had to render for the demesne economy and total wage costs including wage labour would represent up to 33 per cent of seigniorial income. The money tenant farmers received approximately made up for their maintenance costs of the team of draught animals which they had to provide and for food for one person. This compensation was relatively lower than that offered to cottagers as wages as they also received free housing and did not have to bring draught animals.38 As Zs. P. Pach already concluded: “Paid work was (. . .) sometimes burdened with the chains of extra-economic coercion, corvée was occasionally inter- woven with threads of wage labour relations.”39

Forced Labour and the Demesne Economy

The previous section established that East-central and Eastern European demesne lords could rely on various forms of labour to operate their commercial demesne economy. In the light of this, it seems only straightfor- ward to assume that forced labour services were almost never the only source of labour for the demesne economy and very often not even the dominant one. It is far more likely that it worked on the basis of a mixture between (mostly unpaid) forced labour services and (compulsory) wage labour of one or the other form,40 the balance of which was subject to enormous variation between regions and periods. The two poss ible extremes were characterized by GDR-historiography as Teilbetrieb (operation with forced labour services) and Eigenbetrieb (operation with servants and wage labour only). Several structural reasons can account for the variation and degrees of overlap between the individual modes of operation. To begin with, the most valuable variant of labour rents is associated with tenant farmers whose holdings were large enough to supply ploughteams or draught animals for transport services. Thus, during the initial phases of the extension of forced labour services up until the sixteenth century, certain social groups, who were unable to provide these services may have been exempt. Among them were cottagers and lodgers, unless they had been specifically settled as demesne workers during the later Middle Ages, which

38 Stutzer, “Ertrags- und Lohnverhältnisse”. 39 Pach, Agrarentwicklung, 30. See also Rusiński, “Remarks”. 40 Melton, “Gutsherrschaft”, 329; Pach, “Problematik”. regional variation and its causes 201 was not exceptional in some areas, such as in Prussia or in Silesia. In all other regions, in which they had no prior connections to the demesnes, they were only gradually integrated into the demesne economy, first as wage labourers in voluntary but also compulsory wage labour and later for manual services. In Estonia, as late as 1796 and 1798 laws were passed ruling that also cottagers and crofters generally had to supply forced labour services.41 The actual use of forced labour services depended on the extent of the demesne economy and particularly on the proportion of demesne arable land. The ratio between demesne and tenant land varied strongly between territories but also between individual regions. As a consequence, the exact ratios established on the basis of territorial land surveys or cadastres can only serve as an approximation. But even these indicate that it is very difficult to generalize. For instance, in the Brandenburg Kurmark, the average proportion of demesne arable was found to be a third of the total arable land in 1797, but it was as little as 16.1 per cent in the district of Luckenwalde and as much as 43.8 per cent in the district of Uckermark.42 Proportions of more than 50 per cent were not rare on the level of individual estates or even villages, but certainly the exception for entire territories. Thus, the demesne share of about 60 per cent established for late seventeenth-century Swedish Pomerania is an extreme example within East-central and Eastern Europe.43 There, the expansion of demesne at the expense of tenant land paired with the strong decline in tenant numbers as a consequence of the Thirty Years’ War meant that demesne farms did not have at their disposal an adequate number of villagers to provide forced labour services and thus had to turn to wage labour. An unusually high proportion of 25 per cent of demesnes in Swedish Pomerania was operated in Eigenbetrieb in the late seventeenth century.44 Similar developments were also observed in the Brandenburg Uckermark in the same period. It seems as if there was an inverted-U-relationship between the use of forced labour services and the proportion of demesne land. Leaving aside Russia, Belorussia and the Ukraine, a comparative survey on the extent of forced labour services in East-central Europe in the eighteenth century distinguished between three different zones. In the first, tenants had to provide services on a demesne land area that

41 Kahk, Peasant, 32. 42 Harnisch, “Gutsherrschaft”, 230; Müller, Landwirtschaft, 44–5. 43 Schilling, Schwedisch-Pommern, 19–21, 23, 50–1. 44 Rudert, Gutsherrschaft und Agrarstruktur, 118; Schilling, Schwedisch-Pommern, 49–50, 54, 64, 108. 202 markus cerman represented less than fifteen per cent of the land of their own holdings. This would hold for Denmark, much of western and northern Schleswig- Holstein and most areas of Hungary and Slovakia. The maximum burdens occurred in those regions, in which forced labour services required tenants to work on demesne land larger than 40 per cent of the size of their own holdings: eastern Holstein, Mecklenburg, Pomerania, the Branden- burg Uckermark, Greater and Little Poland and Mazovia. Finally, there was a group, in which the work burden amounted to caring for demesne land of between 15 to 40 per cent of their own land, such as in Branden- burg, Upper and Lower Lusatia, Lower Silesia, Bohemia, Moravia, Prussia, Courland and Livonia.45 The type of farming activity could influence the organization of work within the demesne economy. For instance, the importance of viticulture in parts of Hungary, Slovakia and South-eastern Austria limited the use of forced labour services except for transport purposes. As low quality work usually associated with labour services could be damaging, primar ily wage labour was used for demesne viticulture.46 Similarly, seigniorial profits from agricultural exports in the sixteenth century were not based on cereals only, but in Denmark, Schleswig-Holstein, Hungary and some areas of the Polish Commonwealth dairy and oxen exports were far more important. Dairy and pastoral activities involved tasks for which demesne farms usually employed servants in continuous employment. As a last example, in certain areas of Bohemia and Moravia commercial pis- ciculture was more important than arable farming especially in the early phases of demesne expansion. The building and rearing of ponds was carried out by well-paid specialists and there were only few tasks for which forced labour services could be used in this context.47 Finally, local conditions could influence whether the inhabitants of particular villages were called to forced labour services or not. Even in areas with a strong reliance on corvée, individual villages could be entirely exempt and paid cash rents in lieu of services, because they were not close enough to demesne farms. For an estate in Mecklenburg in 1620, the average distance villagers had to cover to reach the demesnes they were assigned to was calculated at five kilometres.48

45 Henning, Dienste, 46–70, 116–19. 46 Horváth, “Der Charakter”, 92–3; Kiss, “Der Agrarcharakter”, 148; Pach, Agrarentwick- lung, 57–9, 74–80, 98–9, 110, 130–31, 133–34, 143–44; Pach, “Problematik”; Zimányi, Economy, 81–2. 47 Míka, “Problém”, 241nn; Míka, Lid; Petráň, Výroba, 161; Vacek, “K dějinám”, 10. 48 Baumgarten/Bentzien, Hof, 221. regional variation and its causes 203

Wage labour of day labourers, farmhands and servants on demesne farms as defined above was always used for the operation of demesne farms. As indicated, the share it represented relative to the total demand for labour varied strongly between regions and over time.49 There is a tendency in most recent research to rehabilitate a rather widespread use of wage labour within the early modern demesne economy in East-central and Eastern Europe in order to overcome traditional generalizations of an overall dominance of ‘serf’ forced labour economies. While several detailed studies succeed in establishing this point for individual cases or in a broader perspective, we need to be aware of the fact that the mere presence of full-time servants on demesne farms cannot always be equated to the absence of forced labour service demands. In these cases, the question needs to be answered, whether the available servants and farmhands were actually employed for the works connected with the dominant demesne activity of arable agriculture, for which forced labour services were usually called. For example, in the Moravian estate of Telec in 1591, only 13 of 50 ser‑ vants on five demesne farms were employed for arable agriculture, because the others were explicitly shepherds or dealt with the other livestock.50 According to the terrier (estate register) of the Bohemian estate of Rum- burk 1587, three demesne farms with an arable area of 200–300 hectares employed six maids for dairy farming and as cowherds (for a total of nearly 200 cattle and 1,200 sheep), nine further maids and only three male farmhands (apart from the three farm managers).51 In the property of Brložec in the Bohemian estate of Toužim in 1634, there were 32 servants on five demesne farms, eight of which were cowherds, six were maids (who very likely also cared for livestock) and two male servants were responsible for building maintenance. The farms kept 100 cows and 129 calves in total. Nine further servants were the administrators of the farms and their wives. In fact, there were only three farmhands for the work in arable agriculture (for a cultivated arable area of more than 300 hectares). In Toužim in 1638, there were eight shepherds, eight cowherds and 19 maid servants

49 Harnisch, Boitzenburg, 170; Harnisch, “Probleme”, 263; Harnisch/Heitz, “Einleitung”, 17; Melton, “Gutsherrschaft”, 329; North, “Lohnarbeit”, 17; Peters, “Ostelbische Landarmut”, 280nn. For demesne operation with servants and day labourers in Poland and Hungary see Geremek, “Problem”; Izydorczyk-Kamler, “Lohnarbeit”; Kamler, Folwark, 65nn; Melton, “Gutsherrschaft”, 304–305, 314; Pach, “Labour control”; Wyczański, Studia, 104. 50 Tiray, “Jak se žilo”, hier 221–22. 51 Hausarchiv der Regierenden Fürsten von Liechtenstein im Schloß Vaduz, Urbar 74; Richter, Geschichte, 33, 36. 204 markus cerman out of 64 servants in total. Four servants were explicitly mentioned to be responsible for the oxen. Apart from the administrators, there were only seven male servants for a demesne area of 480 hectares. Livestock consisted of 164 cows and 205 calves.52 In the estate of Stráž p. R. in Northern Bohemia, 63 servants were employed on six demesne farms in 1630. Only 19, less than a third, were adult male servants, whose duties could have included work in arable agriculture, while 20 female servants must have at least partly or perhaps even mainly cared for the livestock but for the busiest days of harvesting. The rest of the servants was either responsible for the management of the farms or were explicitly listed as shepherds, cowherds etc. A very similar structure prevailed in three demesne farms in the neighbouring estate of Lemberk. There were only eight male and three female servants out of a total of 23 persons who were not explicitly working with the livestock or carrying out other specific duties.53 Around 1600, there were fifty servants employed on five demesne farms in the estate of Frýdlant situated in the same region (Table 1). From twenty male servants, three were responsible for the carriages and four were horse servants, who would probably only partly engage in field labour.

Table 1 Structure of the Servants in Demesne Farms, Estate Frýdlant, 1600/1601 and 1628/1629 Group of employees 1600/1601 1628/1629 Servants N (per cent) N (per cent) Manager 4 (8.0) 6 (16.7) Full male servants 20 (40.0) 10 (27.8) Male youths 1 (2.0) — Full female servants 7 (14.0) 15 (41.7) Shepherds/cowherds 9 (18.0) 1 (2.8) Cooks 4 (8.0) 3 (8.3) Others 5 (10.0) 1 (2.8) Sum 50 36

Source: SOA Děčín, HS, Kart. č. 12a;54 Kart. č. 34.55

52 Ryba, “Hospodářství”, 218–24. 53 Cerman, “Venkovské společnosti”, 372–73. 54 Daß erste jahr. Einnamb (Ausgab) register der Herrschafften (. . .) vom Tage Michaeli des 1600 jhars (. . .) 1601, pag. 19–23. 55 Register uber alle einnamb undt ausgaben des geldes vom July bis zum 30. Decem- ber auff ein halb jahr von der Herschaft Friedlandt 1628; Register uber alle einnamb undt ausgaben des geldes vom 6. Januari bis zum 30. Junij a[nn]o 1629. Auff ein halb Jahr von der Herschaft Friedlandt 1629. regional variation and its causes 205

A comparison with 1628/29 reveals that the proportion of female servants grew strongly, whereas that of males decreased by more than a quarter (a decrease even when one considers three deputy managers newly hired). In total, the number of servants must have remained unchanged (a lower number of demesne farms was recorded in the 1628/29 survey). Until the first half of 1641, the number of servants on the three demesne farms declined to sixteen (a year before, there were still 35 servants); among them only two male hands for the oxen and two maid servants.56 In 1651, eighteen servants were employed on two demesne farms of the estate of Frýdlant. Thus, there was no apparent reduction in the number of servants. Apart from four female servants aged 18 to 23, there were only three male servants who most likely were responsible for the field works. All the other predominantly female servants were either cooks or working with the livestock (cows, calves and pigs). The structure of the servants in fourteen demesne farms held in feof is summarized in Table 2. The table reflects, that apart from a group of female servants only about a third of the workforce was mainly engaged in field labour.57 In 1730, 151 servants were employed on 17 demesne farms in Frýdlant. 38 were maids mainly occupied with rearing the livestock, a further 39 were maids in domestic service and 18 were cooks, all of whom would not have been called for arable duties on a regular basis. In the neighbouring estate

Table 2 Structure of the Servants in Demesne Farms, Estate Frýdlant 1651

Servants Number (per cent)

Full male servants 36 (38.3) Male youths 17 (18.1) Full female servants 23 (24.5) Female youths 4 (4.3) Shepherds 5 (5.3) Cooks 5 (5.3) Nannies 4 (4.3) Sum 94 Source: NA Praha, SM R 109/45, Bl. 10.

56 SOA Děčín, HS, Kart. č. 34, Rechnungk uber die einnahmb undt außgaab der gelder in den rentten der Graff Gallaßischen Herrschafft Friedtlandt vom 29. Decembr: 1640 bieß zum 29. Junij 1641, expenses no. 6; Kart. č. 485, Rechnung uber die einnahmb undt außgaa- ben deß geldes (. . .) vom 17. Martÿ bieß zum 30. Junÿ Anno 1640, expenses no. 14. 57 Národní archiv (NA) Praha, SM R 109/45, Bol. 10. 206 markus cerman of Liberec, 24 servants were employed on demesne farms in the summer of 1704. Two were shepherds, one cared for the horses and three were cooks. Of the remaining individuals, 15 were maids responsible for the 168 heads of cattle.58 An analysis of 100 select demesne farms in Bohemia in 1651 revealed that there were 194 maid servants (among them also kitchen personnel and 72 shepherds and cowherds) among 840 employees in total. Various categories of male farmhands formed the relative majority (44.6 per cent), but more than four fifths of the 185 auxiliary farmhands were younger than 20 years, more than one fifth younger than 14. They could thus not be employed for the more heavy tasks of arable farming.59 Clearly, a more detailed analysis is necessary to fully account for the concrete organization of work on the demesnes. Despite these problems, the literature frequently uses the incidence of a low number of servants on demesne farms as an indicator that forced labour services prevailed in operating demesne farms. Demesne farms in Mecklenburg-Schwerin princely estates on average kept four horses and employed only 4.5 servants at the beginning of the eighteenth century. There were about 1.7 adult farmhands on these farms, so probably one ploughteam could be mustered. With on average nearly seven servants (among them 2.5 grown up males) and nine horses, demesne farms on noble estates were better equipped. But they could usually rely only on a much smaller number of tenant farms rendering labour rents, because the expropriation of tenant farms was stronger on noble estates (there was less re-settlement of tenants after 1650). In four noble administrative districts, there were between 4.5 and 7.5 draught animals on average per demesne farms in 1703. All of the demesne farms also employed servants and farmhands, but it is difficult to determine how many were available for the tasks of arable agriculture. The majority were female—the proportion of female servants was higher than on tenant farms—and they were presumably primarily employed for the livestock.60 The 1703 data prove that demesne organization was highly adaptive to the availability of tenant labour. Thus, princely estates mustered between 17 and nearly 24 tenants but only 4.0 to 5.2 servants per demesne farm. By contrast, in noble

58 Cerman, “Arbeitsrenten”. 59 Further 200 employees were managers and their wives. Maur, “Gesinde nach dem Soupis poddaných”, 122. 60 Heitz, “Die sozialökonomische Struktur”, 26–30, 42–3, 45–50, 70–3; Rudert, Guts- herrschaft und Agrarstruktur, 120. regional variation and its causes 207 estates, only between seven and 9.4 tenant farmers were available, but 6.6 to 7.3 servants per demesne farm. The number of landless wage labourers, which were attached to demesne farms and whose labour could be used was higher in princely estates than in noble ones (3.0 vs. 2.3).61 Of more than 300 demesne farms surveyed in Swedish Pomerania in the late seventeenth century, the average number of servants of 4.6 was as low as in contemporary Mecklenburg. The average number of 9.6 horses and 6.9 oxen (registered for 429 and 237 farms respectively) shows at least that demesnes could supply ploughteams by themselves, which were then probably operated by tenant smallholders who had to provide manual labour as part of their tenurial forced labour service obligations.62 Already in the sixteenth and early seventeenth century, draught power and servants were kept in demesnes in order to avoid overburdening tenant farmers, who were theoretically obliged to perform unlimited service from 1616 onwards. Yet, differences were strong in this respect: In eight princely estates, demesne farms did not have horses as draught animals at all, whereas in six noble estates the average number was 7.7 horses per farm. Again, this may have been for lack of alternatives as many of the smaller (noble) estates probably did not have enough tenant holdings to work their demesne farms with labour rents.63 Even within a relatively small sample of estates of a case study for Upper Silesia, there were profound differences in the management, as some demesne farms were operated entirely with their own draught animals and wage labour. Thus, in the estate of Łubowice, 31 smallholders and cottagers were employed and the demesne farm held 30 draught animals for 189 hectares of arable. In other estates, the number of tenant farmers was not large enough to perform all the necessary tasks of arable farming on the basis of forced labour services. Forced labour was thus complemented by keeping draught animals directly on demesnes and by the wage labour of cottagers.64

61 Rudert, Gutsherrschaft und Agrarstruktur, 101–106, 185–86. 62 Schilling, Schwedisch-Pommern, 24–5, 52, 56, 58, 100–101; Schleinert, Gutswirtschaft, 79, 96–7, 283. 63 Mager, Geschichte, 88; Schleinert, Gutswirtschaft, 92–3, 95–6, 130–37, 283; for the Duchy of Szczecin see Wachowiak, Gospodarcze, 298–311; Wachowiak, Gospodarka, 60–92. 64 Stutzer, “Ertrags- und Lohnverhältnisse”, for the employment of landless in noble demesne farms in early eighteenth-century Mecklenburg see Heitz, “Die sozialökonomi sche Struktur”, 32–4, 52–4. 208 markus cerman

Concentrating on the ‘classical’ period of the rise of the demesnes in Poland in the sixteenth and early seventeenth century, the organization of work was analyzed also in Polish studies. From the evidence, demesnes entirely based on labour services were unlikely for Poland, either. A case study of several hundred royal, church and noble demesnes in three governorships in Little Poland confirms that the majority employed servants and seasonal wage labourers in the period 1525–1642. In some parts of the kingdom, more than 50 per cent were operated in direct management without corvée (Table 3). The proportion of royal demesnes entirely without servants was zero in 1525–53, 13 per cent in 1560–90 and seven per cent in 1611–42. The highest share occurred in the governorship of Lublin with 20 per cent in 1560–90 and 17 in 1611–42 respectively. In the same periods, the proportion of noble and church demesnes operated entirely without wage labour was 24 and eleven per cent.65 Also in other regions, wage labour for demesne farms was important and there was some equipment with draught animals.66 Many demesne lords only owned very small estates and had very little farmland to be worked. Wide- spread commutation of forced labour services was still possible in Little Poland in this period. In 1564, 30 per cent of the villages of royal estates in the Cracow region paid money in lieu of services and a further 40 per cents had only rent duties in cash and kind.67 It is also necessary to consider changes over time in the balance of the use of wage labour and forced labour services. For instance in East Prussian

Table 3 Share of Labour Services on Polish Demesnes, 1551–1580 (in per cent of demesne land operated) R egion Operated with labour Operated in direct services management

Mazovia 79 21 Little Poland 60 40 Western Great Poland 63 37 E astern Great Poland 48 52 A verage 63 37

Source: Adamczyk, “Weltmarkt”, 81; Wyczański, Studia, 109.

65 Adamczyk, Polen, 191; Izydorczyk-Kamler, “Die Lohnarbeit”. 66 Kamler, Folwark, 37–40, 52–76; Wyczański, Studia, 104, 122–38, 163–66; Żytkowicz, Studia, 229–41; Żytkowicz, “The peasant’s farm”, 140–41, 143. 67 Trzyna, Położenie, 222–36. regional variation and its causes 209 princely estates, the formal increase of forced labour services after 1550 was accompanied by a shift towards using the manual labour of smallholders and cottagers at a larger scale.68 With population growth and stagnant agricultural prices demesne expansion stopped around 1600. As a consequence, services which were no longer needed were abolished and converted into cash rents. Also because of slacking grain demand the estate administration tried to increase income from cash rents and used the proceeds to pay waged servants on the demesnes where employment grew.69 A change of preferences between wage and forced labour occurred fairly frequently as the case of the estate of Ostróda illustrates: corvée was converted in 1615 and then reintroduced in 1622. These developments followed economic considerations regarding grain price trends. The area of demesne arable farms reached a peak in the 1570s, was considerably reduced until the 1590s, rose again and then decreased dramatically after 1610. However, when population levels fell, labour rents of the remaining tenants could increase. Ten princely demesnes could use the labour rents of 1,355 tenant farmers in 1610. In 1683, only 530 farmers were still available. Simultaneous with the extension of their land, these demesne farms reduced the number of continuously employed servants over the course of the seventeenth century from 409 to 285. Thus services must have increased also for this reason.70 The estate of the University of Greifswald is another example to illustrate major discontinuities in the estate administration and shifts in the views about efficient management. With a major expansion of demesne farming after 1620 the management shifted from being based primarily on forced labour services of tenant farmers to introducing cash rents in lieu of labour and dividing demesne farmland among tenant farmers between 1670 and 1750. Then after 1750, tenant farms were expropriated, demesne farms rebuilt and labour rents re-introduced. This form of demesne oper ation lasted until the mid-1770s.71 The phenomenon observed above for seventeenth-century Prussia that desertion of tenant farms increased the average burdens of the remaining tenants also occurred in seventeenth-century Poland. In a case study on the district of Korczyn in Little Poland, the average number of service

68 North, Amtswirtschaften, 80–1; North, “Lohnarbeit”, 12–4, 16. 69 North, Amtswirtschaften, 77–8, 81, 100; North, “Lohnarbeit”, 13–4. 70 Kaak, Gutsherrschaft, 284–87; North, Amtswirtschaften, 77–8, 81, 100; Wächter, Ost- preußische Domänenvorwerke, 12, 56–9, 61, 64, 71–4, 143, 147, 153–85. 71 Kaak, Gutsherrschaft, 167–70. 210 markus cerman workdays per household increased from 1.87 per week in 1572 to 2.25 in 1616/1629 and 2.54 in 1660. It was due to a declining number of tenant farms and not due to further demesne expansion or a shift towards a relative increase in the use of forced labour in operating demesne farms.72 In the Brandenburg estate of Boitzenburg the lack of tenants after 1650 did not allow for an intensification of the demesne economy. Instead of leaving the deserted farmland unused, however, it was consolidated into nine new demesne farms until 1700, which were leased out to single tenants, who ran them as Eigenbetrieb, i.e. with their own inventory and with wage labour. When the demesne lord returned to demesne farming in the eighteenth century after the population and agrarian prices had consolidated, some of the farms continued to be managed independently, while others started using forced labour services.73 It was on princely estates of the region, where a rise of labour services was attempted in this period.74 In many regions of eighteenth-century East-central and Eastern Europe a shift towards reducing forced labour services for operating the demesne economy can be observed. While this was often due to an abandonment of demesne farms and their conversion into tenant farmland or due to a temporary reduction caused by agricultural business cycles, in many other cases demesne farms used servants and wage labour to a growing degree. This was caused by population growth, the increasing resistance of villagers to accept new forced labour services resulting from the intensification of demesne farming (partly caused by agricultural innovations) or by landlord awareness that further burdens would threaten the livelihood of tenant farms and thus undermine the whole system. In the Branden- burg estate of Badingen, for instance, eight of eleven villages had to render labour services in 1728. Until the end of the eighteenth century, services per holdings decreased considerably in four villages (by between 15 and 80 per cent from an initially very high level of 298 days per year). Overall, total available draught and manual labour service days fell from 24,615 per year in 1728 to 19,302 in 1763 and 11,457 in 1803 (nearly 55 per cent).75 When demesne farming expanded due to increasing grain demand, wage

72 Muszyńska, Gospodarstwo, 45, 69–72; Wyczański, Studia nad gospodarka, 106, 120–22; see Inglot, Historia, 264, 363. 73 Harnisch, Boitzenburg, 137, 148–49, 158–60, 169; Kaak, Gutsherrschaft, 271–73. 74 Enders, Uckermark, 343–44, 373, 466–67, 469, 518–25. 75 Vogler, “Entwicklung”, 156–62. regional variation and its causes 211 labour was increasingly used in Brandenburg. In the estate of Stavenow after 1763 this was promoted by the fact that the local tenant farmers simply refused to obey the verdicts of the royal chamber court and did not perform the forced labour services they were ordered to. After two decades of costly legal battles over the extent of the services, the court had ruled in favour of the lord—but in vain.76 Similar developments and more conflicts occurred elsewhere, when rising grain prices caused a further intensification of demesne activities.77 Also in eighteenth-century Upper Lusatia, demesne lords increasingly used the wage labour of the rising number of cottagers to work their demesnes78 along with the existing forced labour services. While forced labour services were already abandoned in favour of cash rents in western Holstein in the fifteenth century, in the eastern parts, where demesne farming dominated, reforms started early in the eighteenth century. Under the influence of low grain prices demesne farming contracted and 103 demesne farms and estates had abolished forced labour services for good by the 1720s. Serf resistance occurred where pressures grew again when agrarian prices picked up after 1720/1730.79 Leaseholders of Danish demesne farms did not invest in their own stock of draught animals, but hired tenant farmers and smallholders to work for wages in addition to what could be provided by forced labour services.80 In other areas, the growing number of demesne leaseholders were simply denied access to villagers’ forced labour services. A mixed regime is also obvious in East Prussia, where in the eighteenth century only a part of the demesne farms were operated entirely with forced labour services.81 Demesne lords often had no use for forced labour services in excess of their demands, which automatically occurred as a consequence of population growth, and thus an ever growing share of tenant households paid money instead, which was then used to employ wage labour. In the North- ern Bohemian estate of Frýdlant, around thirty per cent of the households paid cash in lieu of services in the period 1720–1740, while in the neighbouring estate of Liberec only 30 per cent of households were called to

76 Hagen, “Village life”, 181; Hagen, Ordinary Prussians. 77 Harnisch, Boitzenburg, 195; Harnisch, “Probleme”, 263–65; Kaak, Gutsherrschaft, 274–75; Kaak, Bauern, 167–200, 291–304, 338–40; Peters, Lebenswelten. 78 Boelcke, Bauer; Kaak, Gutsherrschaft, 264. 79 Kaak, Gutsherrschaft, 279; North, “Entstehung”, 234–40; Peters, Lebenswelten, 679–80, 715–17. 80 Christiansen, World, 270. 81 Aubin, Geschichte, 157, 166. 212 markus cerman their full service duties in 1703 and a further 44 per cent presumably to some of their duties. While these are significant proportions, the payments only represented between three and less than six per cent of total seigniorial income in Frýdlant until 1765, casting doubt on the general applicability of the theory that such payments attracted demesne lords to abandon the commercial demesne economy in ‘proto-industrial’ areas of demesne lordship.82 By contrast, in areas, where the demesne economy had not fully developed in the first place, resulting payments in lieu of labour services were considerable, sometimes representing a quarter of total landlord income, as seventeenth- and eighteenth-century Lower Austrian examples show. While the value of forced labour services was estimated at 10 per cent for estates of the Lower Austrian princely chamber in 1570, they represented 14 to 21 per cent of total seigniorial income of noble estates in 1750/51, half of which came from cash payments in lieu of services not needed.83 The increasing use of wage labour also in regions where forced labour services were high such as in Poland, Mecklenburg or in the Brandenburg Uckermark finally resulted in labour rents being abandoned or becoming a marginal phenomenon even before their legal abolition.84 Sometimes, there was strong competition for wage labour, which was also needed in tenant households.85 Like many other features, this was neither a uniform nor a linear trend, as Zealand villagers experienced, who faced increases in labour services as late as in the 1790s, partly because demesne farms had been continuously extended over the previous two decades.86 In the estate of Boitzenburg in the Brandenburg Uckermark, two thirds of the demesne farms and three quarters of the demesne land were operated in Eigenbetrieb with wage labour just before the agrarian reforms in 1800, while the others were run on a mixed basis with slightly more than 50 per cent of the total work necessary being supplied by forced labour ser vices.87 Thus, from 1750 at least there was a widespread tendency among demesne lords and leaseholders of demesne farms alike to turn to using more wage labour.88

82 Cerman, “Arbeitsrenten”, 278–82. 83 Knittler, Nutzen, 42–7, 57, 217–18, 222–33. 84 Hagen, “Village life”, 178; Peters, Lebenswelten, 641–42. 85 Lubinski, “Work”. 86 Christiansen, World, 269. 87 Kriedte, Peasants, 112–13; Schmidt, Leibeigenschaft, 104. 88 Bentzien, “Fortschritte”, 151. regional variation and its causes 213

A survey of 657 demesne farms in Swedish Pomerania c. 1700 showed that 421 (61.1 per cent) were operated with forced labour services and 165 (25.1 per cent) were run as Eigenbetrieb (for 71 further farms there was no information). In particular small demesne farms were more likely to be worked with their own means. (Eighty per cent of the demesnes were smaller than 100 hectares.)89 In the Pomeranian case, studies have been cautious to interpret this relatively high proportion in direct management as a result of economically innovative behaviour on behalf of demesne lords. It can rather be seen as a reaction to the lack of tenant farmers due to expropriations and desertions, which forced lords into this mode. On average, each demesne farm could only draw on 5.5 tenant households to perform forced labour services. Rather than agrarian progress and awareness of the weaker productivity of forced labour services, it was the extreme nature of demesne lordship that caused this result.90 Given that the majority of demesne farms operated in this way were leased out, the situation may also have been caused because leaseholders were denied access to forced labour services. The described tendencies in the operation of demesnes therefore oc‑ curred for a mix of reason and a belief in a model of linear progress fitting all regions of demesne lordship would be wrong. Despite the trend, even in the period of peasant emancipation demesne lords were often not convinced of the superior productivity and organization of wage labour when they compared the different forms of work in running their estates. In the 1850s, Estonian demesne lords gave the lack of qualification, unstable employment situation and high fluctuation of workers as reasons that their expectations in the employment of wage labour were disappointed. Apart from paying wages, the necessary investments in draught animals often pushed demesne lords into delaying the introduction of wage labour.91

Conclusion

The development of the rural economy of early-modern East-central and Eastern Europe was traditionally associated with the omnipresence of labour constraints in ‘serf’ societies. Indeed, where labour rents or forced

89 Schilling, Schwedisch-Pommern, 49–50; Rudert, Gutsherrschaft und Agrarstruktur, 118. 90 Schilling, Schwedisch-Pommern, 54, 64, 108. 91 Dorošenko, “Herrenhof”, 199; Kaak, Peasant, 72–7, 89–91. 214 markus cerman labour services (corvée) existed, they belonged to the strongest burdens tenant households faced vis-à-vis their landlords. Arguments, conflicts and resistance against forced labour services, their increase or the way they were commanded occurred almost on a daily basis. In addition, also the obligation to send teenagers into life-cycle service on demesne farms (rather than having them available at home) was frequently experienced as a serious burden. The economic consequences of such services in terms of loss of workforce or draught power available to the tenant agricultural economy has often been analyzed and especially regular burdens beyond two or three days a week meant that households had to maintain additional ploughteams and farmhands. By contrast, this paper focussed on the problems of such generaliz ations that became apparent in most recent research. It is now clear from the evidence that the establishment of forced labour services was far from being a one-sided process driven by seigniorial interests alone. Secondly, it is not always possible to distinguish clearly between individual forms of labour constraints and ‘free’ wage labour in practice. Thirdly and most importantly, the early modern commercial demesne economy could not and did not rely on forced labour services or other forms of labour constraints alone. Rather, the organization of work was based on a complex mixture of forced labour services and wage labour, whose respective importance could shift over time and between places. Also with regard to the organization of work in the demesne sector, therefore, East-Elbian rural societies of demesne lordship seem far from being a uniform area of a (‘second’) ‘serfdom’. Rights and Bondage in Russian Serfdom

Alessandro Stanziani

Introduction

Already in 1921, Marc Bloch warned against the use of the word “serf” and the expression “serf of the glebe”. From this point of view, the “serf of the glebe” was used to identify a largely stylised feudal system and to oppose it to an equally stylised liberal economic system.1 Ever since, medieval studies have widely adopted this conclusion for France and Britain.2 More recently, a similar reassessment has been made for the German second serfdom (notably in criticizing Brenner’s argument).3 Quite surprisingly, discussions on Russian serfdom have focused on its origin (the State4 and/or the landowners5) and profitability6 not so much on the interplay between legal rules and economic activity. The most remarkable contributions along this path are those of Hoch, Wirtschafter, Moon and Melton who have strongly revisited the simplistic definition and functioning of Russian “serfdom”. Its dynamics and rules have been questioned by maintaining that serfdom was never officially introduced7 and that it was much more a set of practises than a system.8 We are willing to push this same way and analyzing rules defining the legal status of actors and their implementation. We shall show that in Russia, in the decades preceding the official abolition of serfdom, a quarter of the whole peasantry, that is, a half of private peasants changed their legal status and

1 Bloch, “Serf de la glèbe.” 2 Bois, La crise; Duby, Les trois ordres; Bonnassie, From slavery; Scott, The peasantries. 3 Brenner, “Agrarian class structure”; Aston, Philpin, The Brenner debate; North, From the North Sea; Melton, “Population structure”; Raeff, The Well-ordered; Shmidt, “über die bezeichnung”; Hagen, Ordinary Prussians. 4 Blum, Lords; Gershenkron, Economic backwardness; Crisp, Studies; Bartlett, “Serfdom”; Hellie, Enserfment; Kolchin, Unfree labour; Bush, Serfdom; Field, The End. 5 Hoch, Serfdom. 6 Koval’chenko, Russkoe krepostnoe; Blanchard, Russia’s age; Moon, The abolition; Den- nison, “Did serfdom matter ”; Melton, “Proto-industrialization”; Melton, “Enlightened sei- gnorialism”; Domar and Machina, “On the profitability.” 7 Barlett, “Serfdom”. 8 Examples of this in Hoch, Serfdom. 216 alessandro stanziani became state peasant, free artisan, free worker, etc. As David Moon has already shown, this means that the official abolition of serfdom has to be put into a longer process, starting at the end of the eighteenth century and that it will be not completed until WWI, when peasants got access to private property while on the labor market criminal punishments were repealed. However, we will show that shifts in the legal status of the peasantry before 1861 were not a purely administrative act (according to a top down reform), but they accomplished through judicial and/or administrative procedures. Peasants and artisans were able to mobilize rules originally conceived for nobles in order to prove their claims on estates. Juridical conflicts within the elites and the reformist attitude of some tsarist leaders made this issue possible. Our work is based upon archival and published sources. We made use of Russian archives, mostly archives of ancient Russia (RGADA) and local archives in Moscow (TsGIAM)(with local law courts decisions). To this one has to add the impressive collection of the Russian laws, decrees and jurisprudence9 which gives the opportunity to capture the incredible number of rules adopted on our topic, their connection with rules concerning nobility and family, but also the main jurisprudential decisions adopted by high courts and Senate in Russia.

Property Rules and the Legal Status of Russian Peasantry

The ways economic and social actor appropriate legal rules make the object of huge works in sociology, economics, history and anthropology, and, of course, in legal studies. In social and economic history, this approach has been developed for different countries and different periods. Historians of Russia have equally developed an ever increasing bibliography on this topic.10 As regards “serfdom”, it is widely accepted that peasants had no but a few legal rights until the reforms of 1861–1864. We argue that, if it is true that peasants had not the right to attack masters in justice, their

9 Polnoe sobranie zakonov Rossiskoi Imperii (hereafter PSZ) (Whole collection of the laws of the Russian Empire), three series: I: 1649–1825, 46 vv. Saint-Petersburg, 1830. II: 1825–1881, 55 vv. Saint-Petersburg, 1830–1884; III: 1881–1913, Spb, 33 vv 1885–1916. 10 Wirtschafter, Structures; Wirtschafter, Social Identity; Martin, Law and custom. Most of the available literature on legal action in Imperia Russia focus on post-1864 period: Wortman, The development; Wagner, Marriage; Burbank, Russian peasants; Frank, Crime; Solomon, Reforming; Frierson, “Rural Justice”; Frierson, “I must”; Czap, “Peasant-class courts; Kritika, special issues, 6,1 (2005) and 7,1 (2006); Pravilova, Zakonnost”. rights and bondage in russian serfdom 217 legal rights increased not only in 1864, but largely before, when volost’ (district) courts where institutionalized in the 1775. These courts were certainly limited in their scope and rights, as landlords controlled them. Most important, peasants had not the right to attack masters for physical punishment. At the same time, peasants could contest the rights of the estate owner not only in requesting their labor service or money fees, but, even more radically, her/his title of ownership. How was this possible? In order to answer this question, we need to carefully examine the legal entitlements of both landlords and peasants. Despite the lack of formal rules, serfdom actually existed and this reflected the power of nobility over state rules.11 Rules adopted in Russia since the middle of the sixteenth century up through the famous Ulozhenie (Statute) of 1649 linked the binding of the peasantry to the establishment of a cadastre in order to improve fiscal income and military conscription, and to settle estates for which different categories of nobles, the crown, the Church and monasteries advanced claims.12 Limitations to peasants’ mobility were not a goal in itself but they entered these relationships between different groups of landowners and the State.13 Nor peasants’ indebtedness14 nor State expan- sion15 was the main cause of such a complex process.16 Indeed, most of the nobility was hostile to territorial expansion, which was viewed a source of instability and was accused to reduced available manpower.17 Thus, measures for delimiting property were not only dictated by tax and military requirements of the State,18 but overlapped with a significant redef‑ inition of the relationships between social groups and the State, namely, the role of land property as a social and political marker.19 The Russian elites welcomed the request of the provincial nobility to legitimate their properties in order to secure an important ally in the fight against the

11 Wirtschafter, Structures. 12 Eaton, “Cadasters”; Voznesenskii, Materialy. 13 Cadastres documents (Pistsovye knigi) are largely available in RGADA in a large number of fonds, among which f. 1239, opis’ 3, chast 17, 69–72, 74, 76, 86–87 (cadastres of Moscow district, 1674–1681); fond 396, opis’ 2, ch. 5 (1616–1732), Smolensk, and several other districts; fonds 1209 (prilozhenie arkhiv premikikh votchnikh del, 1565–1692), opis’ 1, ch. 1–3, opis’ 2, ch. 1–2, opis’ 16–72. Cadasters documents published in Pistsovye knigi Moskovskogo gosudarstva (ed. N.V. Kalachov), St. Petersburg, 1872 and 1877. 14 Blum, “Prices”; Man’kov, Tseny. Smith, Peasant farming. 15 Field, The End; Perdue, “Military.” 16 Hellie, Enserfment. 17 Jones, “The nobility.” 18 Fuller, Strategy; Leonard, Reform. 19 Moon, The Abolition. Zaionchkovskii, Otmena. 218 alessandro stanziani boyari (high aristocracy). Until the first half of the seventeenth century, it was assumed that the binding was temporary; by the 1630s landlords could eventually allow their peasants the right to move, as expressed in document they signed (the otpusknaia gramota).20 This did not stop peasants from moving without authorization. Against this movement, heartland landowners and provincial cavalrymen were those who most complained of the peasants’ flying. Petitions multiplied between the 1620s and the 1640s and the central state responded by lengthening the time to recover fugitives from five, to nine to fifteen years (1637, 1641, 1645, 1648). However, this process was by no means simple, which is evidenced in numerous legal disputes and petitions21 drawn up by noble families against other claimants on their properties (other nobles, merchants, boyari, etc.).22 This game was complicated by the fact that landowners were ready to keep runaways peasants and that the state was unable to enforce the related rules.

This is where the famous Ulozhenie intervention of 1649 entered in. This document actually sought to find a solution to the same problems already regulated in previous rules, that is, the legal identification of a landowner and fair cooperation within this group.23 In particular, many articles of the section 11 were devoted to the documentary evidence the nobles had to provide in order to recover “runaways”. Landlords who accepted to hide fugitive peasants were sanctioned even more than peasants themselves.24 There was no time limitation to recover fugitive peasants. However, as the litigations among landowners and between them and urban merchants clearly show,25 the 1649 did not solve this problem. In the following years and decades, the great landlords became notorious for luring peasants away from smaller estates.26 This game became even more complicated when urban elites (ford fiscal needs) and peripheral authorities (interested in increasing the local population) pushed to keep the “runaways” in place. Negotiations on this occurred at legal, administrative and political level.27 During all the eighteenth century, several new

20 Hellie, Enserfment: 108. 21 RGADA, fond 615 (krepostnye knigi mestnyjh uchrezhdenii XVI–XVIII v), opis’ 1, several dela. 22 Crummey, “Sources”. 23 Man’kov, Razvitie. 24 Ulozhenie, chapter 11, n. 10. 25 RGADA, fond 615, opis’ 1; Fond 294, opis’ 1–3. 26 Melton, “Russian peasantries”: 239. 27 Kahan, The plow: 76–77. rights and bondage in russian serfdom 219 rules were adopted to assess these claims. As in the previous centuries, all these rules took a clear orientation, that is, limiting the definition of “hereditary nobles” (in opposition to service nobles on the one hand, to non-noble landowners on the other hand) and providing the ground for a “fair competition” among entitled landowners. The definitions were made gradually: in 1727, servants and peasants were excluded from civil service; a decree in 1730 prohibited servants, servants in monasteries, slugi and peasants from acquiring land, whether through inheritance or not.28 Next, in 1739, soldiers and lower echelons of the administration who did not own estates lost the right to acquire vol’nye liudy (house servant) and in 1746, the clergy, merchants, urban guilds, Cossacks and raznochintsy (people of various ranks) were prohibited from acquiring and transferring inhabited estates and establishing krepost’ (deed) relationships.29 Several rules adopted between 1754 and 1758 prohibited non-nobles from owning inhabited estates. Finally, in 1762, the factory owners were prohibited from buying or transferring estates with urban or rural workers.30 This did not prevent several merchants and manufacturers from acquiring populated estates and firms,31 so that in 1812, the Senate was forced to confirm that the decree of 1758.32 Least but not the last, the evolution in the legal status of the peasantry was strongly affected by the secularization of ecclesiastic lands in 1763, which placed 20–25% of the peasants under the State authority. The restrictions imposed on non-nobles regarding access to inhabited estates were followed by numerous petitions drawn up by non-noble officers, manufacturers and merchants demanding that a distinction be made between owning workers (urban or rural) and entire villages, which alone should be prohibited to non-nobles.33 This explains why numerous court proceedings also opposed landowners to the administration or landowners to each other concerning the validity of the noble title used by one or the other.34 Disputes occurred as well when estates were transferred or dowries were constituted, in cases of inheritance35 and finally, in view of

28 PSZ (I), vol. 8, n. 5633. RGADA, fond 342, opis’ 1 delo 37, chast’ 2 and 3. 29 PSZ (I), vol. 12, n. 9332, 9367. 30 PSZ (I), vol. 15, n. 11490, vol. 16, n. 11638. 31 RGADA, fond 294, opis’ 1, dela 385, 386; opis’ 2, dela 135, 145, 168. 32 PSZ, II, vol. 3, n. 2378; RGIA, f. 1149, opis’ 2, delo 44. 33 Coquin, La grande commission: 110 and 161–163. 34 TsGIAM, fond 54, opis’ 1; RGADA, fond 615, opis’ 1, several cases; RGADA, fond 342, opis’ 1, dela 64, 119, 120. 35 Several cases in TsGIAM, fond 54. and in RGADA, fond 1209, in particular opis 84. On this Jones, The emancipation, Wirtschafter, Structures: 71–74; Marrese, “The enigma”; Wagner, Marriage. 220 alessandro stanziani admitting the landowners of the newly annexed regions such as Poland to the rank of noble of the Russian empire. These already difficult relations were dramatically sharpened by the 1762 emancipation of the nobility from compulsory service; this made the traditional justification for serfdom disappear; nobles had no more to serve the tsar and they were therefore supposed to manage their estates. In this context, “entitled” nobles seemed to be the more and more interested in selling properties to non-nobles elites, to present a loan as a selling or viceversa, mainly for fiscal reasons, but also in response to financial difficulties, conflicts within the family, etc.36 Thus, in 1799, a merchant bought an estate inhabited by peasants from a noble landowner; the transaction remained valid as long as the merchant did not try and transform his possession into full ownership and therefore pass it on to his children. For in such cases, other interested parties—nobles as well as the peasants living on these estates—could sue. To return to the example of the merchant: he gave part of his estate and serfs as a dowry for his daughter (1810) and bequeathed the rest to his son (in 1817). The peasants protested and brought a lawsuit against him, which they won, and the decision was confirmed by the Senate.37 The reason for the decision was that the owner did not have the right to transfer the estate acquired in 1799. Non-nobles were not the only ones seeking to get around these provisions, however; noble landowners were also interested in doing so for tax reasons or simply to gain access to resources. The solution then consisted in presenting the sale of noble estates (or a part of them) to non-nobles as loan contracts; which was not necessarily false, insofar as the “merchants” often financed indebted nobles and the latter were unable to repay their debts. It was a form of mortgage. What would be normal in mortgage matters in Europe was now prohibited in Russia, for mortgages opened the door to the acquisition of the estate by the creditor; in Europe, this acquisition resulted in full ownership, i.e. transferable, which was impossible in Russia. One solution consisted in transferring the indebtedness of the noble and the seizure of the estate (non-transferable) to the heirs of both the formal owner (noble) and of his creditor (non-noble). To prevent this from happening, a law of 1816 (supported by both part of the nobility and

36 Senat still records a great number of these transactions in 1816. Arkhiv gosudarstven- nogo soveta: 4, vyp. 1, ch. 2, 253–258. 37 RGIA, fond 1149, opis’ 2 delo 20. rights and bondage in russian serfdom 221

Russian officialdom) prohibited the recording of letters of credit in the name of persons who were not entitled to own and transfer peasants.38 Despite these provisions, transactions between nobles and non-nobles continued. To escape the law of 1816, non-nobles used verbal agreements in order to secure peasants’ work or to grant loans to nobles, etc.39 These offences multiplied in the ensuing years; numerous decrees tried to limit the practice of verbal agreements between nobles and non-nobles on the subject of peasants.40 Still in 1836, the “personal” nobles (non-hereditary) were confirmed in their right to own inhabited estates but not to transfer them.41 These provisions were followed by an increasing number of disputes among nobles, service nobles and hereditary nobles, or between nobles and the administration. According to the statistics of the Ministry of Justice, in 1845 alone, 6,400 requests for confirmation of the title of “noble” were submitted, only half of which were validated.42 To a given extent, these nineteenth century rules defining who were legally entitled in possessing and transmitting populated estates were perfectly coherent with sixteenth through eighteenth century rules. They confirmed the distinction between “entitled” owners and the others. At the same time, they sought to answer to a different dynamic, that is, the interest of some groups of “entitled” landowners in establishing transactions with other social groups. The sale of serfs has therefore to be explained in this context made of official rules designing those who were entitled to own and transmit inhabited estates. Many literary sources and a good number of other political or administrative sources mention the sale of serfs. Classical references in historiography such Blum, Madariaga, Hellie, just to mention a few of them, refer of the market for serfs. Demographers currently speak of the prices for serfs,43 while economic historians have also tried to evaluate the trend in serfs’ prices.44 Still, most of these works identify the prices of serfs as the result of statistical regressions and/or proxies of other variables (land prices, fee to pay for mobility, dowries etc.). This

38 PSZ (I), vol. 33, n. 26469. 39 RGIA, Fond 1149, opis 2, several dela. 40 Decrees of 1825, 1828, 1840, 1847. PSZ (I) vol. 40, n. 30407, psz II, vol. 3, n. 1696, vol. 15, n. 13051. 41 PSZ (II), vol. 11, n. 9203. 42 Ministerstvo iustitsii, Otchet: xix. 43 Bushnell, “Did serf owners”. 44 Mironov, The social history. 222 alessandro stanziani necessity is partially due to the fact that sources used in this literature are mostly official sources (ministerial reports and data, reviziia, etc.).45 This actually is an intriguing puzzle: why did official sources did not speak of the price of serfs? Estate sources provide a first answer: registers, accounting books, etc. often indicate the price of serfs; at the same time, when addressed to official institutions, these same estate documents often spoke much less of price of serfs or of sale of serfs than of fees, dowries, transfers, gifts, etc.46 Notarial records offer a quite similar attitude.47 The immediate comment these findings lead to suggest is that for fiscal and/or legal reasons the selling of serfs was made under the cover of other formal practises; however, such a legal fiction would not deny the fact that a real market for serfs existed in tsarist Russia. As Isabel de Madariaga put it, it was a de facto if not de jure slave trade.48 However fiscal reasons alone cannot explain the lack of serf prices and a market for serfs in official sources. Another reason for that, was that Petr I already disliked the sale of individuals, in particular the break up of families. It imposed a tax of 10% of the price in roubles on sales without land (decrees of 1701, 1705, 1720). This was confirmed in 1773.49 Last but not least, the lack of official legal transactions concerning serfs was that all the sets of rules we have mentioned strongly limited those who could own and transmit inhabited estates. They forbade the sale of peasants without lands, of peasants of mortgaged estates, of those of landowners with no heirs, etc. At the same time, the fact that so many transactions of serfs were made despite official rules, was not without consequences about the effective- ness of these practises. Members of noble families, state administrations, but also merchants and ultimately peasants contested titles of ownership.

Changing Legal Status through Judicial Proceedings

In 1836, while confirming that personal, non-hereditary nobles were not entitled to transfer their peasants, a new law encouraged them to redeem

45 For example, Blum, Lords, Domar and Machina, “Profitability”, and Mironov, The Social. 46 RGADA, fond 1287 Sheremetevy, opis’ 3 chast 1, delo 582; fonds 615, opis’ 1. Fonds 615, opis’ 1. 47 RGADA, fond 615, opis’ 1; fond 1273, opis’ 1. 48 Madariaga, “Catherine II”, in particular, 38. 49 PSZ, 19, 13950, 21 febr. 1773. rights and bondage in russian serfdom 223 those peasants in exchange for a compensation paid in part by the State and in part by the peasants themselves.50 The adoption of this law was followed by accelerated voluntary “emancipations”, which also occurred during the years prior to the official abolition of serfdom.51

However, changing legal status also resulted from the increasing number of legal proceedings brought by the peasants themselves against estate owners. In general, illegal enserfment was an offence under the criminal code; while before 1845, the punishment was limited to a fine after that date the usual penalty was imprisonment, in addition to emancipation without compensation.52 Even more than the statutory penalties, legal proceedings and arguments may legitimately be seen as conditioning peasants’ chances of winning their cases. Thus, until the 1770s, there were no special courts in Russia for peasants, who were forced to go through nobles and their courts to request emancipation, which considerably reduced any attempt to initiate proceedings. The situation changed in 1775 with the introduction of peasant courts, accompanied by a clear- cut legal separation between ownership of things and rights over human beings. These courts and their scope of jurisdiction were to be expanded by reforms of 1837–1841, when peasants were given increased rights and the possibility of denouncing invalid ownership of estate and, thus, landowners’ entitlements.53 Throughout the first half of the nineteenth century, conflicts between peasants, peasant-workers and estate owners on both peasants’ obligations and owners’ titles of ownership were on the agenda of the courts. Usually, peasants who won were reclassified either as state peasants or as urban residents.54 The arguments evoked by the plaintiffs testify first of all to the peasants’ knowledge of the legal rules governing the possession of estates and that we mentioned earlier. For example, in many cases, plaintiffs emphasised that the transfer of the estate took place through the intermediary of a non-noble (which was prohibited, as we have seen), or by an emancipated peasant; in other cases, plaintiffs demonstrated that the landowner was not a noble but rather a merchant and as such,

50 PSZ (II), vol. 11, n. 9203. 51 Khovanskii, “Pomeshchiki”, 60, 75, 118. 52 PSZ II, vol. 20, n. 19283. 53 Efremova, Sudoustroistvo. 54 Decisions of Senate in that sense in PSZ (II), vol. 3, n. 2378, vol. 8, n. 6129, vol. 9, 6941, vol. 10, n. 8539, vol. 11, n. 9203, vol. 15, n. 13051, vol. 17, n. 15693, vol. 20, n. 19283, vol. 22, n. 20825. 224 alessandro stanziani prohibited from possessing serfs. In still other cases, the landowner was a non-hereditary noble who transferred his estate and its population to his children, etc.55 In short, in their legal disputes, the peasants took advantage of every case involving the definition of the beneficiaries of rights to own inhabited estates (false nobles, widows, merchants, indebted nobles, etc.).56 Together with the owner’s title of property, the peasant’s obligations could make the object of litigation. For example, in 1815, workers in a factory relied on these provisions to win their case, demonstrating that the widow of their former boss was not registered in the local guild and hence his apprenticeship contract was null and void and along with it, his rights over the workers.57 Coercion in manufactures, shops and cottage industry were quite often sanctioned.58 Other cases of sanctioning the title of the owner and the obligation of the peasant-worker were linked to the civil status of the concerned people. A married peasant, who had fled with his parents before the age of fourteen and therefore could not be considered a peasant of a private estate; she was thus redeemed of any obligation to the estate owner. Similar issues included widows or wives of private peasants who married someone belonging to another legal category; a given peasant who was taken while still a minor in the provinces of Voronezh, Kharkov, Stav- ropol, Don Cossack or the Caucasus; a non-Christian peasant owned by a non-Christian master who converted to the Orthodox religion; a peasant who was made a prisoner of war; a peasant that had denounced his master for treason. The use of classifications was essential in these proceedings. According to the law, servants and peasants of a bankrupt landowner who owned only domestics and peasants without land were also transferred to the category of state peasant. The peasants of a landowner who mortgaged his estate separately from his peasants, to whom he left less than 4.5 desiatine (1 desiatina= 1.10 ha) per family, became state peasants.

55 TsGIAM, fond 54, 1783–1917, opis’ 1, for example: dela 56, 284, 966, 1509. 56 On transfer to married women and involved questions (legal statute): TsGIAM, fond 54, opis 1 et 2; for ex. Opis 1, delo 284; RGIA, fond 1149, opis 2 delo 20 and delo 44. The issue was not always favourable to the peasants and workers. Example in GARF, fond 109 opis’ 3, delo 1885. 57 Quoted in Wirtschafter, Structures of society, 81. Peasant petitions in Materialy dlia istorii. 58 TsGIAM, fond 32 (Moskovskoi gorodskoi Magistrat), op. 1 (1842–1866), dela 5, 13, 46, 118, 153, 284; opis’ 2 (1823–1844) dela 31, 40, 82, 124, 146. rights and bondage in russian serfdom 225

Finally, the peasants of a landowner who had illegally gone abroad were also redeemed.59 The proceedings brought by peasants became so numerous that between 1837 and 1840 the Senate even decided to put a stop to cases involving serfs still living with their masters.60 Overall, between 1833 and 1858, the Senate recorded 15,153 cases of illegal ownership and “illegal bondage” (as listed and qualified above), whereas the provincial courts dealt with 22,000 cases of this type.61 Notaries’ archives and estate archives provide supplementary information. Looking at some of these archives, the impression is that changes in the legal status of private peasants made at the initiative of the estate owner accelerated during the nineteenth century in response to both fiscal support and legal simplification as provided by the tsarist rules.62 According to estimates at the time, the ninth census in 1851 in the twelve provinces (guberniia) resulted in transferring eleven thousand peasants into the category of meshchane.63 The validity of these figures is hard to determine, for, in any case, they do not take into account all the emancipated people that did not register on the urban lists of meshchane. Though not cumulative, relatively detailed data on these individual acts was collected by the Tsarist authorities between 1860 and 1868 in 25 volumes on the state of the population in the Russian provinces and the Empire.64

Naturally, these positive outcomes should not obscure the difficulties confronting peasants who tried to bring proceedings against a noble or someone claiming to be noble. The rulings of local courts often differed considerably and several judges considered peasant petitions inadmissible and even refused to grant them the right to appeal.65 A number of cases of corruption of judges by noble landowners were also recorded. Finally,

59 Svod zakonov, vol. 9, 957 (1842), arts. 544, 1040, 1042, 1078, 1107, 1185–97, 1199, 1201, 1205 (1857); PSZ, I, vol. 30, n. 23806 (1809); PSZ, II, vol. 2, n. 906 and 1144 (1827), vol. 3, n. 1696, (1828), n. 2052 (1828), vol. 7, n. 5425 (1832), vol. 8, n. 6163 (1833), vol. 10, n. 7982 (1835), vol. 11, n. 9203 (1836), vol. 16, n. 14152 (1841), vol. 23, n. 21929 (1848), vol. 24, n. 23508 (1849), vol. 29, n. 28846 (1854). RGIA, fond 1149, opis’ 2; TSGIAM fonds 16, 32, 54, 342. 60 RGIA, fonds 1149, opis’ 2, delo 90. 61 PSZ (II), vol. 20, n. 19283, vol. 22, n. 20825. RGIA, fond 1149, opis’ 3, delo 125. 62 Notarial archives in RGADA, fonds 615, opis’ 1; fonds 1253 and 1274. 63 Keppen Deviataia reviziia: 6, 7, 21, 88, 95–100, 127, 142–144, 152, 159. 64 Glavnyi General’nyi Shtab, Materialy. 65 RGIA, fond 1149, opis’ 2 delo 20. 226 alessandro stanziani the investigations of proceedings went slowly and often took ten years to reach a conclusion. Measures aimed at changing this state of affairs were not adopted until the end of the 1840s, when a new law facilitated the legal proceedings of all those who considered their obligations towards estate owners as illegal.66 However limited they were, judicial procedures ending with a change in the legal status of concerned peasants entered a broader process in which administrative law (or changes from above) strongly supported these shifts.

New General Rules for Local Emancipation The general issue of procedures concerning the legal status of peasants consisted in transferring private estates’ peasants into the category of state peasants, and, sometimes to other categories (mostly raznochintsy, people of various ranks). Since the seventeenth century, and more clearly in the 1720s, peasants of nobles and private estates had been distinguished from state peasants. The latter, which most historians, especially in the West, immediately included in the serf category, only had certain obligations to the State, e.g. to pay an annuity, usually in money, sometimes in product. They also had an obligation to perform work of common interest (which is usually mentioned to justify the existence of forced labour and serfdom in this category). At the same time, state peasants could engage in activities in the cities in trade and in industry, provided they had the necessary documents.67 Catherine II even included representatives of the state peasants in her legislative commission assigned to define the rights and obligations of the main social groups.68 In 1842 state peasants were freed from any obligation (except tax) towards the State. With this distinction in mind, we may now examine the ways a private peasant can be “switched” to another category. Voluntary redemption of peasants’ obligations and thus the change of her/his legal status by the estate owner was already provided for in the seventeenth and eighteenth centuries; a noble landowner could free a peasant from his obligations with or without land. In the latter case—the most widespread—the act could be freely carried out; the civil authorities gave their consent and the will (in the case of emancipation after

66 Law of 1847, in PSZ II, vol. 22, n. 20825. 67 Druzhinin Gosudarstvennye. 68 SIRIO, X, 285–8. rights and bondage in russian serfdom 227 the landowner’s death) was recognised as valid.69 The noble landowner was also responsible for paying the capitation or head tax on his peasants until the next revision, unless a different agreement in the emancipation contract stipulated that the involved peasant would pay his own tax. The latter possibility had been adopted at the time of Catherine, precisely to facilitate changes in the legal status of private peasants70 and it was subsequently encouraged by the laws of 1844 which exonerated freed servants from paying any tax until the next census (a time period reduced to six years for those who registered on urban tax lists as meshchane).71 Indeed, many laws were passed during the first half of the nineteenth century to facilitate administrative procedures for changes of the legal status of private peasants. While responding to contingent situations we are going to detail hereafter, these rules as a whole expressed the particular stakes and worries of a part of the Russian elites. Political stability, economic efficiency, paternalistic criticism to serfdom, concrete economic interests of the State and a part of the nobility, all these variables took part, to a different degree, to these reforms.72 A first reform was adopted by Alex- ander I who ascended the throne in 1801. This same year he allowed merchants, townspeople and state peasant to own unpopulated estates. This was intended as a partial breaking of the nobles’ monopoly on landowner- ship. In 1803, a decree created the svobodnye khlebopashtsy (free farmers), i.e. those who, by will of a landowner, were exempt from any obligation to him, aside from those resulting from the attribution to peasants of plots of land belonging to the owner. The agreements were legally binding for both peasants and landowners.73 The influence of German and Russian academicians, as well as the pressure of some of the noble elites encouraged the tsar’s reforming attitudes. In particular, Alexander was pushed to adopt this provision by count Rumiantsev who sought permission to free some of his peasants of any obligation and sell them land. The positive answer to this request became the ground for a broader reform. This provision was supposed to lead to an improvement of agriculture while being advantageous to estate owners. According to the estimates of Hoch and Augustine, between 1833 and 1858, 58,225 private peasants changed

69 Svod zakonov, vol. 9, art. 674–680, 1833, art. 1148–1184, 1857. 70 PSZ, I, vol. 20, n. 14294 (de 1775). 71 PSZ, II, vol. 19, n. 17977 (12 juin 1844) and n. 17985, 12 juin 1844. 72 Becker, Nobility; Dukes, Catherine; Jones, The emancipation; Lieven, The Aristocracy; Lincoln, The great reforms; Pintner, Russian. 73 PSZ, I, 27, n. 20620 (20 février 1803). 228 alessandro stanziani their legal status on the basis of this law,74 a figure that rose to 114,000 male peasants between 1803 and 1855. However, despite several attempts to go further in this direction, Alex- ander did not adopt other reforms. The Napoleonic wars and the opposition of some groups of nobles and high officials seem to have played a role in his attitude. The greatest worry was not so much the change of the legal status of peasants than their possibility to sell the acquired lands to other “non-noble” owners.75 The next tsar, Nicholas I, also spent some time in projecting reforms before adopting concrete rules. These occurred in the broader realm of Kiselev’s reforms that between 1838 and 1842 radically modified the conditions of state peasants. Peasants’ general and professional education was taken into account, while the fiscal burden and the obligations toward the State were no longer assessed per head, but in conformance with the value of peasant’s land or their incomes from non-agricultural activities.76 Kiselev ordered a general enquiry on agriculture and the peasantry; the final document reached the conclusion that the labor of peasants with no statutory obligations toward the landowners was more productive than existing forms of labor. On the ground of this report, Kiselev and Nicholas I maintained that the present difficult situation of agriculture and the existence of coerced labor was the result of the fact that some landowners had abused their power and extended their seigniorial power over the peasantry. Limitation of the abuses thus became a precondi- tion for deeper reforms. At the same time, peasants had to be given the opportunity to own their own land. In fact, as the experienced reforms in the Western borderland had proved, the end of statutory obligations without land would lead to a worsening of the peasants’ condition and to social and political instability. According to this, in 1842 a new decree was adopted; it was conceived as direct filiations of the 1803 decree, but with the major difference that the contractual agreement between the estate owner and peasants would not be left to the free will of the former but would be regulated by the law. Until when the involved peasants had not redeemed the price of land, estate owners would have retained the right of full patrimonial ownership of the land and the peasant would receive plots of land for use. A particular category of peasants was created; they

74 Hoch, Augustine, “The tax censuses”, 208–217. 75 Moon, The abolition, 42; Mironenko, Samoderzhavie. 76 Druzhinin, Gosudarstvennye. rights and bondage in russian serfdom 229 were the peasants with a temporary and specified obligation (obiazan- nye krest’yanie) towards noble landowners, who fit neither the category of private estate peasants in the strict sense nor that of servants (dvornye liudi). As previous rules, this decree too accurately spoke of peasants, private peasants and rural inhabitants, not of “serfs”. These peasants were subject to other legal rules; they had no more general obligation towards landowners, except those concerning the plot received at the signing of the contract with the estate owner.77 From the adoption of this law until 1858, 27,173 male peasants (and thus their families) were affected by the decree.78 In 1841, peasants that private nobles had freely transferred to monasteries or charity institutions were also re-classified due to the prohibition against ownership and transfer of peasants by these institutions.79 About 8,900 people were thus transferred to the category of state peasants and the institutions to which they belonged were compensated by the State.80 In 1844, two new laws facilitated the re-classification of dvorovnye liudi (servants); the first one freed the master from any responsibility regarding the payment of the taxes of these people, whereas the freed servants were not subject to military service or taxation until the next revision. The second law facilitated the change in legal statute in cases in which the landowner mortgaged his estate to certain credit institutions. According to a study at the time, in 1851 alone, 11,000 meshchane (merchants) from eleven provinces are said to have benefited from these rules (they were previous peasants of private owners).81 A great many worker-peasants attached to private factories were exempted from any obligation toward the factories owners between 1840 and 1851; about 19,000 people were concerned. During the same period, many mines had recourse to “temporary workers” (nepremennye ou urochnye rabotniki) registered as private peasants at the time of the eight revision. The facts were denounced years later and a court decision was made only for the ninth revision in 1851. At that point, 53,900 men working in the mines were transferred to state estates as peasants.82 For both factories and mines the argument behind these provisions was that the

77 PSZ, II, vol. 17, n. 15462. 78 Hoch, Augustine, “The tax”, 410. 79 PSZ,II, vol. 16, n. 14669, 19 June 1841. 80 Mel’nitskii, Sbornik, vol. 2, vyp. 3, 187, 217; vyp. 4, 52, 119. 81 Keppen, Deviataia reviziia, 6–7, 21, 62, 88, 95, 100, 127, 142–144, 152–159. 82 Ryndzinskii, “Vymiralo”. 230 alessandro stanziani owners were not entitled to get statutory (that is non contractual and timeless) obligations and services from the workers. Administrative re-classifications sometimes occurred for other reasons. State and noble elites agreed on this, indebted landowners had to be helped and avoiding their lands would fall into the hands of “speculators” and “bourgeois”. To this aim, in 1841, a new law prohibited the succession of landless nobles or landless owners of peasants. Consequently, between 1841 and 1858, the number of landless nobles dropped from 17,700 to 3,633 and their peasants from 62,000 to 12,045 (even taking into account the possibility of some of the serfs having been sold to other nobles).83 To this same aim, mortgaged estates put up for auction were especially targeted. Local peasants were granted the priority (with the State itself) to acquire lands. Otherwise, the “public treasury” (udel) and, then, the Min- istry of State dominions, intervened and bought estates. In all, between 1833 and 1858, the legal status of 34,3575 male peasants changed because of these operations. Ordinary and administrative law regarding emancipation had a cumulative effect not only on mortgaged estates, but also on the terms of succession. Two new laws adopted in 1836 and 1839 stipulated that when a landowner did not leave any heirs (or heirs that could legitimately have serfs), the peasants would become state peasants whereas the servants could register in cities in the category of meshchane.84 Other provisions admitted changes in the legal statute of peasants as a consequence of political crimes made by landowners. Noble estates of Western Borderlands, Poland, West bank of Ukraina who took part to uprising were confiscated and their peasants (estimated 264,000) entered the category of state peasants.85 Ultimately, beyond peasants voluntary redeemed by their masters and private peasants turned into state peasants following a judicial litigation or an administrative act, massive changes in the legal status of peasants was linked to conscription and military service.

83 Hoch, Augustine, “The tax”, 420. 84 Svod zakonov, vol. 9, art. 462–3 (1857); PSZ, II, vol. 11, n. 9203 (1836), vol. 14, n. 13012 (1839). 85 PSZ, II; vol. 11, n. 9053; See also Ministerstvo gosudarstvennykh imushchestv, Istoricheskoe, part. 2, 80. rights and bondage in russian serfdom 231

The Army As a Factor of Emancipation?

Conscription was the most important source of administrative change of the legal status. Once a conscript had completed his military service, he was classified as raznochintsy and could therefore move about freely, settle in the city, etc. Hoch and Augustine estimate that about 433,750 peasants changed their legal status (from private to state peasants or even urban residents) in this way between 1833 and 1858 alone.86 This outcome was hardly considered favourable by the peasants, however. Indeed, the length of conscription (20–25 years according to the period) implied first and foremost a decrease in workforce for the family, which also had to continue paying a tax for the conscript until the following revision (which might take ten to twenty-five years). Moreover, upon their return, conscripts were relatively old and hence less productive and advantageous for both their families and landowners. Re-classified ex-conscripts (from peasants to raznochintsy or others) were not entitled to land, although most soldiers requested it instead of freedom of movement, which they did not know how to manage very well. Numerous requests and petitions were then sent to village elders, noble lords and local authorities to be fully reintegrated into the rural community, which were nevertheless refused.87 Of course, the perspective of changing the legal statute at the end of the service did not imply that soldiers were allowed to escape before the term. In this case, two crimes were simultaneously acted: desertion and run- away (as a peasant belonging to a given rural community and a given landlord).88 The new social definitions and redefinitions of the legal status of conscripts also affected their families. For example, children born after the soldier entered service were placed under the guardianship of the Ministry of War and destined for military service. This provision and its application gave rise to a number of disputes, petitions, etc. brought by both peasants and nobles who saw themselves deprived of considerable manpower for the benefit of the army.89 Most of these disputes were based on the definition of “entry into service” and “birth”; military leaves were indeed

86 Hoch, Augustine, “The tax”. 87 Some examples in TsGIAM, fond 16, opis’ 1, dela 1293, 1328, 1352, 1801. 88 TsGIAM, fond 16, opis 1, 2, 225, 227 (several files in each). On the importance of desertion at the origin of “runaways”, Kahan, The plow, 76; Stevens, Soldiers. 89 TsGIAM, fond 16, opis’ 2; delo 192; fond 54, opis’ 1, delo 1618; fond 16, opis’ 2, delo 192; fond 54, opis’ 1, delo 1618. Also: TsGIAM, fond 16, opis’ 227, dela 227, 938, 948, 976 980, 986, 1573, 1571, 1588, 1589. 232 alessandro stanziani evoked, as well as the interpretation of the laws, some referring to birth, others to conception. False birth dates, encouraged by both peasants and noble landowners, were commonplace.90 The other subject of dispute concerned the peasants’ illegitimate children; their families and the nobles considered that these children did not fall into the category of those destined for the army and therefore recommended a restrictive interpretation of the rules in this area; the Ministry of War, on the contrary, considered that these children belonged to it. Thus, in 1808, the Senate made an example of Princess Golitsyna due to her rank, finding her guilty of illegally enserfing two illegitimate children of the wife of a soldier from her estate.91 These conflicts intensified with the Napoleonic wars, so that a 1816 decree declared that all children of soldiers (whether legitimate or illegitimate) who had been mistakenly enrolled in the estates of the State or of a noble in the previous revisions would retain their status.92 On the whole, the number of children who changed their legal status by this method has been estimated at about 12,000 between 1833 and 1858. Added to the re-classification of soldiers, the number of peasants who changed their legal status in connection with the status of soldier or child of soldier rose to about 450,000.93 The army thus acted as a powerful source of changing the legal statute of peasantry, mainly private owners’ peasants, who became either state peasants or enter the urban population. As such, this provision joined several other rules the tsarist regime adopted in order to proceed in administrative changes in the legal statute of the peasantry.

In all, between 1833 and 1858, i.e. between the eight and the tenth revision, about a million men were released from the status of peasants of private owners due to administrative acts (the army plus the other cases mentioned), out of a total of about ten million serfs.94 More than a quarter of all male peasants was switched to other categories and became either state peasants or urban dwellers.

90 RGADA, fond 294, opis’ 1, delo 567. 91 Wirtschafter, Social, 578. 92 PSZ, I, vol. 33, n. 26376, PSZ, II, vol. 12, n. 10727. Also TsGIAM, fond 16. 93 Hoch, Augustine, “The tax census”. 94 Hoch, Augustine, “The tax”. rights and bondage in russian serfdom 233

Conclusion: Labour and Economic Dynamism in Imperial Russia

Two main elements are usually called upon to explain “serfdom”: the State and noble landowners. Historians talk about the State in two different ways, which are totally incompatible. For some, serfdom is held to be due to the strength of the autocracy, whereas others emphasise, on the contrary, the weakness of the Tsarist State, which is said to have left the nobles free to enserf the peasants. These two interpretations are partial, if not erroneous, in that they do not take into account the close interaction between the State (or rather, certain administrations and certain leaders) and Russian society. The State was more than an administration and above all, the latter was not limited solely to taxation and the police; the State was also the law, and from that standpoint, the dominant interpretations miss the essential point when they assert that the autocracy imposed a law that was not really law or, on the contrary, that it was incapable of controlling relationships in the countryside. In reality, Tsarist law recognised and relied on other legal sources, including seigniorial law and its courts, and these rules mutually legitimated each other. The disputes between nobles and merchants or the considerable emancipations of peasants through legal and/or administrative channels testify to the importance of this element. However, the law was hierarchical; the various social groups had varying access to the courts and to jurisdictions. In other words, while it is not true that Russian society had no law, it would be a mistake to equate it with the rule of law as defined, let us say, in France or in the United States at the end of the twentieth century. The way law was conceived and implemented shaped social and economic hierarchies. These last cannot be seen as a simple coerced exploitation of “serfs” made by noble landowners.

The peasants on private estates were subject to considerable restrictions on their mobility. However, these restrictions were never definitive and they changed according to the estate and the economic situation. As a whole, these provisions corresponded to a reorganisation of the administration and the introduction of a land registry in order to settle conflicts between nobles and merchants, or even within noble families, over the transfer of estates and the attribution of “noble” status. Our approach leads us to raise doubts on the two main interpretations the historiography provided of 1861 reforms, one stressing their revolutionary 234 alessandro stanziani role and the other undermining their partiality and ultimate failure.95 The first interpretation underestimates previous reforms, exaggerates the importance of serfdom and the development of Russian rural economy after 1861. The second interpretation shares the same perception of a tough and stable system called “serfdom” and puts the accent on peasants’ indebtedness and miserable standards of life after 1861. The first interpretation has difficulties to explain the huge mass of “emancipation” acts we have dealt for, the economic growth before 1861 as well as peasants’ hostile reactions to the emancipation act. The second approach seems unable to take into account the labor institutions after 1861, the economic growth of pre-revolutionary Russia and nobles attitudes towards reforms as well. Indeed, the reforms and judicial acts we have detailed testify to the fact that the reforms of 1861 were part of a long reform process dating back at least to the last quarter of the eighteenth century, which gained strength during the first half of the following century and finally accelerated significantly starting in the 1830s. This dynamic also expresses the complexity of a process that it would be an exaggeration to describe as purely administrative, i.e. as a “reform from the top down”. First of all, because, as we have seen, behind the many laws and decrees modifying the conditions of emancipation stand social groups, both noble and bourgeois. The second reason that leads us to question the argument of “top-down” reforms is that the aforementioned laws and decrees were situated at the crossroads of administrative law and civil law procedure. The administrative authorities adopted measures based on common law rules pertaining to the transfer and mortgaging of estates, and above all, defining those entitled to full ownership of estates and their inhabitants. The reforms of 1861 thus do not mark a break insofar “serfdom” was much more flexible than usually asserted and legal constraints on peasant mobility and peasant labour did not disappear after 1861. In fact, the reform program as it was laid down in 1858 was largely inspired, as regards peasants’ obligations and legal status by the decrees of 1803, 1842 and 1847. As we have seen, these decrees fixed the rules for peasants’ legal emancipation and land acquisitions; they had supported the emancipation of around 240,000 peasant families. The “emancipation act” of 1861 reproduced the main features of these previous rules and only added a broader financial state support to both peasants and landlords in order to facilitate peasants’ access to land.

95 Gershenkron, Economic backwardness; Blum, Lords and peasants; Field, The end. rights and bondage in russian serfdom 235

All the same, between 1842 and 1856, the ministry of state domains assembled impressive quantities of data about life in state peasant villages and the distribution of land between peasants and landlords and among peasants. These surveys covered almost 11 millions peasants and more than 65 millions acres of land in 25 Great-Russia provinces. Least but not the last, the Russian government imposed inventories defining schedules of peasants’ duties to landlords and landlords’ obligations to their bondsmen.96 Our approach equally provides a solid justification to recent estimations of economic growth and demographic trends between 1750 and 1861, as well as from 1861 up through 1914, both having revised upwards.97 These dynamics are much less explained by the possibility of further squeezing peasants under serfdom, than by the coexistence of coercion and flexibility. This was all the more true that before 1861 a large amount of the peasantry had already been transferred into the category of state peasants or urban inhabitants. This widely explains the good performance of Rus- sian agriculture now widely accepted in historiography. Our perspective provides a ground for justifying peasants and nobles reactions to the “emancipation act”. Most of the peasants negatively reacted to the 1861 act in which they saw nothing but a smooth shift from previous conditions. We must to be clear: this does not mean, in our view, that the reforms of 1861 were a “failure” and that they gave landowners a new way for exploiting peasants. Quite the contrary, the reforms went on along the previous path and aimed to strengthen the mixture of coercion and flexibility—mostly to the social appropriation of legal rules—already experienced since the early nineteenth century. Peasants’ had much less duties, a better legal protection and rights and a scheduled access to full property. Still, this was often perceived by the peasantry not as a “revolutionary break” but rather as a further step along the previous reform lines.

What about Landlords? Several scholars have been surprised by the fact that, beyond clearly documented opposition, several nobles largely expressed their assent to reforms.98 This was so not only because of their political weakness and

96 Lincoln, The great reforms, 66–69; Field, The end, 77–83. 97 Gregory, Before command; Gatrell, The Russian economy; Stanziani, L’économie. 98 Field, The end; Moon, The end. 236 alessandro stanziani the social differences among them, but also because nobles perceived that the ongoing reform was in a strong continuity with the previous path. Landlords kept part of peasants’ duties while receiving considerable financial resources from the State. Peasants’ obligations were mostly kept in life while the state financed the transfer of less valuable lands to the peasantry at an extremely interesting price for noble landowners. This for nobles’ support or at least for their lack of reaction to the reforms. At the same time, the 1861 act introduced a new dimension (beyond the financial support the state gave to both peasants and the state) into the reformatory process. This new dimension was at the origin of nobles’ hostility towards the reformatory process. In fact, labor relationships in agriculture were transferred from landlords and estate courts to district courts. This was seen by several nobles as an intrusion of the state into peasants-landowners relationships and marked a considerable change as regards all the previous three and a half centuries. The end of seigniorial justice was not only in Russia, but everywhere in eighteenth and nineteenth century Europe one of the major breaks with the old regime. A second major break and a reason for nobles’ dissatisfaction was that the reform erased any legal relationship between land ownership and social (and political) status of “noble”. By now, no-one was entitled to transmit inhabited estates and, conversely, merchants, kuptsy and urban elites had access to land. Nobles clearly understood this shift and appealed to counter-reforms. Tsarist elites attempted to take into account this last request by strongly limiting the market for land (peasants could not sell their lands to non-peasants or to peasants from other villages), while maintaining an important status and political difference between nobles and merchants (and, of course, peasants and workers). However, these last defences to nobles privileges were progressively abandoned, the latter with the institution of Duma and the former with the introduction of a free market of land at the eve of WWI. Acting as Master and Bondservant: Considerations on Status, Identities and the Nature of “Bond-servitude” in Late Ming China

Claude Chevaleyre

Introduction

In early 1910, following the recommendations of the Commission on Consti- tutional Government, an edict abolishing slavery and trafficking in human beings in the Chinese empire was officially promulgated in the name of the four-year-old Xuantong emperor (r. 1908–1912).1 A few months later, the publication of the new Penal Law of the Great Qing Currently in Use, compiled by the Commission for the Revision of the Code, confirmed the emancipation of all slaves within the empire. A new sub-statute appended to the law entitled “Households should be established according to the registers” (renhu yi ji weiding 人戶以籍為定), stipulated that “household slaves” ( jianu 家奴) were granted permission to redeem themselves, whatever their origins and the number of generations their families had been in bondage. In order to enforce this decision, local authorities were ordered to register redeemed slaves as commoners and to let their sons and grandsons take the official examinations. For those who could not redeem themselves, the law stated that they should be considered as hired laborers. This raised their legal status from that of unfree laborer to that of free laborer, without entirely erasing the personal tie that linked them to their former masters.2

1 For more details on the process that led to the proclamation of the ten regulations for the prohibition of the purchase and sale of human beings, see Edward T. Williams, “The Abolition of Slavery in the Chinese Empire;” see also his translation of the documents related to this emancipation, “Report to the Throne of the Imperial Chinese Commission on Constitutional Government Recommending the Abolition of Slavery.” A former Prot- estant missionary and official translator in China, Edward Thomas Williams (1854–1944) was at that time Assistant Chief of Division of Far Eastern Affairs in the State Department. A short mention in the Political Records of the Xuantong Reign, dated January 1910, confirms that decisions were effectively taken in order to enforce “the prohibition on the old custom of trade in human beings” ( jinge renkou maimai jiuxi 禁革人口買賣舊習); see Da Qing Xuantong zhengji shilu, 28:14a–b. 2 Shen Jiaben, Da Qing xianxing xinglü, 5:4a. 238 claude chevaleyre

Among the variety of forms of personal subjugation (including trafficking in children for immoral purposes) that the collapsing Chinese empire attempted to eradicate, the most widespread was slavery. Yet, this form of slavery was not the same as that abolished earlier by European countries. It was not a form of transnational trade in human beings based on race. It was, rather, the remnants of a practice of domestic bondage that was rooted in the ancient past and that had remained basically the same. Since the Ming dynasty (1368–1644). From its very beginnings, modern Chinese historiography, strongly influenced by Marxist-Leninist—and then Maoist—theories, has shown a great deal of interest in forms of personal bondage, and, more specifically in the late Ming, in slaves, henceforth named “bondservants” in English- language sinology. However, Ming bond-servitude has hardly been studied for itself, despite lengthy and impassioned debates led by Chinese and Japanese scholars who rank among the most prominent in the field of Chinese studies.3 And, since Marxist-influenced views strongly dominated the debate, bond-servitude has remained subordinate to objectives that transcended it. The subject has remained relatively unexplored, either because it was deemed marginal, or because more thorough analyses were likely to introduce undesirable elements into historiographic models.4 Bond-servitude was in fact less regarded as a topic worthy of study per se, than as an instrumental—though critical and well-documented— element in supporting ideologically and politically-oriented historical schemes more concerned with the nature of late-imperial society or the periodization of Chinese history.5 Recent trends in cross-cultural and interdisciplinary studies have now created the conditions for a new dialogue among historians and sinologists

3 Such debates attracted the attention of historians mainly from the 1930s until the 1980s. Pioneering studies on late Ming bondservants and bondservant rebellions are: Xie Guozhen, “Mingji nubian kao” (1925); Chen Shoushi, “Ming Qing zhiji shiliao (1)” (1927); Chen Shoushi, “Ming Qing zhiji shiliao (2)” (1929); and Fu Yiling, “Mingji nubian shiliao shebu,” (1949). 4 For example, despite the abundance of materials, narratives of “loyal servants” barely drew the attention of historians. And as far as I know, while most studies focus on the exploitative aspects of Chinese bond-servitude, no one has studied its place in corporations where technical knowledge was transmitted from patron to apprentice, the latter sometimes being considered and treated as a bondservant. Finally, I would strongly argue in favor of a deep re-examination of the so-called “bondservant rebellions,” especially of their historical significance and even of their “bondservant” nature. 5 For an overview of twentieth century scholarship, see Joseph P. McDermott, “Bond- servants in the T’ai-hu Basin,” 675–677. acting as master and bondservant 239 on universal problems such as personal bondage. However, drawing superficial parallels between past and present or between China and other areas can only be of limited interest.6 If there is any purpose or util- ity in doing history—apart from pure erudition—it is not simply to provide lessons for the future. It is also to establish connections between past and present, to identify dynamics and explain historical evolution. In the case of China’s history, as in other cases, an essential purpose of historical research is to identify continuities and developments in order to analyze and understand long-term changes.7 In the same manner, comparing different areas is not only about establishing broad differences or similarities. Comparison is useless when it is only intended to demonstrate the otherness of one of the entities compared. Historical comparison can nonetheless prove fruitful when it aims at dissecting the complexity of the various mechanisms at work in universally shared human experiences. In the field of the history of bondage, China’s experience can thus be compared with the experiences of other areas in two ways: It can either be studied with the sole purpose to make it fit with Western categories of “serfdom” or “slavery” considered as universal referents; or it can be regarded as another, singular configuration, worthy of attention and likely to enlighten our knowledge of the mechanisms of personal subjugation and bondage. In this regard, the present book should be counted as a forceful incentive to rethink our perception of these topics in the Chinese context. But before we can actually succeed in (re)integrating China’s overlooked experience in a global history of bondage, there is an urgent need for intensified interdisciplinary dialogue and exchange. If collecting sources remains the priority of sinologists, there is an even more urgent need to make Chinese materials accessible to a larger audience of scholars. Such a goal can only be achieved through the translation and publication of additional primary sources. As a preliminary step in this direction, I chose to go back to basics and propose a rough introduction to the status and identity of bondservants in late Ming China.8 I will begin with general

6 This is the case, even though it may be tempting to draw parallels between our global- ized world and late Ming China, be it in terms of acceleration in commercial exchanges, new questioning of social norms, increased competition over resources, financial crisis, etc. 7 For a more persuasive argument, see Pierre-Etienne Will, “L’histoire n’a pas de fin,” 7–8. 8 The useful but vague word “bondservant” is now conventionally used to translate the Chinese terms nubi 奴婢, nupu 奴僕 and the like. However, as a result of the need to think with Western categories, some scholars continue to call these persons “slaves” or 240 claude chevaleyre considerations regarding their legal status, with late Ming social hierarchy and the nature of social degradation. I will then turn to more specific considerations regarding the core nature of Chinese bond-servitude through an examination of the “bond” between master and bondservant as presented in late Ming judicial cases. Bondservants were not the only category of unfree persons in China. But they were the most numerous and, to a certain extent, everything that was related to unfree labor in the late Ming was also, more or less, directly or by analogy, related to bondservant status. As for the focus on late Ming China, this period remains the best starting point for studying bondservants. The seventeenth century was objectively the culmina- tion of a protracted and multifaceted crisis. The most dramatic moment occurred on April 23, 1644, when the rebel leader Li Zicheng (1605–1645) took Beijing and established the short-lived Shun interregnum, which was soon overthrown by the Manchu-Qing who were to rule China until 1911. During the decades before and after this moment, the provinces were confronted with chronic violence and were caught in the midst of political repression, widespread peasant revolts, urban movements, rent resistance, craftsmen uprisings and, long after the conquest, bloody battles between Ming loyalists and troops of the Manchu banners.9 At the same time, bondservants all over China also became a source of social unrest.10 They took up arms (sometimes provided by the local gentry in ill-considered attempts to organize local resistance) and rose up against their masters to try and retrieve their bondage contracts, killing those who resisted them and burning their houses.11 In the blink of an eye, this social

“serfs.” For a case study of bondservants in a late Ming locality, see Harriet T. Zurndorfer, “Servage domestique, conflits sociaux et la transition Ming-Qing dans la préfecture de Hui- Chou, 1644–1646.” In Alessandro Stanziani (ed), Le travail contraint en Asie et en Europe, Paris, FMSH, 2010, pp. 241–264. For other general articles on bondservants and bondservant rebellions in late Ming China see: McDermott, “Bondservants in the T’ai-hu Basin;” Mi Chu Wiens, “Masters and Bondservants;” Masatoshi Tanaka, “Popular Uprisings;” and William T. Rowe, Crimson Rain, 120–128, passim. 9 On the Manchu conquest, loyalism and peasant rebellions see Frederic Wakeman, Jr., The Great Enterprise; Lynn A. Struve, Voices from the Ming-Qing Cataclysm; Jonathan D. Spence and John E. Wills, From Ming to Ch’ing; and James B. Parsons, The Peasant Rebel- lions of the Late Ming Dynasty. 10 Bondservant insurrections broke out in almost every eastern and southern province between the 1630s and the 1650s, but the great majority occurred in the prefectures of the lower Yangzi delta between 1644 and 1646. 11 The “search for contracts” (suoqi 索契), the destruction of masters’ houses and the murder of masters’ families are recurring topoi in bondservant-rebellion narratives. Yet, if every act of violence committed by bondservants is usually and indiscriminately regarded acting as master and bondservant 241 group usually not deemed worth a note in conventional sources jumped from the shadows into the spotlight and became the subject of numerous records and lengthy narratives, which now constitute a crucial corpus of sources on bond-servitude.

Multiple Identities, Social Hierarchy and the Law

Twentieth-century Chinese, Japanese and Western scholarship has amply demonstrated how complex and multifaceted the social identities and conditions of Ming bondservants could be. From the most powerful estate managers, who terrorized the country, to the sexually abused maids described by the prefectural judge Zhang Mingbi (1584–1652);12 from those who submitted their masters to their own will13 to the thoroughly exploited ones depicted by Zhang Lüxiang (1611–1674);14 from the richest

by a majority of Marxist-oriented scholars as the expression of a new “class conscious- ness” and as part of an intensifying “class struggle,” there is still much to do to distinguish precisely those who fought for “emancipation” from more opportunist bandits. Of course, narratives were written by educated elites rarely inclined to commiserate with bondservants and depict them in a favorable light. But such sources bring to light ambiguities and should warn us against misinterpretations. The uprising that shook Liyang County in 1645, for example, is generally classified as a bondservant rebellion. The leaders, the Pan brothers, were indeed bondservants (pu 僕, or cangtou 蒼頭) in the Peng family. Yet, when reading the long week-by-week narrative depicting them as blind rebels plundering and killing rich and poor indistinctly (even allying with the Manchus in a totally opportunist move), one can hardly regard them as idealists fighting for progress and emancipation; see Ming chushi yiming daoren, Laijiang jishi benmo. 12 In a narrative of Jintan County bondservant rebellions, Zhang Mingbi has the following horrific description: “I heard about fierce and cruel masters in Jiangnan, who severely and blindly manage [their bondservants] in accordance with the bondage agreements. Many bondservants have their stomachs empty and nothing to eat, they have no clothes to protect their ankles and their knees, and the skin on their buttocks has disappeared. Before their daughter is married, [the master] takes her virginity. Before their wives have borne children, he ruins their freshness. And when the [master’s] wife is jealous, she even mutilates their vaginas, shaves them, or sews their skin; their cries of pain thus can be heard far away.” See Zhang Mingbi, “Xiaobi ban ji.” Another classic narrative of masters’ cruelty can be found in Jin Zhen’s (1622–1685) 1658 memorial on Guangshan county bondservant rebellions; see Jin Zhen, “Tiaochen Guangshan panpu xiangyi,” 19:16a–18b. 13 Like Chen Peng and Guo Kun, the skillful bondservants of Minister Zhang Fuzhi (1547–1629). According to the Brief Account of the Restoration Society, Zhang totally relied on Chen for writing his official documents and on Guo for the management of his property. They had the reputation of controlling every matter in Zhang’s household and having so much power over him that they did not hesitate to reprimand his elder brother Zhang Yizhi; see Lu Shiyi, Fushe jilüe, juan 13, 516. 14 In a poignant text, Zhang Lüxiang heaps opprobrium on masters who “don’t treat bondservants as humans,” who let them starve and suffer from cold without an ounce of 242 claude chevaleyre to the poorest; from the most educated to the totally illiterate; bondservants assumed nearly every possible position in Ming society. They were maidservants, peasants, shopkeepers, gatekeepers, bookkeepers, apprentices, nurses, estate managers, militiamen, yamen servants, tax collectors, wine shop owners and, in the rarest cases, even minor officials.15 Some lived with their masters and some had their own houses. They could be born bondservants, sold by their parents or kidnapped as children, while others voluntarily offered or sold themselves to a master. Bondservant status could be life-long and hereditary, or only pro tempore. One entered in servitude because of debt or poverty,16 to escape taxation, to avoid punishment, to take advantage of a new masters’ influence, in search for protection, or just because one was not rich enough to pay for a bride17 or a proper interment for one’s parents. And, depending on age, gender, function and even local custom, bondservants were called by a wide variety of names, among which nubi and nupu were both commonly used as well as the legal terms. As a result, and because modern scholars mostly choose to regard concrete social conditions as a key factor in determining status, this very disparity has been the cause of radically divergent, not to say opposing, interpretations regarding bondservant status. While some have focused on the ignominious treatment suffered by a majority of exploited bondservants, others have emphasized the fate of the most powerful, influential and uncontrolled ones. Similarly, when attempting to evaluate to what degree they were integrated or excluded from society, some have come to the conclusion that bondservants—who were also named jiaren 家人 (lit., “family members”)—were deeply integrated into

compassion, who abuse their wives and daughters, who appropriate their goods and who even kill and burn them; see Zhang Lüxiang, Yangyuan xiansheng quanji, juan 19, 575. 15 The Essentials on the Institutions of the Ming mentions 16 officials who were actually of bondservant (cangtou 蒼頭) origin in the 1450s; see Long Wenbin, Ming huiyao, juan 52, 969. 16 Poverty was a frequently invoked reason in bondage contracts. When in 1609 Hong Sanyuan sold himself, his wife and their son to a master, the contract on which he affixed his hand prints clearly stated that he did so because he had not enough to feed his family. In return, Hong received 15 ounces of silver; see Zhang Chuanxi, Zhongguo lidai qiyue huibian kaoshi, 929. 17 Self-sale was often motivated by the lack of resources to pay for a bride. Such contracts were therefore presented as “self-sale and marriage contracts” (maishen hunshu 賣身婚書). Zhang Shenbao’s bondage contract (dated 1574) stated that he voluntarily wanted to become the perpetual bondservant of the Wangs in order to marry Ai Gui, a maidservant of theirs. See the contract held at the Dingling Museum and reprinted in Han Dacheng, Mingdai shehui jingji chutan, 83. acting as master and bondservant 243 a family continuum, while others have strongly asserted that they were socially excluded and marginalized. Every pattern thus proposed is undeniably supported by concrete examples and evidence. Thus, rather than focusing on an impossibly broad range of observable configurations, it seems more fruitful to reconsider the question from a new angle and ask what bondservants had in common, what were the characteristics they all shared. The first characteristic shared by all bondservants was legal status. Bondservant was clearly an acknowledged and legally enforced status. However, when attempting to trace the contours of this legal category, one is soon confronted with the fact that neither the Great Ming Code, nor any other legal or normative source ever provided positive or descriptive definition of any social groups.

Bondservants, Hierarchy and the Great Ming Code

Promulgated in 1397 by the Ming founder Zhu Yuanzhang, the Great Ming Code was mainly a body of penal legislation.18 It operated like a complex legal chart or schedule: It characterized crimes and offences, prescribed the adequate sentences and specified sentence modulations according to the relative status of the offender and the victim. Status was therefore of pivotal importance in the internal rationale of a code of laws that was chiefly intended as a tool to preserve social order and ensure the durabil- ity of the dynasty—rather than to ensure individual rights.19 But identities were neither fixed nor absolute since they were deeply rooted in conceptions of relative personal hierarchical relationships defined by Confu- cian orthodoxy.20 One’s hierarchical position was thus eminently variable

18 It also included statutes that could be regarded as being of administrative or civil nature. As for debates regarding the nature of late imperial Chinese law, see in particular Jérôme Bourgon, “Uncivil Dialogue.” 19 For a good translation and introduction to Ming law, see Jiang Yonglin, The Great Ming Code. See also Edward L. Farmer, Zhu Yuanzhang and Early Ming Legislation, 107; and Farmer, “The Status of the Person in Early Ming Law,” 75–77. 20 Hierarchy derived from the principle of zhengming 正名, or “rectifying names,” according to which denominations have to correspond to realities, the form has to match with the essence. If names were “rectified,” people would accept their lot, and social order would be secured. See Confucius’ Analects XIII:3 and the famous echoing sentence in XII:11: “A ruler ought to behave like a ruler, a minister like a minister, a father like a father, and a son like a son.” It was also reflected in the notion of mingfen 名分, distinguishing (lit. “cutting up” or “dividing”) names or statuses, which was essential in the shaping of social relationships and especially in the engineering of master-bondservant relationships. 244 claude chevaleyre and depended on various factors such as personal relationship, kinship ties, age, rank of birth, gender, or membership in a special category (like scholar-officials or bondservants). The relative value of hierarchical positions was for example embodied in the system of the “five degrees of mourning” (wufu 五服)21 which was intended to determine the close- ness of kinship ties: the tighter the kinship ties, the longer the duration of the mourning, and the greater the gravity of the crime or offence was modulated. In this system, and depending on circumstances, a man was a father, a son, a brother, a husband, a master, etc. Cursing one’s father or master, for example, was considered a more serious offence than cursing a stranger; while cursing one’s own son or bondservant had no legal consequence.22 From our modern point of view, Ming law was obviously discriminatory, in particular against women, juniors and inferiors. But the Code was not the source of discrimination. It only reflected and acknowledged dis- criminations that were constitutive of the social order it was designed to preserve. Reading the Code only informs us on the existence and operation of this social order, not on its core nature. There is no need to scan the whole content of the Code to come to the conclusion that the wider the range of potential hierarchical positions one could assume, the higher one’s social status. In presence of his parents, even the highest official was a junior and inferior, but his position made him superior to ordinary subjects. Depending on the period of his life and on the person he faced, he could experience a wide range of hierarchical positions. If women were inferior in most configurations, even the humblest ordinary woman could also be a mother or the mistress of a bondservant. Her hierarchical position was far less variable than that of a high official. But her inferiority as a woman was, in theory, less absolute than that of bondservants, who were inferior even to ordinary women. The problem, in our perspective, comes from the fact the Great Ming Code was intended for an audience of officials and ordinary people familiar with the concepts, values and categories it referred to. In the same manner as it was not deemed necessary to specify who was a woman and why a woman was inferior, there was no need to explain which persons were bondservants and why they were inferior. As a consequence, one would search in vain for precise

21 See Norman A. Kutcher, Mourning in Late Imperial China; and Ann B. Waltner, Getting an Heir, 53–54. 22 Gao Ju, Da Minglü jijie fuli, 21:3b (law no. 350, following the numbering introduced by Jiang Yonglin in his translation of the Great Ming Code, op. cit.). acting as master and bondservant 245 definitions of inferior categories or the nature of their inferiority in the pages of the Great Ming Code. When dealing with social stratification in late imperial China, one usually refers to the conventional assertion according to which Chinese society was organized along “four categories” (simin 四民): that is, the literati, peasants, craftsmen and merchants (shi, nong, gong, shang 士農工商).23 Slight variants can sometimes be found, such as the assertion in the Ming huiyao presenting the Ming realm as a three-level society: “One [level] is called ‘the people’ (min), another one ‘the soldiers’ ( jun 軍), and another one ‘the craftsmen’ ( jiang 匠).”24 However, it is usually acknowledged that the major line of social discrimination was the line separating scholar-officials (shi)—those who at least had qualified for (or bought) the lowest grade in the Chinese examination system—from the rest of the population—the ordinary people (min 民, pingmin 平民, shumin 庶民 etc.).25 But if the four categories of people represented the great majority of the population and were considered to be “good people” (liangmin 良民, that is to say ordinary or common subjects), then what about the relatively small category of “mean,” or “debased people” ( jianmin 賤民), to which the Code often refers in opposition to the good people, and among which bondservants were clearly the most numerous? The category of mean people was composed of a variety of social groups who lived in very different conditions, whose fate could be either painful or pleasant and who had little in common except the fact that they all belonged to the same legal group. Yet, again, neither the Code nor any other normative source explains who was included in this category and what the source of their inferiority was.26 If we are to believe the statutes in the Great Ming code, they were essentially prostitutes and professional entertainers (changyou 倡優), and bondservants (nubi 奴婢).27 But it is acknowledged that this category also included more marginal groups,

23 This reference to “four categories” dates back to antiquity; Gu Yanwu’s (1613–1682) Record of Daily Knowledge traced its origins back to the Guanzi (seventh century B.C.E.). See Gu Yanwu, Rizhilu jishi, juan 13, 439. 24 Long Wenbin, Ming huiyao, juan 50, 936. 25 To give just an example, in his modeling of social hierarchy in imperial China, Rom- eyn Taylor does not mention inferior categories; see Taylor, “Chinese Hierarchy in Com- parative Perspective.” 26 It shall also be added that some doubts remain about the internal hierarchy among all these groups. 27 The Imperially Endorsed Collected Statutes of the Great Qing clearly stated it: “The four [categories of] people are ‘good;’ bondservants, entertainers and yamen underlings are ‘degraded.’ ” See Qinding da Qing huidian, 17:1b. 246 claude chevaleyre such as the “musician households” (yuehu 樂戶), “fishermen of the nine surnames” (jiuxing yuhu 九姓漁戶), a variety of yamen underlings (zaoli 皂隸), beggar households (gaihu 丐戶), etc.,28 to which statutes on mean people were applied. But clearly, from a legal point of view, bondservants and hired laborers were the most representative groups.29 There might be a good reason for such a situation: Contrary to the other groups, bondservants and, to a lesser degree, hired laborers were engaged in a personal relationship with their masters or patrons. Without a master (either a person specifically named “master” [zhu 主] or “family head” [jiazhang 家長]) and a master’s family to host them, bondservants could not have existed. And, as mentioned above, the Code was designed to preserve a social order in which familial hierarchy was essential. However, if the Code applied harsh discrimination to mean groups, it never explained on what this segregation rested. But the segregation was, in my opinion, far more fundamental and, in theory, more impermeable than the one distinguishing scholar-officials from the common people, and the principal expression of such discrimination was the exclusion of mean groups from the examination system.

Exclusion from the Examination System and Personal Degradation

In his study of the fundamental characteristics of slavery, the French scholar Alain Testart has argued that in every complex society, a slave was always excluded from at least one of the (social) dimensions deemed

28 For more details regarding marginal categories, see Anders Hansson, Chinese Out- casts. Hansson recalls that “there is no extant law or regulation from the Ming or early Qing dynasty stating that the musicians’ households and the other groups were legally mean.” Yet, the one thing they had in common is “the manner in which their legal emancipation was initiated” in the Yongzheng (1723–1735) reign; Ibid., 2. 29 Another category frequently appears in the Code alongside bondservants: hired laborers (gugongren 僱工人). We should also consider law no. 394 entitled “Slaves or hired laborers committing fornication with household head’s wives,” or the numerous statutes associating these two categories. Yet, it is still very difficult to ascertain that hired laborers were also of mean status. They actually suffered a legal discrimination, but there was no provision for crimes committed by hired laborers against commoners outside the family that employed them. Moreover, when they committed a crime against the family head, the sentence was in most cases lighter than the sentence prescribed for bondservants who committed the same crime. And, finally, a new sub-statute was promulgated in 1588 that specified that laborers hired on a contractual and short-term basis would be sentenced like ordinary persons. See Wu Yanhong, “Dui Mingdai beiyou falü diwei de kaocha,” 382–390. acting as master and bondservant 247 essential by the society in which he lived.30 In China, one of the essential prerogatives of which bondservants were deprived was that of attending school and participating in the examination system. In so doing, they were denied a fundamental common (though very theoretical) prerogative shared by all ordinary (male) subjects: the potential access to academic ranks, bureaucratic responsibilities, social prestige and political power. Of course, lining up a series of legal impediments does not suffice to reveal the nature of a social status, but tracing the origins of such incapacities can help understand them. In the Ming, bondservants were not explicitly banned from education, examination or an official career. The Code simply made no such provisions. However, a number of sources clearly suggest that the mean categories (including bondservants) were effectively banned from the examination system.31 A 1384 regulation promulgated by the Ming founder stipulated: “As for the provincial examination, educational officials (xue- guan 學官) and dismissed officials, prostitutes, entertainers and their relatives, yamen runners and their companions, and those who mourn their father or mother, are in no case allowed to take the examinations.”32 In 1449, the Zhengtong emperor (r. 1435–1449) stipulated that only outstanding students could attend district and prefectural schools, and that it was thenceforth “prohibited to improperly accept prostitutes and entertainers, yamen runners and persons of degraded status.”33 Be that as it may, the laments of early and mid-seventeenth century literati over the schools’ decay also suggest that bondservants were included in these dispositions. Wu Yuancui, for example, noticed that: Sons of prostitutes, entertainers and yamen runners are not allowed to enter schools. Yet, recently, the institutions have been undermined. There are [too] many people in schools, among whom are numerous sons of slaves. They are even more numerous in the Wu (Jiangsu) prefectures of Su[zhou], Song[jiang] and Chang[zhou], and in the Zhe[ jiang] prefectures of Hang[zhou], Hu[zhou] and Jia[xing]. In the worst cases they even attain

30 Alain Testart, L’esclave, la dette et le pouvoir, 24. This assertion does not imply that I consider Chinese bondservants as slaves in the strictest meaning. But this approach seems relevant when extended to the search for common characteristics shared by inferiors in general. 31 Considerations on the relation between exclusion from the examination system and personal degradation owe a lot to the pioneering research of Kishimoto Mio. See in particular Kishimoto, “Mindai no shakai shūdan to ‘sen’ no kannen;” and idem, “Maojuan maokao susong yu Qingdai difang shehui.” 32 Da Ming huidian, 77:2b; this regulation is repeated in a slightly different version in Ming Taizu shilu, juan 160, 2468. 33 Ming Yingzong shilu, juan 183, 3560. 248 claude chevaleyre

top places at the examinations, enter the Hanlin academy and hunt for the highest positions. Isn’t that an offence against the Court? Aren’t they a shame for the scholars of our time?34 Wu makes an informative parallel between entertainers and yamen runners on the one hand and “slaves” (nuli 奴隸, a common word for bondservants) on the other hand. None of those categories had their place in schools and they were a shame for contemporary literati, even though bondservants were not mentioned in early Ming legislation. Guan Zhi- dao (1536–1608) shared the same views when he wrote that bondservants were identical to entertainers and yamen runners, and that they could not therefore “mingle with the class of officials”: If we scrutinize laws and regulations, even though there are statutes on [the way] bondservants have to pay homage to the family head35 and on the prohibition against offending the family head, yet, only Dukes, Marquises and officials of the third rank and above are allowed to keep bondservants. [People] sentenced to servitude and whose property has been confiscated are only offered as slaves to families of outstanding officials, but never to [regular] officials. [Ordinary] officials and the common people name them “adoptive sons,” not “bondservants,” because only meritorious and noble [men] can govern ordinary people; ordinary people do not govern their own kind. Those named “adoptive sons” are only one degree inferior to natural sons; they have never been prohibited to study and become officials. Those who are named “bondservants” are identical to prostitutes, entertainers and yamen runners; they cannot be allowed to mingle with the class of officials.36 Then, if bondservants (and the mean categories in general) were mani- festly excluded from the examination system, it may be interesting to ask why and to inquire about the nature of this exclusion. Ming sources are relatively silent, but the evidence leads us to the conclusion that: (1) the mean status was deeply rooted in conceptions of personal degradation;37 (2) that this degradation resulted from a sort of stain or pollution; and (3) that such pollution was either provoked by polluting activities (like that attached to “coroners” [wuzuo 仵作], who were in direct contact with dead bodies), or derived from criminal activities and physical injuries.

34 Wu Yuancui, Linju manlu, 2:5a–b. 35 This may refer to a 1393 edict on “Rituals of salutation among ordinary people” that included regulations regarding the manner bondservants had to pay homage to ordinary people; see Mingshi, juan 56, 1428. 36 Guan Zhidao, Congxian weisu yi, 2:61a. 37 Jian 賤 originally meant “depreciated” in the sense of “low-priced.” acting as master and bondservant 249

In 1522, the newly appointed Shanxi vice Educational Censor, Lu Shen (1477–1544), expelled from school the son of the favorite court musician (yuegong 樂工) of the Prince of Jin. He argued: “It is preferable to have one less student at school, rather than having [a single] one who pollutes the school.”38 Mean categories (here, in Shanxi, including the households of musicians) were regarded as “polluted” (wu 污) and likely to corrupt pure students. And when such impurity was not due to polluting activities, it was intimately related to the degradation generated by criminal activity, legal punishment and physical injury, as suggested in a 1454 proclamation: “prostitutes, entertainers, yamen runners and persons who have suffered punishments, who mourn [a parent], and who have committed faults or crimes, are not allowed to take the examinations.”39 Alongside entertainers and runners, criminals and those who had suffered legal punishment were similarly excluded from the examination system.40 Physical injuries and condemnation resulting from a crime were thus of the same nature as the pollution suffered by mean people. A very interesting source for the study of mean people’s exclusion from the examination system is the 1814 edition of the Qing Imperially Endorsed Complete Guide for Education Commissioners.41 A moral guide, practical manual and compendium of legal jurisprudence on the functioning of education and schools, this text had no equivalent in size and precision in the Ming.42 The importance given to the compilation of such documents covering every aspect of public education can be regarded as a demonstration of the Qing rulers’ strong will to control and reorder (I would say re-interpret) Chinese society, notably by rectifying and clarifying the contours of social status. This can be illustrated by the promulgation of the so-called “emancipation edicts” of the Yongzheng era,43 or the sharp

38 He Liangjun, Siyou zhai congshuo, juan 17, 148. 39 Ming Yingzong shilu, juan 237, 5173. 40 The exclusion of those who had to mourn a parent was different by nature. In order to fulfill the requirements of mourning, officials and students had to retire either from office or from school during a fixed period. 41 For an overview of statuses and examinations as presented in Xuezheng quanshu, see Kishimoto, “Maojuan maokao susong,” 148, passim. 42 Local guides for Educational Commissioners existed in the Ming, such as those collected and reprinted in Yang Yifan and Liu Ducai, Zhongguo gudai difang falü wenxian. However, these are short practical manuals that can compare with the Qing compendium neither in size nor by nature. 43 The problems raised by the Yongzheng “emancipation edicts” are thorny and endless. These edicts, promulgated between 1723 and 1730, were essentially meant to emancipate regional marginal groups such as Shanxi musician households, the “fallen people” of Zhejiang, hereditary servants of Anhui, beggars of Suzhou prefecture, groups of boat 250 claude chevaleyre inflation of sub-statutes on bondservant status in the Great Qing Code. The Qing dynasty’s systematic attempts to deal with status strongly contrast with the relative laissez-faire practices that characterized the more porous and mobile late Ming society.44 Still, I assume that the Complete Guide for Education Commissioners of the Qing promoted principles and ideas that were relevant—though less clearly formulated—in the Ming. Chapter 43 is entitled “distinguishing statuses” (qubie liupin 區別流品). It mainly consists of a series of judicial cases devoted to deciding whether or not a candidate or his ancestors were of mean origins and, as a consequence, if the candidate could be authorized to take the examination. Thus, some Qing documents explain more clearly than any Ming source why the sons and grandsons of emancipated bondservants could not attend school and take the examinations: Would they have been able to do so, they could have become officials and posthumous honors could have been bestowed to their parents or grand-parents (that is to say, to bondservants).45 One pitiful case attracted my attention, as it implicitly confirms that mean people were considered polluted in the same manner as criminals and the physically injured. Sometime after 1800, the Shandong provincial Governor was sent the case of a young boy named Zhang Shanchang for deliberation. In 1797, Zhang had been sexually abused (jijian 雞姦)46 by a man named Zhao Jie. Zhao was sentenced to death by strangulation and executed after the Autumn Assizes of 1800. But the case of Zhang himself still had to be deliberated because he was about to take the examination: [. . .] For students, taking the examinations is the beginning of official career. This is the reason why the established regulation emphasizes before any- people, etc. In my opinion, these edicts were not a first step toward a general emancipation of mean categories, but rather an attempt to clarify social organization by suppressing abusive mean statuses. For an in-depth exploration of the process and the motivations of the Yongzheng emperor, see Hansson, Chinese Outcasts, 163–170; see also Ho Ping-ti, The Ladder of Success, 18–19, who regards it as a complete manumission whose effect was to “open the main avenue of upward social mobility to them.” 44 Porosity and the blurring of social boundaries are hallmarks of late Ming society. See, among others, Wiens, “Masters and Bondservants,” 57; and Ho, The Ladder of Success, 53–125. This feeling of social mobility and a general impression that status was more a question of practice vividly appear in seventeenth century judicial casebooks. My own reading of judicial cases led me to the conclusion that disputes and conflicts over social status were often adjudicated on a case-by-case basis and that concrete configurations prevailed over appellations. 45 See examples in Xuezheng quanshu, 43:4a–5b, 43:20a–22b, 43:24a–b, 43:25a–26b. 46 See Matthew H. Sommer, “The Penetrated Male,” for a study of homosexuality in Qing China. acting as master and bondservant 251

thing the purity of [the student’s] person and family. If [the candidate] has been [physically] wounded during a [legal] punishment or if he has committed a crime or an offence, he may not be permitted to take the examination. Zhang Shanchang, the son of Zhang Mu, a Shandong commoner from Dezhou, has been physically polluted. [Such a demeaning] is in no way different from a wound caused by a punishment. Thus, it would be improper to permit him to take the examinations.47 The case of Zhang Shanchang can certainly not be compared with that of bondservants or other mean people. He did not belong to a mean category but was the son of a Shandong commoner (min 民). Yet the argument of the Governor establishes a striking parallel between criminals and degraded individuals. It clearly states that persons who had committed a crime or an offence were “polluted” (beiwu 被污), in the same manner as those who had suffered the wounds of physical punishments or injuries. As a consequence, all were denied the right to take the examinations, just like mean categories, who were also denied that right for not being “pure” (qingbai 清白, the prerequisite for becoming official). It is tempting (in an assumed syllogistic approach) to regard every group of mean people as suffering a sort of personal degradation comparable in nature to that suffered by criminals and physically or sexually injured persons. Indeed, the fate of many groups of mean people was supposed to be the result of an original condemnation. In Zhejiang province, “fishermen from the nine surnames” were supposed to be offspring of Zhu Yuanzhang’s enemy Chen Youliang (1320–1363),48 and musician households in Shanxi (like the favorite of the Prince of Jin mentioned above) were also deemed to be the offspring of Jianwen loyalists.49 There is one last argument likely to support the idea that bondservants’ personal degradation was of a similar nature and intimately linked with the debasement of criminals. When considering early Ming legislation, it appears that bondservants must have been relatively rare and mainly former criminals. Neither coercion nor marriage could have produced new bondservants. The Great Ming Code clearly protected commoners from being constrained to servitude50 and also prohibited unions between

47 Xuezheng quanshu, 43:23b. 48 Hansson, Chinese Outcasts, 133. 49 The Jianwen emperor (r. 1398–1402) was the grandson and designated heir of the Ming founder. He was deposed in 1402 by his uncle who reigned as emperor Yongle until 1424 (see Hansson, Chinese Outcasts, 60–61). 50 Laws no. 85 and no. 108; see Gao Ju, Da Minglü jijie fuli, 4:11b–12a, 6:6b. 252 claude chevaleyre commoners and mean people.51 Even voluntary subordination must have remained marginal. Only a minority of officials and dignitaries were permitted to possess bondservants,52 and such possession by ordinary people was formally forbidden: Families of commoners who shelter and foster bondservants [will be sentenced] to one hundred strokes, and [the bondservants] will be released and recover [the status of ] good [people].53 As the Code also stipulated that relatives of certain categories of criminal were to be sentenced to enslavement to serve officials (though this was not one of the regular “Five punishments” [wuxing 五刑]),54 it is possible to say that the intention of Ming legislators was to limit the possession of bondservants to officials and dignitaries, and to restrict the origin of bondservants to legally sentenced criminal relatives. This is also confirmed by early Ming decrees promulgated to free people who had been sold (or who had sold themselves) because of poverty or as a consequence of wartime disorder: During the fifth year [of Hongwu, i.e., 1372], it was decreed that ‘[now] that all under heaven has been mostly pacified, rites and customs cannot be [left] un-rectified. All those who have suffered from the disorders and have been enslaved to other people will recover [the status of] commoners.55 Similar decisions that clearly enforced such principles can be found in an “Imperial edict on rectifying rites and customs,” also dated 1372,56 or in The Veritable Records of Ming Taizu.57 In extreme situations, local

51 Law no. 121; see Gao Ju, Da Minglü jijie fuli, 6:33b–34a. A woman’s status was determined by the status of her husband. This is the reason why the law especially focuses on women of commoner origin that married or had intercourse with bondservants. 52 A regulation dated 1391 stipulates that Dukes and Marquises could not foster more than 20 bondservants; officials of the first rank, no more than 12; officials of the second rank, no more than 10; and officials of the third rank, no more than eight. It also states that captured children and families of criminals were mostly offered as bondservants to families of meritorious men. See Long Wenbin, Ming huiyao, juan 52, 969–970. 53 Law no. 84 on the nomination of an illegal heir; see Gao Ju, Da Ming lü jijie fuli, 4:9a- b. In his commentaries, Gao Ju adds that “only the families of meritorious men could raise bondservants;” ibid., 4:11a. However, these regulations were not precise enough to prevent equivocal interpretations; for more details on that problem, see Wu Yanhong, “Dui Ming- dai beiyou falü diwei,” 390, and Wu Zhenhan, “Mingdai de zhupu guanxi,” 147. 54 An example can be found in Fu Fengxiang, Huang Ming zhaoling, 2:7b. It stipulates that those who employed castrated children were to be castrated themselves and given to officials as bondservants. 55 Long Wenbin, Ming huiyao, 950. 56 Fu Fengxiang, Huang Ming zhaoling, 2:4b–5a. 57 Ming Taizu shilu, juan 73, 1352. acting as master and bondservant 253 administrations could even be ordered to redeem children who had been sold.58 However, as is well known, the intentions of the Ming founder soon became obsolete and were unscrupulously violated or by-passed.59 But what remained, despite the loosening of legal or social practice, was a commonly assumed appreciation that mean and bondservant status rested on the idea of personal degradation. The elements described above are not sufficient to explain the difference between bondservants and other groups of mean people: personal degradation, impurity, marginality, exclusion from the examination system and a series of legal incapacities were features that all mean groups (at different degrees) had in common. Yet, what distinguished bondservants from other mean categories was the unique relationship they shared with their masters. In exploring the core nature of their status, we first ought to explore this particular link, rather than focus on the confusing variety of conditions that can be observed when reading the sources.

A Concrete Relationship Based on Reciprocity

A bondservant, whatever his origin, the terms of his contract, the duration of his servitude, his occupation, or the harshness of his condition, was considered a family member subjected, like any other, to the authority of the family head. As such, he or she was considered an inferior or junior (beiyou 卑幼) who was to obey the orders of the family head and other senior or superior family members (zunzhang 尊長). His or her relationship with the master mimicked kinship ties, and when offences or crimes were committed, the statutes in the Code usually cited him or her among the other junior members of the family. In so doing, the law submitted him or her to the highest possible familial authority and to the control of the person one had to respect the most—one’s father. But, of course, bondservants were not their master’s children, they could not have a share in the property, and they would never themselves become senior or superior members of their master’s family. Contrary to the master’s

58 See the decree on Henan situation, in Ming Taizu shilu, juan 177, 2688; or the decree on Xuzhou prefecture (dated 1413) in Ming Taizong shilu, juan 140, 1687–1688. 59 To circumvent the law, commoners sometimes named their bondservants “adoptive sons” and established adoption or marriage contracts rather than bondage contracts (see the quotation from Guan Zhidao, Congxian weisu yi, cited above). These practices were regularly denounced for diverse reasons by late Ming thinkers such as Hai Rui, Shi Tianqi and Gu Yanwu. For more details, see Waltner, Getting an Heir, 86–87. 254 claude chevaleyre son or grandson who, one day, would become master himself, a bondservant would always remain a bondservant. Yet, although the personal relationship between master and bondservant was obviously dissymmet- ric, and though it only mimicked kinship ties for moral and legal purposes, it was nevertheless not simple and absolute subjugation. As will be demonstrated in this second part, this relationship was deeply rooted in Confucian-inspired principles of solidarity and reciprocity, which were not always paternalistic and vacuous justifications, and were sometimes taken very seriously by Chinese magistrates.

“Righteousness” and “Benevolence”: The Basis of Master-Bondservant Relationship

In Ming China, the tie binding master and bondservant was supposed to be based—at least this was the theory—on mutual aid and reciprocity. In an empire where Confucian values were proclaimed as official orthodoxy and played the role of a universal matrix in the shaping of all human relationships (both in family and in society), the pursuit of self-interest was regarded with disdain.60 In this context, the subjugation of others for the sole sake of personal profit could not be considered as morally valid. In other words, exploitation was not openly assumed and was disguised behind a more or less sincerely benevolent (or paternalistic) rhetoric. In a passage of the Jianwen lu, Zhang Lüxiang (1611–1674) inveighed against masters who abused their simple-minded bondservants at the cost of personal solidarity: “They are naïve (shiduan 識短),” he said, “and this often leads masters to disdain solidarity (qu yi 去義) for the [sole] sake of personal interest (jiu li 就利).”61 In structuring the master-bondservant relationship, the principle of yi 義 (often translated as “righteousness”) was undoubtedly essential. Often opposed to self-interest (li 利) and as the corollary of the virtue of “humanity” (or benevolence, ren 仁), yi was a primordial value in the Confucian social order. It mostly referred to appropriateness in human relationships, but also to loyalty and solidarity.62 For example, collective

60 Analects IV:16: “The mind of the superior man is conversant with righteousness; the mind of the mean man is conversant with gain.” Translated in Legge, Confucian Analects, 170. 61 Zhang Lüxiang, Yangyuan xiansheng quanji, juan 31, 894. 62 These are broad simplifications derived from the scholarly and philosophical debates. I only intend here to present how such principles were practiced, simplified and applied in daily life, and how they shaped master-bondservant relationships. For more substantial acting as master and bondservant 255 estates kept by important clans to support clan schools, clan ceremonies or for charitable purposes were named yichan 義產, a term conventionally translated as “charitable estates,” though they were dedicated to solidarity rather than to simple charity. Ming loyalists who refused to submit to Manchu rule were in the same manner named yijun 義軍, literally “yi armies,” that is to say troops of loyal subjects whose solidarity towards the Ming court never failed. And bondservants who remained loyal to their masters in extreme circumstances—sometimes to the point of self-sacri- fice for the sake of one’s master—were named yipu 義僕 or yinu 義奴.63 Persons who had the sense of yi thus acted appropriately; their conduct was in conformity with the role society had assigned to them; they demonstrated a strong attachment to hierarchy and social order, while those who did not were condemned for being “rebellious” (pan 叛) and a source of social disorders.64 Indeed, yi was not only an abstract notion. In practice, it was the cement of contractual or non-blood familial relationships. When an outsider entered a new family as a wife, as an adoptive son or as a bondservant, he or she engaged with the new family head (and his blood-relatives) in a relationship based on reciprocity and on the sense of appropriateness implied by the bond of yi.65 He or she thus accepted a shift of authority and undertook to play a new role, and to act as a wife, as the son of new parents, or as a bondservant.66 Such a shift of identity was not only achieved through the performance of ritual ceremonies, but also through the establishment of a contract.67 A contract was the symbol of the

references on the link between “humanity” and “righteousness,” see Mengzi 6:A:11 or 3:B:9; translated in Legge, The Works of Mencius, 290, 155–160; and see the discussion between Yang Shi (1053–1135) and Cheng Yi on Zhang Zai’s (1020–1078) perception of the relationship between ren and yi, in Yang Shi, Er Cheng cuiyan, 33. 63 We can cite, for example, the courageous Wang Chang, who braved all dangers to get food for his master’s family during the early 1640’s famine; see Zeng Yuwang, Yiyou biji, 7–8; or the case of Xu Da, the bondservant of Ma Shiqi, who, out of gratitude and loyalty, sacrificed himself on the grave of his deceased master after the fall of Beijing in 1644; see Wuxi Jinkui xianzhi, 25:12. 64 Pan 叛 is often used in the rhetoric devoted to condemn late Ming bondservants. It is frequently employed in official sources such as local gazetteers. 65 For example, an adopted son was named yinan 義男, or a “son [bound by] yi;” regarding wives, one of the reasons legally accepted to pronounce a divorce was the “rupture of the yi [bond]” ( yijue 義絕); see law no. 123 in Gao Ju, Da Minglü jijie fuli, 6:37b. 66 This did not imply a total rupture with one’s blood-relatives. See Waltner, Getting an Heir, 53–71, passim. 67 I use the word “contract” here despite the fact that there was no contract law in Ming China. Contracts were thus more comparable to private agreements that could be presented at court, but were not strictly legally binding. 256 claude chevaleyre acceptance of a new identity, rather than a legally binding list of duties (even though material conditions were often specified in such documents). When becoming a bondservant, one accepted not only to assume the duties and respect the conditions listed in the contract to which one had appended one’s hand-prints, but also to act as bondservant, that is to say, to assume the role of a bondservant in such a way as to make the name “bondservant” fit with the expected, corresponding reality. The requirement to make names and titles correspond to realities was indeed central to the Confucian conception of social order. The word that most fits our conception of status in traditional Chinese was the word mingfen 名分 (lit. “name-distinction”), itself a by-product of the idea of “rectification of names” (zhengming 正名), central in Confucian orthodoxy.68 Regarding bondservants, this idea was expressed in the phrase “name distinction between master and bondservant” (zhupu mingfen 主僕名分), which was used to recall that their actions had to correspond with what was expected from someone named bondservant.69 Name distinction and righteousness (or solidarity) were thus complementary and inseparable notions. Such a connection found its ultimate expression in the legal statute entitled “violating names (or roles) and offending against righteousness” (ganming fanyi 干名犯義).70 This statute was not exclusively concerned with bondservants but with role transgression in general. It notably specified the adequate sentence for whoever accused his father or his master, a most serious offence against family hierarchy and solidarity.

68 If the idea of “rectifying names” appears in the Analects, it is more developed and more crucial in the Xunzi (book 22 is specifically entitled Zhengming). Differentiation plays an essential role in Xunzi’s vision of social order and harmony. Distinguishing categories and names or appellations is crucial in that it permits the expression of the complex realities of the world and the maintenance of social order. In book 9, Xunzi extended the meaning of the famous sentence “the ruler ought to behave like a ruler, the minister like a minister, etc.” and added “the peasant ought to act like a peasant, a gentleman like a gentleman,” etc. See the translation in Homer H. Dubs., The Works of Hsüntze, 136. He then demonstrated that the rationale underlying the correspondence between appellations and realities was essential and could be extended to all human relationships. For a deep investigation on “names rectification,” see Michael Lackner, “La portée des événe- ments,” 79–80. 69 The sixth section of Xiao Yong’s Agreement of the Chishan Association, for example, is dedicated to the requirement of “Rectifying Status” (zhengfen 正分); see Xiao Yong, Chis- han huiyue, 7. Guan Zhidao’s Proposals for Safeguarding Customs also includes a text that suggests to “correct master-bondservant [relationship] according to the laws of the Spring and Autumns period” ( yi chunqiu zhi fa zheng zhupu 以春秋之法正主僕); see Guan Zhidao, Congxian weisu yi, 2:60a–63a. 70 Law no. 360; see Gao Ju, Da Minglü jijie fuli, 22:30a–31b. acting as master and bondservant 257

“Name distinction between master and bondservant” implied that bondservants had to act as bondservants. They were expected to work hard, to be obedient in all circumstances and to show respect at all times. Such expectations were frequently repeated in family or clan rules and in contracts.71 Yet, as the master-bondservant tie rested on the principle of reciprocity, masters were also expected to act as masters. First, it has to be recalled that the Great Ming Code offered protection to bondservants against undue physical aggression and murder. Even though crimes or offenses perpetrated against bondservants were less severely punished (as a demonstration of social or familial hierarchy), the principle of protecting bondservants remained in effect, despite a flagrant dissymmetry unfavorable to bondservants. Then, masters were supposed to care for their bondservants. They had to “foster them with benevolence” (enyang 恩養), to protect them from cold and hunger, to provide them with food and clothes and even with a wife or husband when coming of age. In the section entitled “Commiser- ate with Inferiors” (xuxia 恤下) of the rules of the Chishan Association, Xiao Yong ( jinshi 1583) stated: Make sure to provide them with clothes and food, and to marry them in due time. The Law stipulates that they have to be sentenced on the same basis as sons and grandsons, out of respect for the importance of benevolence and righteousness. Considering that they regard their masters as their own father and mother, how could masters not regard bondservants as their own sons?72 Or, as mentioned in a typical text of “family instructions” (jiaxun 家訓): “[you] must provide them with beverages and food in a timely manner, make sure that they are not starving or suffering from cold, and balance their time of labor and leisure.”73 In the Ming, fostering bondservants was even regarded as a natural act of benevolence. According to Wang Mengji’s Family Admonitions: Families with ample resources must keep bondservants. They rely on our provisions and we rely on their labor; when relying on each other we form a household. [Yet], as they can work, can they not work for anyone else? Thus, it would be wrong to pretend that they cannot make a living without us.

71 See, for example, the general clauses in a late Ming “man-sale” model agreement reprinted in Zhang Chuanxi, Zhongguo lidai qiyue huibian kaoshi, 1011; or the “Itemized rules for governing one’s family” (zhijia tiaoyue) in Zhuang Yuanchen, Zhuang Zhongfu zazhu (rare manuscript held at the National Library of Beijing). 72 Xiao Yong, Chishan huiyue, 14. Treating bondservants with commiseration was the main topic of section 21 and also of section 7, entitled “Spreading Humanity” (guangren 廣仁) in the principles promoted by the members of the Chishan Society. 73 Pang Shangpeng, Pang shi jiaxun, 9. 258 claude chevaleyre

There are statutes in the Code that sentence [criminals] to servitude with officials. But who allowed scholars and commoners to keep bondservants? Therefore, male bondservants are called “adoptive sons,” female ones “adoptive daughters-in-law,” and young girls “adoptive daughters.” They are all named after the same way as our own daughters-in-law, sons and daughters. Though there is a distinction between high and low, it is not of the same kind as [the distinction between] human and animals.74 For late Ming scholars, wealthy families had thus “to foster” (xu 蓄) bondservants not out of pure charity, but for the mutual benefit of those who had wealth and those who had their labor. In this typical, ideal and paternalistic vision of social harmony, which probably did not match with the great majority of their contemporaries’ practices, masters and bondservants were regarded as complementary and inseparable; they “relied on each other” (xiangyi 相依) to form a household. The moral demand for reciprocity could hardly be expressed in a clearer way. When stating that “it would be wrong to pretend that they wouldn’t survive without us,” the author of this document even recognizes that masters benefited more of the situation than bondservants. His text also establishes an interesting distinction between criminals sentenced to servitude and those (probably the great majority at the end of the Ming) who were owned by ordinary people or in the private service of scholars.75 He thus confirms the fact that early Ming legislation regarding the possession of bondservants was obsolete. Though of inferior status, private bondservants (who were sometimes named adoptive sons or daughters in order to by-pass the law) were still of the “same kind” (tonglei 同類) as their masters, according to Wang. As such, and because they bore names that mimicked family relationships, they were to be considered as family members, he said, echoing Xiao Yong and many other contemporaries. Ascertaining that bondservants had to be treated and regarded as one’s own children or family members was obviously the expression of a strong paternalistic rhetoric. Such views were intended to provide justification for subordinating others to one’s own authority. In regard to Confucian orthodoxy, which emphasized the virtue of humanity, and to the law, which should have prevented personal servitude, such commonplace

74 Wang Mengji, Wang Mengji jiaxun, 1b. 75 The author distinguishes “officials” (guan 官) and “scholars” (shi 士). This probably corresponds to the distinction between, on the one hand, officials who possessed and used bondservants in their public career, and, on the other hand, scholars inside or outside officialdom who possessed private bondservants. acting as master and bondservant 259 rhetoric was motivated by the need to present acceptable arguments. Such was probably the main goal of narratives that frequently recounted one’s human attitude toward bondservants. Let us cite, for example, another passage written by Zhang Lüxiang’s: When Zeng Wenke was travelling north to take the metropolitan examination, a household servant in charge of guiding his donkey followed him. One night they encountered heavy snow and extreme cold; [Zeng] found it natural to propose that the servant share his room. But his companions refused it. [Zeng] thus replied: “He is also the son of a man. How could we suffer to leave him alone in the cold?” Such were his outstanding virtues.76 Of course, if Zeng’s attitude toward his servant was the expression of an “outstanding virtue” (houde 厚德), this also infers that such concern was unusual or exceptional. However, narratives of commiseration for bondservants’ conditions and allegories of master-bondservant harmony appear frequently in Ming writings. But was it simply pragmatic concern for inferiors who were expected to remain safe enough to provide efficient service? Was it purely paternalistic rhetoric intended only to provide acceptable justifications for personal domination? In most cases, it probably was. Yet, in an innovative and persuasive article, the Taiwanese scholar Wu Zhenhan has provided enough evidences to suggest that, in the Ming, concern for bondservants was to some limited extent sincere. Wu has notably demonstrated that such concern was not simply motivated by paternalism or compassion, but by the emergent conviction that bondservants were of the same nature as their masters and that they were also human. Such ideas mainly developed under the popular influence of Wang Yangming’s (1472–1529) individualistic and equalitarian idea of innate knowledge, according to which all human beings shared the same potentiality to access wisdom.77 This also reflects in other social trends that are the hallmark of late Ming society, such as the development of charitable activities78 or the spread of women’s literacy and education. Nevertheless, such ideas probably attracted only a small (and not quantifiable) minority of masters and had a very limited influence on ordinary practices regarding servitude. But

76 Zhang Lüxiang, Yangyuan xiansheng quanji, juan 44, 1287. 77 Ivanhoe, Ethics in the Confucian Tradition, 121–134, passim. It should nonetheless be added that such conceptions and attitudes were prepared long before Wang Yangming, notably by Song-period thinkers like Zhang Zai (1020–1077), for example; see Ivanhoe, op. cit., 22. 78 See Joanna Handlin Smith, The Art of Doing Good. 260 claude chevaleyre they instilled some substance into the theoretical principle of reciprocity that was supposed to govern master-bondservant relationships. Moreover, they introduced a certain nuance into the usual prism through which bond-servitude has normally been analyzed. In other words, the principle of reciprocity was not completely meaningless in theory, nor was it totally insignificant in practice.

Reciprocity in Practice in Late Ming Judicial Casebooks

New trends in legal history have amply demonstrated how fruitful the study of judicial archives can be in getting closer to economic and social realities. In this regard, reading late Ming judicial cases turns out to be a particularly fruitful approach to explore the nature of the master-bondservant bond and examine the validity of the principle of reciprocity. I will introduce a few examples here, beginning with a case from the Brief Account of Judgments in Songjiang, an anthology of court opinions by Mao Yilu (?–1629), the prefectural judge of Songjiang between 1606 and 1610.79 As a prefectural judge, Mao Yilu was in charge of handling ordinary judicial procedures of revision and appeal transmitted by superior provincial authorities. On the frontline of judicial operations, he was also frequently consulted in cases issued from other jurisdictions because of his reputation as a judicial expert. But he was not simply a legal technician who mechanically applied the codified penal law. When confronted with what we would name civil matters, he had some latitude in pronouncing judgments in a manner closer to civil arbitration than to legal punishment. In so doing, he had to find in the end the best balance between sentiments and law80 in order to get the compliance of the parties, to ease minds, prevent further prosecutions and, ultimately, restore social harmony. Such was sometimes the case in affairs involving bondservants, close to one- fourth of the cases collected in the Brief Account.81

79 Mao Yilu, Yunjian yanlüe. See Pierre-Etienne Will, Official Handbooks and Antholo- gies of Imperial China (forthcoming), “Anthologies of Official Documents by Prefects and Magistrates,” 6.3; see also Jiang Yonglin and Wu Yanhong, “Satisfying both Sentiment and Law.” 80 For a complete presentation of the legal proceedings handled by Mao Yilu and the requirement to balance sentiments and law, see Jiang Yonglin and Wu Yanhong, “Satisfy- ing both Sentiment and Law.” 81 Wu Yanhong counted 42 cases involving bondservants out of a total of 173; see Wu Yanhong, “Dui Mingdai beiyou falü diwei de kaocha,” 392. acting as master and bondservant 261

The Brief Account is not composed of full verdicts enunciating the precise conditions of the affairs and the final sentence. The pieces collected concentrate instead on the motivations of the sentences. Thus, they are invaluable for legal historians trying to penetrate the legal reasoning of Ming magistrates. On the other hand, their sketchy nature makes such documents somewhat difficult to read and forces the reader to speculate about circumstances and incriminations. Yet, it includes some very informative pieces regarding bondservants, such as case no. 10,82 a relatively complex affair opposing the descendants of a master to those of his bondservant.

The Quarrel among the Sun from Qingpu

Some sixty years before the case was brought to Mao Yilu’s attention— that is, in the late 1540s—a young boy named Sun Cheng was sold (yushen 鬻身) to Sun De’s ancestors.83 As time went by, Sun Cheng eventually established his own household (zili menhu 自立門户) and had two sons. In a classic configuration, he first adopted a son he named Sun Qiao before having a natural, though probably unexpected son whom he named Sun Zong. The prosecution started up years after Sun Cheng’s death, when Sun Qiao fell critically ill. As the two brothers had already determined separately, Zong feared that he might not get control over Qiao and so divert his property. Thus, in order to achieve his goal, Zong conspired with the descendant of their father’s master, who made an accusation against Qiao. It is not clear, however, how such a strategy was supposed to enable Zong to get hold of his brother’s property. What is nonetheless clear is that the crux of the case was to decide whether Sun De, the master’s descendant, was himself the master of Sun Qiao, the adoptive son of a bondservant. At the first hearing, the Qingpu magistrate concluded that Sun De was not Sun Qiao’s master. However, considering that the old contract of his father was still extant, he still sentenced Qiao to pay two ounces to Sun De, so as to “demonstrate somehow the significance of the rites.” Yet he also ordered Zong to pay half of the fine, arguing that as Zong was the

82 Mao Yilu, Yunjian yanlüe, juan 1, 415. 83 It should be underlined here that the masters and their bondservants bore the same surname. However, no element in the case can help understand whether Sun Cheng was simply a distant relative of his master’s family, if he had had to change his surname, or if he had been officially “adopted” in order to conceal the fact that he had been bought as a bondservant. 262 claude chevaleyre natural son of the former bondservant, and, despite the fact that he was on the side of the accuser, “the law could absolutely not spare [him] and solely bear on Qiao.” As one might guess, Zong was not satisfied with the sentence. Nor was Sun De, who was in a state of dire poverty and had not received the two ounces. Both being driven by cupidity, the previously allied Sun De and Sun Zong brought new charges with the help of an unscrupulous intermediary. The content of the prosecution that followed and the nature of the accusations are, again, not clear. What is clear is that Mao Yilu regarded their attempts as vain manipulations unlikely to deceive a sagacious prefectural judge. To the demands of Sun De, who wanted to be considered as Qiao’s master in order to have authority over his property, Mao Yilu answered: “Trying to keep people as bondservants generation after generation with the help of a decades-old damaged piece of paper is both greedy and stupid.” And as if this statement were not sufficient to support his argumentation, he added: “If there is actually a connection between Zong and De, when Zong’s father died and when his mother remarried, how did it happen that [De] did not even mention [this relationship]; and he has now uselessly forged a contract!?”84 For the judge, Sun De was only “telling ghost [stories] and dreams!” Finally, considering that there had been no material loss, Mao Yilu only reprimanded Sun De and likewise Sun Zong for his revolting behavior. Yet, the sentence itself is less important than the enlightening (and suppos- edly enlightened) reasoning of the magistrate and prefectural judge. First, it is worth noting that neither the district nor the prefectural jurisdiction raised objections against the possession of bondservants from ordinary birth. The document says nothing about the master’s family status, nor does it explain the conditions that led Sun Cheng to sell himself or to be sold as bondservant. Yet, reading such a case and similar examples, one soon comes to the conclusion that in the late Ming the voluntary selling of persons was regarded as a normal practice, though it was contrary to the spirit of early Ming institutions. And if Mao Yilu rejected the validity of the contract, his motivations were elsewhere. It is also striking to notice that bondservants often adopted sons. If the reasons for such adoptions are rarely mentioned, and if sometimes they clearly result from simple pragmatism,85 they may also denote a strong will among bondservants to

84 It remains unclear whether the contract was authentic or not, since it is once described as an outdated old piece of paper and then as a forged contract; see below. 85 Mao Yilu, Yunjian yanlüe, juan 7, 512. acting as master and bondservant 263 share the same ritual standards as ordinary people. Adopting a son was a pressing need for men who could not father a son and who wanted to perpetuate their lineage and fulfill their ritual obligations towards their ancestors. This could also indicate that bondservants did not consider themselves as totally excluded from the society they lived in, and that they did not necessarily regard their condition as permanently, perpetu- ally and hereditarily binding. Above all, this case confirms that the principles of solidarity and reciprocity underlying the master-bondservant relationship were not pure abstractions. Without concreteness and proper behaviors, the ties that bound them to each other could be considered as relaxed or broken. Before the courts (probably more likely in ordinary, civil cases than in penal affairs), claiming one’s authority over a bondservant with an old contract in hand was not sufficient. One had to prove that one had always acted as a master and never let the link lapse. Thus, when adjudicating cases regarding doubtful bondservant identities, magistrates often resorted to the same argument. If one could show that master and bondservant had established separate households for long enough without a protest on the part of the master, the reality of the relationship could be denied; and so could the formal and moral link that gave a man authority over another one. Indeed, it is not for nothing that, even before recounting the sequence of events, Mao Yilu expounds two facts: (1) sixty years ago, the young Sun Cheng was bought by Sun De’s ancestors; and (2) since that time, “seasons have succeeded one another” and master and bondservant “had established separate households a long time ago.” When reading the first two sentences of the judgment, one can immediately understand the crux of the affair and the way Mao Yilu would handle it. Then, what about the contract and its legal value? In the present case, one cannot ascertain whether the contract was forged or not. It simply seems that the district magistrate who handled the case on first hearing had no reason to consider it was a fake (“Sun Cheng’s old contract is still extant,” he said), while Mao Yilu barely mentions an old piece of damaged paper before regarding it as forgery. Nevertheless, as real as the document might be, a contract was not necessarily a legally enforceable agreement. It was the embodiment of a private agreement and the formalized expression of a mutual commitment; it could be presented in court and its provisions could actually be held against a bondservant who refused to fulfill his obligations. Could it be proved that one of the parties involved had broken the commitments, the contract would be regarded as invalid and totally empty words. In other words, if the reality it was supposed to 264 claude chevaleyre guarantee had evaporated, the contract was no more than an obsolete “old piece of paper.” And such was the case in Mao Yilu’s eyes. Not only had the litigants established separate households for a long time, but the master’s family had also not gotten involved in the funeral and wedding rituals of their former bondservants. Whatever was written in the contract and whatever its authenticity, for the two magistrates who handled the case, the relationship could not be maintained because it no longer rested on a tangible reality. Whatever the precise sentence pronounced by Mao Yilu, he might have followed the initial judgment of the district magistrate and ordered the descendants of the bondservant to pay a sum of money to the descendant of the master in order to “demonstrate the signification of the rites.” If the termination of the solidarity between master and bondservant was mani- fest in the eyes of both officials, the main problem was that it had never been formally established. In consequence, the district magistrate—whose first mission regarding civil matters was to ease minds and prevent further prosecutions—tried to pronounce an arbitration both parties would abide by. In this regard, the sum of money he ordered the two brothers to pay to the master’s descendant was mostly symbolic.86 Materially, it was substantial enough to be regarded as a correct compensation by Sun De, but not to be considered as the price of a regular redeeming. It did not mean that Sun Zong and Sun Qiao were bondservants who redeemed themselves by paying two ounces. Symbolically, the sum was meant to demonstrate “the significance of the rites.” On the one hand, it broke the bond formally and definitively, but on the other, it was meant to remind that as descendants of bondservants, Zong and Qiao remained hierarchi- cally inferior to the family of their father’s master. If the material bond was broken, moral and hierarchical traces still remained. But for the two brothers, such diffuse discrimination and the payment of two ounces were worth the cost for future peace of mind and formal emancipation. Had everyone acted without selfdom motives, the initial arbitration would

86 As a comparison, in case no. 15 of Yunjian yanlüe, a certain Gao Zhong, who was “undoubtedly” a bondservant (even though he was in fact the adoptive son of a bondservant), was finally authorized to redeem himself. Gao Zhong had shown himself to be so reluctant to remain as a bondservant that his master, who in a first move placed a complaint against him, finally accepted to let him go. What could he do with someone who refused obstinately to comply with his orders? In consequence, the master entered into an agreement with the bondservant’s clansmen and accepted a price of 40 ounces (more than 1 kilogram of silver) to redeem Gao Zhong; that is to say, forty times the “compensation” Sun De received for each son of his family bondservants. acting as master and bondservant 265 have morally and materially satisfied all the parties. It is precisely because it was not the case that Mao Yilu showed so much anger against Sun Zong and Sun De: cupidity was not morally acceptable.

The Jin versus Wang Case

As another example of the reprobation against the pursuit of self-interest and the demand for concreteness in master-bondservant relationship, let me refer to case no. 96 in Draft Court Opinions from Cases Personally Tried as Inspector of the Wu Region, another late Ming collection, this one of cases adjudicated by Su-Song Regional Inspector Qi Biaojia (1602–1645).87 As a Regional Inspector, Qi Biaojia held a key position in the process of judicial appeal and review. Some time around 1635, one Wang Qing, his accusers and all the written pieces related to a case of forged homicide were transmitted from Kunshan district to Suzhou and submitted to Qi’s expertise. Examining the ins and outs, he stated: It has been established that Jin Cheng, the grandfather of Jin Zhong and Jin Xiao, long ago sold his own son, Jin Peng, to the Wangs. Since then, [Jin Peng] had always served [the Wangs] well. After the successive deaths of every member with official status, the [Wang] family entered a downturn. As a consequence, [Jin] Zhong and [Jin] Xiao left and established themselves elsewhere, just as if they ignored being of hereditary bondservant status. Everyone since then concealed the fact that they were bondservants of the Wangs, and this is the reason why Xu Shengcai accepted to have [his daughter] married [with one of the Jin brothers]. Yet, in 1627, a Wang clansman, Wang Qing, blamed them for having broken their ritual duties, and had an altercation with them. Thus, Jin Xiao, who was also the most cunning, unexpectedly seized the occasion and accused [Wang Qing] with the hope of washing away the name of “bondservant.” That is where all the judicial troubles began. In the continuation of the document, Qi Biaojia recognizes that the Wangs’ claim of authority over the Jin brothers was legitimate. It was not, he said, an empty claim invented to counter the Jins’ accusations. Moreover, not only did Jin Xiao himself accuse Wang Qing, but he also led his nephews (Ya and Yun) to accuse the Wangs of having murdered his brother

87 Qi Biaojia, An Wu qinshen xigao, 517–518. See Handlin Smith, The Art of doing Good, 248–278, passim; and see Will, Official Handbooks and Anthologies of Imperial China, “Anthologies of Judicial Sentences and Model Cases,” 4.1.5. 266 claude chevaleyre

(Jin Zhong, who had died in the meantime). But the sagacious Regional inspector was not fooled by their tricks: However skillful their counter-attack ploy, they have overlooked the fact that the distinction between master and bondservant still remains. According to the principles and according to the law, they naturally should be sentenced to return and serve [the Wangs]. However, for having falsely accused people of homicide, Jin Ya has already been severely chastised. For the sake of the Wangs, and considering that this bondservant has proved so very rebellious, returning him to serve in their house would be like opening the doors to a bandit, and it must be feared that one day some domestic tragedy might occur. It is not necessary to have him handle housework by himself in order to rectify the distinction between master and bondservant. As to the fair judgment I am pronouncing today onward, Jin Xiao and his nephews shall never place accusations against their master anymore. How- ever, considering that Xiao was likely to arouse people’s attention because he claimed that the Wangs wanted to take advantage of his wealth, there is [only] one way for the Wangs to clarify their motivations: That the persons remain as persons belonging to them, and the property remain as property belonging to the Jins. The Jins are authorized to go to the district magistrate’s office and get a certificate that will guarantee them the perpetual property of all the goods acquired by Jin Zhong and Jin Xiao [and presently] in the Jin family. As literati with dignity, how would the Wangs care for such unrighteous goods? In so doing, the reasons Xiao had to cause trouble and go to court will have been suppressed. [. . .] What should be noted in this case is the care with which Qi Biaojia attempts to determine the status of each person before he tackles the judicial issue. In the same manner as Mao Yilu, he examines who was sold to whom and if there had been a rupture in the relationship between master and bondservants. In this particular case, there was no doubt. The Jins were undoubtedly the hereditary bondservants (shipu 世僕) of the Wangs and the bond had never been formally interrupted until the Wang family fortunes entered a financial downturn. Probably out of resources, the Wangs were constrained to let them leave. The Jins thus enjoyed a state of de facto emancipation and eventually acquired some wealth. When the Wangs attempted to exercise their authority over their former bondservants (and over their newly acquired property), the latter refused to comply. And this was the source of all the subsequent prosecution that Qi Biaojia was in charge of adjudicating. What is striking in such a case is the discrepancy between legal practice and theory in late Ming judicial proceedings. The Code clearly stated that if a bondservant ever accused his master he would be subjected to caning. Moreover, if the accusation proved to be false, he could be sentenced to acting as master and bondservant 267 death by strangulation.88 This provision did not exempt the magistrate from investigating the case, but it offered the accused master quasi-total immunity, whether or not he was guilty of something. As there was no doubt regarding the status of Jin Xiao and his nephews, Qi should logically have rejected the accusations and condemned the bondservants. How- ever, he did not invoke the law, and, rather than pronounce a simple and expeditious judgment, he chose to develop more complex considerations. Given that he mainly aimed at demonstrating the faults of the bondservants, one might deem that, in the end, he respected the spirit of the law; but on the other hand one might also admit that he did not apply the letter of the law. No doubt, in his eyes there were more urgent and more essential problems to deal with, despite the offence committed by bondservants in accusing their master. He knew that reaching the appropriate balance between moral standards and economic considerations was more likely to ease minds than a rough application of the law. The commercial- ized society of the seventeenth century was different from the early Ming social order for which the Great Ming Code was originally designed. The quarrel over the Jins’ status was in fact a pretext for concealing a somewhat banal—though very delicate at that time—problem of property. Considering that the Jins were without doubt hereditary bondservants (shipu 世僕) belonging to the Wangs and in order to satisfy both principles and law (juli jufa 據理據法), Qi Biaojia could not allow the Jins to “wash away the name of bondservant” (xituo numing 洗脫奴名); he had to “rectify the distinction between master and bondservants” (zheng zhunu zhi fen 正主奴之分). Yet, he also acknowledged that over a period of time, the bond had loosened and no longer rested on concrete grounds. In his eyes, this was not insignificant. Thus, he had to clarify whether the Wangs were solely motivated by legitimate moral claims or if they only “wanted to take advantage of [the bondservants’ newly acquired] wealth.” If such was indeed the case, his skillful judgment would teach them a good lesson. On the one hand, the judgment would likely satisfy their moral requirements by rectifying social relationships, and the person of Jin Xiao would remain their property. But on the other hand, it was formulated so as to thwart plans only motivated by self-interest. The Jins could definitely secure (shishou 世守) the property they had acquired, and they would even get an official certificate to enforce the decision. As a consequence, the Wangs could only show themselves satisfied with the

88 Law no. 360; see Gao Ju, Da Minglü jijie fuli, 22:30a–31b. 268 claude chevaleyre judgment. What are a few “unrighteous goods” compared to the rectification of social hierarchy and harmony, Qi asked? Showing any dissatisfaction with the decision would immediately expose their cupidity. And this would probably have other unfavorable consequences. One point cannot be clearly elucidated in this case: Did the judgment imply the dissolution of the hereditary character of the tie binding the Jins to the Wangs? The situation had gone so far that the bond remained only nominally. Xiao and his nephews showed themselves so rebellious that maintaining them in their master’s house would be like “inviting a bandit in one’s house,” Qi said. Then how was the bond supposed to be perpetuated generation after generation if it was only nominal and if it rested on no concrete grounds? Qi also recognizes that the Jins constituted a “household” ( Jin hu 金戶) to which was granted the exclusive use of its property. Yet, bondservants had no personal households and were usually registered under their masters’ names. In other words, how could such a judgment be reconciled in practice with the status of hereditary bondservant? Did it not objectively imply the end of hereditary ownership? In fact, Qi had probably another reason to try to invalidate the hereditary character of the bond. One of the Jin brothers, probably Jin Zhong, had married the daughter of Xu Shengcai, a commoner who ignored the real status of his son-in-law. Since a woman’s status depended on her husband’s, when married to Jin Zhong the daughter of Xu Shengcai would then have become a bondservant herself. This is precisely the reason why the law prohibited weddings between bondservants and commoners’ daughters.89 And this was surely something a father did not want for his daughter—an anomaly Qi Biaojia was very keen to rectify. By restoring the bond only nominally, and, while recognizing that the Jins constituted a household, the Regional inspector rectified both the status of the bondservants and that of Xu’s daughter. Finally, it is worth noting that Jin Xiao and his nephews considered that a judicial offensive against their master’s family could potentially turn to their advantage. Despite their status of hereditary bondservants, they did not consider their condition as totally irrevocable, nor did they regard

89 Law no. 121, op. cit. In a Yuan (1277–1367) encyclopedia for daily use that was very popular in the Ming, weddings between ordinary and mean people are depicted as unions between entities of different nature that were to be dissolved: “Every person has a mate who is to be of the same kind. Considering that commoners and degraded [people] are different, how could they intermarry? Would transgression occur, [the mates] should be separated and [the situation] rectified;” Huang Zongli (?), Jujia biyong shilei quanji, 324. acting as master and bondservant 269 their actions as bound to fail. We will never know on what their calculations rested. But one thing can be taken for certain: At that time, bondservants felt less constrained by the law or by their status than we usually think. The picture emerging from legal cases does not contradict what we already know about Ming bond-servitude, but it is likely to modify the perception we usually have of a rigid and uniform system of personal domination. Of course, Chinese bond-servitude was discriminatory, and it was the result of economic pressures and domination. But it was not the exclusively one-way system of exploitation described with insistence by some historians. It was an unequal game of powers, but in the balance, bondservants were not totally vulnerable and powerless. When the stronger let his attention wander, the situation was likely to turn to the weaker’s advantage. If a master let his power weaken, this was likely to reinforce the bondservant’s position, and, at least in extreme cases, lead to the dissolution of the bond. We should naturally be cautious when making generalizations based on a few judicial cases; in practice, masters probably were not worried by the same moral considerations as magistrates, and it is clear that bondservants did not hesitate to refer matters to courts and make use of every possible judicial trick to get out of their situation. And, in court, social hierarchy and moral domination on the one hand, and economic domination on the other hand, were two different and hierarchized things. Reciprocity was not an empty word, and economic domination could only be the result of a pure moral domination that, in turn, implied obligations on the part of the master. The preservation of social order depended on harmonizing names and realities. On such a scale of values, material considerations were essentially incidental.

Conclusion

Among the various subjects that this chapter has not dealt with, the most telling is probably that of quantitative data. When exploring any form of servitude, historians focus first on two essential preliminary aspects: legal status and detailed figures. In the case of late Ming China, the main problem lies in the fact that both data remain incomplete. Ming law did not give a precise definition of social status and this is the reason why the legal status of bondservants is usually reduced to the discrimination they suffered and to a bare series of legal incapacities. And this is probably the reason why the famous historian Wu Han (1909–1969) could affirm in 1961 that, in the Ming a bondservant was private property with no more 270 claude chevaleyre value than an ox or a horse, considering that the Code prescribed quite the same punishment for sacrificing an ox and for killing a bondservant.90 Hopefully, numerous subsequent studies have demonstrated first that one could not simply regard the seriousness of two crimes or offences as identical only on the basis that they were liable to the same punishment, and, second, that bondservants were not simply dehumanized production tools or chattel slaves. As for figures, it must be acknowledged that, in the case of Ming China, data are rare, most figures being insufficient and unreliable. In a series of conferences held at Collège de France, Timothy Brook has shown how critical the lack of data on prices could be for historians of the Ming period.91 But such a statement is also true when considering population series, especially when considering the distribution of social groups. Ming writers and administrators rarely noted figures. But the task of evaluat- ing the number of bondservants would be even more difficult than eval- uating any group of commoners because many cases of bond-servitude were disguised and hidden behind marriages or adoptions, as mentioned above; and also because many made themselves bondservants in order to evade taxes, subsequently disappearing from population registers. So, what figure shall we accept? How are we to interpret the fact that at least one bondservant was involved in 24% of the cases published in the Yun- jian yanlüe? Shall we rely on the recurring laments of late Ming writers claiming that great clans and families in the lower Yangzi regions had “hundreds,” “thousands,” and even “tens of thousands” of bondservants? Such numbers are less concrete than rhetorical, and they don’t say a word about regional variations.92 Let’s assume that Ping-ti Ho was right when he ascertained that mean people amounted to approximately 1% of the Ming population.93 This would induce a total of mean people contained between 700,000 and 2 million around 1600, among which, probably, a great majority were bondservants.94 This is a plausible figure, but how far

90 Wu Han, Dengxia ji, 76. 91 Brook, “Valeurs et prix dans la Chine des Ming.” 92 Most of the documents presented in this paper focus on the Jiangnan region, that is to say the prefectures of the rich and developed Yangzi River delta. As a consequence, I shall warn the reader against generalizations on China as a whole. If bondservants had the same legal status all over the empire in the seventeenth century, practices may have greatly varied from one part of the realm to the other. 93 Ho, The Ladder of Success, 19. 94 Depending on ways of calculating, population figures for 1600 range from 66 to 230 million. See Brook, The Troubled Empire, 42–45. acting as master and bondservant 271 is it from the reality? This remains a pending question, as must the question of how long-term trends evolved. Some scholars have advocated a model of general decrease in the use of bondservants through the Ming.95 Yet, one cannot ignore the recurring narratives of late Ming and early Qing writers who lamented the increasing number of bondservants all throughout the Ming.96 In the same manner, shall we accept that late Ming insurrections were such a social trauma that they lead to the quasi disappearance of bondservants as soon as the Kangxi era,97 or just after the Yongzheng reforms?98 Or shall we simply trust the numerous imperial archives that remain to testify that crimes or offences committed by or against bondservants kept judicial authorities busy at least until the beginning of the nineteenth century? What can be ascertained is that bond-servitude and trafficking in human beings persisted long after the fall of the Ming and probably after the establishment of the Republic of China, some 250 years later.99 The present chapter thus attempted to demonstrate that China, this civilized, Confucian and too often atemporal historical object, was also part of the world history of human bondage. As such, it should less be regarded as exotic than as another configuration worthy of attention to shed light on the universal experience of human bondage. In many respects, most of its characteristics, though encapsulated in so-called Confucian values and in specifically Chinese practices, can be fruitfully compared with the characteristics of bondage practices throughout the world. As in many places in the world, voices were sometimes heard to denounce the evils of the

95 Elvin, The Pattern of the Chinese Past, 235. However, Elvin’s analysis focuses on the progressive disappearance of the “manorial order with serfdom and a serf-like tenancy,” in which bondservants were included and not totally differentiated from other forms of “serfs” and “tenants.” For him, this structure “continued to dominate the countryside, though with diminishing vigour as time passed by,” to finally disappear in the course of the eighteenth century. 96 Xie Guozhen estimated that bond-servitude had quite disappeared after the Song dynasty, before a new period of steady growth in the Ming. See Xie Guozhen, “Mingji nubian kao,” 195. 97 See the quotation of Shengguo jiwen stating that “under Kangxi, great families didn’t dare to raise bondservants anymore,” in Xie Guozhen, “Mingji nubian kao,” 211. 98 Ho, The Ladder of Success, 19. 99 If Chinese bond-servitude can in no way be compared with forms of servitude based on the traffic in human beings, illegal human trade networks, however, existed. See, for example, the early Qing case adjudicated by Yu Chenglong, who dismantled a trade in children between Guangdong and Guangxi provinces. More than 200 boys and girls had been sold for “entertainment” by a network involving some 80 accomplices. See Yu Chenglong, Yu Chenglong pandu jinghua, 37–39. See also the studies by James Watson on slavery remnants in Guangdong under the Republic; in Watson, “Transactions in People”, 223–250. 272 claude chevaleyre servile institution and to call for more humane treatments and benevolence. But, in fine, and despite limits imposed by the principle of reciprocity and the requirement to make social status match realities, Confucian values—as many other civilized or humanist doctrines—did well with the servile institution. Considering that time has come to reintegrate the Chinese experience into the world history of bondage, focusing on the late Ming dynasty seemed a good starting point as many of the characteristics of the Chi- nese servile institution as it was denounced—both by Western observers and by Chinese modernists and revolutionaries—at the end the imperial era were shaped at that time. To decide whether the Chinese servile institution must be qualified slavery, bond-servitude, or whatever else would seem appropriate is of secondary importance. What is important is that the Chinese experience is no less informative than other experiences to understand how universally shared institutions such as slavery operate. Public Works and the Question of Unfree Labour

Chitra Joshi

Studies on Indian labour till recently were framed within a teleological logic which associated capitalist industrialization with notions of freedom. Within such frames, forms of coercion and unfreedom appeared as signs of the incompleteness of processes of modernization and transformation. Discussions around bondage and slavery were important to writings on agricultural labour but the focus of labour history was primarily on factory industries. Within modern industries state legislation was seen as a move from punitive to protective measures: laws limiting working hours, regulating wages legalizing trade unions marked this shift.1 State intervention was thus seen as formalizing, legalizing the network of industrial relations. The ways in which state legislation in fact sanctioned the private coercive powers of capital over labour were not studied. Recent writings on law and labour question earlier frameworks in two ways: One, they push the origins of labour policy to the late eighteenth and early nineteenth century;2 two, they move beyond the realm of the formal and suggest how the state was deeply implicated in creating a network of informal systems of control and coercion which lay outside the field of law. Coercive, ‘unfree’ forms were actively created by the state and were not residual or a product of the absence of laws.3 The first half of the nineteenth century was a period when the East India Company was in the process of expansion and consolidation of its empire. Mapping, exploring and building of road networks to allow a rapid movement of troops was part of this process.4 Some of the major road projects, the Grand Trunk Road extending over fifteen hundred miles from Calcutta to the North-West, the Agra-Bombay road extending over 700 miles,

1 On this shift see Anderson, “Work Construed”; Anderson, “India, 1858–1930”, in Hay and Craven (eds.), Masters, 402–54. 2 Ahuja, “The Origins”. 3 Arguing this very forcefully, Prabhu Mohapatra points to various legislative measures that gave additional private powers to employers for instance, Act. VI of 1865 gave additional penal powers to planters to discipline labour. Mohapatra, “Regulated Informality”. 4 On the importance given to road building by administrators like Bentinck and Auckland see Singha, A Despotism, 230–32, 258–61; Sharma, Famine, 160–71. 274 chitra joshi the Delhi—Allahabad, road (around six hundred miles) the Western Road in Madras Presidency, the Deccan Road from Mirzapore to Jabalpur the Hindustan-Tibet road, were undertaken in this period.5 Projects like the Western Road employed on an average 1598 labourers per day, running up a total of around 40,000 in some months; the Hindustan-Tibet road between Dugshai and Simla employed on an average 1747 labourers per mile (a total of 86000 labourers) and between Simla and Narkunda, 1965 workers per mile.6 Who worked on the roads? How were workers mobilized for work on the roads? Labour on the roads embodied some of the contradictions of modernity. Roads, in a sense, symbolized modernity. Yet they were con- structed by forms of labour that could be not legitimated through the rhetoric of modernity. How do we understand this? Work on large public works had certain imperatives and these are important to understanding the strategies of labour mobilization and discipline. The construction of roads, canals and other public works required large bodies of labour at particular sites for a fixed time period. The demand for labour was a spatially shifting one, moving from one construction site to another. The demand was temporally bounded: once a project was complete the demand fell. This created problems. It was difficult to get an adequate and stable supply when it was needed, and a spurt of demand in one place could push up wages. So officials on public works had to organise a supply that could meet the concentrated demand of labour without raising wage costs. Obtaining local labour in the different places that the roads passed through was particularly difficult. During plowing, sowing and harvest times field labourers worked on the land. They could not be attracted to public works without offers of high wages.7 Moreover landed

5 Contrary to colonial official claims however it is clear from recent studies that road works in the nineteenth century did not always entail the opening up of completely new and uncharted routes but the re-making of old in accordance with the new transport requirements. See for instance Ravi Ahuja’s important recent publication, Pathways, 119–51. 6 Report on the Western road by W.K. Wurder, 30 October 1839, Home Judicial Proced- ings (Progs), 1 March 1841 (National Archives of India, hence NAI); Briggs, Superintendent, Hindustan Tibet road to Elliot, Secretary, Foreign Department, Foreign Consultations, 7 March 1851, no. 196 (NAI). 7 In the 1850s, Public Works officials in Madras Presidency pointed out that except in districts where traditional castes associated with earthwork or castes, considered ritu- ally impure within the Brahamanical order—like Pallars or Pariahs for instance—were abundant, it was impossible to get labour to work on the roads. To resolve this problem they proposed the idea of forming a ‘Permanent Corps of coolies’ for work on roads. See public works and the question of unfree labour 275 elites were often apprehensive of roads that absorbed village land, and cut through cultivated tracts. They were interested in smaller local roads that connected the villages to the grain markets, not highways. Inter-regional roads, in fact, undercut the authority of landed elites. In many regions local lords resisted the projects of public works departments.8 Historically, in different conjunctures, several different forms of coercive labour were employed to meet the demands of road building. Three kinds of labour commonly employed in road works in the nineteenth century were begar,9 convict and famine labour. If in the early decades of the nineteenth century, convict labour was predominant in India, other forms were more important later. Canal and road building projects in the late nineteenth century recruited large numbers of workers from famine stricken areas. Public works in many areas became synonymous with famine relief.10 The construction of irrigation works and roads proceeded most rapidly during times of famines. Starving peasants were paid famine wages to build the public infrastructure that became one of the signs of modern India.11 What explains the shift from one form to another?

Tribute Labour

On projects like the Hindustan Tibet road undertaken in the 1850s, the demand for labour was much higher than roads in plain areas. At any point of time there were from 1700 to 2000 workers on each mile of the road. A large proportion of the labour employed on the Hindustan Tibet road was categorized officially as ‘unpaid’ labour. A 132-mile stretch of road on the route between Kalka and Chini employed around 1,164,644 workers in the year, of which nearly 27 per cent (312598) was ‘unpaid’. In the official logic, labour classified as ‘unpaid’ was considered as part of the tribute hill states were obliged to pay for the protection they received from the colonial government. In the official rationale, tribute labour supplied by Hill states was not like ‘begar’ or the customary form of unpaid

Capt W.H. Horsley, to Capt J.H. Bell, 29 September 1853, Selections from Government Records, Madras, No. VI (Government Press: Madras 1855), 1–4. 8 See for instance R. Ahuja, “Opening up”. 9 Begar was a customary form of unpaid labour demanded by dominant groups in a region. 10 In the famine of 1838, labour from famine affected areas was employed on road works in many areas. See Sharma, Famine, 161–66. 11 See for instance, Henvey, A Narrative, 99–128; CER Girdlestone, Report, 47–50, 99. 276 chitra joshi labour. The Hill states were expected to give revenue remission to those who were working on the roads. In official perception the injustices of this system were mitigated by its longer term benefits. Road work could ultimately do away with what in official eyes was the worst form of the system—the use of coolies for porterage. Roads would allow a freer movement of animal and cart traffic. In practice, tribute labour on mountain roads intensified the practice of begar.12 While forms of ‘begar’ labour were customarily employed in building village roads, the use of begar in public works was in many ways distinct from its customary use. Work on public roads was located away from the village. Often it meant working dangerously on treacherous terrain and staying out in the open without shelter for days. Begar on distant roads was doubly exacting because workers had no access to subsistence resources from the village: they had to buy grain at high prices from the market. Zamindars from Kulu protested against labouring on roads: as our interests suffer by this order we presume humbly to approach your lordship with this petition we are poor men, and our country is ill populated we have roads of our own not less than one hundred and eleven kose long to look after. . . . we are charged with the construction and maintenance in thorough repair of this long line of road. This duty leaves us little or no leisure. Ordinarily we live in separate families with no more than one adult male to a family. Our agricultural occupation and our public duties on the roads and for conveyances of baggage are got through with difficulty. If in addition we are obliged to work on the Grand road our interests will suffer. As a people we are remarkable for attachment to our own hills. From choice we have never consented to leave them. The Government orders we obeyed without a murmur. But now that your lordship has honoured our country with your presence we account the present a favourable opportunity for settling our grievance before your lordship’s consideration.13 Contrary to the directives of the Governor General the hill roads project could not be executed on the basis of tribute labour alone.14 Yet the use

12 The use of the bethi labour in Orissa in the nineteenth century is somewhat similar. See Ahuja, “Opening Up the Country”, 73–130. 13 Petition of Zamindars of Koolloo to Gov Gen of India, Public Works Department, Hill Road Papers, Hindustan Tibet Road Papers, 6 June 1851, no. 29 (NAI). 14 Note for instance a minute of the Governor General in 1850, ‘I acquaint the Super intendent of Hill states that without further information I decline to give my sanction to the measure he has proposed, for employing a large number of hired labour in the construction of the new road now in progress from Kalka through the Hills. The application of the Superintendent involves a wide departure from the conditions on which I gave my assent to the commencement of this work, and entirely alters its direction.’ Foreign Dept (Political), 13 September 1850 (NAI). public works and the question of unfree labour 277 of tribute labour from Hill states served two functions: it cut total costs involved; it also tended to push down the market rate for labour. The official rate of wages on the Tibet road was fixed at 2–2 1/2 as. per day. However, there were areas where workers could not be mobilized even at double the rate. On regions bordering on Ladakh for instance, road building authorities were anxious: ‘I doubt very much whether a single Lad- hakhee could be got to work on the road for 4 as. a day.’15 Apart from the question of costs, there was also the issue of moral legitimacy. The state was complicit in sanctioning a system of forced labour within Hill states without being directly implicated in deploying a system of forced labour. The terms within which discourse on such forms of tribute labour was framed is different as we shall see from discussions on convict labour. Unlike convict labour which bore the marks of its unfreedom, force in the case of tribute labour is masked: the coercive mechanisms through which labour is deployed from the supplying region is not recognized. In the case of both forms however, the idea of public good provides a basis through they are legitimated within colonial discourse.

Convict Labour

Some of the important road projects undertaken in the 1830s and 1840s covering vast stretches employed large bodies of convict labour. The construction of public works in India in this period was supervised by the Military Board and the practice of using coercive forms of labour for works commissioned by military authorities was common. Military engineers supervising road works had anywhere between 1000 to 3000 prisoners working under them. On the road between Delhi and Allahabad (approximately 600 miles) around 4686 prisoners distributed over 15 gangs, were employed.16 Usually convicts were employed on work

15 WM Hay, Superintendent Hill States to Secretary, Government of India, 1 Nov. 1854, Hindustan Tibet Road Papers, Public Works Department, 19 January, 1855, no. 512–516 (NAI). 16 A gang consisted of some 300 prisoners who were sub-divided into groups of 20–30 to do specific tasks. A hierarchy of guards: one burkandauze (an armed guard, usually with a matchlock) to every three prisoners and four duffadars (originally a person in charge of a small body of troops) over every 100 burkandauzes kept watch while the convicts worked. A gang of 300 prisoners had over 90 burkandauzes guarding them. The heavy work of digging and raising of embankments was done by large gangs of prisoners. For purposes of surveillance large gangs were preferred. For the work of metalling and making bridges, smaller groups had to be employed which made guarding more difficult. I.G. Drummond, 278 chitra joshi which required large bodies of workers—like digging and raising embankments. For metalling roads or making drains and bridges which required smaller parties of workers, hired labour was preferred.17 In the Western road in Madras Presidency, on the basis of a daily average of 1598 convicts, the total number of convict days involved peaked to over 40000 in the rainy month of July.18 Convicts doing earthwork had to lift a minimum of 50 cubic feet of earth by midday. Usually it took much longer and even by ‘straining every nerve’ it was 2–3 pm by the time they finished.19 Superintendents in a rush to complete projects and corrupt guards made heavier demands on convicts.20 Medical officers enquiring into the reports of mortality on the roads drew direct connections between burden of work and mortality on the roads. An important feature characterizing the convict labour regime on the roads was the elaborate system of surveillance and control. On the Delhi— Allahabad road for instance, there were a total of 1338 guards watching 4686 prisoners. There were in all 15 gangs of prisoners working on this road. Each gang had over 300 prisoners with around 90—burkandauzes overlooking them.21 Gangs were further subdivided into smaller parties of

Superintendent, Delhi and Allahabad to Capt Gowan Officiating Secretary, Military Board (MB), 21 March 1834, MB Proceedings 27 June 1834 (NAI). 17 ‘All that belongs to the marking out and metalling of roads, and the construction of drains and bridges should be done by hired labour. It is necessary to employ small parties for executing works of the description mentioned and it appears that prisoners cannot with security be so unless by increasing the strength of guards at an expense that would perhaps exceed the value of their work. Nicolson, to Debude, MB Proceedings, 25 November, 1836 (NAI). 18 Report on Western Road by W.K. Wurder, 30 Oct 1839, Home Judicial Proceedings, 1 March 1841 (NAI). 19 Drummond, Superintendent, Delhi and Allahabad to Capt Gowan Officiating Magis- trate, 21 March 1834, MB Proceedings, 27 June, 1834 (NAI). 20 Burkandauzes tended to be partisan forcing more work on some and letting off others. Reports like the following were common ‘. . . one of the convicts named Golamully made a complaint to me stating that his basket was loaded with much more earth than the rest wore, as he had not given the usual present.’ Jenkinson, to Thomson, 21 February 1833, Judicial Criminal Proceedings, 15 March, 1833 (NAI). 21 Drummond, Superintendent Delhi and Allahabad road to Gowan, Offg Secy Mili- tary Board, 21 Mar, 1834, MB Progs, 27 June 1834. In selecting burkandauzes, caste was an important issue. Prisoners from an upper caste background—Brahmins, Thakurs—were not placed under the supervision of burkandauzes from a lower caste background. MB Pro- ceedings, 11 November, 1834 (NAI). Burkandauze: armed guard usually having a musket. public works and the question of unfree labour 279

20–30 prisoners with assigned tasks to complete. In addition there were around four duffadars and two jamadars22 over every 100 burkandauzes.23 Convicts had to work in fetters in the day and at night groups of around twenty were bound together by a chain which ran through the centre link of each individual leg chain. Although prisoners were supposed to be relieved of their irons in case of extreme sickness, this was rarely done. A doctor reporting on the state of convicts working in road gangs in India in the 1830s was surprised to find: ‘. . . that the sick [convicts] even in the last stages of disease are kept in fetters . . . in every instance even when the prisoners were reduced to the appearance of a skeleton and unable to move from side to side the irons had not been struck off . . .’24 On the Burdwan road, prisoners attacked by cholera, remained chained all night. A doctor visiting them was almost driven back by the insufferable stench of excreta emanating from the tent.25 Surveillance and repressive structures of control are not peculiar to the convict labour system. In plantations and mines, managers closely regulated the body of the worker. The space of work in mines and plantations was bounded. Coolies trying to escape from plantations were often flogged sometimes to the point of death.26 The powers of planters extended beyond the workspace to the ‘coolie lines’: they regulated and controlled the private lives of workers. Mines were known for the severity of measures against theft of mine property. Miners songs in the Kolar gold mines in South India, recorded the extremes to which checking against pilferage went: they lampooned mine authorities for forcing them to bend over like cattle to for signs of hidden gold in their orifices.27 In contrast to mine or plantations workers who are kept under surveillance within a circumscribed work space, convicts on road work have to

22 Duffadar: originally a military word for a person in charge with a small contingent of troops. Jamadar: used usually for the leader of a group of individuals. 23 Thomson, Supdt roads to Goween, Officiating Secretary, MB, MB Progs, 11 November, 1834. 24 MB Progs, 2 August, 1836 (NAI). 25 Report of the Committee on Convict Labour, MB Progs, 2 August, 1836 (NAI). 26 See for instance, Rana P. Behal, and Prabhu P. Mohapatra, “ ‘Tea and Money versus Human Life’: The Rise and Fall of the Indenture System in the Assam Tea Plantations 1840–1908”, Journal of Peasant Studies, 19 (1992): 142–72. 27 It was suggested that women should be made to jump up when they left the mine premises, to check if gold was hidden in their private parts. Nair, Miners, 52. On systems of surveillance and control in South African mines see: Worger, South Africa’s, 110–46; Crush, “Power”. 280 chitra joshi be on the move. The paradoxical character of the convict labour regime lies in its creation of a labour force that it is on the one hand fixed and immobile and on the other flexible and mobile: the location of work is a continuously shifting one. This mobility enhances the need for fettering. Unlike other coercive systems in which labour mobility is controlled and regulated through the burden of debt or penal contracts, convicts workers are physically immobilized.28

The Logic of a Fettered Regime

Officials in the Military Board engaged in the building of roads had to continuously engage in a dialogue with critics of convict labour. What were the terms of this debate? What accounts for the use of such a coercive system of labour? Crucial to the discussion on convict labour in the early nineteenth century India as in debates over free/unfree labour elsewhere was the issue of costs.29 Did employing convict labour work out cheaper than hired labour? What were the calculations involved? In official discussions on the relative advantages of convict versus other forms of labour, three issues are raised: the difference between average wage rates for hired labour and costs of employing convicts; the costs of keeping prisoners in jail and on the roads; the productive efficiency of convicts and hired labour. Military engineers executing vast road-building projects were insistent that prisoners could provide labour which was in the long run more cost efficient than hired labour. They responded sharply to critics of the system, questioning the basis of their evidence. In regions like Bengal, in India, lines were sharply drawn between the military and civil authorities over the question of employing convict labour. Here the high mortality among convicts on road work in the 1830s, led to mounting attacks on the system.30 Judicial officials gave evidence to prove that it was in the long run more expensive to employ convicts

28 On the question of mobility and ‘forced’ labour see also, van der Linden, “Forced Labour” in Brass and van der Linden (eds) Free and Unfree Labour, 351–62. 29 On the debate on this in the transatlantic context, see Drescher, “Free Labor” in Engerman ed. Terms 50–86. 30 In Bengal Presidency the average mortality among prisoners working away from the prison, under executive officers, was around 7.3 per cent in the period between 1833–36, as against 3.38 per cent in North-West Provinces. Report of the Prison Discipline Committee, 1838, Home Legislative Progs, September, 1838, no. 47–48, (NAI). public works and the question of unfree labour 281 on road work. According to estimates drawn up by the Prison Discipline Committee set up in 1838, the extra costs involved in employing convicts to do road work was around Rs.2 a month, but the same work could be done by contract labour at 2/3rds and in some places at 1/2 the price.31 Besides, convicts they argued worked fewer days a month: on an average, around 21 days, and this had to be taken into account in calculating costs.32 Military engineers were not convinced by these arguments. Irvine, a Superintending Engineer in Central Provinces, asserted that the estimates of the Prison Discipline Committee had no general validity, derived as they were from limited, unreliable and questionable evidence. They were in fact based only on the experience of Captain Murray in one single area in Bengal, in Noakhali. Ridiculing the estimates drawn by Murray, Irvine wrote: Captain Murray may probably have got a ditch round his compound at the rate of 13 as. per 1000 cubic feet, but this is very different kind of work from making a road 30 ft wide . . . In no part of Bengal could work of this kind done by hired labour on contract at the rate of 13 as per 1000 cubic ft. Unfor- tunately Capt Murray has done no road work by hired labour so that I am unable to bring forth any of his Bills to show what 1000 cubic ft of road would really cost at Noacally.33 He went to absurd lengths to argue that fetters did not really restrict the convict’s ability to dig earth or move as fast on level ground as free labourers. In regions like Bombay, those arguing for convict labour system claimed that absolute costs of hired labour in the Province were higher than other regions. In Salsette, in Bombay Presidency for instance, military officials responded to the objections of civil authorities arguing: ‘. . . labour here . . . is so much dearer than in the Upper Provinces we cannot here hire labour under Rs 5 a month’.34 Military engineers in Bombay emphasized on the specific local situation to justify the need for employing convicts at a time when the Bengal Government was trying to legislate against their use.

31 Prison Discipline Committee Report, 57. 32 Convicts did not work on Sundays and other local holidays. 33 Irvine to Military Board, MB Progs, 8 December, 1840 (NAI). 34 R Foster to IP Willoughby, Secretary to Government, Bombay, 9 September, 1839, Home Judicial Progs. 13 January, 1840 (NAI), Irvine, Superintendent Engineer, to Military Board, MB Progs, 8 December, 1840 (NAI). 282 chitra joshi

Others asserted the advantages of putting convicts on road work by comparing the relative costs of keeping them in jail. Estimates drawn up by the Revenue Department in Madras for instance showed that employing convicts on road work cost the state around 13 per cent more than hired labour. This relative disadvantage, Madras officials argued, was more than compensated by the savings in prison expenses. Keeping them in prison was more than three times the cost of employing them on roadwork.35 The question of costs was also linked up with the issues of labour discipline and efficiency. Critics of the convict labour system tended to argue that ‘free’/ hired labour was characterized by a motivation to labour, absent in convict labour.36 Morever, the wearing of fetters, they argued, restricted the mobility and productivity of convict labour. Advocates of the system countered assertions about the productive efficiency of ‘free’ labour by pointing to the problems of managing and disciplining hired workers. Hired labour had to be mobilised through the contract system and this always had its risks. The majority within the Military Board were opposed to giving the contract system a trial. Local contractors, they were convinced, used poor quality material and did not fulfill terms of contract. Although regulations drafted in 1814 and 1816 gave employers the right to file complaints against workers who left before completing their contract, these were considered inadequate.37 Employers continued to demand greater safeguards against absconding workers.38 Convict workers could be employed directly under executive officers without the mediation of contractors. Military officials like Vetch, involved in the construction of the Burdwan–Benares road saw this as a chief advantage: ‘To Executive officers labour by convicts afford much satisfaction.

35 Minute on Employment of Convict Labour, by Chief Secretary, 22 February, 1840. Home Judicial Progs, 1 March, 1841 (NAI). 36 This assumption was important in some of the exchanges between defendants of slavery and abolitionists in the late eighteenth and early nineteenth century. In reality abolitionists found it difficult to sustain this argument, especially when sugar production in the Caribbean collapsed in the aftermath of abolition. However these assumptions continued to underline debates over convict labour in India. On this debate see Drescher, “Free Labor vs Slave Labor”, 50–86, S. Engerman, “Slavery and Emancipation in Compara- tive Perspective: A Look at Some Recent Debates” Journal of Economic History, 46 (1986): 317–39. 37 The law of 1814 gave powers to Justices of Peace to issue warrants and conviction notices against workers who left work in violation of terms of contract. The regulation of 1816 empowered the Justices of Peace to take action against workers attempting to organise in defense of higher wages. Legislative Dept Progs, 12 March, 1852, no. 47 (NAI). 38 Legislative Dept Progs, 12 March, 1852, no. 47 (NAI). public works and the question of unfree labour 283

Extension work by daily labourers requires a system of native contractors who are in general an endless source of vexation to superintendents.’39 Moreover, convicts could be mobilized to work on distant jungle roads over long stretches of time necessary for road work, and in places where hired labour was difficult to find.40 However, if we closely examine the context in which it was used, it is quite clear that costs alone do not provide a sufficient explanation. In India, South Africa and in other countries the convict labour system points to the connected histories of labour and punishment. The regime of Bentinck the Governor – India in India, associated with grand road building schemes in the 1830s, coincided with a period when utilitarian ideas of legal reform were being deliberated. Imprisonment and new modes of prison discipline were part of a programme of penal reforms in this period. In the absence of a prison infra-structure and buildings, local magistrates saw outdoor work for prisoners as a short term solution to the problem of prison space.41 A spurt in the demand for convict labour coincided with a rise in the number of convictions in the 1830s and an absence of prison infra-structure.42 But extra-mural work became a contentious issue for advocates of penal reform. When the Prison Dis- cipline Committee was appointed in 1836, there was a growing body of opinion against extra-mural work. The Committee members saw outdoor work particularly on roads at a distance from jails as a threat to a system of prison discipline.43 To judicial officials, a continuation of the convict labour system was dependent on its deterrent effect. Work in fetters on the roads was mean- ingful if it shamed the criminal. But did it? Officials differed, turning the arguments continuously around, but the theme of shame recurred in different ways. For some, fettered work on the roads meant a mode of public

39 Capt. G.A. Vetch, Superintendent, Burdwan roads to E.P. Gowan Acting Secy, Mili- tary Board, MB Progs, 27 June, 1834 (NAI). 40 Lt Col Craigie, Military Board to Auckland, Governor, Bengal, 12 Aug., 1836, MB Progs, 12 Aug., 1836 (NAI). 41 Before the Law Commission put forward its proposals for penal reform, judicial officials were instructed by the Governor General to send prisoners with sentences exceeding three months to labour on public works. Macsween to Hunter, Judicial Criminal Progs, 30 Dec. 1833 (West Bengal State Archives); See also David Arnold, ‘The Colonial Prison’, Subaltern Studies VII (New Delhi: Oxford University Press, 1994), 148–87. 42 Prison Discipline Committee Report, 9; F.J. Mouat, Journal of The Statistical Society of London, 25 (1862): 183. 43 Prison-Discipline Committee Report, 61. On the changing ideology of punishment see, Garland, Punishment, 111–55. 284 chitra joshi shaming—the fettered body was not hidden away within a prison; it was visible, made the object of public gaze and censure. Upper caste convicts, in fact, were doubly shamed: first being exposed in irons and then being seen working with other ‘lowly’ folk. Other officials saw no straightfor- ward connection between caste and notions of stigma. ‘Natives’, they felt, were devoid of any sense of shame.44 Besides, reports on prisons like the one in 1838, point to a preference for muscular, outdoor over indoor work among male prisoners. Spinning and grinding flour were seen as feminine activities and therefore more demeaning.45 This was seen as undermining the deterrent value of extra-mural work. Once outside, masculinity was exercised and displayed in a variety of ways. At the worksite, it was often the weaker ones who were forced to do the strenuous tasks, while the stronger and macho bodies could bully supervisors and get away with less. Within a seemingly contradictory logic, masculine attributes were reaffirmed by demonstrating the ability to do the least muscular activity. The judicial critique of convict labour in India developed in the context of a larger conflict between civil and military authorities over questions of governance. The use of openly coercive methods conflicted with liberal ideas of rule and notions of legitimacy important to the colonial state in the nineteenth century. These tensions between the civil and military authorities recur in discussions within the Committee on Convict Labour. In the scathing indictment of the system by the Committee it is the liberal, reformist discourse which has an edge. The voice of military officials arguing for an intensification of control and surveillance is increasingly marginalized. By the 1850s the legitimacy of the system was corroding. The influence of liberal and utilitarian, arguments against the use of such forms of labour became more important. This was also a period marked by an expansion of the prison establishment, and a questioning of the idea of road work as a form of punishment among the prison authorities. As scarcities and famines recurred with untiring regularity from the 1870s, the supply of

44 The appendix to the Report of the Prison Discipline Committee includes written statements from different regions on the question of caste and punishment. Some officials were convinced that labour in irons was not considered ‘disgraceful’ by local people. See for instance evidence of the Magistrate of Farukhabad, Prison Discipline Committee Report, 223. 45 The Magistrate of Humeerpore noted: I do not think that labor on the roads in irons is felt as a greater disgrace than laboring in irons within the Jail, at spinning thread or grinding atta- perhaps not so much, for these occupations are looked down upon as fit only for women.’ Prison Discipline Committee Report, 224. public works and the question of unfree labour 285 cheap famine labour also expanded. Yet there remained a current of opinion within judicial discourse which favoured the employment of convicts on public works.46 By the second half of the nineteenth century the idea of public good that gave a moral sanction to the convict labour system was itself under question. Supporters of the convict labour system operated with two sets of assumptions: one, physical labour had a redemptive potential: it puri- fies the body and redeems the individual; two, labour on public works is doubly virtuous. It has a symbolic reciprocal function of giving back to the public; of paying for the economic and moral costs of criminality. By the late nineteenth century reformers were arguing against such assumptions.

Public Works and Coerced Labour

The colonial state was involved in distinct and different ways in the production and reproduction of different forms of coercive labour. In tea plantations in India state legislation formalized the penal powers of planters; it also sanctioned a wide range of ‘informal’ private powers of planters over labourers. The wider system of debt peonage was made possible by colonial legislations on contract. But state intervention in the working of these systems was indirect: it created the legal framework which allowed the allowed them to operate. In employing systems of tribute and convict labour, the state was more directly implicated in the working of a repressive regime of labour.47 State strategies of mobilizing tribute labour provided a basis for the intensification of begar in Hill states. The specific historical context in which the state came to play a more interventionist role in regulating and disciplining labour is important. State involvement in recruitment and disciplining strategies in the late eighteenth and nineteenth century India took place against recurring problems of creating and retaining a pliant labour force. In eighteenth century Madras for instance, militarist forms of control over labour emerged

46 The Report of the Prison Conference of 1877 was firmly of the opinion that the system of employment on public works was a valuable adjunct to the system of jail administration: ‘It affords a means by which over-crowding may be relieved and space made available for the construction of separate sleeping accommodation for the main body of prisoners. It is also an important extension of our available means for providing labour at once penal and profitable.’ Home Judicial Progs, October 1882, nos. 121–152 (NAI). 47 Behal and Mohapatra, “Tea and money”; Mohapatra, “Assam and the West Indies” in Hay and Craven (eds.) Magistrates, 455–80. 286 chitra joshi against a background of East India Company wars against regional powers and its struggles over scarce local labour resources. Ravi Ahuja elaborates how this situation was complicated by ritualized modes of collective with- drawal and desertion by labour prevalent in the region.48 Coercive labour strategies could act as a regulating mechanism, putting a downward pressure on the price of hired labour in the market. State intervention thus shows the processes through which the labour market was segmented. It was not just limits to labour mobility which created segmented pockets, but state regulation created a segmentation of the market. As with other coercive forms the use of convict labour allowed employers to regulate the labour market by pushing down prices. The very presence of convicts on construction sites, officials pointed out, could keep a check on local wage rates.49 In the late nineteenth century when the practice of extra-mural work for convicts was practically discontinued, many officials still argued for convicts on public works. The Sirhind Canal in Punjab in the 1880s was built largely with convict labour. The use of convict labour ensured a steady supply of labour which was not subject to harvest time fluctuations. It could be secured in the harshest of conditions; in times when hired labour was impossible to get. Note for instance the satisfaction with which an official employed on the Sirhind canal project reported: ‘On cold wet mornings it is practically impossible to collect a large body of free labourers at short notice . . . whereas a single message is sufficient to bring a jail full of prisoners at a moment of need on to an endangered bund’.50 On dangerous and difficult stretches of the Hindustan-Tibet road, convict labour was requisitioned.51 Tunnel building in terrain where blasting was not possible required steady and concentrated labour for which convicts were preferred.

48 Ahuja, “Labour Unsettled”; Ahuja, “Labour Relations”; Lucassen, “The Brickmakers’ Strikes”. 49 Note for instance a statement by Inspector General of Prisons, Lahore: ‘this value [of convict labour]is due, not only to the amount of work done by them, but also to the fact that their presence tends to prevent the rates of local labour from rising excessively, and that their numbers do not fluctuate with sowing and harvesting seasons.’ W.P. Dickson, Report on the Extra Mural Employment of Prisoners on the Sirhind Canal at Rupar, Punjab between 1869–81 (Lahore 1884), 11. 50 Sirhind Canal Report, p. 11. 51 In making a tunnel through hard rocky terrain, north of Simla, 10000 convicts were employed. Report on Hindustan Tibet road, 1 Oct. 1852, Foreign Department (Foreign Con- sultations), 22 December, 1852, no. 284 (NAI). public works and the question of unfree labour 287

The history of labour on public works brings out the close association between coercive labour regimes and public works. Large bodies of labour were required across vast spaces and over a sustained period of time. In the official logic, convict labour, begar and other forms of impressed labour could provide a compliant labour force which could do arduous tasks at lower than market prices. Yet there were limits within which such systems worked. It was not possible for Public Works authorities to draw entirely on tribute labour. There were regions which refused to comply with state demands. On difficult stretches of road other coercive labour strategies were deployed. The convict labour regime was one of the ways in which capital and the state try to deal with the problem of worker ‘insubordination’ but like all coercive forms it had a troubled existence. Problems of escape and mortality threatened its basis.52 We need then to look beyond binaries of freedom and unfreedom and unravel the tangled pasts that lie behind these categories. The margins of closed and ‘unfree’ regimes were continuously pushed and remade, and it was in these margins that possibilities of transcendence and breakdown also lay. The history of labour on public works in India brings out the tenuous links between modernity and freedom. Roads- symbolic of unfettered mobility—hide behind them the marks of fettering and unfreedom. I do not wish to end with the suggestion that these coercive forms were peculiar to colonial modernity. The specificities of colonial forms are often developed through unproblematic contrasts with idealized models of capitalism and modernity. The colonial appears in these constructs as an imperfect articulation, or deformation of the pure form which unfolds in the west.53 Recent writings suggest that even in the west the growth of the market and capitalism was not synonymous with ideas of freedom. Steinfeld and others powerfully demonstrate how relations between employers and workers in the ‘modern’ factory continued to operate outside formal languages of contract.54 Employers continued to use penal sanctions and various coercive forms to subordinate labour into the late nineteenth century. Regulations, often, formalised these coercive relations. We need to recognize that coercion and violence is integral to the history of capitalism, and that it is inseparable from the history of modernity.

52 For more on this see “Fettered Bodies”. 53 See for instance, Chakrabarty, Rethinking. 54 Steinfeld, Coercion; Reddy, Money.

References