contents i

Technology, Skills and the Pre-Modern Economy in the East and the West ii contents Global Economics History Series

Series Editors Maarten Prak, Utrecht University Jan Luiten van Zanden, Utrecht University

Editorial Board Gareth Austin, Graduate Institute of International and Development Studies, Geneva Şevket Pamuk, Boğaziçi University, Istanbul Kenneth L. Pomeranz, University of Chicago Tirthankar Roy, London School of Economics and Political Science Peer H.H. Vries, University of

VOLUME 10

The titles published in this series are listed at brill.com/gehs contents iii Technology, Skills and the Pre-Modern Economy in the East and the West

Essays dedicated to the memory of S.R. Epstein

Edited by Maarten Prak and Jan Luiten van Zanden

LEIDEN • BOSTON 2013 Cover illustration: painting in China. Porcelain was a very popular product in Europe, which led to variousiv attempts to imitate the qualitycontents of Chinese porcelain. Such attempts were ultimately successful in the 18th century. Source: Walter A. Staehelin, Das Buch vom Porzellan: Herstel- lung, Transport und Handel von Exportporzellan in China während des 18. Jahrhunderts, erläutert anhand einer zeitgenössischen Folge chinesischer Aquarelle (Berne: Benteli, 1965); reproduced with the pub- lisher’s permission.

Library of Congress Cataloging-in-Publication Data

Technology, skills and the pre-modern economy in the East and the West : essays dedicated to the memory of S.R. Epstein / edited by Maarten Prak and Jan Luiten van Zanden. pages cm. -- (Global economics history series ; volume 10) Includes bibliographical references and index. ISBN 978-90-04-24535-8 (hardback : alk. paper) -- ISBN 978-90-04-25157-1 (e-book) 1. Guilds--History. 2. Skilled labor--History. 3. Industrialization--History. 4. Technological innovations--History. 5. Economic history. I. Epstein, Stephan R., 1960-2007. II. Prak, Maarten Roy, 1955- III. Zanden, J. L. van

HD6456.T43 2013 330.9--dc23

2013009558

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 1872-5155 ISBN 978-90-04-24535-8 (hardback) ISBN 978-90-04-25157-1 (e-book)

Copyright 2013 by Koninklijke Brill NV, Leiden, The . 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 v

CONTENTS

Foreword by Patrick O’Brien...... vii Acknowledgements...... xiii About the Authors...... xv List of Tables, Figures and Illustrations...... xvii

Introduction: Technology, Skills and the Pre-modern Economy in the East and the West...... 1 Maarten Prak, Jan Luiten van Zanden

PART I REGIONAL PERSPECTIVES

1. Transferring Technical Knowledge and Innovating in Europe, c.1200 – c.1800...... 25 S.R. Epstein†

2. Apprenticeship and Industrialization in India, 1600-1930 . . . . . 69 Tirthankar Roy

3. Skills, ‘Guilds’, and Development: Asking Epstein’s Questions to East Asian Institutions...... 93 Kenneth Pomeranz

PART II INDUSTRY PERSPECTIVES

4. Mega-structures of the : The Construction of Religious Buildings in Europe and Asia, c.1000-1500...... 131 Maarten Prak

5. The Technology and Teaching of Shipbuilding, 1300-1800. . . . . 161 Richard W. Unger vi contents 6. Moving Machine-makers: Circulation of Knowledge on Machine- building in China and Europe between c. 1400 and the Early Nineteenth Century ...... 205 Karel Davids

7. Guilds and Apprenticeship in China and Europe: The Jingdezhen and European Ceramics Industries ...... 225 Christine Moll-Murata

8. Labour Relations, Efficiency and the Great Divergence: Comparing Pre-industrial Brick-making across Eurasia, 1500- 2000...... 259 Gijs Kessler and Jan Lucassen

9. Explaining the Global Distribution of Book Production before 1800...... 323 Jan Luiten van Zanden

Bibliography of the Published Works of S.R. Epstein...... 341 Index...... 345 foreword vii

FOREWORD

Our dear friend Professor Stephan (Larry) Epstein, then Head of the Department of Economic History at the London School of Economics, died without warning on 3 February 2007 at the age of 46. His field—the eco- nomic history of medieval and early modern Europe—demands more credentials to become recognized as a scholar than almost any other branch of the humanities. Thanks to schooling in Switzerland, higher edu- cation in Italy and postgraduate research at Cambridge, Larry almost ef- fortlessly acquired the palaeography, a full range of languages, the comprehension of separated but rich histories of several Western countries, a firm historiographical grasp of Smithian, Marxist, Weberian, Schumpet- erian meta-narratives, as well as the rigour acquired from critical reading in economics, sociology and, latterly, cognitive psychology to be much more than a mere expert. Indeed his peers were beginning to see that Larry Epstein would, in the fullness of time, take his place among a pan- theon of scholars that includes the likes of Henri Pirenne, Marc Bloch, Munia Postan, Frederick Lane and Fernand Braudel. As promise, that view emerged in early career in the form of cum laude degrees, fellowships from Trinity, Cambridge, a British Academy Readership, prizes from foundations in Switzerland, Italy and England, accolades from academic journals, as well as visiting professorships at New York University, the Dibner Institute at MIT, Harvard and the Wissenschaftskolleg in Berlin. Meanwhile his smooth transition to a chair at the London School of Economics occurred before his fortieth birthday. At the moment when his vulnerable body gave out, Larry Epstein had been publishing for a little over two decades, but had already written three books (two of seminal significance), 16 articles (in prestigious journals) and 26 essays in collections in the company of Europe’s most distinguished medieval and early modern historians. All that, and with so much more on the stocks and in his sights, validates the claim that Larry Epstein’s untimely death represented something of a calamity for scholarship in economic history and for our field’s potential for advances in economics, historical sociology and global history. Furthermore, and as those fortunate enough to be taught by him will testify, as a teacher Larry Epstein success- fully combined high expectations with a vigorously critical academic viii foreword engagement and an ability to inspire, relate to and communicate with his students.

My perusal of his historiographical essays suggests that the range of his intellectual interests was nothing less than catholic. Larry Epstein’s own writings are already acclaimed and likely to remain as exemplars for all social scientists and historians who engage with the core concern of our subject, namely ‘The Wealth and Poverty of Nations’ for generations to come. After several decades when the best minds in their discipline became fascinated with the mathematical theology of general equilibrium theory and the construction of macro production functions (prompted by the rediscovery of German knowledge for which Douglass North obtained a Nobel Prize), one of our subject’s absent parents, economics, now takes it as axiomatic that institutions are behind those proximate determinants of long run growth (land, labour, capital, technology and other inputs) as specified, re-specified, measured and re-measured in whole libraries of increasingly sophisticated exercises in production function analysis. Theo- ries formulated to include institutional variables fill the journals. Mecha- nisms have been elaborated and taxonomies imposed to expose their role in explanations of the glaring fact that after more than three centuries of modern economic growth, some countries are rich but most of the world’s population continues to live in societies that remain poor, and even, in the interiors of successful economies like China and India, subsist in condi- tions of poverty. Larry, along with the rest of his academic tribe, became delighted when economists and sociologists brought their mathematical skills, penchant for theory and knowledge of canonical writings in social science to bear on two tenets that have always remained foundational for the discipline of economic history; namely that institutions (notwithstanding the recent return to geographical reductionism) have always been ‘highly’ significant for long-run growth; and that their construction and destruction could only be explained as part of a process that has been fundamentally political, geopolitical and historical in origin and evolution. While he kept up with relevant theory from economics, sociology, an- thropology and cognitive psychology, Larry became more sceptical about the insights that those disciplines offered as to where, when, how and why history conditioned the formation of institutions that either promoted or restrained economic growth. As he saw it, the way forward resided in his- foreword ix toricizing particular contexts in which specific institutions functioned, which might (as his own work exemplifies) be local, national, continental or even global in scale; and then systematically comparing those histories of institutional development and their failures to develop over long spans of time. He remained entirely optimistic that economic historians, once reliable maps of relevant data were in place, could navigate the social ­sciences towards more grounded and plausible general explanations and theories for material progress. Furthermore, as several chastened speakers in his presence may recall, Larry emphatically rejected any suggestion that it was the mundane role of economic historians to simply test (and prefer- ably validate) whatever ideologically charged theory or anachronistic and Copernican-style hypotheses that other social sciences lobbed in their direction. For him the ontological realities and contingencies of time and place really mattered for histories of economic progress, stasis and regres- sion.

At the moment of his untimely death, his enviable credentials, extraordi- nary range of historical knowledge and capacity for rigorous analysis have left us with three programmes for future work in the subject he loved. These unfulfilled plans anticipated that historical perspectives on Europe’s expe- rience with the construction and development of institutions (pursued with respect to Marc Bloch’s admonitions for reciprocal comparisons) could reveal sustainable generalizations concerned with leads and lags in passages to modernity in the West. They could also include valid sugges- tions as to the nature of European exceptionalism in a recently revealed world of surprising institutional and economic resemblances that charac- terized the advanced maritime economies of Eurasia before the Great Divergence between occident and orient set in after the French Revolution. Chapters for a seminal narrative that Larry might well have been writing by now are already anticipated in two of his books and present in several articles emerging from a decade of fruitful involvement as the leading fig- ure in a European network for the historical study of human capital forma- tion. His views are spelled out with brevity and clarity as contestable generalizations in his essay, ‘The Rise of the West’, in honour of Michael Mann’s acclaimed contribution to the social theory of power.1

1 Stephan, R. Epstein, ‘The Rise of the West’, in John A. Hall and Ralph Schroeder (eds), An Anatomy of Power: The social theory of Michael Mann (Cambridge: Cambridge Univer- sity Press, 2006), 233-62. x foreword Indeed, almost all of Larry’s opus addressing both Smithian and Schum- peterian forms of economic growth across the regions and countries of pre-industrial Europe is suffused with a concern for the complex connex- ions between power and material change. His last book, Freedom and growth: The rise of states and markets in Europe 1300-1750 reveals that the historical record exposes no clear associations between constitutions for liberty with economic efficiency and growth.2 Furthermore, as we now appreciate, the economic thought, ideology and governmental policies of that period held no such expectations either.3 His first prize-winning book on Sicily and classic articles in Town and Country in Europe 1300-1800 warns us against urban bias, common to his- tory and economics, of representing cities as poles de croissance.4 Unless and until towns were forcibly integrated by powerful centralizing states into wider territorial units, they could also exploit their hinterlands, resist the extension and coordination of markets, erect barriers to trade or simply succumb to the predatory geopolitical forces that surround them.5 For some two decades before his death Larry had been at the centre of a network of economic and social historians concerned to investigate the political, legal, cultural and institutional conditions that led to the forma- tion of skills across the countries, regions and industries of Western Europe. Apart from a few fact-averse recidivists, historians now recognize that this wide-ranging and impressive programme of historical research has effec- tively degraded the view, long based upon little more that canonical polem- ics in the history of economic thought (and examples that are trivial), that the guild based regime for the accumulation of human capital in the West simply imposed rent-seeking upon an otherwise imminent emergence of competitive markets for manufactured commodities and labour.6 Across

2 Stephan, R. Epstein, Freedom and Growth: The rise of states and markets in Europe, 1300-1750 (London: Routlege, 2000). 3 Sophus Reinert, Translating Empire: Emulation and the origins of political economy (Cambridge: Harvard University Press, 2011), 1-71 4 Stephan, R. Epstein, An island for itself: Economic development in Late Medieval Sicily (Cambridge: Past and Present Publications, 1992) 5 Stephan, R. Epstein, Town and country in Europe (Cambridge: Cambridge University Press, 2001). Michael Lipton had issued similar warnings to economists as early as 1977: Michael Lipton, Why poor people stay poor: A study of urban bias (London: Temple Smith, 1977). 6 Stephan R. Epstein, with Gerhard Haupt, Carlo Poni, and Hugo Soly (eds.), Guilds, economy and society (Madrid, 1998) and Stephan R. Epstein, and Maarten Prak (eds), Guilds, innovation and the European Economy 1400-1800 (Cambridge: Cambridge University Press, 2007). foreword xi that Eurasian world of otherwise surprising resemblances, the hypothesis that Europe’s primary comparative advantage over regions to the East re- sided in an institutionalised and persistent and serious commitment to the formation of human capital begins to look increasingly plausible. The com- mitment by states, aristocracies, urban oligarchies and merchant elites to sustain guilds for the accumulation of skills and for the control of the qual- ity as well as the protection of their members turned out, on balance, to be functional not merely for the upgrading of urban workforces, but also for the discovery and diffusion of useful, reliable and ultimately systemic knowledge is now turning into the most probable and sustainable of all historical narratives constructed to account for economic divergence between the West and Rest.7

All three of the meta-narratives that engaged his mind are common prop- erty, but they were advanced into the realm of plausibility by a unique scholar and cosmopolitan intellectual. Alas, Larry Epstein will no longer be with us as economic historians trace its canons of Baconian evidence upon those rigorous, but all too parsimonious theories of institutional de- velopment that are simply without ontological foundation in European and Asian history.8

Patrick O’Brien London School of Economics

7 Joel Mokyr, The Enlightened economy: An economic history of Britain 1700-1880 (New Haven: Yale University Press, 2009). 8 Peer Vries, ‘Why the West Rules for Now: the patterns of history and what they reveal about the future’, Journal of Global History 7 (2012), 143-48. xii foreword

Contents CONTENTS v FOREWORD vii ACKNOWLEDGEMENTS xiii ABOUT THE AUTHORS xv LIST OF TABLES, FIGURES AND ILLUSTRATIONS xvii INTRODUCTION 1 TECHNOLOGY AND HUMAN CAPITAL FORMATION IN THE EAST AND WEST BEFORE THE INDUSTRIAL REVOLUTION 1 Maarten Prak and Jan Luiten van Zanden 1 Global History and Technology 2 Technology, Skills, and Science 5 Collective Invention 8 Technological Leadership and Clustering 11 Human Capital Formation 13 Circulation of Knowledge 15 Organisation of Labour 18 Comparative History and the Great Divergence 19 Conclusion 21 PART ONE 23 REGIONAL PERSPECTIVES 23 CHAPTER ONE 25 TRANSFERRING TECHNICAL KNOWLEDGE AND INNOVATING IN EUROPE, C.1200-C.1800 25 Stephan R. Epstein 25 Introduction 25 Acquiring Experiential Knowledge 28 Collective Knowledge and Technical Heuristics 35 Knowledge Sharing 35 Predictability, Codification and Innovation 42 Drawings and Models as Heuristic Devices 46 Experimentation 49 Spatial Transfer of Technical Knowledge 52 Texts and Patents 52 Transferring Skilled Technicians: When and Why Did Craft Guilds Oppose Technical Innovation? 55 Transferring Skilled Technicians: How Did It Work? 58 Conclusions 63 CHAPTER TWO 69 APPRENTICESHIP AND INDUSTRIALIZATION IN INDIA, 1600-1930 69 Tirthankar Roy 69 Craft Culture in Early Modern India 72 Craft Culture and Apprenticeship in Colonial India 77 Apprenticeship and Modern Industry 84 Conclusion 92 CHAPTER THREE 93 SKILLS, ‘GUILDS’, AND DEVELOPMENT: ASKING EPSTEIN’S QUESTIONS TO EAST ASIAN INSTITUTIONS 93 Kenneth Pomeranz 93 Introduction 93 The Rise of Guilds in Late Imperial China: Trade-based and Native Place-based Organisations 98 Guild Functions, Economic and Otherwise 105 City and Countryside 106 Skill Formation and Technological Creativity 108 Did Guilds Obstruct Technical Learning? 112 Migration and Skill Diffusion in More Recent Times 115 Skill Acquisition in Rural Industries 117 Japan 121 Conclusion 124 PART TWO 129 INDUSTRY PERSPECTIVES 129 CHAPTER FOUR 131 MEGA-STRUCTURES OF THE MIDDLE AGES: THE CONSTRUCTION OF RELIGIOUS BUILDINGS IN EUROPE AND ASIA, C.1000-1500* 131 Maarten Prak 131 Introduction 131 Church Building 134 The Workforce 138 Knowledge 142 Apprenticeship and Other Sources of Information 150 Conclusions 157 CHAPTER FIVE 161 THE TECHNOLOGY AND TEACHING OF SHIPBUILDING 1300-1800 161 Richard W. Unger 161 Introduction 161 Developments in Ship Design 164 East and West: Cross-fertilization of Techniques 176 Creation and Transfer of Knowledge 182 Transferring Knowledge with the Written Word 187 Disincentives for Technical Advance: Guilds and Market Structure 191 Incentives for Technical Advance: Guilds, Profits and Training 199 Conclusions 201 CHAPTER SIX 205 MOVING MACHINE-MAKERS: CIRCULATION OF KNOWLEDGE ON MACHINE-BUILDING IN CHINA AND EUROPE BETWEEN C. 1400 AND THE EARLY NINETEENTH CENTURY 205 Karel Davids 205 Introduction 205 Circulation of Knowledge on Machine-making in China, Late Fourteenth – Early Nineteenth Centuries 206 Circulation of Knowledge on Machine-making in Europe, Late Fourteenth – Early Nineteenth Centuries 214 Conclusion: Comparative Perspectives on the Circulation of Knowledge in China and Europe 221 CHAPTER SEVEN 225 GUILDS AND APPRENTICESHIP IN CHINA AND EUROPE: THE JINGDEZHEN AND EUROPEAN CERAMICS INDUSTRIES 225 Christine Moll-Murata 225 Introduction 225 Porcelain: Definitions and Early Exchange between China, the Middle East, and Europe 226 Jingdezhen: General Setting and Supply for the European Market 227 Number of Kilns and Size of Workforce 229 Total Output 231 Population Figures 231 Professional Training in the Industry 232 Alternatives to Guild-organised Apprenticeship 238 Delft: General Setting 240 Number of Enterprises and Size of Workforce 242 Guilds and Apprenticeship in Delft 245 Starting from Scratch: The Meissen and Vienna Porcelain Manufactures 249 Conclusion 253 Appendix 257 Jiangxi Province with Jingdezhen and Home Places of the Common-origin Guilds. 257 CHAPTER EIGHT 259 LABOUR RELATIONS, EFFICIENCY AND THE GREAT DIVERGENCE: COMPARING PRE-INDUSTRIAL BRICKMAKING ACROSS EURASIA, 1500-2000 259 Gijs Kessler and Jan Lucassen 259 The Historical Dynamics of Co-operation in Production 261 A Global, Comparative Perspective 264 Brickmaking in Western Europe from the Late Middle Ages until the Nineteenth Century 267 Division of Labour and Co-operation in Production 268 Production and Productivity 281 Russian Brickmaking between the Late Sixteenth and the Early Twentieth Centuries 287 Division of Labour and Co-operation in Production 290 Production and Productivity 302 Indian Brickmaking from c. 1800 307 Division of Labour and Co-operation in Production 309 Conclusion 319 CHAPTER NINE 323 EXPLAINING THE GLOBAL DISTRIBUTION OF BOOK PRODUCTION BEFORE 1800 323 Jan Luiten van Zanden 323 Introduction 323 Three Diffusion Processes 324 Quality 329 Font Size 330 Relative Prices 332 Size of the Market 335 Conclusion 338 Bibliography of the books and articles published by S.R. Epstein 341 INDEX 345 foreword xiii

ACKNOWLEDGEMENTS

The chapters in this book were first presented as papers in a conference at the London School of Economics, dedicated to the memory of S.R. (Larry) Epstein, who died tragically on February 3rd, 2007, aged 46. At the time of his death, Larry was Head of the Economic History Department at LSE, and working on a major research project concerning technology and techno- logical development in pre-modern Europe. It was therefore decided to combine this interest with another that he cared about deeply: the varia- tions in economic development between the regions of Eurasia. The confer- ence organisers challenged the contributors to write not only about their own region of expertise, but to also look at one or more other regions in Eurasia, and highlight this comparative perspective. The contents of this volume reflect that ambition. After three surveys, concentrating on Europe (Larry’s final unpublished work), India and China respectively, the second half of the book offers comparative discussions of a variety of industries.

The conference was held on 19th–21st June 2008, and was generously fund- ed by the London School of Economics, Suntory-Toyota International Centre for Economics and Related Disciplines (STICERD), The (British) Economic History Society and Utrecht University. It had been organized by the two editors, together with Patrick O’Brien and Patrick Wallis, both from the Economic History Department at LSE. Patrick Wallis has also supported the transformation of the papers into the chapters in this book in various ways.

For a variety of reasons papers delivered by Ted Collins, Philip Hoffman, John Langdon, Chan-Hui Mau, and Prasannan Parthasarathi have not been included in the volume. We want to acknowledge, however, their contribu- tion to the conference, as well as those of the conference commentators Maxine Berg, Chris Evans, Simona Valeriani, Oliver Volckart, and Patrick Wallis, and session chairs Gareth Austin, Kent Deng, Janet Hunter, Colin Lewis, Patrick O’Brien, and Chris Wickham. Finally, Sarah Carmichael, as assistant to the series editors, has been regularly reminding us that this was another project that had to be finished, and last but not least, Winny Bierman prepared the index. xiv foreword foreword xv

ABOUT THE AUTHORS

Karel Davids is professor of economic history at the Vrije Universiteit Amsterdam.

S.R. Epstein was professor of economic history at the London School of Economics, until his death in 2007.

Gijs Kessler is senior research fellow and head collection development Russia at the International Institute of Social History in Amsterdam.

Jan Lucassen is senior researcher at the International Institute of Social History in Amsterdam, and professor of social history at the Vrije Univer- siteit Amsterdam.

Christine Moll-Murata is research fellow at the Department of Sinol- ogy of the Eberhard Karls Universität Tübingen.

Patrick O’Brien is Centennial professor in economic history at the London School of Economics.

Kenneth Pomeranz is University professor of modern Chinese history at the University of Chicago.

Maarten Prak is professor of social and economic history at the Univer- siteit Utrecht.

Tirthankar Roy is reader in economic history at the London School of Economics.

Richard Unger is emeritus professor of economic history at the Univer- sity of British Columbia.

Jan Luiten van Zanden is faculty professor of global economic history at the Universiteit Utrecht. xvi foreword list of tables, figures and illustrations xvii

LIST OF TABLES, FIGURES AND ILLUSTRATIONS

Tables 2.1 Skill premium in the crafts...... 84 2.2 Comparative labour-capital ratios, 1931...... 86 2.3 Skill premium in the factories...... 87 7.1 Kilns and workforce of Jingdezhen...... 232 7.2 Taxed and registered population of Fuliang district...... 232 7.3 Fayence potteries and the workforce of Delft...... 244 7.4 Population of Delft, Delfshaven and its jurisdiction, 1600-1795. 244 7.5 Workforce at the Meissen Manufacture, 1765-1880...... 250 7.6 Workforce at the Meissen Manufacture, 1765-1880...... 251 8.1 Remuneration modes among brickmakers working in gangs, North-Western Europe c. 1600-1900...... 271 8.2 Workers blacklisted by the Leiden based Rhineland brick-­ makers guild, 1754-1818...... 279 8.3 Production capacities and labour productivity, Europe, c. 1600- 1900...... 282 8.4 Brick production in Moscow and St. Petersburg, 1670-1904. . 303 8.5 Production costs per 1000 bricks 1880 (Moscow) and 1904 (St. Petersburg)...... 307

Figures 1.1 Established craft guilds, Italy and Netherlands ,1100-1800. . . 34 1.2 Skill differentials in the European building industry, 1300-1799 (by city)...... 34 1.3 Integration of the skilled builders’ market 1400-1799...... 44 7.1 Jiangxi province with Jingdezhen and home places of the common-origin guilds ...... 257 8.1 Manual brickmaking - stages in the production process...... 265 8.2 Collective subcontracting units in brickmaking in Western Europe, Russia and India, 1670-2000 ...... 320 9.1 Diffusion curve of printing in China and Europe (total imprints per decade) ...... 327 xviii list of tables, figures and illustrations Illustrations 5.1 The sea-going junk in the Liu-Chhiu Kuo Chih Lüeh of about 1757...... 166 5.2 A Malaysian jong other Malaysian craft ...... 167 5.3 A sketch of a cog from the Venetian book on shipbuilding of the first half of the fifteenth century by Michael of Rhodes ...... 169 5.4 Fleet of Vasco da Gama ...... 172 5.5 Dutch fluyts at anchor ...... 177 5.6 A Kuwaiti baghla from the 1950s with a stern very similar to European sea-going sailing ships...... 178 5.7 & 5.8 Progress of ship construction in the past...... 194

8.1 Overview of Cathrinesminde Teglvaerk, a Danish brick factory at Iller Strand bordering at the Flensburger Förde. . . . 269 8.2 Gang of Lippe brick makers in Sunday dress in front of a roofed oven without chimney, Oberndorf an der Oste (Northern ), 1903...... 274 8.3 St. Petersburg 1904—handmoulder at work...... 289 8.4 St. Petersburg 1904—horse-drawn pug mill and drying sheds. 291 8.5 Mixing sand and clay in an iron pug mill at the Akra brick factories, India ...... 311 8.6 A moulder at work at the Akra brick factories, India ...... 315

Contents Figure 1.1. Established craft guilds, Italy and Netherlands 1100-1800. 34 Sources: Angelo Moioli and Jan Lucassen, per litteram. 34 Figure 1.2. Skill differentials in the European building industry, 1300-1799 (by city). 34 Sources: Building masters’ wages from Robert Allen’s website: http://www.economics.ox.ac.uk/Members/robert.allen/WagesPrices.htm. 34 Figure 1.3. Integration of the skilled builders’ market 1400-1799. 44 Source: See figure 2. 44 Table 2.1. Skill premium in the crafts (lower end of the wage scale = 100). 84

Table 2.2. Comparative capital-labour ratios, 1931. 86

Table 2.3. Skill premium in the factories (lower end of the wage scale = 100). 87

Illustration 5.1. The sea-going junk in the Liu-Chhiu Kuo Chih Lüeh of about 1757. Joseph Needham, Wang Ling and Lu Gwei-Djen. Science and civilisation in China. Volume 4, Part III, Civili Engineering and Nautics, Cambridge [Eng.]: University Press, 1971, p. 400. 166 Illustration 5.2. A Malaysian jong other Malaysian craft from Willem Lodewycksz., D'eerste boeck, Historie van Indien, waer inne verhaelt is de avontueren die de Hollandtsche schepen bejeghent zijn, Amsterdam, 1598. Reprinted in Eerste schipvaart der Nederlanders naar Oost-Indië onder Cornelis de Houtman, 1595-1597; journalen, documenten en andere bescheiden, ed. G.P. Rouffaer en J.W. IJzerman. ’s-Gravenhage, M. Nijhoff, 1915, page 132, plate 28. 167 Illustration 5.3. A sketch of a cog from the Venetian book on shipbuilding of the first half of the fifteenth century by Michael of Rhodes. The Book of Michael of Rhodes: A Fifteenth Century Maritime Manuscript, Volume 1, edited by David McGee, published by The MIT Press, fol. 182b. I am grateful to Alan Stahl for supplying me with the image. 169 Illustration 5.4. Fleet of Vasco da Gama, from the Livro de Lizuarte de Abreu, between 1558 and 1565, Pierpont Morgan Library, Manuscript M.525, fol. 18v-19. 172 Illustration 5.5. Dutch fluyts at anchor. Wenzel Hollar, print, 1647 Naues Mercatoriae Hollandicæ vulgo VLIETEN. 177 Illustration 5.6. A Kuwaiti baghla from the 1950s with a stern very similar to European sea-going sailing ships. Dionisius A. Agius, Classic ships of Islam: from Mesopotamia to the Indian Ocean. Leiden; Boston: Brill, 2008, page 315. 178 Illustrations 5.7 & 5.8. Progress of ship construction in the past from Nicolaes Witsen, Aeloude en hedendaegsche scheeps-bouw en bestier of 1671 but copied from Fernando Oliveira's Liuro da Fábrica das Naus of 1580. Leiden University Library, 678 A 15, plates XVIII and XIX. 194 Table 7.1. Kilns and workforce of Jingdezhe. 232

Table 7.2. Taxed and registered population of Fuliang district, of which Jingdezhen forms part. 232

Table 7.3. Fayence potteries and the workforce of Delft. 244

Table 7.4. Population of Delft, Delfshaven and its jurisdiction, 1600-1795. 244

Table 7.5. Workforce at the Meissen Manufacture, 1765-1880. 250

Table 7.6. Workforce at the Vienna Manufacture. 251

Source: http://www.maps-of-china.com/jiangxi-s-ow.shtml. 257 Note: Provincial borders should be ignored, as they were subject to change over time. 257 Figure 8.1. Manual brickmaking—stages in the production process. 265 Illustration 8.1. Overview of Cathrinesminde Teglvaerk, a Danish brick factory at Iller Strand bordering at the Flensburger Förde. Since the late 1840s brickmakers from Lippe were engaged in factories along this coast, characterized by horse-drawn clay pits and ovens under a roof without chimneys (note the smoke escaping through the tiled roof). 269 Source: Copy of an oil painting in a private collection, circa 1860, Museet på Sønderborg Slot (). 269 Table 8.1. Remuneration modes among brickmakers working in gangs, North-Western Europe c. 1600-1900. 271

Illustration 8.2. Gang of Lippe brick makers in Sunday dress in front of a roofed oven without chimney, Oberndorf an der Oste (Northern Germany), 1903. 274 Source: Photograph Landesarchiv NRW Abt. OWL, Detmold (Germany), D 75 Nr. 2670. 274 Table 8.2. Workers blacklisted by the Leiden based Rhineland brickmakers guild, 1754-1818. 279

Table 8.2. Cont. 280

Table 8.3. Production capacities and labour productivity, Europe c. 1600-1900. 282

Illustration 8.3. St. Petersburg 1904—handmoulder at work. 289 Source: T.F. Sanotskii, Kirpichnoe proizvodstvo na r. Neve i ee pritokakh (St. Petersburg: 289 A.E. Vineke, 1904), between pp. 44-45. 289 Illustration 8.4. St. Petersburg 1904—horse-drawn pug mill and drying sheds. Of the eight to nine workers who served a pug mill in operation, four can be seen at work in the pit on this photograph. Two of them are transporting the clay they have just dug out to the pug-mill on wheelbarrows, one is taking care of the mill and one is tending the horses. Of the other four men of the team, who supplied the moulders with clay, one can be seen at the top of the pit. 291 Source: T.F. Sanotskii, Kirpichnoe proizvodstvo na r. Neve i ee pritokakh (St. Petersburg: A.E. Vineke, 1904), between pp. 44-45. 291 Table 8.4a. Brick production in Moscow and St. Petersburg, 1670-1904. 303

Table 8.4b. Moulder productivity in Moscow and St. Petersburg, 1670-1904. 304

Table 8.5. Production costs per 1000 bricks 1880 (Moscow) and 1904 (St. Petersburg) 307

Illustration 8.5. At the Akra brick factories south of Kolkata, established in the mid-nineteenth century and still owned by the West Bengal Government, a gang of 35 men is engaged in transporting the river sand to the moulding pit, digging the clay, mixing sand and clay in an iron pug mill and delivering it to the moulders’ gangs. On this photograph, taken on February 21st, 1997, five gang members at the end of this chain are visible, plus a little boy keeping the oxen-drawn mill beam down. The gang, hailing from Murshidabad (250 kilometers north of Akra) receives a task wage. One of them, the primus inter pares called sardar collects the total wage sum from the subcontractor who has rented the factory from the Government. On average the gang members earn 1,300 rupees per month. To the right of the pit four women are waiting to carry the mixed clay to the moulders, as a rule their husbands. 311 Source: IISH, collection Jan Lucassen on Indian brickmaking, factory 10, photo 12-17. 311 Illustration 8.6. The moulder on this photograph is the husband of the woman in the orange dress on the previous picture. Both, possibly together with their daughter (left) form one of the 75 moulders’ groups who receive piece wages per unit. A moulders’ couple, migrant labourers from other parts of Bengal or Orissa, may earn, depending on its productivity, between 70 and 80 rupees per day. Living costs of adults are estimated at some 20 rupees per day. Besides clay mixers and moulders, the factory also employs 150 Biharis from Ranchi (400 kms) and also Patna (600 kms) and Chaibasa to carry dried bricks to the ovens. The carriers are exclusively women and children (both carry the bricks on their head) and boys using yokes, and they are paid individual piece rates. 315 Source: IISH, collection Jan Lucassen on Indian brickmaking, factory 10, photo 13-OA. 315 Figure 8.2. Collective subcontracting units in brickmaking in Western Europe, Russia and India, 1670-2000 (*Moscow 1880 = hand-machine moulding, the most widely used moulding technique at the time). 320 Figure 9.1. Diffusion curve of printing in China and Europe (total imprints per decade). 327 Sources: McDermott, ‘Ascendancy’, 58; Buringh and Van Zanden, ‘Charting the ‘Rise of the West”’. 327 introduction 1

INTRODUCTION

TECHNOLOGY AND HUMAN CAPITAL FORMATION IN THE EAST AND WEST BEFORE THE INDUSTRIAL REVOLUTION

Maarten Prak and Jan Luiten van Zanden

In the history of technology, as in much of the rest of human history, the Industrial Revolution stands as one of the great watersheds. As far as tech- nology is concerned, the transformation was from a situation of random effort to one of focussed effort. As Joel Mokyr, in his enormously helpful survey The lever of riches (1990) reminds us, pre-modern technology was primarily the result of ‘drift’, i.e. more or less accidental improvements that occurred to artisans in the course of their daily routines. From the eigh- teenth century onwards, on the other hand, a systematic search emerged for inventions, improvements—in other words for new technology that would make the production process more efficient.1 What has come to be known as ‘R&D’, research and development, was itself an invention of the nineteenth century, when firms started to set up laboratories. Thomas Edison’s Menlo Park, set up in 1876, was one of the first facilities for the pursuit of dedicated R&D activities,2 but the development of European porcelain in the first half of the eighteenth century, discussed in this book by Christine Moll-Murata, can be seen as an early example of the same process. Because pre-modern technology seemed such a random process, it has long been ignored by economic historians, or treated as a curiosity rather than a vital ingredient of their story. This has changed in recent years. A major reason for this change is that, from a comparative point of view, the ‘drift’ does not look so random after all. New technology—i.e. techniques that help push the production-function outward, or in non-economist par- lance: methods of production that give you more products for the same amount of inputs—did not appear just by coincidence, but originated in

1 Joel Mokyr, The lever of riches: Technological creativity and economic progress (Oxford: OUP, 1990), 12-13. 2 Randall E. Stross, The wizard of Menlo Park: How Thomas Alva Edison invented the modern world (New York: Crown, 2007); we owe this reference to Gerben Bakker (LSE). 2 maarten prak and jan luiten van zanden unusually large numbers in quite specific regions during specific periods of time. As a result, the balance between regions shifted.3 As Mokyr re- minds us, technologically speaking Europe was a backwater in 900 com- pared to China and the Middle East. It had made up for its deficit in 1500, and by 1750 some regions in Europe clearly outstripped all their rivals.4 The essays in this volume investigate the similarities and differences between various regions of Eurasia and try to find out what changed those balances. The first part consists of three regional surveys that address a similar set of questions for Europe, India and China. The second part of the book offers essays on specific industries—ship-building, porcelain, instrument-mak- ing, brick-production and the building industry—in which two, and some- times three or even four areas of Eurasia are compared in terms of their outputs and the knowledge, skills and technology used to create that out- put. The final chapter offers a comparative analysis of the availability of printed information in the major regions of Eurasia and hence of the pos- sibilities to store and distribute technical knowledge. The purpose of this Introduction is to set the individual chapters in the framework of a broad- er range of debates among historians of the pre-modern economy, pre- modern technology, and pre-modern labour relations.

Global History and Technology

For reasons that are so obviously related to recent developments in the world economy that it would be tedious to spell them out in greater detail, global history has become arguably the most exciting new development in the reconstruction of past human societies. Moving beyond the sometimes crude celebrations of Western superiority of previous generations of world historians, modern global history has nonetheless been forced to address the issue of differential development, and the possible path-dependencies that might explain such divergences between various regions.5 Much of this debate has focused on Eurasia, for the two simple reasons that initial indications seemed to suggest that around AD 1000 various regions of Eur- asia were functioning on roughly equivalent levels of economic develop- ment, and that at various times various regions of Eurasia seemed to be out-performing other regions around the world in terms of productivity,

3 Cf. Arnold Pacey, Technology in world civilization (Cambridge, MA: MIT Press, 1990). 4 Mokyr, The lever of riches, 57. 5 Patrick O’Brien, ‘Historiographical traditions and modern imperatives for the restora- tion of global history’, Journal of Global History 1 (2006), 3-39. introduction 3 innovation, and technological sophistication.6 These shifts of the centre of economic—and one must assume also social and political—progress have fatally undermined older arguments about the inherent superiority of Western societies, their cultures and religion, and at the same time fore- grounded the historical specificities of combinations of factors that al- lowed societies to push forward at some point in time, and the mechanisms that later slowed them down. Among economic historians this debate has revolved around the pro- posals of one single work, Kenneth Pomeranz’s Great Divergence, first pub- lished in 2000.7 In his book Pomeranz launched two very precise claims. The first was that the socio-economic differences between the most ad- vanced areas of China (the Lower Yangzi delta) and Europe (especially the British Isles) on the eve of the Industrial Revolution could not explain the ‘great divergence’ that emerged between them after 1800. This was the re- sult, according to Pomeranz, of two specific circumstances: coal and colo- nies. Together these provided the British economy with massive windfall energy resources that allowed the British economy to overcome the con- straints of the Smithian growth pattern that had previously been charac- teristic of economic development everywhere in Eurasia. In Pomeranz’s book technology was discussed at considerable length.8 Three claims were made. The first is that fundamental differences in over- all levels of technology between China and Europe will be difficult to sub- stantiate. The second is that Chinese technological progress was achieved particularly in agriculture, where impressive increases in productivity were achieved. And thirdly, Europe’s break-through invention, the steam-engine, depended on a fortuitous availability of accessible coal in Northern Eng- land. In an equally influential earlier work about China’s economic trajec- tory, Mark Elvin had suggested that Chinese technology was caught in what he called a ‘high equilibrium trap’. Its achievements were concentrated in agriculture, where output levels were such that very little progress could be made before the arrival of artificial fertiliser. In the meantime, produc- tivity levels in agriculture were difficult to match in other sectors, creating a disincentive for technological innovation.9

6 Gregory Clark, A farewell to alms: A brief economic history of the world (Princeton: Princeton University Press, 2008), 140. 7 Kenneth Pomeranz, The Great Divergence: China, Europe, and the Making of the Mod- ern World Economy (Princeton: Princeton University Press, 2000), 43-68. 8 A short introduction is provided by Daniel R. Headrick, Technology: A world history (Oxford: Oxford University Press, 2009) 9 Mark Elvin, The pattern of the Chinese past (London: Eyre Methuen, 1973), 298-315. 4 maarten prak and jan luiten van zanden Recent contributions to the debate, including Prasannan Parthasarathi’s book on India and the Great Divergence, have also addressed technology as a separate issue. Parthasarathi, too, claims that in general—and insofar as we will ever be able to measure this with any accuracy—the level of technology in India must have been in the same range as that in Europe. Contemporaries were astonished by the quality of Indian cotton cloth, and that cloth was competitive in markets more or less around the globe. In- dian shipbuilders had no problems building European-design ships, again suggesting that their levels of skill matched those of European shipwrights. Moreover, European ships did not push Indian—or for that matter other Asian-type—ships out of their home markets, another indication that in this industry at least the differences that no doubt did exist, were not fun- damental.10 Jean-Laurent and Bin Wong, who returned to Pomeranz’s initial riddle of China versus Europe, agree with Pomeranz and Par- thasarathi about technological parity in general, but mark out one par- ticular area where Europe was in a different position vis-à-vis its competitors. Due to its never-ending feuding, Europeans had built up a head-start in weapons technology. That was significant for two reasons. First of all, it allowed them to make inroads in non-European markets with the help of force rather than by economic means. But secondly, weapons technology made European gun smiths experts in the type of fine metal crafts that would eventually spill over into the technology necessary to build a steam-engine.11 By reframing the debate in this particular way, Rosenthal and Wong also highlight a perennial problem of the Great Divergence debate. Because what is it exactly that we are trying to explain? If it is a forging ahead of the European economies on a broad front, then clearly we have to be look- ing for clues in late medieval and early modern European societies that set them apart from their rivals.12 If, on the other hand, it is the arrival of the steam-engine that needs to be explained, a much narrower set of explana- tions seems in order.13

10 Prasannan Parthasarathi, Why Europe grew rich and Asia did not: Global economic divergence, 1600-1850 (Cambridge: Cambridge University Press, 2011), esp. ch. 7. 11 Jean-Laurent Rosenthal and R. Bin Wong, Before and beyond divergence: The politics of economic change in China and Europe (Cambridge: Harvard University Press, 2011). 12 Cf. Jan Luiten van Zanden, The Long Road to the Industrial Revolution. The European Economy in a Global Perspective 1000-1800 (Leiden: Brill, 2009). 13 See esp. Robert C. Allen, The British Industrial Revolution in global perspective (Cam- bridge: Cambridge Univerisity Press, 2009). introduction 5 Technology, Skills, and Science

One of the big issues confronting any history of technology is to identify the bottlenecks. Three can be distinguished. The first is knowledge itself. Even though knowledge will almost by definition always be imperfect and incomplete, there can be no doubt that significant progress in this area has been made in the course of human history.14 So was it the case that pre- modern societies simply lacked the requisite knowledge to raise their lev- els of technology? Or is the problem rather one of skills? It is quite likely that the great majority of artisans were left out of the academic debates that we usually have in mind when we talk about new knowledge. So per- haps the bottleneck was rather the training of the skilled workforce; where this was organised in an effective way, levels of technology would be pushed upwards. A third explanation focuses on the transfer of knowledge from the scientific to the practical realm. Even though important scientific breakthroughs had been made, these might never reach the producers. The argument here is, in other words, about application. Useful knowledge was lying around, but nobody knew what to do with it. These three interpreta- tions deserve to be further elaborated. S.R. Epstein’s contribution to the debate about technology in pre-mod- ern societies has been first and foremost an emphasis on the role of the anonymous craftsman. As is shown in his opening contribution to this volume, he underlined that crafts were an important—in his view probably the single most important—source of technological innovation. He did not deny that much of this innovation was quite un-spectacular. On the contrary, he insisted that the nature of technological progress in the Middle Ages and Early Modern period was incremental. But it made a huge differ- ence nonetheless. By 1700, on the eve of the great break-through inventions that would change Western industries, and ultimately the world in general, European technology had changed massively from what it had been 500 years earlier. Most of this change, Epstein argued, was the result of small improvements made by craftsmen, changes that in themselves were insig- nificant, but taken together were able to increase productivity and take European economies to unprecedented levels of development already be- fore the Industrial Revolution kicked in.

14 For the period covered by this book, see Mokyr, The lever of riches, ch. 4; also Peter Burke, A social history of knowledge: From Gutenberg to Diderot (Cambridge: Polity Press, 2000). 6 maarten prak and jan luiten van zanden Emphasising the role of craftsmen, who were active in massive numbers in all Eurasian societies, immediately raises the question, of course, what it was that set Europe apart from China, India, or the Near East, not to mention Japan or South-East Asia, all societies where at one point or an- other impressive feats of artisanal engineering had been performed. In his by now classic 1998 article for the Journal of Economic History, Epstein claimed that Europe’s advantage lay in the peculiarities of its guild system.15 As he argues more explicitly in the chapter published here, European guilds stood out amongst the competition because they had unusually low en- trance barriers, and recruited on a non-ascriptive basis. In China, for ex- ample, membership qualifications were usually a combination of technical expertise and regional origin. Access to skills was thus limited to a much narrower group of practitioners. Because competition was more limited, the pressures to innovate were less keenly felt, and the pool of new ideas would be anyway smaller. Whether these claims are indeed correct, is one of the questions guiding the chapters in this book. The most important defender in recent years of the position that knowl- edge precedes innovation has, no doubt, been Joel Mokyr. In his Gifts of Athena from 2002—that has become a classic in the same way as Pomer- anz’s work on the Great Divergence—Mokyr distinguishes between prop- ositional or Ω-knowledge, and prescriptive or λ-knowledge. The first provides an answer to ‘what’-questions, the second to ‘how’-questions. Progress in λ-knowledge must ultimately rest on Ω-knowledge, Mokyr claims.16 He has amplified this point of view in his most recent book, with the programmatic title The Enlightened economy (2009), where scientific developments and novel ideas are an important explanation for the Indus- trial Revolution. A similar point of view is defended in Patrick O’Brien’s recent survey article about the role of science in the Great Divergence. Rejecting such alternatives as cosmologies, institutions, career-incentives, or family structures as at this point unsubstantiated explanations, O’Brien focuses on the Scientific Revolution as a watershed in Europe’s regime of knowledge production.17 Like Mokyr, O’Brien sees the breakthrough of

15 S.R. Epstein, ‘Craft guilds, apprenticeship, and technological change in pre-industrial Europe’, Journal of Economic History 58 (1998), 684-713, reprinted in S.R. Epstein and Maarten Prak (eds), Guilds, innovation, and the European economy, 1400-1800 (Cambridge: CUP, 2008), 52-80. 16 Joel Mokyr, The gifts of Athena: Historical origins of the knowledge economy (Princeton: Princeton University Press, 2002), 4-5. 17 Patrick K. O’Brien, ‘The Needham question updated: A historiographical survey and elaboration’, History of Technology 29 (2009), 7-28. introduction 7 ‘modern’ science as the result of societal changes, like new attitudes among elites, in turn partly due to the Reformation, but also Europe’s confronta- tions with the non-Western world resulting from the voyages of reconnais- sance, all of which helped prevent new ideas being nipped in the bud.18 Nobody would of course want to dispute that science was a vital ingredient for the development of the steam-engine. The question is: can this claim be expanded to argue that Europe’s trajectory of technological innovation was science-based? A third type of the argument about the interactions between technol- ogy, skills and science has been emerging only recently. It focuses on the interplay between the worlds of science and crafts.19 Scholars like Lissa Roberts and Pamela Smith have been exploring the artisan dimension of the Scientific Revolution, in the process showing how the worlds of science and crafts intersected in much more complex—and intensive—ways than was previously imagined.20 A recently published volume explores this pro- cess by concentrating on materials, showing how metallurgy, ceramics, gunpowder, dye and ink production in the early modern period were all displaying this interaction of ‘learned’ and ‘craft’ knowledge to push the boundaries of quality and efficiency to new levels.21 Ursula Klein has high- lighted the role of laboratories as sites where such meetings took place; Simona Valeriani has pointed to the role of what she calls ‘in-between ob- jects’, like architectural models, where construction challenges were worked out through the combined inputs of scientists and builders. This literature considerably expands the social basis of knowledge production, and in ways that impact directly on the core issue for this book, because it

18 Patrick K. O’Brien, ‘Towards a global perspective on the emergence of a Western European regime for the discovery, development and diffusion of useful and reliable knowl- edge’, unpublished paper 2011; see also Joel Mokyr, ‘The Market for Ideas and the Origins of Economic Growth in Eighteenth Century Europe’, Tijdschrift voor Sociale en Economische Geschiedenis 4 (2007), 3-38. 19 This part has been inspired by conversations with Simona Valeriani from the URKEW team at LSE, and Ann-Sophie Lehmann from Utrecht University. 20 Lissa Roberts, Simon Schaffer, Peter Dear (eds), The mindful hand: Inquiry and inven- tion from the Late Renaissance to early industrialisation (Amsterdam: Edita, 2007); Pamela Smith, The body of the artisan: Art and experience in the Scientific Revolution (Chicago: University of Chicago Press, 2004). See also Ian Inkster, ‘Potentially global: “Useful and reliable knowledge” and material progress in Europe, 1474-1914’, The International History Review 28 (2006), 237-86. 21 Ursula Klein, E.C. Spray (eds), Materials and expertise in early modern Europe: Between market and laboratory (Chicago: University of Chicago Press, 2004), esp. the chapters by Pamela Smith, Hanna Rose Shell, Adrian Johns, Ursula Klein, Seymour Mauskopf and Agustí Nieto-Galan. 8 maarten prak and jan luiten van zanden pictures craft producers of objects as simultaneous producers of knowl- edge. Its deconstruction of craft practices has, however, not led to a new thesis about Europe’s specific trajectory. Indeed, as Dagmar Schäfer’s study of the works of Chinese scholar Song Yingxing (1587-1666?) demonstrates, this convergence of theoretical and practical knowledge was probably in itself not a uniquely European process. The Ming dynasty’s interest in crafts stimulated ambitious scholars to pay close attention to craft processes. Song’s Works of Heaven, first published in 1637, made a great effort to doc- ument craft production and the technology used in it. Detailed illustrations provided good working drawings for those who might want to build the equipment for themselves. At the same time, Song was keen to identify general patterns and broaden his observations to more general levels. In his methodology, Schäfer insists, he resembled Robert Boyle.22 The knowledge-dimension of the Great Divergence is currently being investigated by a group of scholars at the London School of Economics. Therefore the production of theoretical knowledge or its interactions with the artisan world are not covered by this volume that concentrates, in Epstein’s footsteps, on the skills angle.23

Collective Invention

An important implication of the ‘craft perspective’ of innovation is that it shifts the emphasis from individual to collective forms of invention. The term was coined by Robert Allen in 1983, in a description of blast furnaces built for the iron-smelting industry in the Cleveland district in Northern England during the nineteenth century. Allen was able to uncover an in- cremental process of improvements to the furnaces’ design, which created greater capacity while using significantly less fuel per unit of output. Fur- nace builders in the area freely shared the results of what was in essence a series of experiments.24 More examples from the era of the Industrial Revolution have been documented by Macleod (English clock and instru-

22 Dagmar Schäfer, The crafting of the 10,000 things: Knowledge and technology in seven- teenth-century China (Chicago: University of Chicago Press, 2011), 10-11, 94-108, 129-30. 23 The project, funded by the European Research Council under the title ‘Useful and Reliable Knowledge in East and West’, or URKEW for short, is led by prof. Patrick O’Brien, with contributions from researchers dr. Mina Ishizu, dr. Khodadad Rezakhani, dr. Anjana Singh, dr. Simona Valeriani, and dr. Ting Xu. See http://www2.lse.ac.uk/economicHistory/ Research/URKEW/aboutUrkew.aspx (accessed December 2011). 24 R.C. Allen , ‘Collective invention’, Journal of Economic Behaviour and Organization 4 (1983), 1-24 introduction 9 ment makers), and Nuvolari (Cornish steam-engine), and the topic has been attracting increased attention because of parallels with the develop- ment of open-source software.25 For the pre-industrial period Pérez has charted the targeted innovation by the Lyon silk-industry, which had a deliberate policy of stimulating innovations, and making them available to all its members.26 Similarly, the related issues of secrecy and openness have been ques- tioned, especially by historians of science. They argue that the two catego- ries are usually not so clear-cut in practice, and that whereas access to information may have been restricted in some media, that same informa- tion could be available in others. ‘Something can be open or secret to different extents’.27 During the period covered by this book, two contradic- tory trends emerged in Europe. The creation of craft guilds led to greater claims of intellectual property rights, by individuals as well as guilds them- selves. A famous example is the Venetian glass industry, but that also dem- onstrates two important caveats to the issue of secrecy. Attempts to limit access to craft knowledge were not individualised, but ‘communal’. The Venetian authorities encouraged the free flow of information within the craft, but attempted to prevent it from moving to other places. These at- tempts, however, were bound to fail, because of the numbers of temporary migrants in Venice who would be able to access the trade, but also because other communities offered Venetian glass-makers very attractive condi- tions of settlement. As a result, Venetian experts set up shop elsewhere, first in other Italian towns, later across Europe.28 The ‘mystery’ of the craft was, in other words, not necessarily some sort of a closely guarded secret, but rather the tacit knowledge shared between masters as well as with

25 C. Macleod, Inventing the Industrial Revolution: The English patent system, 1660-1800 (Cambridge: CUP, 1988), 113; Alessandro Nuvolari, ‘Collective invention during the British industrial revolution: The case of the Cornish pumping engine’, Cambridge Journal of Eco- nomics 28 (2004), 347-63; James Bessen, Alessandro Nuvolari, ‘Knowledge sharing among inventors: Some historical perspectives’, LEM Working Paper October 2011 26 Liliane Pérez, ‘Inventing in a World of guilds: Silk fabrics in eighteenth-century Lyon’, in: Epstein and Prak (eds), Guilds, 232-63. 27 Koen Vermeir, ‘Openness versus secrecy? Historical and historiographical remarks’, British Journal for the History of Science 45 (2012), citation from the pre-publication version, p. 6. 28 Pamela O. Long, Openness, secrecy, authorship (Baltimore: Johns Hopkins University Press, 2001), 90-95; also Francesca Trivellato, Fondamenta dei vetrai: Lavoro, tecnologia e mercato a Venzia tra Sei e Settecento (Rome: Donzelli, 2000), ch. 8; and for the Dutch Repub- lic Karel Davids, The rise and decline of Dutch technological leadership: Technology, economy and culture in the Netherlands, 1350-1800 (Leiden: Brill, 2008), 388-416. 10 maarten prak and jan luiten van zanden apprentices.29 Moreover, craft knowledge became increasingly available in manuscript and later printed manuals, especially from the fifteenth cen- tury.30 The type of incremental innovations typical for many pre-modern in- dustries, combined with the formal and informal organisation of crafts, must have made collective innovation and invention the norm rather than the exception during this period. Take for example the ship-building in- dustry, discussed in Unger’s contribution. In Europe, shipbuilders were responsible for at least one macro-invention, the Dutch flute (fluyt) type of ship that dominated world trade during much of the seventeenth cen- tury and was a major contributing factor in the expansion of global trade during that era. Rather than being the product of a stroke of genius by a single inventor, the flute was the result of many piecemeal improvements made by many different shipyards along the North Sea coast during the preceding centuries. Typically, there is much speculation that the first flute was built by a shipyard in 1595 in the town of Hoorn, but there is no cer- tainty about either place or date, let alone a single name connected to this momentous event.31 As Unger reiterates in his chapter, there were no pat- ents or secrets in shipbuilding that protected and rewarded improvements. New ideas were quickly imitated, and cultural barriers counted for little in the industry in the sense that Indian shipbuilders, for example, were quite comfortable with producing European-style ships when the demand for them emerged. The experimental side of this collective process is highlighted by Prak’s contribution to this volume, where he discusses what was happening on cathedral building sites in the Middle Ages. The size and complexity of European cathedrals increased dramatically with the introduction of the Gothic style, and much of this was due to on-site experiments as long as the mathematical models to predict the behaviour of large structures were not available, which only happened in the nineteenth century. Therefore, even the building of St. Paul’s Cathedral in London as late as the second half of the seventeenth century, designed by one of the best scientists in the world of his day, still took place as an experiment. In the case of St. Paul’s we know exactly who designed it, and there is massive documenta-

29 Lon R. Shelby, ‘The “secret” of the medieval mason’, in: Bert S. Hall and Delno C. West (eds), Pre-modern technology and society: Studies in honor of Lynn White, Jr. (Malibu Ca.: Undena, 1976), 201-19; cf. also S.R. Epstein, Maarten Prak, ‘Introduction: Guilds, innovation and the European economy, 1400-1800’, in: Epstein and Prak (eds), Guilds, 14. 30 Long, Openness, ch’s 6 and 7. 31 Richard W. Unger, Dutch shipbuilding before 1800 (Assen: Van Gorcum, 1978). introduction 11 tion of Christopher Wren’s contribution to the process, but for most other landmark churches of similar size this is not the case. That, Prak argues, is not so much the result of a loss of information, but of the building process itself. It took in most cases many generations to build a cathedral and sig- nificant design changes were made during the process. Outside experts might be brought in to give advice or comment on certain aspects of the plan. The building of these large structures, in other words, was a collective process, that could not be reduced to the essential input of one single in- dividual. There can be no doubt that the absence of individual incentives and rewards, as well as the high entry barriers and low levels of protection of- fered by pre-modern patenting systems, significantly slowed down the pace of technological progress. That progress was very real nonetheless, can be seen from the long list of major innovations produced during this period.32 The tower mill, which allowed the operator to turn merely the top, rather than the whole mill-structure towards the wind, was developed by Italian and Dutch millwrights in the sixteenth century. It allowed the development of a major industrial concentration along the river Zaan, to the north of Amsterdam. It was in this same area that the ‘Hollander’ was developed, to beat rags into pulp for the paper industry, much more efficiently than previous methods. Italians massively improved the efficiency of blast fur- naces, and significantly improved the quality of hydraulic engineering. The originally hand-operated spinning wheel became foot-operated. It is im- possible to ascribe the invention of this, or any of the other examples, to an individual, or even a specific date. This was not the result of incomplete sources, but of the fact that they emerged slowly from a large community of practitioners working together on the improvement of their techniques and products.

Technological Leadership and Clustering

One important feature of this collective dimension of technological prog- ress was—and in fact still is—the regional clustering of advanced technol- ogy. Epstein underlined the significance of the fact that in Europe technological leadership shifted, from Italy to the Low Countries, and sub- sequently to England, in the course of the early modern period. These

32 The examples come from Mokyr, Lever of riches, ch. 4. 12 maarten prak and jan luiten van zanden shifts, which were first analysed by Karel Davids in 1995,33 were seen by Epstein as proof of the importance he himself attached to the circulation of knowledge and competition between European states for craft technol- ogy. As, for example, Jan de Vries pointed out, the problem of growth in an essentially Smithian economy is the dead-weight of its more traditional sectors. In pre-modern Europe, above-average levels of economic growth were achieved in relatively small regions—and even these had problems sustaining the dynamic.34 But while it lasted, spill-overs from one sector to another were clearly part of the momentum.35 Clusters, or agglomera- tion economies, have become very influential as a concept among eco- nomic geographers to understand patterns of innovation since it was first articulated by Michael Porter in 1990.36 Economic historians have so far been reluctant to integrate clusters as an explicit category into their analy- ses.37 Clusters were defined by Porter as ‘geographic concentrations of interconnected companies, specialized suppliers, service providers, firms in related industries, and associated institutions (e.g. universities, stan- dards agencies, trade associations) in a particular field that compete but also cooperate’.38 Porter identified four areas where the geographical con- centration of certain industries might benefit those industries: spatial con- centration would have a positive influence on quantities and costs of inputs, provide a habitat for specialised suppliers and supporting indus-

33 Karel Davids, ‘Shifts of technological leadership in early modern Europe’, in Karel Davids, Jan Lucassen (eds), A miracle mirrored: The Dutch Republic in European perspective (Cambridge: Cambridge University Press, 1995), 338-66. 34 Jan de Vries, ‘Economic growth before and after the Industrial Revolution: A modest proposal’, in: Maarten Prak (ed.), Early modern capitalism: Economic and social change in Europe, 1400-1800 (London: Routledge, 2001), 177-94. Sidney Pollard had already emphasised the regional nature of the Industrial Revolution itself in his Peaceful conquest: The industri- alization of Europe, 1760-1970 (Oxford: Oxford University Press, 1981). See also Pat Hudson (ed.), Regions and industries: A perspective on the Industrial Revolution in Britain (Cambridge: Cambridge University Press, 1989). 35 This concentration of technological change does not show up in the population density approach that Greg Clark applied to whole continents; his figures therefore seriously underestimate the impact of technological change: Clark, A farewell to alms, 140. 36 Michael E. Porter, The competitive advantage of nations (Basingstoke: Macmillan, 1998). 37 An exception is Patrick O’Brien, ‘Reflections and mediations on Antwerp, Amsterdam and London in their golden ages’, in: Patrick O’Brien, Derek Keene, Marjolein ’t Hart, Her- man Van der Wee (eds), Urban achievement in early modern Europe: Golden Ages in Antwerp, Amsterdam and London (Cambridge: Cambridge University Press, 2001). 38 Michael E. Porter, ‘Location, competition, and economic development in a global economy’, Economic Development Quarterly 14 (2000), 15. introduction 13 tries, sustain a sophisticated consumer demand, and create an environ- ment conducive to appropriate strategies. In terms of innovation, clusters would allow firms to capture the benefits of a specialised workforce, knowl- edge spill-overs, and last but not least, acute competitive pressures.39 Sili- con Valley is probably the best example of the advantages of clustering in the modern world. Northern Italy, the seaboard provinces of the Low Coun- tries, and later the British Midlands are all obvious examples from an ear- lier age. One important incentive for clustered innovations to emerge, has been the level of real wages in a particular region. Famously, the rise of Dutch windmill sawing in the seventeenth century coincided with the soaring wages of the hand-sawyers who had traditionally prepared the elements for the construction of ships and buildings. The economic boom that start- ed in the western provinces of the newly founded Dutch Republic in the 1590s, created a huge demand for both ships and buildings, driving up saw- yers’ wages in the process. The saw-mill not only halted that increase, but ultimately destroyed hand-sawing as a line of trade altogether.40 Instead, the new technology helped create a new industrial district on the river Zaan.41 In Robert Allen’s recent interpretation of the Industrial Revolution, the rise of spinning technology in Britain was a similar reaction to high levels of real wages.42 The combination of wage levels and cluster effects seems to offer a potent explanation for the emergence of, and shifts in technological leadership.43

Human Capital Formation

European guilds recruited their apprentices and members usually outside the circles of immediate relatives. From Caen, in France, a figure of 45 per cent relatives is reported, but in Dijon it was less than 20, and in Paris

39 Ibid., 23-24. 40 Jan de Vries and Ad van der Woude, The First modern economy: Success, failure and perseverance of the Dutch economy, 1500-1815 (Cambridge: Cambridge University Press, 1997), 301, 633-34. 41 Davids, Rise and decline, 394-98. 42 Allen, British Industrial Revolution, ch. 8 and passim; see also the same author’s Global economic history: A very short introduction (Oxford: Oxford University Press, 2011) for a global application of this perspective. 43 Michael Porter, ‘Why does a city grow? Specialisation, human capital or institutions?’, Urban Studies 47 (2010), 2027-50. 14 maarten prak and jan luiten van zanden roughly one third.44 In the Netherlands one third was the maximum, and often the numbers were substantially lower.45 In sixteenth-century London the highest percentage of masters’ sons admitted to a sample of sixteen different guilds was thirteen; in most companies the percentage was well below ten.46 In India, as Tirthankar Roy reports in his contribution to this book, crafts recruited through the family. European travellers repeatedly remarked on this in their accounts of the region, and portrayed it as differ- ent from their own experience. Roy at the same time insists that this system was not at all watertight; in other words, Indian crafts—the word guild remains problematic in the Indian context—were probably more open to outsiders than standard images of the caste-system suggest. Despite a high degree of informality that was also characteristic of the Indian situation, apprenticeships were long and stable, which seems to imply that even without the formal arrangements created by guilds in Europe, the Indian system in fact worked quite well. This might also be deduced from some of the outcomes of training in India, as has been argued by Parthasarathi in his recent contribution to the Great Divergence debate. Numerous contemporary commentators testified to the high levels of accomplishment of Indian workers.47 The impressive quality of Indian textiles could not have been achieved without matching levels of skill. The same applies to the Chinese porcelain industry discussed in Christine Moll-Murata’s chapter for this book. Porcelain became a very popular product in seventeenth-century Europe, and was widely imitated by European producers. Perhaps its best-known European equivalent in the seventeenth century was the ‘Delft blue’ produced in Holland, which emerged as a substitute during the Ming-Qing transition, when Chinese porcelain exports came to a virtual standstill. However, it took European manufacturers almost a century of frantic efforts before they were able to

44 Michael Sonenscher, Work and wages: Natural law, politics and the eighteenth-century French trades (Cambridge: Cambridge University Press, 1989), 107-08; Clare Haru Crowston, Fabricating women: The seamstresses of Old Regime France, 1675-1791 (Durham: Duke Uni- versity Press, 2001), 335. 45 John Michael Montias, Artists and artisans in Delft: A socio-economic study of the seventeenth century (Princeton NJ: Princeton Univerisity Press, 1982), 152; Bibi Panhuysen, Maatwerk: Kleermakers, naaisters, oudkleerkopers en de gilden (1500-1800) (Amsterdam: IISG, 2000), 169-70; Maarten Prak, Republikeinse veelheid, democratische enkelvoud: Sociale veran- dering in het Revolutietijdvak, ’s-Hertogenbosch 1770-1820 (Nijmegen: SUN, 1999), 99-100. 46 Steve Rappaport, Worlds within worlds: Structures of life in sixteenth-century London (Cambridge: Cambridge University Press, 1989), 293. 47 Parthasarathi, Why Europe grew rich and Asia did not, 217-19. introduction 15 produce ‘the real thing’, i.e. a product with the same refined quality as had been coming from China. Chinese producers in Jingdezhen lacked formal training and they received, moreover, few stimuli from the merchants who were exporting to the rest of the world. Whereas in Delft one local manu- facturer produced a detailed manual for the industry in the late eighteenth century, an equivalent document has not been found for Jingdezhen. Still, the Delft product never managed to have the same impact in global markets as its Chinese competitor. Moll-Murata’s observations are confirmed in Karel Davids’ chapter deal- ing with the machine-makers in Europe and China. In this industry, that was to become crucial during the Industrial Revolution, the internal or- ganisation was less formalised in China. Exams set by European guilds defined levels of expertise for those aspiring to work as a machine-maker. Printed information was more readily available to European producers and circulated more easily in the West. Ken Pomeranz in his contribution expands Davids’ and Moll-Murata’s observations to a broader scale, and in doing so also confirms Epstein’s intuitions about Chinese guilds. Pomeranz identifies two specific areas where Chinese guilds may have been less effective than those in Europe, next to the regional affiliations that restricted access to the craft. The first is the absence of formal training mechanisms in Chinese guilds. They clearly did provide apprenticeship, but whereas these were regulated in Europe, no such thing seems to have happened in China. In the absence of regulation, the likelihood is that training was provided by relatives, and hence a relatively narrow group of experts, instead of the much wider train- ing opportunities provided by many European guilds. Secondly, in the ab- sence of ‘monopolies’, the temporary advantages allowed to European innovators were probably missing in China.

Circulation of Knowledge

In Epstein’s view of pre-modern technology, knowledge was very much an embodied phenomenon and the circulation of knowledge therefore an issue of labour migration. He claimed that European guilds, by stimulating and supporting the movement of labour through various labour markets, especially for young practitioners, allowed new ideas not only to circulate, but at the same time to form new combinations with best practices avail- 16 maarten prak and jan luiten van zanden able in a variety of production centres.48 Plausible as it sounds, increasing evidence is emerging that labour migration was equally common in other parts of Eurasia. Indeed, much of the Chinese guild system was based on the migrants’ place of origin, and the so-called native-place guilds were as prominent as purely professional forms of organisation. Francesca Bray has argued that in Ming China technological progress resulted, not so much from inventions, but from the wider application of existing techniques.49 Migration seems to have been an important element in that diffusion pro- cess. Several essays in this book highlight the fact that Europe was different from the other regions because of the availability of many practical, ‘how- to’ type of manuals that reduced search costs for production procedures, and at the same time set standards for the industry that could act as points of departure for improvements. Moll-Murata, for example, shows how the Delftware industry, even though it was already in decline, produced a hands-on description of the production process, written by somebody with first-hand experience of the sector. In China, on the other hand, descrip- tions of the Jingdezhen porcelain works were clearly not written with the producers themselves in mind, and lacked precisely the type of information that somebody trying to produce porcelain would require to help him or her through the various stages of the production process. Davids makes a similar observation for the tools and machine industries, where codified knowledge was much more prevalent in Europe than in China. This observation is set in a broader context by Van Zanden’s contribution about the book industry. This chapter is somewhat different from the oth- ers in that it does not investigate the quality of the workforce as such, but concentrates on the technology and its application. In his contribution Van Zanden makes two major points. The first is a quantitative confirmation of the dramatic changes that took place in the European book industry after the introduction of moveable type from the middle of the fifteenth century, compared to the availability of information in the Middle East and China. Problems of transforming Arabic script and Chinese characters into standardised type clearly worked against book producers in these areas

48 S.R. Epstein, ‘Labour mobility, journeyman organisations and markets in skilled labour Europe, 14th-18th centuries’, in: M. Arnoux and P. Monnet (eds), Le technicien dans la cité en Europe occidentale 1250-1650 (Rome: Ecole française de Rome: 2004), 251-69; Liliane Hilaire-Pérez and Catherine Vena, ‘Dissemination of technical knowledge in the Middle Ages and the early modern era’, Technology and culture 47 (2006), 536-65. 49 Francesca Bray, Technology and society in Ming China (1368-1644) (Washington DC: American Historical Association, 2000), 65. introduction 17 who might be willing to copy the European invention. But, as Van Zanden can demonstrate, the shift from written to printed reproduction of texts was primarily a supply-driven process, and mechanisation in Europe can be explained as the outcome of the cost of labour. The knock-on effect of the introduction of moveable type was, however, a dramatic reduction in the cost of information. The emergence of craft manuals in Europe in the century after the invention of moveable type was merely one dimension of the information revolution that was sweeping through Europe at the time, a revolution that was affecting other domains, most notably religion, in perhaps even more dramatic ways. There can, however, be no denying that practitioners and experts in crafts as much as in medicine or naviga- tion all of a sudden had access to a vast range of information that in oral form would just have been very time-consuming and therefore expensive to collect.50 The question remains what the implications were. It seems quite plau- sible to relate this cheap access to useful and reliable knowledge, both of the practical and the theoretical varieties, to the breakthrough to steam- power and the Industrial Revolution—and some very convincing examples have been produced in recent years.51 Davids adds more weight to such examples by his systematic discussion of the instrument and machine manufacturers in China and Europe, and finds much evidence for an im- portant role of printed documents in the development of the European industry. At the same time, several of the essays in this volume continue to raise questions about this connection. How could the illiterate porcelain makers in Jingdezhen continue to out-perform their European competitors well into the eighteenth century in terms of both quantity and quality, as is shown in Moll-Murata’s chapter? Likewise shipbuilding, as discussed by Unger, remained stubbornly regional when it came to design. Chinese junks, Indian dhows, and European fluyts were happily plying the same Asian waters side by side, even though shipbuilders had access to all three types, and were in some cases building ships of the various models in one and the same shipyard on the Indian coast. Elements of the designs crossed over from one type to the other, with possibly the Europeans as the most active copyists, but the three types remained fundamentally unchanged, not least, Unger claims, because they were finely balanced in quality. Obvi- ously, it was European ships that went to Asia, and not the other way

50 This seems to be underestimated in Bray, Technology and society in Ming China, 16-17. 51 The best example is Margaret Jacob, Scientific culture and the making of the industrial West (New York: Oxford University Press, 1997). 18 maarten prak and jan luiten van zanden around. The quality of the ships themselves was, however, not the reason for this imbalance of trade relations. The absence of a convergence in ship designs therefore remains a problem for arguments that portray European technical knowledge as either inherently superior, or as merely cheaper to access and therefore more easily disseminated.

Organisation of Labour

One of the issues that has not been taken up in the debate about pre- modern technology so far, is the status and organisation of labour. Some regions of Europe had already during the late Middle Ages reached levels of wage labour of fifty per cent and more.52 While much of the debate on guilds focuses on people who could hope to become an independent mas- ter one day, such levels of proletarianisation raise questions about its im- pact on human capital formation and skills training. Gregory Clark has recently identified poor coordination as one of the main reasons for low levels of labour productivity outside Europe.53 Gijs Kessler and Jan Lucas- sen address this issue in their wide-ranging investigation of the brick in- dustry in Western Europe, Russia and India. They find that labour regimes were by and large the same in all three areas, and consisted mainly of self- regulation by the work-units themselves. Remuneration was organised in such ways that groups of workers were forced to collaborate and together achieve the best possible results. Knowing how to collaborate to maximum effect, they observe, was therefore a vital skill in its own right. This skill had to be learned on the job, and was as tacit in character as any other aspect of the work. In Europe teams consisted normally of non-kin, in India kin played a more significant role. Guilds in the brick industry—insofar as they existed—were not organisations of the workforce, but of the employers. Their role was to supervise and control a fordist type of industry, where work was organised on a production line and output volumes were all- important.54 Such industries were to become much more common in the industrial period, but already in the economies covered by this book, large-

52 Bas van Bavel, ‘Rural wage labour in the 16th-century Low Countries: an assessment of the importance and nature of wage labour in the countryside of Holland, Guelders and Flanders’, Continuity & Change 21 (2006), 62. 53 Clark, A farewell to alms, ch. 17. 54 They were thus similar to the type of organisation described in Sheilagh Ogilvie, State corporatism and proto-industry: The Württemberg Black Forest, 1585-1797 (Cambridge: Cam- bridge University Press, 1997). introduction 19 scale production was emerging in a range of industries, including ship- building and public construction. Some of these large-scale production sites had been known in earlier periods and in other places. Was it forced labour that made those in charge indifferent to questions of efficiency? Tantalisingly, Moll-Murata’s contribution on Chinese porcelain seems to suggest that self-regulated teams were the norm in the Royal Manufactures. A fascinating comparative history of collaboration in the work place seems to be waiting to be written.

Comparative History and the Great Divergence

Global history has two major challenges: to pin down a proper objective, and to choose adequate units of analysis. To point up global interactions, or identify similarities and differences between societies in various regions, is ultimately no more that repeating the obvious. To move beyond that point, and make sense of such interactions, similarities and differences, we have to know what their implications were. As comparative social scientists have discovered quite some time ago, it helps when there is something that can be explained as the outcome—or dependent variable—of such pro- cesses.55 Economic historians have been able to borrow such an outcome from the economists, and focus on National Income, Gross Domestic Prod- uct, and similar indicators of output and, by implication, prosperity—all preferably measured in per capita terms. Although we are still in the he- roic stage of reconstructing the basic numbers, economic historians of the pre-modern period have been making steady progress in the production of data that are reliable and verifiable, and therefore acceptable to the scholarly community as indicators of societies’ progress along the trajec- tory of growth and prosperity. The debate among economic historians was galvanised by the publica- tion of Pomeranz’s book on the Great Divergence, precisely because he outlined a chronology of the trajectories of Europe and China respectively, while at the same time proposing a set of theses that explained those tra- jectories. Pomeranz still had to work without the benefit of quantitative data, beyond the initial figures proposed by Angus Maddison, which be- come increasingly unreliable as one moves away from the nineteenth-

55 Charles Ragin, The Comparative Method: Moving Beyond Qualitative and Quantitative Strategies (Berkeley: University of California Press, 1987). 20 maarten prak and jan luiten van zanden century and essentially Western core of his dataset.56 To reduce the problems of missing data and multiple causation, Pomeranz sensibly fo- cused on two specific regions: England in Europe and the Lower Yangzi delta in China, as the most economically advanced in their respective environments. Other works have followed in Pomeranz’s footsteps, most notably Timur Kuran’s book on the Middle East and Prasannan Par- thasarathi’s about India. Both are regional studies in the sense that they focus on one part of Eurasia to investigate its specific trajectory of eco- nomic development, always with reference to what was happening else- where, and particularly in Europe. Neither Kuran nor Parthasarathi has, however, identified a region within the Middle East or South Asia similar to England or the Lower Yangzi delta. This in itself raises questions about developments in the Middle East and South Asia. In the absence of con- centrated areas of industrial enterprise and innovation, the type of tech- nological spill-overs that we discussed above, would have been even more difficult to achieve.57 Still, questions can be raised about the suitability of regions even of the size of England and the Lower Yangzi delta, and more specifically about the number of variables that will inevitably come on board in any region- al comparison. Some of the research that was triggered by the publication of Pomeranz’s book has therefore further narrowed the search for clues, by focusing on one particular aspect, like for instance real wages, or urbaniza- tion rates, as indicators of change and development.58 In this book, a similar limitation is applied by first looking at skills in a regional context— discussed in the first part of the book, and then more specifically focussing on a range of industries, which are the topic of the second part of the book. In that second part, moreover, we have asked the authors to look not only at the region they know best, but to apply their expert knowledge in at least one other area of Eurasia as well. Some authors, most notably Moll-Mura- ta, as well as Kessler and Lucassen, have undertaken archival work in Asia and in Europe. Other authors are relying to a greater extent on the second-

56 Angus Maddison, The world economy: A millennial perspective (Paris: OECD, 2001); Maddison, Contours of the world-economy: Essays in macro-economic history (Oxford: Oxford University Press, 2007). 57 de Vries, ‘Economic growth’. 58 Robert C. Allen, Jean-Pascal Bassino, Debin Ma, Christine Moll-Murata, and Jan Luiten van Zanden, ‘Wages, prices, and living standards in China, 1738-1925: in comparison with Europe, Japan, and India’, Economic History Review 64 (2011), 8-38; Maarten Bosker, Eltjo Buringh, Jan Luiten van Zanden, ‘From Baghdad to London: Unravelling urban devel- opment in Europe and the Arab world 800-1800’, forthcoming in Review of Economics and Statistics, 95, 2013. introduction 21 ary literature, but have addressed this literature with a consistent set of questions that allows them to identify with greater precision the similari- ties and differences between the regions they investigate.

Conclusion

When added together, the papers collected in this volume seem to suggest three significant conclusions for the Great Divergence debate. First and foremost, in several important industries no obvious differences could be identified between the main regions of Eurasia. In the brick-making, build- ing and shipbuilding industries there were variations in the knowledge and technology applied to the production process, but none seems to have been obviously superior to the others. Tellingly, familiarising Indian shipwrights with European construction methods did not induce them to change the design of their own crafts. For centuries, Chinese, Indian and European ship designs existed next to each other, aware of the alternatives, but not converging on one particular model. In some areas Europe was ahead of the others—machine-making clearly being one of them. In others, how- ever, other regions were able to produce superior quality and dominate world markets for centuries: think of Indian cotton, or Chinese porcelain.59 This seems to confirm the Rosenthal-Wong view of the Great Divergence, that within a broader pattern of over-all similarities Europe had edged ahead well before 1800 in just a handful of key areas. Next to machine- making we might add weapons technology, as was demonstrated in a recent article by Phil Hoffman.60 Secondly, in the domain of knowledge formation it is difficult to reach a final verdict. Epstein’s strong claims about Europe’s unique institutions for skills formation are to some extent confirmed by the chapters on India and China, where family and regional backgrounds seem to have con- strained the circulation of knowledge. Roy and Pomeranz are, however, also pointing out that the systems in India and China were quite flexible and would therefore permit more of the elements that Epstein tended to

59 For Indian cotton, see also: Giorgio Riello, Prasannan Parthasarathi (eds), The spin- ning world: A global history of cotton textiles, 1200-1850 (Oxford: Oxford University Press); Giorgio Riello and Tirthankar Roy (eds), How India clothed the world: The world of South- Asian textiles, 1500-1850 Global Economic History Series, vol. 4 (Leiden: Brill, 2009); for Chinese porcelain: Robert Finlay, The pilgrim art: Cultures of porcelain in world history (Berkeley: University of California Press, 2010). 60 Philip T. Hoffman, ‘Prices, the military revolution, and western Europe’s comparative advantage in violence’, Economic History Review 64 S1 (2011), 39-59. 22 maarten prak and jan luiten van zanden label as uniquely European. But thirdly, in one area Europe was indeed unique, as is confirmed by almost every paper in this volume: the easy ac- cess to written—and more specifically printed—information.61 As Van Zanden’s chapter makes clear, the emergence of movable type created a tsunami of cheap information. Even when most of that was religious, enough remained of what we might call ‘professional literature’. In the sixteenth century, and even more so in the seventeenth, a large number of ‘how-to’ books was published by craft experts, and sold at prices well with- in the range affordable for ordinary craftsmen. Song Yingxing’s Works of Heaven, on the other hand, published in 1637 and reprinted in the 1650s, came out in print-runs of fifty or so copies. Even if it was affordable—and let’s not forget that all pages were custom-made from separate wood- blocks—it is hardly surprising that the work never reached a wide audience and was quickly forgotten.62

61 Now also Karel Davids, Religion, technology and the Great and Little Divergences: China and Europe compared, c. 700-1800 (Leiden: Brill, 2012). 62 Schäfer, The crafting, 249-52, 259. introduction 23

PART ONE

REGIONAL PERSPECTIVES 24 maarten prak and jan luiten van zanden transfer of technical knowledge and innovating in europe 25

CHAPTER ONE

TRANSFERRING TECHNICAL KNOWLEDGE AND INNOVATING IN EUROPE, C.1200-C.1800

Stephan R. Epstein

Introduction

The role of technology in the transition from premodern, ‘Malthusian’ to modern economies in late eighteenth- and nineteenth-century Europe is among the major questions in economic history, but it is still poorly un- derstood.1 In particular, the view that technological change before c.1800 was close to zero due to poorly specified property rights to knowledge and pervasive rent seeking by guilds is hard to square with the fact that the surge of technological innovation in the eighteenth century occurred with- in institutional frameworks not too dissimilar to those of 1300.2 A plausible explanation of premodern European technological develop- ment and industrialisation must account for three established facts. First,

1 [This paper was found on the author’s personal website at the London School of Economics at the time of his death. As a matter of fact, another version of the same text was available on that same website as well, in the Working Papers Series from the project, led by prof. Mary Morgan, ‘The Nature of Evidence: How well do “facts” travel?’. The version printed here, with the permission of Mary Morgan and the author’s partner dr. Rita Astuti, is the more extended of the two, and therefore presumably the latest edition produced by Larry Epstein. It is also the last of his works to appear in print. In some places obvious mistakes have been rectified. Information that was later added is between square brackets. The editors] Versions of this paper have been presented at the Economic History Association Annual Conference, San Jose, 10-12 September 2004; the 4th Global Economic History Network Conference, Leiden, 15-16 September 2004; and the Conference on ‘Endogenous Institutional Change’, Stanford Institute for Economic Policy Research, 4-5 March 2005. The research on which this paper is based was supported by a British Academy Research Readership, a Senior Fellowship at the Dibner Institute for the History of Science and Technology, an E.S.R.C. Research Award (no. RES-000-22-0031), and a Fellowship at the Wissenschaftskol- leg zu Berlin; I am very grateful to all these institutions. Particular thanks for comments go to Mary Morgan and Simon Schaffer. 2 D.C. North, Structure and change in economic history (New York: W.W. Norton, 1981); J. Mokyr, The gifts of Athena: Historical origins of the knowledge economy (Princeton: Princ- eton University Press, 2002). 26 s.r. epstein in the early thirteenth century Europe was still a technological backwater by comparison with the great Asian civilisations. Only a process of small- scale incremental innovation in metallurgy and instrument making, min- ing, building and shipbuilding, chemical process and cloth production, can explain the technological and industrial success of steam power—the most salient European contribution to premodern technical knowledge— six centuries later. The most striking feature by comparison with other coeval societies, however, is not so much that technological progress in premodern Europe occurred at a faster rate than elsewhere, but that progress was per- sistent and uninterrupted. By contrast, technological development in the great Asian civilizations of India and China experienced comparatively short periods of efflorescence, lasting a few centuries at a time, which were regularly followed by long phases of near-stagnation. Second, the geographical location of technological leadership in pre- modern Europe moved over time. Between the eleventh and the nineteenth centuries, Europe’s technological frontier shifted increasingly north-west: from the east-central Mediterranean to northern Italy during the thirteenth and fourteenth centuries, to southern Germany and in the late fifteenth, to the southern Low Countries in the sixteenth, to the Dutch Republic and finally to Britain during the seventeenth and eighteenth.3 Each new regional leader added the innovations of its predecessors to its local technical stock and recombined them for further technological ad- vances. Although leadership was temporary, falling prey over time to tech- nological sclerosis, declining marginal returns, and rent seeking by producers and elites, loss of leadership did not lead to a technological dead end.4 The existence of an increasingly integrated European market for skilled labour with a great deal of ‘ecological’ variation in demand, and of many polities whose rulers’ peaceful and military competition created spatial and temporal variation in demand for skills, generated the market and institutional conditions for new technological growth poles to take over.

3 K.A. Davids, ‘Shifts of technological leadership in early modern Europe’, in: J. Lucassen and K. Davids (eds) A miracle mirrored: The Dutch Republic in European perspective (Cam- bridge: Cambridge University Press, 1995), 338-66. 4 One might speculate that similar processes of slow, incremental technological dif- fusion and recombination under changing social, economic and institutional conditions are less apparent in premodern Asia. Instead, technological leadership seems to have persisted in the same regions (south-eastern China, western India) over very long stretches of time—significantly raising the likelihood of long-run equilibrium (or in a more pessi- mistic scenario, technical sclerosis due to ‘Cardwell’s Law’). transfer of technical knowledge and innovating in europe 27 Last, the technical knowledge of premodern craftsmen and engineers was largely experience-based.5 Thus, practically all premodern technical knowledge—which I define simply as knowledge of how to make things, and get them right—had to be transferred in the flesh. The shifts in re- gional technical leadership I just described could therefore only occur if technicians could take their knowledge elsewhere. This was arguably more easily done in Europe than in other parts of Eurasia, because European technicians were not members of ascriptive (kin-, religion- or locality- based) communities, and because they benefited from competitive bidding for technical expertise across a fragmented political and economic system.6 The implications for premodern economic history of the basic cognitive limitations to how technical knowledge can be expressed, processed and transmitted have yet to be examined in any detail. This paper asks how premodern European societies were able to generate incremental technical innovation under three headings: How was established and new knowledge transmitted? How was premodern technical knowledge stored to avoid loss? How were tacit, visual, verbal, and written means of transmission used heuristically? In answering these questions, I aim to sketch a model of endogenous technological progress that incorporates and explains the three stylised processes outlined above. I focus mainly on the period before 1700, in order to emphasize the similarities with better-known eighteenth- century conditions. Section 2 discusses the nature of experiential knowl- edge and its intergenerational transfer. Section 3 addresses knowledge transfer between peers, including technical codification and heuristics. Section 4 discusses technological transfer across space. Section 5 con- cludes.

5 A.S. Reber, Implicit learning and tacit knowledge. An essay on the cognitive unconscious (Oxford: Oxford University Press, 1993). 6 Although ascriptive forms of membership were not insuperable hurdles to mobility in China (barriers were higher in India), China may have lacked the kind of economic pull factors that underpinned technicians’ mobility in Europe, because their most technologi- cally advanced industries were concentrated in and around workshops under imperial control. China also lacked the kind of non-ascriptive institutional support, such as craft guilds, that lowered the costs of absorbing technical information from immigrant techni- cians. Consequently, the average cost of technical transfer was probably lower in premod- ern Europe by comparison with other societies. 28 s.r. epstein Acquiring Experiential Knowledge

In discussing the experiential knowledge of premodern technicians (crafts- men and engineers), I take as premise the fact that intelligent behaviours, long associated with the overt and conscious domain of cognitive function- ing, are better understood as the result of both implicit and explicit ca- pacities. Thus, experiential knowledge includes implicit or tacit knowledge; non-propositional and non-linear knowledge, including imagery, which has both implicit and explicit components; and explicit, propositional knowledge, which is linear and verbal or mathematical. Implicit knowledge equates to knowledge that is acquired largely independently of conscious attempts to learn, and largely in the absence of explicit knowledge about what was acquired. Implicit knowledge relies on rule finding and abstrac- tion, and is the basis for the acquisition of skills. Thus, the distinction be- tween implicit and explicit knowledge is hazy, and they form part of a continuum; but the implicit component is consistently greater than the explicit. Also on this definition, the boundaries between experiential knowledge in technical activities and in the sciences are far fuzzier than assumed by standard claims that technical practice and experimentation is ‘non-sci- entific’ because it lacks an underlying conceptual or propositional frame- work. There is no scope here to enter the debate on the relative significance of scientific and non-scientific thinking and practice for the Industrial Revolution, recently rekindled by Joel Mokyr, but it may be useful to set out this paper’s underlying assumptions on the matter. Following an estab- lished tradition of studies of scientific practices, I assume that the major distinction between scientific and technical practice for the purposes of economic history is neither cognitive nor epistemological, but resides in differences in the aims and forms of codification. In this view, scientists’ main aim is to codify knowledge, for codification is essential both to com- municate, convince, and establish credentials, and to establish a shared base for further advance. Technicians, by contrast, codify only as a means to an end, their end being to make things work reliably and well. They do not avoid codification in principle, but they generate it less systematically than scientists and they do so largely as a result, rather than independent- ly, of the production process itself. At the same time, both scientists and technicians rely on experiential knowledge that is hard or impossible to codify. Experiential knowledge is a good, and its exchange and diffusion demand that those who have it take transfer of technical knowledge and innovating in europe 29 deliberate action to share it through face-to-face communication. These operations are costly to implement, and have relied historically on different institutional solutions. Analytically, it is useful to break down the question how technical knowledge was transferred into the issues of inter-genera- tional transmission and transmission between skilled peers. The first stage in acquiring technical knowledge was through a long-term relationship of pupilage based on formal or informal sanction, in other words through apprenticeship, which is the most widespread arrangement for transmitting technical knowledge outside the family devised by human societies. Parents or guardians (including people acting for religious found- ling institutions) would usually present a child for apprenticeship between the ages of 13 and 15; but not all apprentices were adolescents, and guild statutes never specified the maximum age at which the indenture could begin. Most statutes set the minimum term of service, proportionate to the craft’s skill requirements and to its expected returns. Thus the average length, which appears to have increased slowly over time, was variable; the English Statute of Artificers (1563, repealed 1814), which prescribed a na- tional norm of 7 years terminating at age 24 or older, was unique. Even in England, however, the actual length of service was negotiated individually on the basis of the apprentice’s age and prior experience, of the premium (if any) the parents’ could advance, and of the master’s reputation. Most statutes required longer terms for outsiders than for sons of members, who would have experienced some basic induction to the craft in their father’s shop. Apprenticeship years could also be bought out at a later date, or condoned if the trainee could demonstrate sufficient skills. Duration was further influenced by the fact that before the dissolution of craft guilds apprenticeship was not just a traineeship for a skilled occupation, but also a means for socializing children and adolescents into adulthood and the world of work, so that the term was longer the younger the age at entry. The duration of training does not capture the intensity of resources expended during it. Apprenticeship training was costly, because skills and expertise take time and effort to acquire. Expertise depends on two main processes: heuristic search of problem spaces, and the recognition of cues that access relevant knowledge and suggest heuristics for the next step. Experts store thousands of ‘chunks’ of information in memory, accessible when they recognize relevant cues. Experts use these recognition process- es to achieve unusual feats of memory, reorganize knowledge into complex hierarchical systems, and develop complex networks of causally related information. The knowledge of less skilled individuals, in contrast, is 30 s.r. epstein encoded using everyday concepts that make the retrieval of even their limited knowledge difficult and unreliable. It consequently takes about 10 years of focused training to acquire top-level expertise in activities as di- verse as chess, dog training, wine tasting, playing and composing music, sports, and, possibly, language acquisition.7 There is no reason to believe that the length of training would be any different in areas of more practical expertise—a fact plausibly reflected in the lengthy technical apprentice- ships of premodern Europe. Secondly, apprenticeship was costly because most craft knowledge was experiential.8 Consequently, craft statutes and labour laws never specified the content of the training regime. Crafts were not learned prescriptively, because training was in the master craftsman’s head and hands; instead, craftsmen and women tested the quality of training by examining its out- come. The acquisition of technical expertise was sanctioned through a mastership. Starting in the late thirteenth century and with increasing frequency from the late fourteenth, candidates to mastership in the most highly skilled crafts had to prove their skills through examination or by making a masterpiece.9 The masterpiece combined a physical embodi- ment of collective knowledge and individual creativity and virtuosity (‘ge- nius’). It was a demonstration of skill and of self-confidence that the proposed product could be constructed and would work; and it established the expert as someone who had assimilated tradition so well that he could adapt, modify and transcend it. Expertise also made it easier to formulate non-verbal practices and heuristics explicitly, as Salviati, on the first day of Galileo’s Discourses, famously remarks: ‘The constant activity which you Venetians display in your famous arsenal suggests to the studious mind a large field for investigation … for … all types of instruments and machines are constantly being constructed by many artisans, among whom there must be some who, partly by inherited experience and partly by their own

7 K. Anders Ericsson, R.Th. Krampe, and C. Tesch-Römer, ‘The role of deliberate prac- tice in the acquisition of expert performance’, Psychological Review, 100 (1993), 363-406. 8 The salience of implicit knowledge and experience provided an inbuilt advantage to employing family members, who had been socialised early into the craft and generated higher levels of trust, particularly in the most technically advanced industries like mining and metal-working, ship- and high quality edifice building, and clock and instrument mak- ing. For similar reasons, highly specialised craft knowledge and techniques was transmitted through craft lineages; see e.g. J. Brown, Mathematical instrument-makers in the Grocers’ Company 1688-1800 (London: Science Museum, 1979). 9 W. Cahn, Masterpieces: Chapters on the history of an idea (Princeton, NJ.: Princeton University Press, 1979). transfer of technical knowledge and innovating in europe 31 observations, have become highly expert and clever in explanation’.10 Ex- pertise, in other words, was also a precondition for the ability to teach, and teaching apprentices helped solve the conundrum of making tacit techni- cal knowledge public. Standard economic theory explains why apprenticeship was needed. Since future human capital cannot act as collateral, resource poor but po- tentially able workers may be incapable of bearing the costs of their invest- ment in skills, leading to a socially suboptimal supply of skilled workers. Premodern apprenticeship allowed trainees to exchange subsidized train- ing for below-market wages after training was concluded. However, masters would have still supplied suboptimal amounts of training if the trainee could quit before contract expired because the training masters could not capture the full return to their investment. Trainees with transferable skills (which are neither entirely general nor wholly specific to one firm) would be poached by masters that did not have to recover the training costs and that could pay them less than their marginal product but more than the wage paid by the original master. For apprenticeship to be viable, poaching had to be constrained through legally enforceable indentures, which allowed the masters that provided the training to appropriate the full benefits in the immediate post-training period. In premodern Europe, this enforcement was provided largely, though not solely, by craft guilds; for in the absence of compulsory schooling, su- pra-local legislation, and efficient bureaucracies, formal or informal craft associations were best suited to enforce apprenticeship contracts and rules outside the family.11 Craft guilds overcame externalities in human capital formation by supervising job performance, work conditions, and quality of instruction; enforcing contracts through compulsory membership, stat- utory penalties, and blackballing; and protecting apprentices against poor training in craft specific skills within oligopsonistic labour markets. Live-in

10 G. Galilei, Discorsi e dimostrazioni matematiche intorno a due nuove scienze attenenti alla mecanica et i movimenti locali (Leiden: Elsevier, 1638), 1-2. 11 Large numbers of children were also never apprenticed because they were trained within their parents’ homes, or because some crafts (particularly those involving trade) did not require formal training; this fact accounts for the low number of apprentices with practicing masters relative to the number of masters and journeymen that were needed to reproduce trades over time, and for the low number of working girls recorded. Conversely, apprenticeship could exist outside guild structures, although it faced the problem of enforce- ment outside a formal institutional framework. For these reasons and because of the nature of the skills involved, apprenticeship was mainly an urban, craft-based phenomenon, although in seventeenth- and eighteenth-century England it was also undertaken by the children of the rural poor under the remit of the national Poor Laws. 32 s.r. epstein apprentices had the right to be lodged, fed, housed, clothed, and heated on a par with members of their master’s family, but they were equally sub- jected to his disciplinary rule as a surrogate father. Even those apprentices who lived at their parents’ home (which they did in increasing numbers from the late seventeenth century) were expected to pay unquestioning obedience to the master’s orders and respect the craft’s rules. The appren- tices’ minority status explains the universal ban on marriage, and why breaches of the rule were treated severely, while the contract’s education- al features explain why younger trainees were set longer terms. The master controlled the apprentice’s work-time, and could offer the apprentice’s labour to another guildsman; the apprentice had to work to the master’s benefit and profit, and the guild enforced the master’s right to keep the apprentice on after his training had been completed so as to repay the master’s training costs. Guilds, however, were more effective in banning poaching by their mem- bers than in stopping apprentices from quitting before their term ended. Masters tried to reduce this by demanding entry fees (de facto bonds post- ed to ensure the apprentice’s commitment for the full term), by setting apprentices’ wages on a rising scale for the contract’s duration, and by promising a pay-off upon completion, but there was little they could do to fully stem the hemorrhage. The rate of attrition in early modern England has been estimated at 30 to 50 per cent in sixteenth and seventeenth-cen- tury London, Bristol and Norwich. Although a significant proportion of apprentices who quit early were simply unable to cope, were mistreated, or moved to another occupation, many left in search of work in the rural and small town provinces where skill requirements were lower: crafts in pre modern towns acted as training centers for their regional or even, in the case of London and other capital cities, national hinterlands, which they provided with a constant flow of skilled and semi-skilled labour. Such high rates of defection would suggest that masters were unable to fully recover their training costs, and that the reason they were nonetheless willing to train is that they had ex post monoposony power arising from their superior information about their employees’ abilities.12 Rising rates of defection might also explain attempts to extend the minimum length of the apprenticeship contract, although as explained above, statutory lengths were easily evaded.

12 D. Acemoglu and J.-S. Pischke, ‘Why do firms train? Theory and evidence’, Quarterly Journal of Economics 113 (1998), 79-119. transfer of technical knowledge and innovating in europe 33 Many of the departing apprentices had originally immigrated to the town from the urban hinterland. This gave rise to problems of adverse selection and asymmetric information, which guilds and governments addressed by stipulating entrance requirements that signaled the labour- er’s quality or provided surety against misbehaviour, such as place of resi- dence, family income, or the father’s occupation; the Statute of Artificers specified all three. In some highly specialized and cyclical industries, like mining and iron-making, shipbuilding and high-quality masonry, skills training was often kept within closely knit kin networks, possibly because the higher risks of those industries restricted the supply of apprentices. Evidence that apprenticeship achieved its stated purpose and was not simply a means to exclude workers from the market and a source of rents for craftsmen is twofold. First, practically all crafts’s jurisdiction extended only as far as their town or city walls, so there was ample scope for more efficient means of training to develop in the surrounding countryside, or in the many towns and cities where the rule of craft guilds did not apply. Although as mentioned, craft guilds were not the only way of enforcing apprenticeship contracts, before the nineteenth century the crafts’ pri- macy in training highly skilled labour was largely unchallenged. Second, the sharp rise during the fourteenth and fifteenth century in the number of guilds (Figure 1.1), together with a growing emphasis on formal appren- ticeship, anti-poaching rules, and final examinations, appears to have pro- duced a substantial increase in the supply of skilled labour after 1350. Evidence of this comes from the sharp—and most importantly, perma- nent—fall in the wage ratio between skilled and unskilled labour in the building industry (the only one with adequate data) after the Black Death wage, from c.2.2-2.3 : 1 (1300-25) to c.1.5-1.7 : 1 (1500-25) (Figure 1.2).13 In sum, craft guilds restricted the mobility of workers so that masters could earn rents on trained workers. This may have restricted the efficient allocation of workers to firms, but it did supply critical institutional support for the provision and transmission of skills.14 In this sense, crafts helped the European economy achieve a higher-level equilibrium, and also ex- plains the extraordinary longevity of European craft guilds from the late eleventh century to the early nineteenth.15

13 [Jan Luiten van Zanden, The Long Road to the Industrial Revolution. The European Economy in a Global Perspective 1000-1800 (Leiden: Brill, 2009), ch. 5]. 14 Acemoglu and Pischke, ‘Why do firms train?’. 15 S.R. Epstein, ‘Craft guilds, apprenticeship and technological change in preindustrial Europe’, Journal of Economic History 53 (1998), 684-713. 34 s.r. epstein

Figure 1.1. Established craft guilds, Italy and Netherlands 1100-1800. Sources: Angelo Moioli and Jan Lucassen, per litteram.

Figure 1.2. Skill differentials in the European building industry, 1300-1799 (by city). Sources: Building masters’ wages from Robert Allen’s website: http://www.nuffield.ox.ac.uk/ General/Members/allen.aspx. transfer of technical knowledge and innovating in europe 35 Collective Knowledge and Technical Heuristics

Knowledge Sharing Although apprenticeship contributed substantially to the collective or ‘dis- tributed’ nature of premodern technical knowledge, which was an essential feature of technological progress, the inter-generational transmission of knowledge was less important than knowledge sharing—including ‘collec- tive invention’—between skilled peers.16 Technical knowledge sharing between peers took place on site and through migration. Although practices in making, repairing and running machines, building ships and edifices, digging mines, making clocks and watches and so on were necessarily common or accessible knowledge (not least because technicians could not keep reinventing the wheel,17 evidence of on-site sharing is more sporadic than for sharing via migrants. The avail- able evidence also relates mostly to ‘hi-tec’ industries in which competitive pressures and the advantages of cooperation were greatest, and which were therefore most likely to employ foreign workers with new techniques. In the course of the fifteenth century Venetian glassmaking became one of the most advanced industries in Western Europe, comparable in terms of capital investment, specialisation of production, and rate of process and product innovation with shipbuilding, large-scale edifice building, and luxury cloth production.18 Discussion of the Venetian guild of glassmakers dwells for the most part on the truculent craft and government policies towards emigrating craftsmen, but this misrepresents the situation in sev- eral important ways. First, relations between glassmakers in Venice (Mu- rano) and the outside world were not, and could not be, foreclosed. Already in 1271 the guild statutes deal with the issue of ‘foreigners’ practising the craft of glassmaking in Venice, and the issue persisted into the seventeenth century when Venice finally lost its quasi-monopoly over crystal glass. Sec- ond, the production process required an annual closure of 3-4 months during which the workers were in practice free to find work outside the city. Despite the reiteration of fines and even prison terms for glassmakers

16 R.C. Allen, ‘Collective invention’, Journal of Economic Behavior and Organization 4 (1983), 1-24. 17 G. Hollister-Short, ‘Invisible technology, invisible numbers’, ICON 1 (1995), 132-47 18 W.P. McCray, Glassmaking in Renaissance Venice: The fragile craft (Aldershot: Ashgate, 1999); McCray, ‘Creating networks of skill: technology transfer and the glass industry in Venice’, Journal of European Economic History 28 (1999), 301-34. 36 s.r. epstein who left Venice during the dead season, ‘such notices often contained the same names of glassmakers over the years which does not give an indica- tion that the Venetian state’s policies designed to prevent seasonal worker migration were very effective’.19 Worker migration eventually led to the diffusion of Venetian technologies to other European courts and cities; the reason commercial industries took so long to develop elsewhere (in Ant- werp, Amsterdam, Paris and London) lay with the problems with transfer- ring technical knowledge that was still unable to control consistently for chemical content, and was therefore highly contextual, rather than the Venetians’ ability to monopolise their secrets. Third, Venetian glassmakers made systematic use of codified experimentation in response to consum- er demand. Following the decision to ‘rezone’ the industry from the city of Venice to the small island of Murano in 1291, the persistent circulation of skilled workers among the master glassmakers (but the prohibition of poaching), and the speed with which technical innovations were shared and standardized for foreign export, prove that craft technologies, innova- tions and skills were viewed and acted upon as collective goods.20 Beginning again in the fifteenth century, instrument-making (for horol- ogy, navigation, land surveying, weighing and measuring, drawing, gunnery and architecture) gradually became one of the most distinctive and tech- nologically innovative industries of premodern Europe. For the most part, these trades were organised into craft guilds. But ‘there was a priori no reason why the corporations should frown upon innovation, and they do not seem to have. Major innovations in the structure of clocks and watch- es such as the introduction of the fusee in watches, the pendulum in clocks, the balance-spring in watches, the jewelling of bearings in watches, new escapements and the development of thermal compensation systems all occurred with little or no guild comment, let alone opposition. Similarly the making of new instruments such as telescopes, spyglasses, microscopes and barometers, or new adaptations to old ones, provoked no more reac- tion than did innovations in methods of manufacture such as the diffusion of wheel-dividing engines and gear-cutting machines, or the invention of

19 McCray, Glassmaking, 44-45. 20 However, craft statutes did not regulate the sharing of know-how; this remained proprietary in certain key respects, as shown by differences in the surviving family recipes. The latter offer proof of the kind of ‘competition within cooperation’ sustained by other ‘appropriability institutions’ through time: R.P. Merges, From medieval guilds to open source software: Informal norms, appropriability institutions, and innovation. Paper presented at the annual Economic History Association Conference, San Diego, September 2004. transfer of technical knowledge and innovating in europe 37 a method for the polishing of multiple spectacle lenses in a single operation’.21 The most salient feature of early modern instrument making was in fact the guilds’ systematic resistance to individual patenting. One form taken by opposition to patents was to ignore them. In late 1656 the Dutch natural philosopher Christiaan Huygens (1629-95) completed the first clock to em- ploy the pendulum as a regulator. In order to exploit the design he ex- plained it to the Hague clock-maker Salomon Coster, permitting him to take out a an octroy or privilège (the equivalent of a patent) for it on 16 June 1657 that gave Coster the exclusive manufacturing rights for 21 years. By then, however, knowledge of the new timepiece had already spread within the Low Countries and to Paris and . In early 1658 a Rotterdam clock-maker, Simon Douw, circumvented Huygens’ patent with such suc- cess that Huygens abandoned the attempt to enforce the patent in the Dutch Republic. Nonetheless he did try to profit from his invention by obtaining a privilège in France. The request was refused three times by the chancellor, Pierre Seguier, with the comment each time that he did not want to have ‘all the master clockmakers of Paris crying after him’. At the same time in London, Ahasuerus I Fromanteel was also constructing pen- dulum clocks on Huygens’ pattern, which he advertised for sale in late 1658. The French response points to the second form of craft opposition, namely resistance to any patent concession itself. Instrument-makers did not object to innovation, but to giving one of their number a perceived unfair advantage over the others. The standard objection was that an in- novation ‘was not new, nor his [the patentee’s], nor of the use claimed by him’, as the London Spectacle-Makers put it in a court case in 1694 against John Marshall, who had got the Royal Society to sponsor his new technol- ogy. Fortunately, a few months later Marshall decided to share his innova- tion with his peers. Between 1685 and 1755, the London Clockmakers’ successfully blocked four, and unsuccessfully opposed three out of nine British horological patents; the main objection was that not every innova- tion was a genuine invention.22 With 143,000 inhabitants in 1789, Lyon was the second largest town in eighteenth-century France. The Grande Fabrique, run by silk merchants in

21 A. Turner, ‘“Not to hurt of trade”: Guilds and innovation in horology and precision instrument-making’, in: S.R. Epstein, and M. Prak (eds) Guilds, innovation, and the European economy, 1400-1800 (Cambridge: Cambridge University Press, 2008), 269. 22 See also L. Hilaire-Pérez, ‘Invention and the state in eighteenth-century France’, Technology and Culture 32 (1991), 916. 38 s.r. epstein close association with the town authorities, employed nearly a quarter of the population. By changing patterns and fashions on a yearly basis, the Fabrique played an essential part in the success of the silk industry on international markets. But technological progress was also a major concern of local elites. Lyon artisans, who accounted for at least 170 of the nearly 900 inventors who applied to the French national administration for a privilege of invention, were strongly encouraged by the local municipality to develop new technologies, especially new looms. From 1711 the town government, the guild and a representative of the state, the intendant, col- lectively administered a special fund for inventors, the Caisse du droit des éttoffes étrangères, paid by a tax on foreign silk entering Lyon. The fund financed innovation from the research stage to the training of expertise right through to the stage of commercialization.23 The main principle underlying the fund during the eighteenth century was that inventions were a collective good. Most artisans were expected to invent new looms or improve existing ones, display them publicly and sell them to their Lyonnaise peers with no private protection. Exclusive privi- leges (the Ancien Regime equivalent of patents of invention) were few. Inventing was considered a service to the town and this assumption lay at the basis of the examinations jointly administered by guild officials, mem- bers of the Lyon Academy of Sciences and weavers. Then, after a reward was granted, the looms were deposited in the guild’s office and artisans would have to create their masterpieces on newly invented looms. Inven- tors also had to teach their know-how and were rewarded according to the number of pupils they would train. Most grants were indexed by a bonus system on the numbers of new looms actually diffused in town; evaluation and reward were based upon the users’ verdict, which encouraged inven- tors to commercialize their mechanical devices. These networks were the basis for patterns of innovation in Lyon. Inven- tive artisans, both weavers and not, were quickly informed of new devices and were constantly striving to improve on them. Indirect evidence that invention was a collective activity is that the new drawing looms, from Falcon’s loom to Jacquard’s, had compatible programmes resulting in cu- mulatively compatible technology. Vaucanson’s programming cylinder was inspired by the Falcon looms that had paper boards passing round a prism (1742). In 1777, a certain Rivet signed one of Dardois’ certificates; a few months later Rivet presented a new loom of the same kind. For the build- ing of his second loom, Falcon called upon a weaver, Allard, who in 1763

23 L. Pérez, ‘Inventing in a world of guilds: Silk fabrics in eighteenth-century Lyon’, in: Epstein and Prak (eds) Guilds, innovation, and the European economy, 232-61. transfer of technical knowledge and innovating in europe 39 registered an improvement; Jacquard’s invention was also much improved by a certain Breton, a mechanic from the town of Privas. In the context of the earlier discussion of guilds’ hostility towards pat- ents, it is interesting to note that in Lyon day-to-day business practices ranged from free exchange to theft both of skilled workers and ideas. The free circulation of knowledge, including if necessary stealing, was idealised by one of the major eighteenth-century inventors, Philippe de Lassalle, but it seems likely that he was expressing a more widespread opinion. He claimed that he did not condemn the theft of patterns or inventions and that he was pleased when his printed silk cloth was copied and his workers enticed by rivals: ‘more than twenty of my colleagues employ hand-paint- ers and entice mine every day as soon as they are trained and they get from them colours and even my own drawings; but I do not complain about these events if they can help to prove that all prejudice against new styles is useless for the common weal and for private business’.24 The most extensive evidence of technical sharing over time and space, however, is associated with large building sites, which early on drew skilled workers and engineers from across Europe. For example, the master build- er or cleric Villard de Honnecourt stated in his book of drawings (c.1215-20) that he settled points with other masters inter se disputando—the techni- cal expression for formal debate that had long been standard in the uni- versity schools—to underline the fact that his art too rested on firm intellectual principles that could be applied in systematic argumentation. In 1459 master and journeyman masons involved in building major church- es across Central Europe met at Regensburg and stipulated that no-one should be taught for money—with the implication that information should be freely shared.25 Similarly, the habit of competitive bids for artistic and building projects, well established by the late fourteenth century in Italy and common elsewhere by the sixteenth, assumed that applicants pos- sessed a common core of technical competencies, which patrons could only assess indirectly. Public displays by engineers—which their peers would understand, even if laypeople could not—are recorded from the late fourteenth century, when Giovanni de’ Dondi of Padua put his astronomi- cal escapement clock on public show; in the sixteenth century, craftsmen from Augsburg and Nuremberg made rival displays of technical prowess. And, in a letter to Mersenne dated 7 December 1642, Descartes describes the ingénieur Etienne de Villebressieu as ‘a very curious man who knew

24 Ibid., 257. 25 A. Black, Guilds and Civil Society in European Political Thought from the Twelfth Cen- tury to the Present (London: Methuen, 1984), 9. 40 s.r. epstein many of those little chemical secrets which are exchanged between mem- bers of the craft’.26 The strongest proof of on-site knowledge sharing is nonetheless indirect. Once again, some of the most systematic evidence arises in the records of large-scale religious and secular building sites that gave rise to the most complex technical challenges. Church and cathedral building in particular demonstrates both the considerable degree of structural innovation that did take place, and some of its inherent limitations. The complexity of Gothic cathedrals made it common practice, already in the twelfth century when the first new cathedrals were struck, to call on outside experts to consult on major structural issues. This fact stimulated experimentation—in the use of buttresses, the width of aisle and the height of nave, the height of pier-buttresses and pitch of the roof—that persisted after 1500 when the Gothic style went out of fashion. One measure of such experimentation is the slenderness ratio, that is, the ratio between height and width of the main supporting piers—the higher the ratio, the ‘lighter’ the final structure. The ratio for the cathedral of Chartres, finished in 1194, was 4.4; thirty years later, at Amiens and Beauvais, the ratio had doubled; by c.1350, at the cathedral of Palma, the master-builders achieved a remarkable ratio of 13.8.27 As cathedrals grew in height, however, builders faced increasing struc- tural problems. The lower nave, clerestory and roof were subject to in- creased outer thrust and wind forces, and the foundations were subject to increased vertical pressure and settlement. Since builders lacked a work- able theory of structural force before the nineteenth century, they had no means of predicting the structural effects of increased scale. The most frequent solution was to build in modules and to build slowly, observing the evidence of stress over time and making repairs and innovations as needed. The flying buttress was a crucial structural innovation introduced along these lines; ‘all flying buttresses in the great northern [French] churches prior to the second half of the twelfth century seem … to have been added as casual expedients only after weaknesses had become ap- parent or … the vaults had already pushed the walls aside and collapsed’. On other occasions, like the building of Brunelleschi’s Florentine dome, ‘new structural ideas were deliberately tried out on a smaller scale’.28

26 Charles Adam and Paul Tannery (eds), Œuvres de Descartes (Paris: Cerf, 1897–1913), vol. III, 597–98. 27 R. Mark, ‘Structural experimentation in ’. American Scientist 6 (1978) 542-50. 28 R.J. Mainstone, ‘Structural theory and design before 1742’, Architectural Review 143 (1968), 305. transfer of technical knowledge and innovating in europe 41 Achievement of expertise requires the ability to display flexibility with the rules. Major changes to plans were made as the need for them arose, in response to changes in the commission or to structural problems. Thus, when Brunelleschi did not provide workers with a 3-dimensional model for the Florentine Spedale degli Innocenti, the masons and carvers devi- ated from his design. Originally conceived as a block (cuadro) on its shelf in majestic isolation from other buildings, the design of Philip II’s palace of the Escorial was gradually extended to include various outbuildings. Twenty years after the start of the building works, ‘the artisans were still unsure whether the sanctuary was to be rectangular or apsidal, and [the master mason Herrera] was asked for drawings to clarify the question.’ In 1577 ‘grave doubts arose about the stability of the dome support where the stones were showing fractures. It is reported that public fears caused Her- rera reluctantly to reduce the height of the dome’s pedestal by 11 ft., and to eliminate the niches, which reduced the mass of the pillars’.29 At about the same time, Venetian architects and masons refused to approve a single plan for the construction of the Rialto Bridge, which was therefore built in stag- es, with each stage receiving a different plan.30 A century later, Christopher Wren ‘adapted the design [of St.Paul’s Cathedral] as defects occurred, or his widening experience suggested improvements’. Although as a natural philosopher he developed a wrong theory of arches, as a practical engineer employing little or no calculation he was highly successful, because he employed the heuristics of practical building and engineering.31 In conclusion, there is strong evidence that craft guilds—particularly in the more specialised trades—promoted collective knowledge sharing and invention (we shall examine under what circumstances they opposed innovations further below). Conversely, they opposed patenting as a means to privatise technical knowledge and ‘damage the trade’. These conclusions raise questions about the type of innovation promoted and hindered by guilds that require additional research. First, we would like to know more about the sources of innovation. For example, individual scientists or sa- vants, who turned to guilded craftsmen to turn their invention into a work- ing machine or to commercialise it more widely, devised some of the most

29 G. Kubler, Building the Escorial (Princeton: Princeton University Press, 1982), 82, 98. 30 D. Calabi, and P. Morachiello, ‘Le Pont du Rialto: un chantier public à Venise à la fin du XVIe siècle’, in: T. Ruddock (ed.), Masonry bridges, viaducts and aqueducts (Aldershot: Ashgate, 2000), 109-32. 31 S.B. Hamilton (1998). ‘The place of Sir Christopher Wren in the history of structural engineering’, in: W. Addis (ed.), Structural and Civil Engineering Design (Aldershot: Ashgate, 1998), 193-208. 42 s.r. epstein important innovations in horology. However, most of the other innovations discussed here were devised within the craft itself. Second, the relation between guild-based and patented innovations is not entirely clear. Guilds tended to oppose process and product innovations in sectors in which they had competence and which could be viewed as trade secrets, and were more accepting of mechanical inventions in power production (milling, hydraulics and heating), which had high sunk costs and indivisibilities and low reproduction costs. However, the distinction was not clear-cut: Vene- tian glassmakers, for example, were able to patent some product innova- tions without major craft opposition. Third, since craft guilds in different urban centers never formally cooperated, technological spillovers had to occur informally through the markets for labour and intermediate goods. It is an open question, if and how these mechanisms constrained the de- velopment of more dynamic industrial districts.

Predictability, Codification and Innovation A less charitable view of the rule-bending described previously suggests a lack of codification, that is, extreme empiricism and a poor ability to predict. For example, the solutions to structural concerns in cathedral building I described were, inevitably, strongly related to the cathedral’s dimensions, such as the ratio of height to width of the nave, and the height and angle of the clerestory and the roof. Gothic dimensions were based on geometri- cal criteria, which, in northwest Europe, seem to have been largely derived from simple manipulations of the square. Although the rules or algorithms were never fully formulated, they gave rise to specific engineering problems and, thus, to quite specific technical solutions. Although the development in Gothic building of heuristic ‘rules of thumb’ or algorithms provided reasonably safe and economical solutions, while reducing computation and design time, it also tended to establish a conceptual identity between building structure and form.32 This made it hard to transfer the structural theory developed in one Gothic building lodge or lodges in one region to somewhere that had a different ideal form. An instance of the conceptual and technical problems that could ensue occurred at the building site of the new cathedral at Milan at the turn of the fifteenth century. The difficulties arose because Milan at the time was an architectural backwater, and local building skills were inadequate. From the start, therefore, the Milanese asked experts from Central Italy—then

32 Mainstone, ‘Structural theory’. transfer of technical knowledge and innovating in europe 43 architecturally and technically more advanced, yet still peripheral to the Gothic powerhouses further north—to advise them on the form and struc- ture of the new church. Importantly, the plan drawings were based on simple manipulations of the triangle—with the result that the nave and roof of the cathedral were both lower and broader than in the Gothic heart- land over the Alps. Structural problems soon arose, however, so the Milanese brought in North European experts to advise them—with explosive effects. In 1400, Jean Mignot, a master-builder from northern France, insisted on applying his own geometrical design principles to the cathedral’s buttresses. ‘He argued passionately that only high flying buttresses—a rigorous solution based on scientia, that is, on geometrical proportion—could yield a stable structure: “mere craft [ars] without rigorous knowledge [scientia] is useless”’.33 The Lombard masons rebutted that scientia without ars, with- out the practical knowledge gained from experience, was equally useless. But the discussion was not, in fact, concerned with either theory or practice taken individually, but rather with the practical links between the two. For Jean Mignot, form (based on scientia) defined structure (built through ars)—and there was only one legitimate form, derived from the geometri- cal permutations of the square he was trained in. The disagreement arose because the Milanese preferred another form, derived from a different, albeit equally ‘scientific’, geometrical procedure. However, they lacked the well-trained, skilled labour to build the related structure and were forced back onto their own local judgment and experience. The problem of combining or synthesizing different empirical traditions that did not clearly distinguish between building structure and form could be addressed in different ways. One way was to codify existing traditions. In the late fifteenth and early sixteenth centuries, several German master masons (Matthäus Roriczer, Lorenz Lechler, and others) drafted detailed notebooks or handbooks that reproduced the square-based configurations of form. The reasons for doing this are not entirely clear, but one relevant factor was probably the increased circulation of masters, journeymen and trainees between Central European building lodges, which must have given rise to confusion and conflict over which lodge tradition would pre- vail (see Figure 1.3 for evidence of strong integration of building wages in

33 A. Grafton, Leon Battista Alberti, master builder of the Italian Renaissance (Cambridge, MA: Harvard University Press, 2000), 268; O.G. von Simson, ‘The Gothic cathedral: design and meaning’, in: Addis (ed.), Structural and Civil Engineering Design, 159-70. 44 s.r. epstein

Figure 1.3. Integration of the skilled builders’ market 1400-1799. Source: See figure 1.2.

Northwest Europe during the pre-1550 era of Gothic building). Although we do not know if the German master masons were trying to synthesise different lodge traditions or if they were simply codifying their local lodg- es’ practice, their actions seem to have been essentially reactive. The encounter of different technical and design traditions could, how- ever, also generate cognitively new procedures. In sixteenth-century Spain, where tension between Gothic and Italian Renaissance building traditions was particularly lively, the master builder Rodrigo Gil de Hontañon at- tempted to systematize the design process by creating a sequence of codi- fied procedures to be followed in large church-building projects. Gil’s algorithms, drafted around 1540, had three objectives. They aimed to com- bine Gothic and Classical proportion-based design methods, and to prove their basic identity. They also tried to establish an independent ‘science’ of structural design. Finally, they attempted to establish new collective heu- ristics for on-site builders to work with. In pursuing this effort to synthesize and codify two seemingly incompatible aesthetic and building traditions, Gil was led to experiment with Gothic practices on classical arches, and to ‘apply new arithmetic procedures to Gothic rib vaults’.34

34 S.L. Sanabria, ‘The mechanization of design in the 16th century: the structural formu- lae of Rodrigo Gil de Hontañon’, in: Addis (ed.), Structural and Civil Engineering Design, 1-14. transfer of technical knowledge and innovating in europe 45 An assessment of craft and engineering heuristics must distinguish be- tween well structured problems, in which situations, operators, and goals tests are all sharply defined, and little specific domain knowledge is need- ed; and ill structured problems, which require extensive experiential knowledge to be solved effectively through a combination of inductive and deductive processes. Designing buildings, for example, is a poorly struc- tured task. The tests of success are complex and ill defined, and are often elaborated during the solution process. The solution requires flexibility that will often manifest itself as a lack of precision, a ‘good-enough’ and make-do approach that mathematically grounded theoreticians find dis- concerting. Premodern shipbuilding appears superficially more structured than edifice building, but in other ways it was similarly open-ended: criti- cally, it could not proceed, like building, by testing individual modules as they were built, because success could only be ascertained after the ship was actually launched. The heuristic tools of ship- and edifice building were nonetheless remarkably similar. Like masonry builders, shipbuilders achieved structural stability through a shared, mnemonically rich ‘geomet- ric discipline’ that legitimized experience gained from building similar structures, and a ‘wider tacit or intuitive understanding of the conditions of static equilibrium’ based on two components, ‘spatial and muscular’.35 Venetian shipwrights, for example, based their dimensions on a module that was normally the beam of the proposed galley; this was multiplied in a fixed proportion to give the deck-length, and a fraction of this in turn gave the length of the keel. In addition, the Venetian, or Mediterranean system of module building, was carvel-built. Between the late fifteenth and the early sixteenth century North Atlantic ships, which were previously clinker-built, began to be built according to the Mediterranean system. As the technology migrated, first to Portugal and Spain, thereafter to England and the Hanse area, it changed from its purely tacit and demonstrative form, which employed no graphical support, to a system that relied increas- ingly on graphical design. The Venetians had written up their shipbuilding schema already in the fifteenth century, followed by the Portuguese in the mid-to-late sixteenth, but these drawings were purely descriptive and were not used for planning purposes. Proportional design for future planning seems to have been in- troduced by the Englishman Matthew Baker in the 1580s, spreading from the 1630s together with 3-dimensional modeling and becoming the norm

35 R.J. Mainstone, ‘Stability concepts from the Renaissance to today’, in: Addis (ed.), Structural and Civil Engineering Design, 171-92. 46 s.r. epstein in England after the Civil War. The French, spurred by Colbert’s build-up of the navy, introduced design slightly later but with more sophisticated geometrical methods and tools. These innovations appear to have had two practical implications. On the one hand, planning design may have intro- duced greater building flexibility. It did not entirely break the link between structure and form, because designers still lacked adequate hydrostatic and hydrodynamic theories; modeling new ships on the basis of experimental drawings was therefore very risky. In the English case, moreover, only part of the hull was designed; the rest was still derived geometrically in the dockyard. Yet even with these limitations, scaled design did offer a more effective way than the algorithm-based Mediterranean system of keeping track of experimentation in the absence of material constraints.36 On the other hand, the use of scaled design made it possible to plan ships with more complex shapes. In the Mediterranean system, a single mould was sufficient to define the whole hull shape (except for the ends). This mould was used literally at midship section and at all sections between midships, while the end stations (about 10 percent of the ship length from the ends) were constructed on the basis of a rule of curvature or interpo- lant. Thus the variety of shapes was governed by the chosen midship sec- tion and by the few parameters of the longitudinal interpolant, which created section shapes that were close cousins of the midship section and did not permit much curvature. The introduction and improvement of scaled design allowed the English to introduce two interpolants, and the French to design ships with two or more (the number of interopolants defined the number of times the curve of the hull could be changed). This was a typical example of how technological latecomers could benefit from, and improve their predecessors’ experience.

Drawings and Models as Heuristic Devices Comparison between Venetian and Portuguese ship-drawings, whose sole purpose was to depict established building proportions for non-practitio- ners, and English and French scaled drawings, which aimed to establish new proportions for master-builders, suggests that we should not take the nature and purpose of design for granted. Consider the aesthetically stun- ning plans of Gothic cathedrals, the first of which depicts Rheims cathedral

36 D. McGee, ‘Ships, science and the three traditions of early modern design’, in: W. Lefèvre, J. Renn, and U. Schoepflin (eds), The power of images in early modern science (Basel: Birkhäuser, 2003), 24-46. transfer of technical knowledge and innovating in europe 47 in the mid-thirteenth century, and which seem at first glance to offer re- markably detailed building directions. In fact, many of these plans were presentation copies, drawn after the building was finished; others were drawn for the building commission, and thus differ substantially from the final product; none appear to have been actual working copies, used by the building lodge for practical purposes, because none were actually drawn to scale. There were two major obstacles to the practical use of Gothic drawings for building purposes. One was the use of geometrical rules in design. This had the advantage of being easily ‘portable’, since it did not rely on fixed measurements, but the method also generated irrational numbers (such as the diagonal of a square) that could not be easily reproduced on arith- metically proportioned plans. The second obstacle to the use of drawing was, paradoxically, the rediscovery by Filippo Brunelleschi of 3-point per- spective in early fifteenth-century Florence, which led his friend Leon Bat- tista Alberti to emphasise the use of ‘illusionism in architectural rendering’. As Alberti recognised, however, the perspectival method was of no use to planners and builders. It took three generations of Italian draftsmen to find out how to draw ‘plans and elevations, not according to the perspective method but by orthogonal projection, which … permits every element to be shown at the same scale, so that the carpenter and the mason can work from it’.37 But Alberti’s technical effort had another, more desirable conse- quence (from his point of view), which was to replace the master mason’s traditional role as surveyor and planner with the far more prestigious figure of the architect-designer. Plans, which avoid distortions whilst representing the spatial elements of the object so that it can be reproduced, were nonetheless practically unknown outside architecture before the seventeenth century. In particu- lar, the pictorial or illusionistic method persisted in the drawing of ma- chines. Although the degree of sophistication of machine representations grew markedly over the period between the early thirteenth-century sketches by Villard de Honnecourt and his colleagues, the fourteenth cen- tury designs by Guido da Vigevano, the fifteenth century drawings by Brunelleschi, Francesco di Giorgio Martini and Leonardo, and the six- teenth-century representations of mining machinery in Georgius Agricola’s

37 J.S. Ackermann, ‘Introduction’, in: W. Lotz, Studies in Italian Renaissance Architecture (Cambridge, MA-London: The MIT Press, 1977), xviii-xix. 48 s.r. epstein De re metallica, they were all in one way or another ‘false plans’, inasmuch as they left size, proportions and many essential details, undefined.38 The first systematic, measured plans of machines are, as we saw, those of English ships. Yet, as with architectural drawings, the development of graphic design in shipping may have been more a strategic element in the cultural and functional separation between designers and builders, than a genuine cognitive advance in the making of premodern ships. Certainly, the analogy raises the question—which cannot be addressed here—of the cognitive significance of graphic design for technological progress. One may simply note, that although the introduction of planning design un- doubtedly allowed greater flexibility in designing form, be it the form of buildings or the form of a ship, it is not self-evident that design effected a clear improvement for innovation in structure. From the late Middle Ages technicians were more likely to use 3-dimen- sional models in wood, clay, and gypsum to convey information about ma- chines (including buildings), and to test their performance. Like drawn plans, 3-dimensional models have two distinct uses: 1. to store information and to help communicate it from one person to another (e.g. designer to client, builder or supplier); 2. to help produce in the engineer and client the necessary level of confidence that the proposed structure will work and can be built.39 Although the use of 3-dimensional building models is at- tested as far back in time as Babylonian Mesopotamia, it became a more regular documented practice only in fourteenth-century Tuscany; a cen- tury later the use of models for building purposes was mentioned as a matter of established practise in architectural treatises by Leon Battista Alberti, Antonio Averlino, and Francesco di Giorgio Martini, with Martini making the cognitive aspects of model-building explicit: ‘Whereas it is difficult to demonstrate everything through drawings, nor is it at all pos- sible to express many things in words, … so it is necessary to make a mod- el of nearly every object’.40 Soon after 1500 the usage of building models spread to southern Germany and France, with the English following about a century later.

38 W. Lefèvre, ‘The limits of pictures: cognitive functions of images in practical mechan- ics – 1400 to 1600’, in: Lefèvre, Renn and Schoepflin (eds), The power of images, 69-89 39 W. Addis, ‘Introduction’, in: Addis (ed.), Structural and Civil Engineering Design, xiii-xliii. 40 F. di G. Martini, Trattati di architettura ingegneria e arte militare, ed. by Maltese, Corrado (Milan: Il Polifilo, 1967), 1, 142. transfer of technical knowledge and innovating in europe 49 Far less is known about the related practice of making scaled-down models of working machines. The earliest reference to a mechanical mod- el is found in a late fifteenth century description of a new wire-drawing machine invented in late fourteenth century Nuremberg.41 A few years later, in May 1402, the master masons at Milan cathedral were asked to inspect sketches submitted in a contest to find the best mechanical device for sawing stone blocks ‘without manpower’; the most promising design was then to be realised in the form of a wooden model in reduced size, suggesting a well-established combination of sketch-based and 3-dimen- sional mechanical planning, experimentation, and demonstration of ex- pertise.42 By the early 1500s scaled-down models were being used both in engi- neering competitions and for applications for technical patents. Models were commonest until the mid-sixteenth century in the two most advanced industrial regions of the time, north-central Italy and southern Germany, but thereafter they began to be used also in Spain and France. In the early decades of the sixteenth century a Nuremberg craftsman made a ‘nice wooden design for the king of England, about one Ellen long, in which one water wheel drove mechanisms for grinding, sharpening, polishing and fulling’, but this may have been an article for the king’s private collection; 3-dimensional models are first recorded in English shipbuilding in the early seventeenth century.43

Experimentation Despite the documented use of model machines from the 1300s, evidence of technical experimentation in premodern Europe is irregular and rarely indirect; some of it was reported previously in discussing building prac- tices. It was exceedingly rare for inventors, tinkerers, or technicians to write in any detail about their activities (as opposed to their speculations, like

41 B.C. Blake-Coleman, Copper wire and electrical conductors: The shaping of a technol- ogy (Chur: Harwood Academic Publishers, 1992). 42 M. Popplow, ‘Models of machines: a ‘missing link’ between early modern engineer- ing and mechanics?’, Max-Planck Institut für Wissenschaftsgeschichte Preprints, 225, 2002. 43 Ibid., 12. After the late sixteenth century models of machines increasingly became collectors’ items in Kunstkammern and articles for mechanical demonstration in the private homes of engineers and the public estabishments of scientific academies and engineering institutions. Model-based testing was central to the work of eighteenth-century engineers like Christopher Polhem (1661-1751), Antoine de Parciewux (1703-68) and John Smeaton (1724-92). In the same years, in a curious inversion of their origins in craft and engineering practice, reforming technical institutions briefly adopted machine models as a means to teach apprentices craft skills without submitting them to craft-based training. 50 s.r. epstein Leonardo) before the eighteenth century. However, two unusual sixteenth- century texts do shed light on kinds of experimental practice that under normal circumstances left no material trace, namely machine and chemi- cal testing. Some of the earliest evidence of individual testing and experimentation comes from the Venetian glass industry. According to local tradition, An- gelo Barovier was a Venetian glass-worker who during the 1450s invented new kinds of glass—crystal, lattimo, chalcedony and porcellano— which rapidly became the base for the siuccess of the Venetian glass industry throughout Europe. In fact, he seems to have been ‘an owner of a glass shop who carried out experiments purposefully intended to produce new glass compositions’. He was a friend of Paolo da Pergola, a humanist-philosopher who taught in Venice and lectured on ‘the combinations and transforma- tions of metals’; and a book published in 1500 states that Angelo took ‘the fruit of this speculation and put it into practice’. There is also documen- tary reference to what appears to be a series of experiments by the Barov- ier, Mozetto and d’Angelo families into these new types of glass in 1457 and 1460. As this suggests, Angelo Barovier’s successes were not the work of a solitary genius, but the outcome of a series of small-scale innovations stretching over the preceding century that included, most crucially, the purification of the alume catino ash that increased the amount of sodium (Na2O), and the discovery of an as yet unknown material that reduced the problem of cristallo glass corrosion.44 Although it is unclear if these early experiments were recorded in writing, the first known reference to a reci- pe book dates from 1446, and we know—because they survive—that by the sixteenth century it was normal practice for family-based glass-making firms to keep their own books of recipes or ‘secrets’. The description by Giuseppe Ceredi, a Paduan engineer, of his invention (or rediscovery) of Archimedean water-screws for drainage and irrigation purposes contains what may be the first suggestion in print to build mod- els at different specifications in order to optimize machine-building. Here is Ceredi’s description: ‘I was able to fabricate a great many models, small and large, adding, changing, and removing various things according to the condition of the material, or the grouping of many primary and secondary causes, or the variety of the mediums, or the proportions, or the force of the movers, or many other obstacles that hinder the thing sought. For it is well known by scientists [scientiati] that when things are put in operation,

44 McCray, Glassmaking, 98-100, 115. transfer of technical knowledge and innovating in europe 51 so numerous and great a heap of observations need to be kept in mind all together to hit on any new and important effect that it is almost impossible to fit them all properly together’. Having found that no uniform rules could be found concerning the optimum construction of water-screws, he ulti- mately determined that the best procedure would be to use a screw about 8 m. long, to raise water about 5 m. Ceredi was aware of scaling problems with machines, and proceeded accordingly. ‘To put this into execution’, Ceredi stated, ‘and have it based firmly on experience as guided by reason, it was necessary to make a large number of models, both small and large, now with one length and height of channels and now with another, in order to be able to proportion the whole to the mover [the screw] and to its organ [the crank]’.45 At about the same time, the French potter Bernard Palissy described how, over ten years, he slowly mastered how to combine the quality of clay, the pot’s thickness, the melting point, type, quality and colours of the enamel, the level and constancy of fire, and the pot’s position in the kiln to make Italian-style enamel (Fayence).46 Although narrated in the form and with the tropes of Reformed Christian salvation, the tale of Palissy’s struggle to control for the many variables of pot-making rings true in re- minding us that in chemical processes, visual and 3-dimensional models were of little use. Positive results could only be gained through an approach on the borderline between alchemical and craft practice, exemplified also by the recipe books for Venetian glassmaking. It is all very well to define the ‘scientific method’ as ‘accurate measurement, controlled experiment, and an insistence on reproducibility’. As Palissy noted, the problem with this ideal, to which in principle he subscribed, was to know what to measure and experiment with—something scientists would be no better at defining for nearly three centuries thereafter. So recipes were the solution—but recipes, as opposed to machines, were hard to transfer, because their results depended critically on a combination of material ingredients, and atmo- spheric and other conditions that could not be easily controlled for, and thus, easily reproduced. In sum, evidence of technical heuristics, codification, and appropriation shows some of the ways how existing and new craft and engineering knowl- edge was shared or ‘distributed’. However, knowledge sharing was more

45 G. Ceredi, Tre discorsi sopra il modo d’alzare acque da luoghi diversi (Parma: Seth Viotti, 1567). 46 B. Palissy, Oeuvres completes, ed. by M.-M. Fragonard (Mont-de-Marsan: Editions InterUniversitaires, 1996). 52 s.r. epstein likely and more intense within large-scale, hi-tec or high-value sectors like ship- and edifice-building, mining and metalworking, the making of clocks and scientific instruments, gold-smithing and silk weaving, and glassmak- ing, industries which displayed strong division of labour and advanced levels of coordination and where cooperation provided either clear econ- omies of scale and scope or marketing advantages—many of which are also notable for having played the most technologically innovative role in the Industrial Revolution.

Spatial Transfer of Technical Knowledge

Texts and Patents Thus far we have focused on how premodern technical knowledge was codified and shared. In order to fully answer the initial question of how premodern technical innovation was generated and sustained, we must also address the matter of how technical knowledge travelled. In theory, technical knowledge could be disseminated across space in three ways: through publicly available texts, through patents, and through migrating individuals. In practice, published, ‘disembodied’ technical knowledge did not disseminate well, as John Harris, a lifelong student of technological transfer between eighteenth England and France, concluded: ‘the craft nature of virtually all the technologies … meant that written descriptions and plans and drawings were only marginally useful’.47 Premodern technical writers seldom practiced what they described, and so typically overestimated the role played by explicit, propositional knowl- edge in craft and engineering practice. Written manuals were incomplete and sometimes misleading; they might contain technical details not actu- ally applied in solving the problem; and they left out crucial practising ‘tricks’. Such problems were compounded by the difficulties faced by ex- perts in describing what cues they responded to and what factors contrib- uted to their decisions. An investigation on the training of Spanish ship pilots for the Indies defended their alleged incompetence as follows: ‘even though a person is not very resolute in responding to the theory, [yet] he understands it well, and he who has experience understands it if he acts correctly, and there are many who don’t know how to propose or explain how

47 J.R. Harris, Industrial espionage and technology transfer. Britain and France in the eighteenth century (Aldershot: Ashgate, 1998), 549. transfer of technical knowledge and innovating in europe 53 to use an instrument, but with one in their hand use it very well’.48 The large tacit and non-linear component of experience-based knowledge explains why equally skilled experts in the same field disagreed on how to do their job,49 and why not a single premodern innovation was transferred through print alone. The most popular and sophisticated manuals, architectural treatises, were searched for formal motifs rather than for building techniques. The woodcuts in the most famous and extensively copied treatise, by Andrea Palladio (published 1570), were drawn in orthogonal projection and there- fore may have made it possible for architects to study building proportions; however, they gave little indication of construction methods or the use of materials, for Palladio like other treatise writers assumed that architects and builders would adapt his designs to local building traditions and to the availability of materials.50 Part of the popularity of Palladio’s treatise arose from this inherent flexibility. By contrast, most readers would have found the technical information on construction difficult to decipher from the illustrations alone. The English architect Inigo Jones, for example, learned the design principles of the orders and the fundamental planning issues of domestic architecture on his own; since he was not trained as a mason or carpenter, however, he needed to speak with workers and architects in order to learn practical building techniques. Between 1613 and 1614 he trav- eled to Italy for this purpose; on meeting the architect Vincenzo Scamozzi, Jones asked him for help with the technical aspects of vaults, noting in his diary: ‘Friday the first of August 1614 spoake with Scamozo in this matter and he hath resolued me in this in the manner of voltes’. Premodern patents faced similar technical and cognitive problems. Pat- ent law was first established at Venice in 1474 and spread rapidly either in law or in practice to the rest of Italy and northwards, first to the German principalities, then to France, Spain and the Low Countries, and subse- quently to England.51 By contrast with their modern counterparts, how- ever, premodern patent laws did not require novelty and originality; most

48 A.D. Sandman, Cosmographers vs. pilots: Navigation, cosmography, and the state in early modern Spain. Unpublished PhD-dissertation, University of Wisconsin-Madison, Madison, 2001, 276; emphasis added. 49 E.H. Ash, ‘“A perfect and an absolute work”: Expertise, authority, and the rebuilding of Dover harbour, 1579-1583’, Technology and Culture 41 (2000), 239-68. 50 L. Trogu Rohrich, Le tecniche di costruzione nei trattati di architettura (Monfalcone Gorizia: Edicom, 1999). 51 M. Frumkin, ‘Early history of patents for invention’, Transactions of the Newcomen Society 26 (1947-49), 48-56. 54 s.r. epstein patent descriptions were generic and did not remotely approximate a mod- ern blueprint; and innovations were seldom examined systematically be- fore the eighteenth century. Although some administrations, like Venice from the early sixteenth century, demanded a working model of patented machinery, inventors working on models were frequently unable to over- come scaling problems with full-sized machines, as noted by Giuseppe Ceredi in 1567.52 The problems arose particularly for large-scale mechani- cal inventions involved in power generation (milling, hydraulics, heating). In practice, patents were a means for towns or rulers to encourage the in- troduction of a new machine or process in their jurisdiction, by conceding a contingent monopoly over exploitation. Patents were also used as a means of commercial advertisement. Since patents tended to require cost- ly lobbying and upfront fees, and placed the entire burden of proof and investment risk on the inventor’s shoulders, barriers to entry to the technol- ogy market via patents were generally high. The propensity to patent was also affected by other factors. Many product and process innovations were never patented because they were better protected as trade secrets or be- cause they were part of the collective knowledge of a craft; for example, the makers of watches, clocks, and astronomical and other scientific instru- ments, most of who were organised in guilds, opposed patents that tried to privatise knowledge that was already in the craft’s domain or that were perceived to restrain trade.53 Consequently, premodern patent rights seem not to have played a major role in innovation before 1800.54 The assumption that patent rights to invention were necessary for pre- modern technological innovation rests on the view that intellectual cre- ation is non-rivalrous, and that once in the public domain, it can be copied at no additional cost. This fact may be true but is economically irrelevant, since what matters is the application of the new idea, which has learning and physical costs. In premodern manufacture, the costs of application arose from the largely implicit nature of technical knowledge, which cre-

52 Ceredi, Tre discorsi, 52; S. Drake, ‘An agricultural economist of the late Renaissance’, in: B.S. Hall, and D.C. West (eds), On pre-modern technology and science: A volume of studies in honor of Lynn Whyte, jr. (Malibu: Undena Publications, 1976), 53-73. 53 Epstein and Prak (eds) Guilds, innovation, and the European economy, 1400-1800. 54 C. MacLeod, ‘Accident or design? George Ravenscroft’s patent and the introduction of lead-crystal glass’, Technology and Culture 28 (1987), 776-803; MacLeod, Inventing the Industrial Revolution: The English patent system, 1660-1800 (Cambridge: Cambridge Univer- sity Press, 1988); L. Molà, ‘Il mercato delle innovazioni nell’Italia del Rinascimento’, in: M. Arnoux, and P. Monnet (eds), Le tecnicien dans la cité en Europe occidentale, 1250-1650 (Rome: École française de Rome, 2004), 215-50. transfer of technical knowledge and innovating in europe 55 ated the need for one-on-one training and meant that technological in- novations had to be transferred by travelling craftsmen and engineers.

Transferring Skilled Technicians: When and Why Did Craft Guilds Oppose Technical Innovation? In practice, technological transfer could only be successfully achieved through human mobility. However, successful transfer faced several ob- stacles. The most oft-cited, trade secrecy and guild opposition to innova- tion, were also the least important. As the previous discussion of technical heuristics makes clear, most so-called craft secrets were in fact open to anyone willing to train in the relevant craft and engineering practices. For example, although ‘Gothic’ geometrical principles for drawing elevations—developed around Paris between mid-twelfth and mid-thirteenth centuries—were said to be the closest guarded masons’ ‘secret’, they were actually shared by every trained mason north of the Alps. The application of Gothic principles was simply a practice that distinguished trained masons from everyone else, and there is no evidence of technical exclusivism.55 Similarly, the distributed char- acter of technical knowledge—institutionalized through apprenticeship, guild practice and division of labour, and the systematic circulation of skilled labour—meant that genuine technical secrets were hard to keep, if they were deemed useful. The belief that crafts were vowed to secrecy and exclusivism appears to have originated in the seventeenth century among the ‘new scientists’ and natural philosophers. Fascinated by technicians’ proven empirical knowl- edge of the material world, empirically-oriented intellectuals between the late fifteenth (Leonardo) and the early seventeenth century (Bacon, Gali- leo, Descartes) wrote admiringly about craft practices and craft knowledge. But their admiration was tinged with suspicion. They were unable to understand technical knowledge without extensive practice, and being unaware of the cognitive reasons for this, they found it hard to believe that illiterate or near illiterate technicians could know more about nature than they did. Thus, for example, reports of Royal Society experiments never

55 L.R. Shelby, ‘The “secret” of the medieval masons’, in: Hall and West (eds), On pre- modern technology and science, 201-22; E. Fernie, ‘A beginner’s guide to the study of archi- tectural proportions and systems of length’, in: E. Fernie and P. Crossley (eds), Medieval architecture and its intellectual context: Studies in honour of Peter Kidson (London-Ronce­ verte: Hambledon Press, 1990), 229-38. 56 s.r. epstein name the technicians who actually made and maintained the instrumen- tation and performed the experimentation.56 Second, the new scientists wished to distance themselves forcefully from the long-standing tradition of alchemy, which they associated not wholly justifiably with a strong desire for secrecy and with social and technical exclusivism.57 In this the new scientists followed the Scholastics, for whom ‘knowledge of [alchemical] secrets was strictu sensu impossible: they could be experienced, and could be found out ‘experimentally’, but they could not be understood or explained according to the canons of logic and natu- ral philosophy’.58 During the sixteenth century alchemists such as Paracel- sus, Girolamo Cardano, and Andreas Libavius deliberately associated their practices with craft activities and methods in order to emphasize their empirical, non-scholastic approach. Seventeenth-century new scientists were thus offered a ready-made conceptual framework, which stressed secretiveness and unreliability, into which to slot craft practices, and which moreover drew attention to the scientists’ self-declared intellectual open- ness. The third strand in the scientists’ emerging theory of craft practice arose from the new scientists’ concern with establishing a readily transportable method, whose principal aim was to codify the facts of the natural world into a universal language. This set them explicitly at odds with technicians, who they described as having no method at all: this was of course a mis- representation, for codification was also important for technicians, albeit as a means to the end of making things that worked rather than an intel- lectual end in itself. The claim that guilds systematically opposed outside innovations is also problematic. One reason is that it is excessively generic. If it is meant to say that guilds never innovated, it is, as we have seen, demonstrably false. In addition to European instrument-making, Venetian glassmaking, and

56 S. Shapin, ‘The house of experiment in seventeenth-century England’, Isis 79 (1988), 373-404. 57 W.R. Newman, ‘Alchemical and Baconian views on the art/nature division’, in: A.G. Debus and M.T. Walton (eds), Reading the book of nature: The other side of the Scientific Revolution (Kirksville, MI: Truman State University, 1998), 81-90; Newman, ‘Alchemical symbolism and concealment: the chemical house of Libavius’, in: P. Galison, and E. Thom- son (eds), The architecture of science (Cambridge, MA.: The MIT Press, 1999), 59-77; Newman, ‘Alchemy, assaying, and experiment’, in: F.L. Holmes, and T.H. Levere (eds), Instruments and experimentation in the history of chemistry (Cambridge, MA.: MIT Press, 2000), 35-54 58 W. Eamon, Science and the secrets of nature: Books of secrets in Medieval and Early ModernCulture (Princeton: Princeton University Press, 1994), 53. transfer of technical knowledge and innovating in europe 57 Lyonnais silk cloth production, studies of premodern guild subcontracting, and of Dutch painting in the Golden Age are equally conclusive on this point.59 Patterns of patenting in sixteenth-century Italy also show that guilds were in the forefront of testing and introducing technical innova- tions.60 If, on the other hand, the claim is meant to say that guilds would at some point become technically conservative, it loses any predictive value. The argument is also methodologically naive. Although it assumes that all innovations that were refused were better than current practice, the record seldom reveals whether guild opposition was driven by rent seeking or by an objective assessment of the innovation’s merits. Individual instances of resistance to change tell us little about relations between the guilds and technological progress in general. A theory of guild innovation must identify both the technical and the political criteria that dictated the choice of technology and established a given technological path. In principle, one would expect the crafts to prefer technology that privileged skill-enhancing, capital-saving factors. Despite a lack of system- atic research, evidence from patent records indicates that this was pre- cisely the kind of innovation that prevailed in England before the mid- to late eighteenth century, when the country’s guilds were still very active. Between 1660 and 1799, labour saving innovations accounted for less than 20 percent of the total, whereas innovations aimed at saving capital (espe- cially working capital) and at quality improvements accounted for more than 60 percent. There is no reason to believe that patterns elsewhere in Europe were very different.61 The response to innovation by individual crafts depended primarily on political rather than market forces. There was a fundamental difference in outlook between the poorer craftsmen, who had low capital investments and drew their main source of livelihood from their skills, and who there- fore (frequently in alliance with the journeymen) opposed capital-inten- sive and labour-saving innovations, and the wealthier artisans who were less threatened. Relations between the guild’s constituencies and the state also affected decisions. On the one hand, the wealthier and more innova-

59 C. Lis, and H. Soly, ‘Sub-contracting in guild-based export trades, thirteenth-eigh- teenth centuries’, and M. Prak, ‘Guilds and the development of the art market during the Dutch Golden Age’, both in: Epstein and Prak (eds) Guilds, innovation, and the European economy, 81-113 and 143-71 respectively. 60 Molà, ‘Il mercato’. 61 MacLeod, Inventing the Industrial Revolution, ch. 9; T. Griffiths, P.A. Hunt, and P.K. O’Brien, ‘Inventive activity in the British textile industry, 1700-1800’, Journal of Economic History 52 (1992), 892-95. 58 s.r. epstein tive masters were more likely to influence government policy, and under normal circumstances authorities seem to have allowed them to circum- vent guild regulations. On the other hand, city councils were more willing to meet the small masters’ concerns if labour saving innovations coincided with a serious economic downturn, both to ensure social and political stability and to restrain unemployed craftsmen from leaving the town. In other words, guilds were most likely to act as ‘recession cartels’ when eco- nomic circumstances took a turn for the worse, but they still required po- litical support to enforce cartel restrictions successfully against free riders and competing guilds. Thus, Dutch guilds began to resort systematically to restrictive policies when the country entered a long phase of stagnation after the mid-seventeenth century—but only after obtaining municipal approval.62 Similarly, craft guilds reacted very differently to the introduc- tion in the seventeenth century of a major technological innovation, the silk ribbon loom, according to their political situation within town and region, the competition from neighbouring industries, the degree of inter- nal stratification and their regulatory capacities. Their response to a labour saving and deskilling technological innovation was depended on external and internal market structures.63

Transferring Skilled Technicians: How Did It Work? Although most technical knowledge remained either unformulated or un- recorded, one should not confuse the absence of written texts detailing technical practice with technician’s fundamental commitment to secrecy. Rather, the absence of texts is evidence that writing (including, for many purposes, drawing) was a highly ineffective mode of transmission. As Pal- ladio’s work suggests, useful or experiential knowledge—knowledge that works—is, in principle, local. This does not mean that it is necessarily secret, or that it remains in an individual’s head: premodern technical knowledge was extensively socialized and shared. Some elements of expe- riential knowledge—in shipping, and to a lesser extent in building—were increasingly codified in writing. A partial result of written codification was

62 J. de Vries, and A. van der Woude, The First Modern Economy. Success, Failure, and Perseverance of the Dutch Economy, 1500-1815 (Cambridge: Cambridge University Press, 1997), 294, 340-41, 582; R.W. Unger, Dutch Shipbuilding before 1800. Ships and Guilds (Assen: Van Gorcum, 1978), ch. 5. 63 U. Pfister, ‘Craft guilds and technological change: The engine loom in the European silk ribbon industry in the seventeenth and eighteenth centuries’, in: Epstein and Prak (eds) Guilds, innovation, and the European economy, 1400-1800, 172-98. transfer of technical knowledge and innovating in europe 59 to make local knowledge less local, accessible both to the emerging profes- sional categories of designers and, in principle, to makers outside the original community of practitioners. Other experiential knowledge was embedded in objects, and objects could travel and be observed: ships could be seen, clocks could be taken apart, imported Chinese porcelain could prove that something deemed impossible, or unknown, could in fact be done. Strong evidence as to the effectiveness of technological transfer through migration comes from the observation, discussed previously, that techno- logical leadership moved over time from southern to northwestern Eu- rope—from Italy (1200-1450), to the southern Rhineland and southern Netherlands (c.1450-1570), to the Dutch Republic (1570-1675) and finally to Britain after c. 1675—largely thanks to skilled individuals trained by guilds or by other communities of specialized technicians (miners, builders, ship- builders etc.). Between c.1300 and c.1550, European craft guilds and polities devised institutional arrangements that sustained skilled workers’ mobility and raised the potential rate of technological innovation. Skilled migrant work- ers included mainly apprentices and journeymen, who travelled on a sea- sonal basis, or established masters, whose migrations were more often permanent. More systematic apprentice and journeyman mobility was an outgrowth out of the temporary skills shortages that followed the plague epidemics of 1348-50. By 1550 tramping was common in much of Western Europe, although it was only fully institutionalised in German-speaking Central Europe from the sixteenth century and less extensively in late sev- enteenth- and eighteenth century France. In England, independent jour- neyman organisations seem to have been formed after the decline of London as a national training centre from the 1680s. Since the main pur- pose of organised tramping was to coordinate information and allocate skilled labour more efficiently across regions, formal organisations never arose in densely urbanised regions like northern Italy and the Low Coun- tries where information costs were low.64 Apprentice and journeyman mobility helped develop and diffuse tech- nical knowledge within areas that were on the whole institutionally, eco- nomically and culturally similar or adjacent. The main source of innovation

64 S.R. Epstein, ‘Labour mobility, journeyman organisations and markets in skilled labour in Europe, 14th-18th centuries’, in: Arnoux and Monnet (eds), Le technicien dans la cité, 251-269; D.E. Wildasin, ‘Labor-market integration, investment in risky human capital, and fiscal competition’, American Economic Review 90 (2000), 73-95. 60 s.r. epstein in the late Middle Ages was Italy, and the main initial recipients, southern and central German-speaking territories. Cotton weaving, for example, was transferred to Germany from northern Italy in 1363, and by 1383 already its wares were being sold in large quantities on north European markets. One of the first cotton weavers—a ‘cotton-maker’ (Parchantmacher)—is men- tioned in Nördlingen in 1373, and ‘Milan’ and its declensions are frequent among the earliest weavers’ names—although the transfer was probably also facilitated by German merchants or by the return home of German weavers who had learned the craft in Genoa, Venice and Lombardy.65 Fol- lowing the craft’s speedy diffusion in upper Germany, regional industries there established the central European standards in cloth types and qual- ities, to which east German production conformed following the large-scale migration of upper German weavers to Leipzig between 1471 and 1550. Many East German towns adopted the Augsburg ordinances on cotton, and it is said that the flourishing of guilds in the region dates from the time ‘when the Swabians came flocking’.66 A first phase in diffusing the public clock occurred in 1370-80; by 1400 all major towns had their public clocks; and by 1500 the innovation had spread across the whole of Europe, albeit entirely thanks to migration of technical experts.67 The diffusion of papermaking in central Europe also relied on help from central and north Italian craftsmen. The hugely suc- cessful spread of book printing—which had been a purely German affair until 1465—was based on wandering printers and craft experts; already by 1472 Germany was importing Italian book characters via returning German printers. In the sixteenth century, thanks to Venetian migrants, transparent (‘Venetian’) glass began to be produced throughout Central Europe. The chances for apprentices and journeymen of accumulating technical skills and knowledge probably stood in direct relation to the length and radius of the tramping experience. Journeymen who travelled widely learned about regional differences in work organisation, and came to rec- ognize different practices, raw materials and products. The clearest evidence that itinerant journeymen could acquire additional technical skills comes from the existence of bans on migration, as among Venetian glassmakers and Nuremberg metalworkers. Nuremberg tried to protect its

65 W. von Stromer, Die Gründung der Baumwollindustrie in Mitteleuropa: Wirtschaft- spolitik im Spätmittelalter (Stuttgart: Hiersemann, 1978), 20, 31, 142. 66 G. Aubin, and A. Kunze, Leinenerzeugung und Leinenabsatz im östlichen Mittel- deutschland zur Zeit der Zunftkäufe (Stuttgart: Kohlhammer, 1940), 34ff. 67 G. Dohrn-van Rossum, 'Migration-Innovation-Städtenetze: Ingenieure und tech- nische Experten', in: Arnoux and Monnet (eds), Le technicien dans la cité, 291-307. transfer of technical knowledge and innovating in europe 61 technical primacy in the metal industries by banning any kind of emigra- tion: apprentices had to swear not to practise their craft anywhere else, journeyman tramping was forbidden, and to avoid the poaching of workers, masters and employers had to ply them with work and ‘not allow them any holidays’, or if necessary, provide them with holiday pay. Every so often Nuremberg’s town council proceeded against crafts like the wirepullers, which allowed journeymen to be lured by outsiders to whom they divulged manufacturing secrets. Over time, however, the lockout became counter- productive, inasmuch as it hindered Nuremberg craftsmen from traveling and acquiring new techniques elsewhere. Nascent monarchies and territorial states made it a point to attract new skills and technology from beyond adjacent regions. International compe- tition for skilled workers, for example for master cathedral builders, ex- isted already during the Middle Ages, but it increased markedly during the early Renaissance (c.1450-1550) in the western Mediterranean, and after the Reformation in north-central Europe, when European rulers made it policy to attract displaced craftsmen from enemy lands. The expulsion of the Jews from Catholic Spain and southern Italy in the late fifteenth cen- tury; of Walloons and Nederduits from the Habsburg Netherlands between the posting of Luther’s Theses (1517) and the Treaty of Westphalia (1648), which scattered about 100,000 skilled technicians and merchants across northern France, England, Germany, , Scandinavia (especially Swe- den), and the Dutch Republic; and of the Huguenots from France to, espe- cially, Geneva and England after the Revocation of the Edict of Nantes (1685) are just some threads in a complex web of religiously and politi- cally driven technical diffusion.68 From the mid-seventeenth century, mercantilist states engaged in an increasingly systematic promotion of domestic industry via industrial es- pionage and more deliberate and focused immigration policies; attempts by guilds and political authorities to stop skilled workers from migrating were stymied by weak administrations, state competition, and the in- creased circulation of correspondence, men and equipment.69

68 H. Schilling, ‘Innovation through migration: The settlements of Calvinistic Nether- landers in sixteenth- and seventeenth-century Central and Western Europe’, Histoire Sociale/Social History 16 (1983), 7-33; W.C. Scoville, ‘Minority migrations and the diffusion of technology’, Journal of Economic History 11 (1953), 347-60; Irene Scouloudi, Returns of strangers in the metropolis: 1593, 1627, 1635, 1639 (London: Huguenot Society of London, 1985). 69 D. Roche, Humeurs vagabondes: De la circulation des hommes et de l’utilité des voyages (Paris: Fayard, 2003); J.R. Harris, Essays in industry and technology in the eighteenth century: England and France (Hampshire/Brookfield: Ashgate, 1992). 62 s.r. epstein Each passing of the technological torch set in motion a period of rapid innovation in the new regional leader. Although technological leadership is hard to establish for this period, two measures are available. One is the technology of energy production, as suggested by Karel Davids, which ex- panded and improved systematically during the period we are concerned with, from timber (Ancient and medieval Mediterranean) to advances in water power (fifteenth-sixteenth century Southern Germany), from the extensive use of peat and, especially, wind (seventeenth-century Dutch Republic) to the systematic use of coal (seventeenth- and eighteenth-cen- tury Britain).70 Another measure of technological leadership is the produc- tion of scientific and timekeeping instruments, which followed roughly the same course, from Italy northwest to Britain—with a detour through six- teenth and seventeenth-century Paris in the case of instrument making. Being at the right place at the right point in time could be transforma- tive. Britain, for example, was a one-way technological debtor up to the late seventeenth century; between 1600 and 1675 it imported from the Conti- nent the most advanced techniques in metal smelting and forging, in the making of glass, pottery, guns and watches, scientific instruments, gold- smithing, wool, linen and silk cloth, and in hydraulic engineering and ag- riculture.71 The country’s position of dependence began to be reversed after c.1675, and already by 1720, the English Parliament had become so confident in native technical abilities, and so worried about international competitors, that it passed a law banning the emigration of resident tech- nicians. The two main impediments to technological transfer were thus informa- tion and transport costs, which restricted labour mobility, and the absence of a local skills base that could successfully apply incoming techniques. Exogenous innovation could be absorbed only if an adequate supply of technicians able to apply the new techniques was available: a major hurdle with transferring British coal-based technologies to non-coal based Con- tinental economies in the eighteenth century, for example, was the incom- patibility of the associated intermediate goods, parts and skills.72

70 Davids, ‘Shifts of technological leadership’, 340. 71 G. Hollister-Short, ‘Leads and lags in late seventeenth-century English technology’, History of Technology 1 (1976), 159-83; D. Mitchell (ed.), Goldsmiths, silversmiths and bankers: Innovation and the transfer of skill, 1550 to 1750 (Stroud / London: Alan Sutton / Centre for Metropolitan History, 1995). 72 J.R. Harris, ‘Attempts to transfer English steel techniques to France in the eighteenth century’, in: S. Marriner (ed.), Business and businessmen: studies in business, economic and accounting history (Liverpool: Liverpool University Press, 1978), 199-233. transfer of technical knowledge and innovating in europe 63 Transmission of the most up to date knowledge could therefore be excru- ciatingly slow. It took over a century to transfer Hollander paper beaters from the seventeenth-century Netherlands to eighteenth-century France because of a lack of good machine makers and repairers; eighteenth-­ century French metalworkers—who, significantly, were not organized in guilds—had no knowledge of high quality steelmaking that had been prac- tised in Germany, northern Italy, Sweden and England for up to two cen- turies before.73 Bottlenecks to technical transfer were relaxed over time by falling infor- mation and transport costs, which can be proxied reasonably accurately by trends in urbanisation, and in financial and other market integration.74 The most salient example of the correlation between technological leader- ship and urbanisation is premodern England, which was transformed be- tween 1650 and 1750 from a technological and under-urbanised semi-periphery to the most technologically innovative and urbanised coun- try in the West. The most plausible reasons for the correlation are the stan- dard Marshallian ones: economically successful towns attract skilled workers, whose pooling stimulates the growth of specialised intermediate goods industries; knowledge spillovers among firms increase; and reliable knowledge improves and increases with use. This model fits well with the evidence that premodern regional technological leadership followed com- mercial leadership, with a certain lag.75

Conclusions

Notwithstanding the absence of much written evidence, evidence from technical practice suggests that premodern non-scientific technical knowl-

73 L.N. Rosenband, Papermaking in eighteenth-century France: Management, labor, and revolution at the Montgolfier mill 1761-1805 (Baltimore: The Johns Hopkins University Press, 2000); C.S. Smith, ‘Introduction’, in: A. Grünhaldt Sisco (ed.), Réamur’s Memoirs on steel and iron: A translation from the original printed in 1722 (Chicago: University of Chicago Press, 1956), vii-xxxiv. 74 P. Bairoch, J. Batou and P. Chèvre, La population des villes européennes 800-1850: banque de données et analyse sommaire des résultats (Geneva: Droz, 1988); S.R. Epstein, ‘Town and country in Europe, 1300-1800’, in: S.R. Epstein (ed.) Town and Country in Europe Between the Fourteenth and the Eighteenth Centuries (Cambridge: Cambridge University Press, 2001), 1-29; L. Neal, ‘How it all began: the monetary and financial architecture of Europe from 1648 to 1815’, Financial History Review 7 (2000), 117-40; K.G. Persson, Grain markets in Europe, 1500-1900: Integration and deregulation (Cambridge: Cambridge Univer- sity Press, 1999). 75 Davids, ‘Shifts of technological leadership’. 64 s.r. epstein edge expressed significant degrees of abstraction, experimentation and cumulation. There is also strong evidence that premodern technicians codified heuristic rules in response to growing pressure for standardization and rising mobility of skilled workers. Finally, the process of codification was dynamic, in two ways. On the one hand, the technology of codification was improved and its usage vastly extended over the period under consid- eration. Printing played a role in this, but it was arguably less important than the falling cost of paper to draw on. On the other hand, to an extent we still barely understand, the process of codification was cumulative. Drawings, models, recipes, and lists of proportions could circulate inde- pendently of their authors and outlive them, although it still required ­experiential knowledge to interpret them. Premodern technical progress was sustained and limited by the manner by which generic technical knowledge was codified and by ‘collective invention’.76 Premodern technical codification faced three important cog- nitive limitations, which it shared in several ways with contemporary natural philosophy. First, premodern technicians, like seventeenth and eighteenth century natural philosophers and their modern counterparts, faced the problem that tacit knowledge—both ostensive knowledge, and knowledge inexpressible in natural language—is ubiquitous and unavoid- able; thus, written codification was, by definition, always incomplete. Sec- ond, premodern technicians, like natural philosophers, faced the problem that some kinds of knowledge were more easily codified and transferred— via proportions and ratios, diagrams, models and ‘recipes’—than others. Thus, technical knowledge related to chemistry and metallurgy was hard- er to mobilize, because the character and quality of inputs was more vari- able, and because the final product could not be easily ‘reverse engineered’ to reveal its underlying manufacturing process. Lastly, premodern technol- ogy’s empiricism made it hard for technicians to distinguish clearly be- tween theoretical structure and form; a similar difficulty may explain the inability of most premodern natural philosophy to generate technologi- cally fungible science. Technicians extrapolated experiential knowledge from empirical observation of what worked within a given set of material circumstances and practices. They produced second order codifications of practice, rich in information, able to capture a high degree of variance in information, but possessing limited predicitive powers. Although prac- tices and practice-based algorithms gave broad scope for technical im-

76 Allen, ‘Collective invention’; Epstein, ‘Labour mobility’. transfer of technical knowledge and innovating in europe 65 provements, they offered little information on how a set of rules with different premises would affect a known technical process. In other words, each set of rules came with a corresponding bundle of practices. In principle, the weak distinction between structure and form, between rules and practice that we saw at work in cathedral and shipbuilding, raised the costs of switching from one set of rules to another. In practice, how- ever, these constraints were less serious than those coming from restric- tions to information flows, for there is no reason to believe that most premodern technologies, based on empirical practices and available ma- terials, had reached their technical frontier even by 1800. The most severe restrictions to premodern technological reliability and innovation arose from the high information and reproduction costs related to experience- based knowledge. The principal source of diminishing returns to technical knowledge seems to have been the cost of communication between dis- persed craftsmen and engineers, rather than the narrowness of the pre- modern crafts’ epistemic base. Although in principle tacit knowledge should have raised the appropri- ability of rent streams from invention, in practice appropriability was rather low, because the system of apprenticeship training and the use of a mobile skilled labour force made it difficult for individuals to protect tech- nical secrets. Since patent laws and patent concessions were commonplace but ineffective, and displayed high barriers to entry, incentives for indi- vidually driven innovation were rather weak. Most technical knowledge within industrial regions or districts with integrated skilled labour markets would have been shared, but technological transfer over long distances was inherently rivalrous, because it required non-local patterns of expertise to be applied successfully. A distinctively European technological system emerged from the late eleventh century, based on craft-based apprenticeship training, non-­ ascriptive membership of craft associations, and, increasingly, inter-state competition for skilled workers. These three elements defined a set of ­necessary and sufficient endogenous conditions for the generation, codi- fication and circulation of reliable technical knowledge.77 The craft guild enforced the rules of apprenticeship against free riding and exploitation. Second, it offered institutional, organisational and practical support to the migrant apprentices, journeymen and masters who transferred their tech-

77 S.R. Epstein, ‘The rise of the West’, in: J. Hall and R. Schroeder (eds) An anatomy of power: The social theory of Michael Mann (Cambridge: Cambridge University Press, 2006), 243, 257-58. 66 s.r. epstein nical knowledge from one town and region of Europe to another. Third, it supplied incentives to invention and knowledge sharing that the patent system did not by enforcing temporary property rights over members’ in- novations. Notably, only the first effect was the outcome of deliberate policy; the other two were unintended consequences of the club goods that the craft supplied its members. Lastly, and critically, the crafts’ jurisdiction was limited in space. Consequently, even the most notoriously restrictive crafts, the glass guild of Venice and the metallurgical guilds of Nuremberg, were unable to stop their members from migrating elsewhere.78 In the long run, Europe derived its unusual technological momentum from the mobility of its skilled labour. Mobility was the result of three forces.79 First, there was a great deal of ‘ecological’ variation in demand across Europe. Second, there were many polities whose rulers’ peaceful demand centred on court consumption and somewhat disconnected re- sources created spatial and temporal variation in demand for skills—thus ensuring a high rate of rotation and a form of technological competition. Third, the same polities were also in persistent and long-term military competition. However, the first factor is not distinctive of Europe, and the second factor would have most probably resulted in a long-term equilib- rium. On the other hand, despite its short-term costs, warfare within a system of competing states maintained the economic system in a process of dys-equilibrium. The periods of most damaging conflict—the late me- dieval Hundred Years’ Wars, the sixteenth-century Wars of Religion, the seventeenth-century Thirty Years War—generated huge shocks to indi- vidual regions and drove large numbers of skilled technicians away from their homes. Significantly, these periods of more intense warfare coincided with surges in technical innovation and in the transfer of technological leadership between regions. Growing state competition and urbanisation also reduced the costs of technical dissemination over time. Urbanisation offered increased oppor- tunities for exchanging knowledge, higher average quality of labour, a greater likelihood of matching skills to demand, and stronger incentives for the codification of knowledge. Although it is not a priori clear whether high urbanisation attracted skilled migrants, or whether migration (driven by exogenous factors like war) caused high urbanisation, the evidence

78 McCray, Glassmaking; Turner, ‘“Not to hurt of trade”; H.-P. Lanz, ‘Training and work- shop practice in Zürich in the seventeenth century’, in: Mitchell (ed.), Goldsmiths, silver- smiths and bankers, 36. 79 I owe this formulation to Jean-Laurent Rosenthal. transfer of technical knowledge and innovating in europe 67 points to the primacy of the former, pull factors, specifically of urban com- mercial success. Migration by skilled workers allowed new technological leaders to shift rapidly to the technological frontier, recombine foreign with domestic knowledge, and innovate further. The acceleration of technical innovation during the eighteenth century is more likely to have been caused by increasingly mobile and better-informed technicians sharing both propositional and prescriptive knowledge than by an intellectually driven ‘Industrial Enlightenment’.80

80 These conclusions are thus partly at odds with Mokyr’s recent argument that the Scientific Revolution and its cultural expression in the ‘Industrial Enlightenment’ were at the root of the first Industrial Revolution, because they provided the forms of ‘propositional’ knowledge that technicians lacked [Joel Mokyr, The Enlighted Economy: An Economic His- tory of Britain, 1700-1850 (New Haven: Yale University Press, 2009)]. Although I am sceptical about the significance of eighteenth-century natural philosophy for contemporary techni- cal progress, I agree with Mokyr about the importance of information flows for inducing technicians to travel, and possibly— though less testably—to increase their rate of exper- imentation. 68 s.r. epstein apprenticeship and industrialization in india 69

CHAPTER TWO

APPRENTICESHIP AND INDUSTRIALIZATION IN INDIA, 1600-1930

Tirthankar Roy

Vocational training represents a well-known case of market failure.1 Train- ees are often unable to self-finance instruction, and, if the skills are trans- ferable between employers, employers may be unwilling to instruct.2 This prospect explains important features of pre-modern apprenticeship sys- tems, in particular the role that corporate organizations played in regulat- ing apprenticeship, employment, and mobility between jobs. In the past, corporate regulation of apprenticeship came in many variants, and grew stronger or weaker, depending on market conditions, industrial organiza- tion, and available means of enforcement. Two stylized models of training- cum-regulation, however, can be distinguished. One of these involved relatively open recruitment of apprentices, and regulation by means of an indenture contract recognized by the state or guilds. The other system involved recruitment restricted by ethnicity and kinship, and regulation of younger workers by their intimate relations. This distinction is stylized, and the two models overlapped often. And yet, the distinction was not a trivial one. It was visible, for example, in the relation- ship between the states and the corporate bodies whose rights to regulate the contract the former upheld. In the first model, the relationship was legalistic and statutory. In the latter, the relationship was predicated on a moral right of communities to govern themselves. Arguably all large man- ufacturing regions in the early modern world, India included, accommo- dated both types and variants between them. However, relatively speaking, the latter exercised a wider scope in India.

1 I am indebted to Patrick Wallis, Maarten Prak, and an anonymous reader, for many helpful comments and suggestions, which led to significant improvements on the earlier drafts. 2 The distinction between general and on-the-job training, the source of this problem, owes to G.S. Becker, Human Capital: A Theoretical and Empirical Analysis, with Special Reference to Education (Chicago: Chicago University Press, 1993), 34-35. Trainees tend to pay for on-the-job training by accepting a lower-than-market wage. 70 tirthankar roy With indenture contract, although in some cases the apprentices did contribute to the cost of training, in most cases, the nature of training al- lowed a length of time when the productivity of the apprentice exceeded the wage paid, and the training cost could be recouped by the master.3 At that point, of course, the apprentice would entertain a strong desire to escape. Enforcement of the indenture involved the agency of the state or the guild. A well-known case of statutory enforcement in England was the late-sixteenth century Tudor legislation, the Statute of Artificers. Penal clauses of the Statute could be in principle used to put the unqualified practitioners out of business, though in practice the law served to inflict a bribe on them.4 In western European cities, on the other hand, guilds reg- ulated labour market transactions, monitored and certified quality of train- ing, and regulated competition in the commodity markets.5 Why were corporations powerful? In one view, ‘the states of early modern Europe grew much faster than the economies that sustained them’, a situation forcing them to offer various concessions to corporate bodies.6 Recruit- ment of apprentices was wide open in both cases. In the English case, apprenticeship was one of the main routes that the under-employed agri- cultural population took in order to acquire industrial skills.7 In both cases, furthermore, the regulatory apparatus suffered from weaknesses. Statutes were either powerless or not invoked often.8 Merchants exercised choices

3 I rely on E. Perroy, ‘Wage Labour in France in the Later Middle Ages’, Economic History Review, 8 (1955), 232-39; and Jane Humphries, ‘English Apprenticeship’, in: Paul David and Mark Thomas (eds), The Economic Future in Historical Perspective (Oxford: Oxford University Press, 2003), 73-102. The difference between the statutory length of pre-modern apprenticeship, seven years in many cases, and the time that shop-floor training took in the early twentieth century, 5-6 years, is a rough indicator of the reward accruing to the master. Patrick Wallis disputes the conventional wisdom that training and implicit repay- ment were sequential, suggesting that the masters rewarded themselves in a different way, ‘Apprenticeship and Training in Premodern England’, Journal of Economic History, 68 (2008), 832-61. 4 Margaret Gay Davies, The Enforcement of English Apprenticeship 1563-1642 (Cam- bridge Mass.: Harvard University Press, 1956). 5 For two contrasting interpretations of the externalities of corporate regulation, see Sheilagh C. Ogilvie, ‘Institutions and Economic Development in Early Modern Central Europe’, Transactions of the Royal Historical Society, 5 (1995), 221-50 and S.R. Epstein, ‘Craft Guilds, Apprenticeship, and Technological Change in Preindustrial Europe’, Journal of Economic History, 58 (1998), 684-713. 6 Ogilvie, ‘Institutions’. 7 Humphries, ‘English Apprenticeship’, 93-5. 8 On causes of decline of English apprenticeship, see K.D.M. Snell, Annals of the Labouring Poor: Social Change and Agrarian England 1660-1900 (Cambridge: Cambridge University Press, 1987), 228-69. Snell discusses two views on the high and increasing inci- apprenticeship and industrialization in india 71 between alternative suppliers, at times preferring the un-apprenticed lo- cated in the countryside over those qualified. The time-serving dimension of indenture was not necessarily conducive to good quality training. Finally, societies populated by European settlers presented an ambiguous case, containing indentured servitude in less skilled work, and an un-­ apprenticed and unregulated labour force in the more skilled work.9 Pre-modern apprenticeship, then, can be usefully compared on three points —patterns of training, corporate control, and stability of the system. When we compare India with Europe on the first two points, we observe that, with a few significant exceptions, the apprentices were members of the masters’ households in India.10 By being a part of the family economy, apprenticeship became informally ‘bounded’. Means of enforcement can only be inferred, and apparently involved some form of withholding of community goods such as, protection, inheritance, marriage, or honour. One apparent outcome of the system was, the term of apprenticeship was longer in India. Indeed the term could not be easily defined at all. Sources on the Indian crafts remained conspicuously silent on particular categories of young workers that figure prominently in the Europeanist literature, such as ‘runaways’, ‘dropouts’, and in a different context, ‘paupers’.11 This difference too can be explained in terms of the dominance of the family and community in India. It is not easy to run away from one’s relations. Beyond this difference, India and Europe were similar in some respects. In both regions, kings felt obliged to offer privileges to corporate bodies of skilled professionals. There is a great deal of evidence showing that the right of endogamous communities in possession of valuable skills to govern themselves was respected by the states and enshrined in books on state- craft.12 And in both India and Europe, premodern regulation ultimately had to change or adapt as migration, commercialization, and factories dence of dropouts and runaways, an optimist one suggesting that demand for workers encouraged dropouts, and a pessimist one suggesting that uncertain job prospects did. 9 Bernard Elbaum, ‘Why Apprenticeship Persisted in Britain But Not in the United States’, Journal of Economic History, 49 (1989), 337-49. 10 The term ‘apprentice’ occurs mainly in a contractual setting, and is perhaps not altogether appropriate in a family economy. Youth workers or trainees may be preferable. However, I will continue to use ‘apprenticeship’ in order to avoid confusion and keep the comparative project in focus. 11 The runaway problem informs interpretations of apprenticeship, see Wallis, ‘Appren- ticeship and Training’, and of urban society, see Ilana K. Ben-Amos, ‘Failure to Become Freemen: Urban Apprentices in Early Modern England’, Social History, 16 (1991), 155-72. 12 T. Roy, Company of Kinsmen: Enterprise and Community in South Asian History 1600- 1950 (Delhi: Oxford University Press, 2010). 72 tirthankar roy introduced new paradigms of training. Just as in other old world societies, vestiges of former training-cum-regulatory institutions carried over into modern India. And yet, the carryover of craft culture was modified by a number of circumstances, the most important one being that the work- force in the early Indian mills consisted of peasants in the main, and not artisans. The hierarchy of workers inside the factory was influenced by social segmentation, but it had little to do with the craft culture of appren- ticeship. The present essay illustrates this systemic transition in industrial educa- tion. The next three sections deal with, respectively, descriptions of com- munity mode of training in early modern India, craft culture of training in colonial India, and the new paradigm of training and regulation in the modern factories.

Craft Culture in Early Modern India

European visitors to Indian port cities in the seventeenth and eighteenth century observed that significant social interaction was denied the most skilled and wealthy occupational classes in India. ‘No one marries but in his own trade or profession’, François Bernier noted of North Indian society in the time of Emperor Aurangzeb (reign 1658-1707), ‘and this custom is observed almost as rigidly by Mahometans as by the Gentiles, to whom it is expressly enjoined by their law’.13 Thomas Bowrey remarked in the con- text of the seventeenth century Bengal, that the professions ‘are not admit- ted to marry one Occupation with another. A Merchant must marry a Merchant’s Daughter, a Weaver a Weaver’s Daughter, a Taylor with a Taylor’s Daughter, A Gold Smith with a Gold Smith’s Daughter, and soe the rest’.14 In contemporary Surat, J. Ovington saw the same rule at work, [e]ach single Trade is diversified by some particular Opinions; the Goldsmith, and Scrivan, the Joyner, Barber, and Merchant, & c. as they have different Employments, so are they of divers Sentiments, and distinguished in the Ceremonies of their Worship; and mix no more in their Sacred Sentiments of Religion, than in their Civil Arts. Therefore all their Arts are Hereditary, and their Employments confin’d to their own Families. The Son is engag’d in the Father’s Trade, and to maintain the Profession of it in his Posterity, it is transmitted always to the succeeding Generations, which is obliged to

13 Travels in the Mogul Empire A.D. 1656-1668 (London: Humphrey Milford, 1916), 259. 14 Thomas Bowrey, A Geographical Account of Countries round the Bay of Bengal, 1669 to 1679 (Cambridge: Haklyut Society, 1895), 31. apprenticeship and industrialization in india 73 preserve it in a lineal descent, uncommunicated to any Stranger. Upon this account, all Marriages are restricted to one Sect, and contracted only between persons of the same Perswasion and Profession.15 In the seventeenth century, Pietro Della Valle (Italian traveller), John Fryer (an East India Company servant) and Alexander Hamilton (a Scottish sea captain), wrote about the Indian merchants and artisans in similar terms.16 In the late eighteenth century, the Carmelite priest Paolino da San Bartolo- meo saw that in Travancore society, ‘the painters; ... dyers of cloth ... gar­ land-makers; ... smiths; ... coblers; also the weavers, taylors, carpenters, silver-smiths, clockmakers, and other artisans ... form separate classes, the members of which cannot eat with each other, and much less intermarry’.17 Bernier, Bowrey, San Bartolomeo, Niccolao Manucci, Luke Scrafton, and many others besides used the word ‘law’ when discussing these social con- ventions. The usage was not accidental. Surely, merchants and artisans elsewhere in the world at times connected marriage with professional in- terests. In India, the connection appeared to them to have a particular moral force. Not marrying within the community invited punishments, and contemporary literature provides evidence on such practices. The word also referred to the juridical autonomy of communities in the eyes of the local states. San Bartolomeo described urban society to be segregated, and each segment, containing a cluster of interrelated families plying one trade, in possession of the supervision of ‘one of their countrymen, ... who is obliged to preserve peace and good order among them, and to take care that they do not transgress the laws’.18 I do not wish to address the question, how authentic these reports were as descriptions of Indian society at large. Clearly, the European observers saw, and came in close contact with, only a small slice of society, popu- lated by those in possession of useful skills highly valued in the world economy of their time, and nearly all were residents of trading-manufac- turing towns. They rarely had occasion to watch the peasantry, or indeed had any interest in doing so. In this essay, however, the skilled artisanate is of interest, which was part of the observed set of people. The remarks sug-

15 A Voyage to Surat in the year 1689 (London: Humphrey Milford, 1929), 165. 16 A New Account of the East Indies being the Observations and Remarks of Capt. ­Alexander Hamilton, vol. 1 (London, 1739), 316. 17 A Voyage to the East Indies, containing an account of the Manners, Customs, & c. of the Natives With a Geographical Description of the Country. Collected from Observations made during a Residence of Thirteen Years, between 1776 and 1789 (London: J. Davis, 1800), 308. 18 Ibid., 8. 74 tirthankar roy gest that the commentators found in this world a practice of recruitment and training quite different from that prevailing in their own milieu, and one that was organically linked with norms of social interaction. They found professional guilds that were also endogamous. Several of these commentators were struck by the effect that this exclu- sivity had on the vocational training of young people. Bowrey wrote, ‘[e] very man must Consequently traine their Sons up to the Occupation he is of himselfe, and not assume any Other.’ A century later in a different part of India, San Bartolomeo observed, ‘every child receives the same educa- tion that was given to his father’. Some of his remarks show that he consid- ered the Indian system of education to be distinct from the western European one on the rarity of general training in India. All learning was specific to a trade. ‘Hence it happens that the Indians do not follow that general and superficial method of education by which children are treated as if they were all intended for the same condition, and for discharging the same duties; but those of each cast are from their infancy formed for what they are to be during their whole lives’.19 On the other hand, the apprentice- ship that the Indians received, being bound up with their status as the younger members in the masters’ family, took an interminably long time, and ‘enabled farther to improve [skills] and bring them nearer to per­ fection’.20 According to these observers, then, one effect of the family-bound ap- prenticeship system was that it could on occasions deliver exceptionally long periods of training, and therefore, very high levels of dexterity in using a fixed set of tools. The learning period for the younger generation never formally ended while the head of the family firm continued to work. In this fashion, the family could produce a highly efficient form of training in complex manual work. The dexterity aspect was a subject of admiration and wonder to both traders and scholars alike, and can be understood with reference to hereditary apprenticeship. It had a downside, however.21 Family-bound training tended to be conservative about technology insofar as the use of unconventional tools, materials, and ideas would have

19 Ibid., 257, 267. 20 Ibid., 268. 21 ‘In India rare skills transmitted through customary and family practices were not easily reproduced’, Maxine Berg writes, in order to explain the persistence of comparative advantage of early modern Indian craftsmen in supplying luxuries to world trade, ‘In Pursuit of Luxury: Global History and British Consumer Goods in the Eighteenth Century’, Past and Present, 182 (2004), 85-142. apprenticeship and industrialization in india 75 demanded relatively easy access to extra-family training.22 By the same logic that the family-bound training delivered deterity, the absence or weakness of extra-family training would reinforce conservatism and sim- plicity of the tools, another feature noted by the contemporary observers. In design-intensive crafts, the family system could excel, both by develop- ing designs and by keeping these secret quite effectively. Where extra- family training was more usual, Western Europe for example, innovation on tools and instruments were more likely than it was in India. These sources did not comment in enough detail on a different pattern of recruitment and regulation that characterized the karkhanas of Mughal Indian cities. Karkhanas were large workshops sponsored by the political elite, which manufactured goods for direct consumption by the elite as well as for overland trade.23 About 1650, the major cities in the Ganges-Indus flood-plains were home to about 1.4 million people. If a third of this popu- lation were connected with the crafts, we have a rough idea of the scale in which these workshops operated. By any benchmark, they catered to one of the richest clientele in the world, who had command over between one- fourth and one-third of the taxed income of the empire. These workshops did not only make goods for the local elite. In some cases, they also made money for their owners. Although they produced a variety of goods, the ones that provided scope for structured master-apprentice relationships were the ones that involved complex craftsmanship, such as carpets, em- broidery, inlayed metals, shawls, lace, and designed woven stuffs. These production sites saw concentrations of male wage-workers, who were trained and supervised by master artisans. It can be presumed that an exclusive recruitment strategy of the kind that characterized the fam- ily and community setting was neither necessary nor practical in this sphere. It was not practical because the scale of operation was often larger and dictated by the consumers. Having to recruit apprentices only from the immediate family would mean a constraint. It was not necessary be- cause the endogamous guild was redundant as a regulatory device and as embodiment of civil law. The masters belonged in a quasi-political elite themselves. Formal professional guilds were absent too. The masters were recruited directly, and if there was a club of some kind, it served only to

22 On the argument that hereditary learning posed an obstacle to innovation on tools, see Tirthankar Roy, ‘Knowledge and Divergence from the Perspective of Early Modern India’, Journal of Global History, 3 (2008), 361–87. 23 More on the karkhanas, see T.Roy, ‘The Guild in Modern South Asia’, International Review of Social History, 53 (Supplement) (2008), 95-120. 76 tirthankar roy regulate training and had otherwise restricted jurisdiction. Possibly be- cause of its position straddling politics and industry, the karkhana world has left little documentary evidence of its own on employment and wages. These systems of training and regulation were transformed by Indo- European trade of the eighteenth century, the end of the trade in the early nineteenth century, the collapse of the Mughal empire, and the rise of new centres of power and industrial patronage. At present, we know too little about the implications of these processes for the skilled artisans. The dominant analytical narrative on the period of transition is known as ‘de-industrialization’. The de-industrialization school claims that the crafts, and with it the embodied knowledge and the institutional legacy thereof, disappeared under the onslaught of imported manufactures from Britain in the first half of the nineteenth century.24 This perspective now stands revised. While there is evidence that many artisans had changed jobs or migrated away in the nineteenth century, a large number (possibly 10 million) also continued in business at the end of the century. Survival on such an enormous scale suggests that globalization did not destroy the crafts, but polarized them into a vulnerable segment and a resilient one. The former consisted of producers of intermediate goods such as metals, cotton yarn, dyes, wool, silk; and village crafts serving the needs of the peasant. Technological standard in these spheres was falling too quickly too far behind the European one from the early nineteenth century. Ma- chinery, whether installed in India or in Europe, destroyed intermediate goods production in the craft workshop; and changing standards of con- sumption weakened the village artisans. On the other side, many amongst the ‘survivors’ were skilled, that is, employed craftsmanship in innovative new ways in fields of production that offered scope for such innovation. Decline here was muted and restricted to remote towns serving local ­patrons. The extent of urban concentration increased, and there was a qualitative change in markets. From political bosses, their market shifted towards export, the expatriates, and the rising Indian middle classes.25

24 For general statements, see A.K. Bagchi, ‘Deindustrialization in India in the Nine- teenth Century: Some Theoretical Implications’, Journal of Development Studies, 12 (1976), 135-64. Irfan Habib, ‘Colonialization of the Indian Economy, 1757-1900’, Social Scientist, 3 (1975), 20-53. 25 Reinterpretations of crafts history can be found in T. Roy, Traditional Industry in the Economy of Colonial India (Cambridge: Cambridge University Press, 1999); ‘Acceptance of Innovations in Early Twentieth Century Indian Weaving’, Economic History Review, 55 (2002), 507-32; and Douglas Haynes, ‘The Logic of the Artisan Firm in a Capitalist Economy: Handloom Weavers and Technological Change in Western India 1880-1947’, in: B. Stein and apprenticeship and industrialization in india 77 The consequence of this transformation was a retreat of the household. But the retreat was a protracted one, change and evidence on craft organi- zation until well into the twentieth century continued to be dominated by the family, as we see in the next section.

Craft Culture and Apprenticeship in Colonial India

Between 1961 and 2001, the proportion of family workers in industrial work- force fell from 60 per cent to 15 per cent. If we apply a slightly lower rate of change to the early twentieth century, about 1900, 80-90 per cent of indus- trial workers worked inside households, the rest being accounted for by factories and small workshops. Industrial occupations were followed by 14.3 million males. Male dependents in such households numbered an estimated 8 million. Assuming half of this number consisted of youth in the apprenticeship age-range, the proportion of apprentices in the work- force should be in the neighbourhood of 22 per cent.26 The number is similar to the ones that Leonard Schwarz reports for London in the seven- teenth century (19-25 per cent), but considerably higher than London fig- ures for 1700 (12).27 A direct comparison would be misleading, however, for the proportion for India is the proportion of youth workers, not all of them may have actually received training. At any rate, at the end of the nineteenth century, the family was strong and the main vehicle of training. There were two divisions of labour inside a family—one based on sex and the other based on age. Inside a weaving household, one would generally see adult men working as weavers, adult women on winding and sizing, and children as assistants in both weaving and winding according to gender. In designed garments, one task was usu- ally reserved for children, manipulation of the drawing thread and the dobby (but not working the jacquard). The houses of more substantial silk weavers were usually two-storied, the upper story containing the workshop. In one such workshop located in a western Indian textile town, a visitor in 1900 found ‘three or four children .. seated underneath the loom in a ring, laboriously spelling Urdu sentences, while their father wove the fabric

S. Subrahmanyam (eds), Institutions and Economic Change in South Asia (Delhi: Oxford University Press, 1996); ‘Artisan Cloth-Producers and the Emergence of Powerloom Manu- facture in Western India 1920-1950’, Past and Present, 172 (2001), 170-98. 26 The proportions are derived from Census occupational statistics. 27 ‘London Apprentices in the Seventeenth Century: Some Problems’, Local Population Studies, 38(1987), pp. 18-22. 78 tirthankar roy above their heads, pausing now and then in his task to explain to them some crucial point’.28 While doing such auxiliary work, ‘through observa- tion and practice they gradually pick up the skill of weaving’.29 Both boys and girls in weaver families did such work. But boys were groomed for weaving, whereas the girls stopped working, just before both reached their teens. In silk, boys started working a loom when they were about 12, and contracted on their own when they were 20.30 Having a larger family size with many boys was helpful to weavers; there is confirmation that some urban weaving communities did have larger families. Work under parental authority has been described in very dark language by one author writing in the 1920s. ‘[P]arents are indeed veritable exploit- ers ... and this exploitation, however merciless, inhuman and unjust, is shielded by so-called parental authority and filial devotion. The evil de- creases in intensity and barbarity as the earnings of the families increase’.31 There was an alternative to working under the direction of a parent, and that was working in the home of a family friend. Neighbourhood hiring at all ages involved only boys. In South Indian silk weaving in the 1920s, such apprenticeship was established. Insiders explained it as a means to disci- pline unruly children of relations. Without supervision through work, the children would grow into ‘disorderly youth and men, predisposed to drunk- enness and brawls’.32 In Dharmavaram silks, for example, boys were em- ployed outside the family to learn the craft. Weavers who had more than one son let the younger sons work for another weaver who needed assis- tance. In the first six months of training, the boys received nothing. There- after they took a token wage, about a tenth of adult family earnings a month. On attaining adulthood, around the sixteenth year, the father and prospective employers discussed whether the boy was reliable and skilled enough to deserve adult wage. If he was, the boy then started saving up to buy looms. In a year or two the apprentice could begin to contract on his own.33 An important point added in connection with the silks of

28 S.M. Edwardes, A Monograph upon the Silk Fabrics of the Bombay Presidency (Bom- bay: Government Press),, 23. 29 . India, Draft Report on the Himroo Weaving Craft of Aurangabad (Maharashtra) (New Delhi: All India Handloom Board, 1955), 12. 30 Among the silk weaver caste Patvegars of Deccan, ‘sons, many of whom attend school in their early years, commence to learn the details of weaving and dyeing when they are 12 years old, and are experts at the age of twenty.’ Edwardes, Monograph upon the Silk Fabrics, 51. 31 . N.G. Ranga, The Economics of Handloom Industry (Bombay: Taraporevala, 1930), 113. 32 . Ranga, Economics of Handloom, 110. 33 . Ibid., 81-2. apprenticeship and industrialization in india 79 Peddapuram and Uppada in the same region of South India is that the more complex the work the more pains did the master take on the novice, and the smaller was the apprentice wages. Offering smaller wage when the training was more rigorous is consistent with theory. Yet, such payment was not a general feature. Two factors, la- bour-shortage and training costs, often pulled in different directions. In the fine cotton weaving of Salem town we find reference to the contrasting situation, so great is the demand [for labour] that employers will give an advance of Rs. 100 to Rs. 150 to a boy’s father in order to secure the boy’s services. No interest is charged but the advance is regarded as a debt due from the boy.. By this system it becomes possible for a thrifty weaver to make money out of a large family.34 Clearly, in these cases of rapid growth, productivity rose so quickly that the masters competed for apprentices. In examples like these, an exchange of apprentices signified the transformation of a small weaver neighbourhood into a large manufacturing cluster. There is no evidence yet of the appren- tices leaving the cluster of families where their parents belonged. The ex- change was limited to the community of weavers interrelated by marriage or caste ties. Community-bound apprenticeship was stretched but not broken in such growing urban centres, Sholapur in the early twentieth century being an example. Located in a cotton growing region in the western Deccan, Shol- apur became a mill town and a cotton trading point in the decades after the railways connected it with Bombay. From the 1890s, Sholapur received a large number of migrant weavers from the Telengana region 200-300 miles east. Some of them entered the mills. But the majority settled down as handloom weavers. In the interwar period, the town had 7-8,000 looms located in small factories owned mainly by the Telugu-speaking Padma- sali weavers. These weavers dominated capital, enterprise, and labour.35 By 1930, there was a mixture of small family firms with 3-8 looms and large factories with 20-40 looms. The second type relied on migrant weavers staying in dormitories or as paying guests of weaver families. For a long time thereafter, the Sholapur cluster managed to maintain a combination

34 Ibid., 136. 35 On Padmasali migration into Sholapur and the history of handloom factories in this town, see Douglas Haynes and Tirthankar Roy, ‘Conceiving Mobility: Weavers’ Migration in Precolonial and Colonial India’, Indian Economic and Social History Review, 33 (1999), 35-67, and Haynes, ‘Logic of the Artisan Firm’. 80 tirthankar roy of wage-labour and community-based learning. Initially the migrants were adult males, who had been trained inside their own families before migra- tion. Even when younger persons and trainees were employed, the hand- loom factories recruited mainly from within the Padmasali caste. If they did share the cost of training, the caste-based hiring ensured that job mo- bility would not be easy for anyone without the consent of the employer. Family and community control was usual outside textiles. About 1900, wood-carving in Southwestern India ‘is strictly confined to the caste [of Gudigars]; the boys are initiated into it at the age of 12 years commencing with a two years’ course of drawing.36 Gold and silver wire workshops of Yeola, a silk weaving town in western India, employed workers and were owned by men of the goldsmith caste. ‘Like so many of the artisans in this country their craft is more a matter of hereditary skill than technical training’.37 In a smaller carving tradition, Monghyr in Bihar, the caste basis of hiring was again maintained. Masters were more than masters, they were also successful capitalists. Wages seemed too ‘low .. for the skilled labour necessary for work of this kind’, while the master artisan took ‘the lion’s share of the profits’.38 Community control faltered in at least three contexts. In a city with larger-scale workshops, it was difficult to recruit only from community members. Sometimes merchant choices shifted custom away from com- munities and towards non-traditional groups. And in new industries, the old institutional choices did not work so well. In the cities of Gujarat, wood-carving, while ‘still a caste occupation’, was not monopolized by one caste group. The lines demarcating castes was thinner here, and so was that between the skills of the carver and the car- penter. The largest carpentry workshop in Ahmedabad employed more than a hundred people, and apprenticeship in such a workshop could not be confined to one caste. In Surat, there was again both intermixing of castes and shifting application of skills. We see a further variation in Bom- bay, with its huge market for furniture. Here the carving industry was over- whelmed by and absorbed in the furniture industry. Carpenters were immigrants from Gujarat, many being carvers by background. Carpenters and carvers were in great demand, for the furniture that commanded a

36 J.A.G. Wales, A Monograph on Wood Carving in the Bombay Presidency (Bombay: Government Press, 1902), 5. 37 J. Nissim, A Monograph on Wire and Tinsel in the Bombay Presidency (Bombay: Government Press, 1909), 11. 38 C.O. Ghilardi, A Monograph on Wood-carving in Bengal (Calcutta: Government Press, 1905), 4. apprenticeship and industrialization in india 81 great deal of business involved some ability to reapply Indian carving de- signs on Europeanesque goods. Likewise, the wire and bar drawers of Yeo- la and other small towns were leaving their craft for work in the big city, giving rise to the fear that the knowledge was at risk ‘unless efforts are made to train others as apprentices in their craft’.39 Again, when the same indus- try reappeared in Bombay, the workers ‘are not confined to any special caste or creed’.40 Similar departures from tradition were visible, perhaps much more com- mon, in northern India. The north Indian city Lucknow was the home of an embroidery industry, ‘the exceptionally great skill of the Lucknow work- men’ having developed through the patronage of the Awadh rulers. This craftsmanship successfully overcame the end of the local custom, by reach- ing out to ‘a new outlet in external markets in Bengal and the south of India’.41 Almost as soon as this long-distance trade began, merchants tried saving labour costs by transferring the work to women of impoverished families. Embroidery suited this choice particularly well. In northern and western India, needlework had long been ‘more a part of the Mussalman girl’s education than the “three R’s”’.42 The commercial branch of the craft moved from male professionals to women household workers thanks to this upbringing. A similar process of transformation occurred in contem- porary Punjab cotton embroidery, phulkari, again a female domestic pur- suit that met the market with remarkable success.43 In the craft of the smelter and blacksmith, the cleavage between tradi- tional practice and new urban markets was wider. In iron-making, there are few reports of recruitment or the training of children. Indigenous iron- making was done by semi-nomadic groups that tended to set up temporary camps near ore deposits.44 These groups almost always consisted of men. ‘The womenfolk never appear to render any assistance’; nor did children. While this craft, along with a large part of the blacksmith tradition supply-

39 Nissim, Monograph on Wire and Tinsel, 8 40 Ibid., 11. 41 William Hoey, A Monograph on Trade and Manufactures in Northern India (Lucknow: Government Press, 1880), 26-27. 42 W.S. Hadaway, Monograph on Tinsel and Wire in the Madras Presidency (Madras: Government Press, 1909), 10. Gift-exchanges between families during marriages involved, with the southern Muslims, home-crafted cloth bags filled with nuts, spices, or money. 43 H.C. Cookson, Monograph on the Silk Industry in the Punjab (Lahore: Government Press, 1887), 7, Appendix B by F.A. Steel. 44 For a study, see T. Roy, ‘Did Globalization Aid Industrial Development in Colonial India? A Study of Knowledge Transfer in the Iron Industry’, Indian Economic and Social History Review, 46 (2009), 579-613. 82 tirthankar roy ing agricultural tools, disappeared with the importation of English and Swedish iron, a new urban demand for cutlery grew rapidly at the turn of the century. Blacksmiths engaged in meeting the new demand required a higher degree of skill than was necessary with agricultural implements. ‘This higher class of work is carried out in larger shops and the workers are the employés of the master cutler’. Skill commanded a wage premium. The average monthly income of Rs. 15 in 1900 placed the cutler at the top of the urban wage scale.45 In factories such as these, references to the commu- nity as an agency in the hiring and training processes become rare. In the urban skilled crafts of northern India, an alternative to the fam- ily system could be seen at work. Major urban centres of carpet weaving in India—Amritsar, Srinagar, and Agra—saw a revival in the interwar pe- riod with American interest in Indian carpets. The industry here was al- ready established in small workshops run by master weavers who had migrated from Kashmir during episodes of famine. Now it saw large facto- ries with as many as 300 looms each. These new enterprises were owned by foreign carpet trading firms having operations in Persia and , which had been destroyed during the First World War. Correspondingly, their India trade grew and the Indian factories became larger. The employment contract in the factory was not a wage-contract. It involved three parties, the owner, masters, and workers. The owner entered a piece-rate contract with the masters, also called ustad, master, or karigar, artisan. The latter recruited his team, which consisted of adults and chil- dren. The young apprentices were called shagirds or disciples. The adults sometimes appear to be partners of the ustad, sometimes his employees. The second group was almost always engaged on time-rates. The tasks they did were wide-ranging. In the large factory, the ustad was an intermediary. In smaller factories, the ustad and the owner (called karkhanadar) often became indistinct. The apprentice was a student, and not ‘the household drudge of the master’.46 A man, frequently a weaver himself, brought his son or grandson to a master, and took a loan from him to be repaid from the wages of the boy. The boy worked as an apprentice for a few years and, if a good worker, was considered an adult entrant into the labour force like any other.47 In

45 E.R. Watson, A monograph on iron and steel works in the province of Bengal (Calcutta: Government Press, 1907), 33. 46 Census of India, 1961,Uttar Pradesh, Woollen Carpet and Blanket Industry of Uttar Pradesh, XV (VII A) (Allahabad: Government Press, 1964), reference to Shahjahanpur, 19. 47 Mukhtar, Report on Labour Conditions, 16. apprenticeship and industrialization in india 83 most cases, the talented worker could repay the loan. But more often, he persuaded the next employer to take it over. In no other craft were children more extensively employed outside the family. The task of knotting piles was simple and not too exhausting. And by a few ingenious devices, it did not need the reference of the whole carpet. One such device was the talim. Repeatable sections of a design were coded and read out loudly to the weavers.48 The task of coding needed separate and literate groups of work- ers. Once these tasks were separated from weaving, the weaving itself could be done by persons skilled only in tying knots. These were boys between nine and fourteen. A normal-sized loom in an Amritsar factory involved four persons. Two of them were boys under twelve years, the third a boy in his late teens, or a senior apprentice, and the fourth the master himself.49 Apprentices were divided into a hierarchy according to whether they were domestic or hired, and according to the age of entry and duration of stay. The usual age of entry into apprenticeship was nine, though six year olds were not unknown. The boys earned an income about half that of an adult weaver. On becom- ing an expert weaver, a distinction attained in the late teens, they earned a wage somewhat above the average in textile crafts. As they reached this stage, chances increased of their leaving for a new job, ‘just as they are beginning to be really useful’. Masters did try to hold wages down as far as possible, and certainly apprentice wages did not seem to be correlated with increased proficiency.50 However, there were examples of productivity- adjusted apprentice earnings. Good times induced quicker graduation of apprentices into contractors, if not into ustadhood.51 Table 2.1 collates some numbers indicating wage differential in the crafts. All categories of workers reported in the table were wage workers. In work that was still dominated by the family system, wood and stone carving, for example, wage workers were a rarity, or a new system. In such

48 For a sample text, see Indian Carpets (Bombay, undated), section titled ‘Kashmir’. 49 India, Report of the Royal Commission on Labour in India (London: HMSO, 1931), 97. 50 Royal Commission on Labour, Vol. 2, Part 2, Oral Evidence, Punjab, evidence of masters Muhammad Ramzan and Rajbai, 104-5. Some aspects of the arrangement we will be dealing with below, are also described in a book on Pakistan, where the typical interwar Indian con- tracts seem to have persisted longer than in India. See S.M. Shah, Hand-knotted Carpet Industry of Pakistan (Peshawar: Pakistan Council of Scientific and Industrial Research, 1980), 8-40. 51 For example, ‘the great demand for labour makes the instruction less thorough and, very often, raw workmen are turned out’, who competed with their masters, Kunwar Jagdish Prasad, Monograph on Carpet Making in the United Provinces (Allahabad: Government Press, 1907), 9. 84 tirthankar roy contexts, the earning differential tended to be small in extent. On the oth- er hand, in work that employed mainly wage-workers, such as metal-work and carpets, the difference was relatively wide. There was an approximate- ly twelve-year distance between the two wages in such cases, the period that it took for a boy to establish himself as a contractor.

Table 2.1. Skill premium in the crafts (lower end of the wage scale = 100). Artisans

Low-wage High-wage Wood-carver, Gujarat, 1900 100 130 (‘ordinary workman’ in (wage in a factory Ahmedabad) manufacturing European style furniture) Blacksmith (cutler), Bengal, 1900 100 200 (average earning) (‘skilful polisher’) Wood-carver, Bihar, 1900 100 170 (‘boys’) (‘men’) Stone-carver, Gujarat, 1900 100 140 (‘second class’ artisan) (‘first class artisan) Punjab carpet weaver, 1929 100 200 (boy apprentice, age 9-12) (master)

Sources: Watson, Monograph on Iron and Steel, p. 33; Tupper, Monograph on Stone Carving; Ghilardi, Monograph on Wood-carving, 4; Wales, Monograph on Wood Carving, p. 5.

Even the widest wage-gap in the crafts tended to be smaller in extent than that inside the factories, which should suggest that the modern mecha- nized mills relied on other paradigms of recruitment and industrial educa- tion. It is to this sphere of work I now turn.

Apprenticeship and Modern Industry

Nineteenth century cotton textile mills practiced two models of industrial training, a decentralized artisanal system, where the supervisor-cum-fore- man was also the skill-leader, and effectively in charge of training junior workers; and a centralized system where the employers arranged for on- the-job training. England would answer to the former description rather well, whereas Japan would answer to the latter. India, in the view of early apprenticeship and industrialization in india 85 twentieth-century observers, did not follow any centralized employer- supplied training system. ‘The mill-owner’, one foreman remarked in 1917, ‘deliberately ignores all industrial training’.52 A 1945 report on apprentice- ship in the jute textile mills concluded that ‘there has never been estab- lished any common method of teaching recruits or of teaching young workers’.53 Did the mills in Bombay and Calcutta, then, follow the artisanal system? In the weaving department, which was populated and monopolized by handloom weavers from one region in north India, training needs might have been taken care of by means of apprenticeship and already trained adult migrants, though there is little detailed information. By and large, documentation on training institutions reveals a great degree of informal- ity and uncertainty, to the point of hinting that there was no system of any kind at all. The Calcutta report suggested that individuals learned the work by observing a ‘friendly’ worker already employed, and at times, they learned work different from their own by observing ‘surreptitiously’ what another person was doing. There is a view among historians that the em- ployers could afford to be casual about training because ‘the task of struc- turing a labour force was .. not a question of skill formation, training, or efficiency’.54 The bigger problem was gathering a large number of people together and making them stay in the city. The cited statement was made in reference to Calcutta jute mills. In the case of Bombay cotton mills, ‘instruction appears to be largely in the hands of the jobber’.55 Morris believes that the ‘skills were elementary’ in Bombay, so much so that the jobbers (labour agents) could ‘pass on what little training was necessary’.56 The main responsibility of the foreman, on the other hand, was to ensure adequate staffing of the machinery, and his incentive payments were tied to the performance of this service. There are two problems with the argument that technical inefficiency did not matter. First, if inefficiency did not matter so much in the nine- teenth century, when Indian mills competed with mills operating in high- wage labour markets, it did matter in the interwar period, when competition

52 ‘The Textile Operative in Bombay’, Indian Textile Journal, 27 (April 1917), 195-6. 53 Cited in Dipesh Chakrabarty, Rethinking Working-Class History: Bengal 1890-1940 (Princeton: Princeton University Press, 1989), 92-3. 54 Chakrabarty, Rethinking, 93. 55 Witness was Acting Principal of Victoria Jubilee Technical Institute, Bombay, Report of the Indian Tariff Board (Cotton Textile Industry Enquiry) 1927, vol. III (Evidence of Local Governments, etc.), 156. 56 Emergence of an Industrial Labor Force, 130-31. 86 tirthankar roy had shifted to low-wage East Asia. Mills in Japan and China had a better record of technical efficiency (Table 2.2), though whether that fact owed to better discipline or better training remained a point of heated debate among contemporaries.57 Bombay’s factory neighbourhood was a world far apart from that of the dormitories in Japan with their proverbial neat- ness and regimentation.58 Secondly, as Rajnarayan Chandavarkar reminds us, it is misleading to assume that skill levels were uniformly low.59 Tasks such as doffing or winding that tended to employ women and children in the nineteenth century were not very skilled, and the wage progression reflected that fact (Table 2.3). On the other hand, task-wise wage difference was wider in the mills, and these differences persisted. The persistence of task-wise differential indicates the presence of segmentation by skill. Weav- ers recruited workers from a different pool than did the doffers.

Table 2.2. Comparative capital-labour ratios, 1931. Japan China Bombay Island Looms/weaver 6 4 2 Wage per day (Rs.) 2.25 1 2 Labour cost per loom (India = 100) 37.5 25.0 100.0 Spindles/Ringsider 600 380 181 Wage per day (Rs.) 1.90 0.90 1 Labour cost per spindle (India = 100) 57.3 43.0 100.0

Source: Evidence by H.P. Mody, Manmohandas Ramji, S.D. Saklatvala, T. Maloney, of Bombay Mill-owners Association, before India, Royal Commission on Labour in India, I(2), Bombay Presidency, Oral Evidence, 1931, p. 303. These figures reflect approximate shop- floor practices. Exchange rate in 1931 18d/Rupee or Rs. 13.3/£.

57 The Millowners Association believed that machines worked more slowly on average in India because few operatives had the skill or the authority to repair faults and stoppages on their own, Royal Commission on Labour, 401. Official enquiries conducted in the interwar period revealed instances of migration, absenteeism, turnover, ‘loitering’, and indiscipline in the Bombay textile factories. 58 A book published in 1925 distinguished four classes of mill-hands in Bombay. In addition to the ordinary full-time workers, usually ‘steady, intelligent men’, there were the ‘budliwallahs’ or substitutes, ‘athawadas’ or seven-days-men who worked for seven days and relaxed for the next seven, the ‘mawali’ who led ‘a gay life’ working when they pleased, and the ‘dadas’, or hooligans, often in the pay of jobbers and their female counterparts the naikins. A.R. Burnett-Hurst, Labour and Housing in Bombay (London: S. King and Son, 1925), 54-55. 59 Origins of Industrial Capitalism in India: Business Strategies and the Working Class in Bombay 1900-1940 (Cambridge: Cambridge University Press, 2003), 334. apprenticeship and industrialization in india 87

Table 2.3. Skill premium in the factories (lower end of the wage scale = 100).

Bombay cotton textile 100 110-120 mills, 1890 (Girls, age 9-11, winding) (Women, age 30-40, winding) Bombay cotton textile 100 200-250 mills, 1890 (Girls, age 9-11, winding) (Weavers, age 30-40) Bombay cotton textile 100 235 mills, 1934-5 (‘doffers, men’) (‘two-loom weavers’) Ordnance factory 100 400-500 Calcutta, 1900 (‘ordinary tool-shop worker’) (‘trained mechanic’)

Sources: Report of the Indian Factory Commission (London: HMSO), 1890, various pages, Chandavarkar, Origins of Industrial Capitalism, 311.

Such segmentation was not peculiar to India.60 What was peculiar was that the task-wise segmentation was correlated in India with social segmenta- tion. The mill labour force came mainly from peasant background, and was very heterogenous in terms of language, region, caste, and community. Historians of labour often deal with a phenomenon they call ‘caste cluster- ing’. Certain social groups populated certain jobs and tasks inside the fac- tories. Most workers were recruited horizontally, and within the mill vertical job mobility was very restricted. Instituting any training system that superseded these clusters would not have been either easy or consid- ered appropriate. Let us consider a few examples of such clusters. There were four domi- nant social groupings in Bombay’s cotton mills, the Marathas of peasant stock mainly from Konkan, the ‘untouchables’, the Bhayyas or North In- dian Muslims, and the Pardeshis or North Indian Hindus. Above all of them were the foremen or jobbers who were usually Marathi-speaking upper castes, or of Maratha stock. Above them still were the Masters, Assistant Masters, and Managers who, in the nineteenth century, were Lancashire men or Parsis. A good example of how these groupings mattered in main- taining segmentation on the job is that of the Maratha weavers in Bombay who successfully resisted the entry of untouchables into the weaving sheds. Chandavarkar demonstrated that not only did such exclusion succeed, but also that it succeeded because of ‘caste consciousness that extended be- yond the workshop’.61 In Kanpur, ‘Mistris [foremen] recruited people from their own village or community... In the late nineteenth century, Julahas

60 On early nineteenth-century Lancashire, see H.M. Boot, ‘How Skilled were Lancashire Cotton Factory Workers in 1833?’, Economic History Review, 48 (1995), 283-303. 61 Ibid. 88 tirthankar roy and Koris were predominant in reeling and weaving, Ahirs in carding and Brahmans in the weaving preparation section [in the textile factories].’62 In the early twentieth century in Calcutta, ‘Telugus, mainly from the Coro- mandel coast and Ceded districts, worked in mills and factories, as porters and carriers. Oriyas worked in mills, both as skilled and unskilled workers, on the construction of rails and roads and did earthwork. Tamils, from southern Madras, worked as middle-class clerical employees, as unskilled agricultural workers and in rice mills. Hindustanis, from UP, both Hindus and Muslims, often took service as peons or watchmen, and many started as petty vendors. The majority of the Bengalis who migrated to Burma were employed as clerks in government, railway or other services, or offices. Some started as shopkeepers or traders, and many were mechanics. Last, there were groups of Punjabis working as artificers, turners and overseers- Gujaratis and Suratis.’63 Joshi reminds us that ‘caste and community ties ... were not peripheral but integral to the lives of workers in Kanpur’. Chan- davarkar would agree, ‘social organization in the workplace and in the neighbourhood were interdependent’ in Bombay.64 Although we use the word ‘caste’, these social segmentations cannot really be understood in terms of ‘caste’, defined as jatis that formed of shared rules of intermarriage and interdining. A variety of identities could bring people to cooperate with each other, and form clubs that defended their interests against those of other workers and employers. In India, such cooperative strategies almost always involved the language of or even claims to castehood. Some of the major designations of worker groups in Bombay—Pardeshi, Bhayya, or untouchable—did not conform to any caste name. Bhayya and Pardeshi signified region and language rather than caste. In Kanpur counterpart terms indicated clearly marked territorial divisions within the city. In this respect, Morris’ point that ‘caste’ was not a significant influence on occupational choices and production decisions, is appropri- ate. But it is a limited point. Cooperative clubs formed of a variety of social markers, and it was perfectly sensible for these clubs to derive legitimacy by drawing upon the idiom of caste. These were quasi-castes, or recon- structed castes.

62 Chitra Joshi, ‘Bonds of Community, Ties of Religion: Kanpur Textile Workers in the Early Twentieth Century’, Indian Economic and Social History Review, 22 (1985), 251-80. 63 Arjan De Haan, ‘Migration in Eastern India: A Segmented Labour Market’, Indian Economic and Social History Review, 31 (1995), 51-93. 64 Chandavarkar, Origins of Industrial Capitalism, 227. apprenticeship and industrialization in india 89 The pattern of informal training in the Indian factories cannot be un- derstood without reference to social segmentation. But how did the two influence one another? In one view, the identity of the worker affected the level of effort. In the 1950s, American sociologists invoked the notion of ‘commitment’ of the individual worker, anticipating a later Marxistante argument that there was a carry-over of a pre-capitalist mentality into the capitalist work-place.65 Richard Lambert wrote, ‘where the centre of grav- ity is the family and not the individual, the major determinants of com- mitment may not lie within the power of the factory’.66 Susan Wolcott and Gregory Clark revive the view that ‘the effort level of the Indian worker’ was poor by international comparison.67 Cultural approaches to labour, however, had been received with justifiable scepticism.68 In another view, social segmentation reduced the worker’s incentive to learn, because learn- ing did not necessarily enable one to change specialization.69 This, in my view, is a valid but incomplete argument. It is incomplete because it does not address the key source of market failure in training, the teacher’s incen- tive to teach. In a third view yet, the labour aristocrat had limited interest in improve- ments in technical efficiency. The basis for the argument can be found in statements made by the technical cadre of the mills in the interwar peri- od.70 The labour agent or contractor was a powerful figure in Bombay and Calcutta, as indeed elsewhere in the early stages of industrialization. Unlike the other regions, in India, the agent was also useful in communicating with and supervising workers who did not speak the language of the man- agers. They were an integral part of the factory social system. These head- men were informally in charge of training workers, and ensuring skill acquisition. But their ability and willingness to perform these tasks were limited, for two reasons.

65 Charles Myers, Labor Problems in the Industrialization of India (Cambridge Mass: Harvard University Press, 1958). See also Chakrabarty, Rethinking Working Class History. 66 ‘Labor in India’, Economic Development and Cultural Change, 8 (1960), 206-13 67 ‘Why Nations Fail: Managerial Decisions and Performance in Indian Cotton Textiles, 1890-1938’, Journal of Economic History, 59 (1999), 397-423. 68 M.D. Morris, ‘Caste and the Evolution of the Industrial Workforce in India’, Proceed- ings of the American Philosophical Society, 104 (1960), 124-33. 69 Lalita Chakravarty, ‘Emergence of an Industrial Labour Force in a Dual Economy— British India, 1880-1920’, Indian Economic and Social History Review, 15 (1978), 249-327. 70 I discuss and develop this argument in ‘Labour Institutions, Japanese Competition, and the Crisis of Cotton Mills in Interwar Mumbai’, Economic and Political Weekly, 43 (2008), 37-45. 90 tirthankar roy First, in a socially segmented labour force, to be successful as an agent one needed social leadership more than skill-leadership. Jobbers were sometimes recruited on the basis of seniority and proficiency, but not nec- essarily so recruited.71 ‘Their educational or technical qualifications are just as good or just as bad as those of the workers who work under them’.72 They had their positions of authority because they could exercise control on the workers outside the factory premises, and control on the workers demanded social and political capital. In some cases, and especially in Calcutta and Kanpur, the intermediary was distinctly a headman. In Bom- bay, they derived their authority from connections with the managerial class or positions of authority in the neighbourhood and local civic bodies. Second, intermediaries suffered from a selection bias. They discrimi- nated the skilled workers who were not jobbers themselves. The original authority of the intermediary was buttressed by the power conferred on them by the management to create their own work-teams and hire and fire workers under them. This privilege created a moral hazard. An agent feared another skilled worker as a potential competitor. The Bombay jobber in the interwar period thrived on commissions collected from job-seekers, and workers who believed they had the prospect of getting a job elsewhere refused to pay the commission.73 In the factory, ‘an efficient worker is con- scious of his ability and does not always countenance the frowns of the head-jobbers … the former class of men are thus under a perpetual risk of receiving all sorts of punishments from the supply of inferior or inadequate raw material down to ... dismissal’.74 These statements suggest a conflict of interest, and a weakness in the pattern of control the intermediary exer- cised. They also suggest that, in those areas of recruitment the intermedi- ary still controlled, there was a bias for recruiting ‘raw hands’ and dependents rather than experienced hands. On the other hand, in situa- tions where the best workers were dependent on the jobber, the former tended to withhold effort. The more proficient worker ‘has to keep his pro- ficiency ... confined to himself as he entertains no hope of getting higher posts or rank or increment of pay’.75

71 In Ahmedabad, the fresh technical graduates in the supervisory divisions could take lessons from the jobbers, M.S. Bhumgara, Karachi, Royal Commission on Labour, 511. 72 Textile Labour Union, Royal Commission on Labour, 2, 296. 73 The most detailed description of commissions, including these practices, is available in the representation of the trade unions. See, in particular, Bombay Textile Labour Union, Report of the Indian Tariff Board (Cotton Textile Industry Enquiry) 1927, vol. III (Evidence of Local Governments, etc.), 438-39. 74 Royal Commission on Labour, 298. 75 Girni Kamgar Mahamandal, Royal Commission on Labour, 459. apprenticeship and industrialization in india 91 There was a third factor, connected with the second one that could insert a conflict between efficiency and labour regulation. Since they received a commission on the numbers supplied, the intermediaries had an incentive to get work done with a bigger team of ordinarily trained workers rather than with a small team of well-trained ones.76 Hence Bombay’s technical inefficiency. Not all factories in colonial India followed the decentralized system of apprenticeship that existed in the cotton mills. The metallurgical and en- gineering factories, for example, resembled the cotton mills in hiring rela- tively few traditional blacksmiths (one railway workshop in Bihar was possibly an exception in this respect); on the other hand, these factories seemingly gave foremen hired from Europe a greater responsibility of on- the-job training. In steel, ‘the higher class of employés [were] always ob- tained from Europe’.77 The principal gun factory of British India, located in Calcutta, reported that ‘the labour as recruited is generally quite un- trained’. The factory created its own system of training workers. Factory- trained mechanics, however, were too scarce relative to demand, so that the wage differential between the ordinary tool-shop worker and the skilled mechanic tended to be very wide (Table 2.3).78 The mechanic left ‘frequent- ly ... to find employment with private firms’. In both cotton mills and in other factories, the solution to the training problem came from progressive institutionalization of technical training. Colonial India only saw the beginning of this process. In cotton mills, some recruitment did occur horizontally from technical schools. Civil engineer- ing was introduced as a university subject in the 1870s. For a long time af- terward, the students in the university degree programmes numbered a few dozen and all joined the Public Works Department upon graduation. But these programmes had an externality. The engineering schools had a mechanical shop affiliated to them, which recruited men of artisanal class as helpers. There were also, at 1900, about half a dozen provincial technical schools in Bengal. These schools offered short courses in machining. In this way a few boys of ‘the mistry class’ did receive formal training.

76 This syndrome was widely discussed in late-interwar Bombay. See Report of the Textile Labour Inquiry Committee, Volume II (Final Report) (Bombay, 1940), 338. See also Morris, Emergence of an Industrial Labour Force, 129-30. 77 Watson, Monograph on Iron and Steel, 52. 78 Ibid., 48. 92 tirthankar roy Conclusion

The chapter began with the proposition that, by and large in precolonial India, apprenticeship was regulated by social institutions such as the fam- ily and the community, giving rise to a world of varied livelihoods that appeared to outsiders as a world also socially segmented. Community elders and parents ruled in matters of training and the progression of apprentices to masterhood. In comparison with Western Europe, the Indian situation was characterized by a relatively high proportion of apprentices, a longer apprenticeship, and rarity of ‘runaways’. In the nineteenth century, craft culture was characterized by institu- tional continuity, with three qualifications. First, the long-term tendency was in the direction away from the family. Migration, technological change, urbanization of the consumers, and increasing wage work pushed the boundaries of traditional recruitment. At the same time, some relatively wealthy artisans tried to resist the trend. They re-organized themselves and strengthened collective bonds, by forming caste associations and investing in club activities such as community-bound educational projects.79 As a recreated guild, these projects were perhaps limitedly successful in the long run. It is significant nevertheless that they drew on the moral strength of an old institutional legacy. Third, from the last quarter of the nineteenth century, the growth of modern mechanized mills created other paradigms of recruitment and education. The new work-force had rather little contact with the craft tra- dition. Most workers came from the peasantry. The huge ethnic mix of the factory work-force induced the mill-owners to avoid dealing with the work- ers and deal with the headmen instead in matters of staffing, training, and disciplining the workers. The indirect-control model compromised ap- prenticeship and on-the-job learning, until technical schools began filling in the gap.

79 Wealthy artisan groups, for example, were known to set up schools and engineering colleges to educate their own children. These movements remain poorly researched. For one example, see Tirthankar Roy, ‘Capitalism and Community; A Study of the Madurai Sourashtras’, Indian Economic and Social History Review, 34 (1997), 437-63. On the uneasy transition to a recreated fellow feeling to serve a new commercial world, see also Douglas Haynes, ‘Just Like a Family? Recalling the Relations of Production in the Textile Industries of Surat and Bhiwandi, 1940-60’, Contributions to Indian Sociology, 33 (1999), 141-69. skills, ‘guilds’, and development 93

CHAPTER THREE

SKILLS, ‘GUILDS’, AND DEVELOPMENT: ASKING EPSTEIN’S QUESTIONS TO EAST ASIAN INSTITUTIONS

Kenneth Pomeranz

Introduction

Similar names can hide huge differences, and what we call ‘guilds’ in late imperial and twentieth-century China often differed greatly from the Eu- ropean guilds on which S.R. Epstein focused for much of his all-too-brief career. One could even argue that the use of ‘guild’ to translate various Chinese terms is an artifact of two bad assumptions: first, that late impe- rial and early twentieth-century China belonged to the same developmen- tal ‘stage’ as medieval Europe, and second, that guilds were both symptoms and causes of ‘backwardness’. Since I have argued at length against the first proposition, and Epstein against the second, it might seem ill-advised to undertake a comparison of East Asian and European guilds on this occa- sion. But it would be a strange historian indeed who did not wish to under- stand how some of the general themes that Epstein illuminated through his path-breaking analyses of guilds in the larger context of European eco- nomic development played out in his or her own field. These themes in- clude the growth of markets; changing relationships between sovereign power and private interests, and between groups and individuals; and cir- cumstances affecting the transfer and enlargement of useful knowledge. And exploring those questions requires exploring the organizations in which merchants and artisans were grouped together—whatever we choose to call them. It would also be a strange historian—though perhaps a somewhat less unusual economist—who did not recognise in Epstein’s work an important methodological lesson: that beginning from a theory-driven understanding of what kinds of institutions have proved conducive to economic growth in modern times, and then measuring the distance between such institu- tions and those that existed in another time and place, may blind us to 94 kenneth pomeranz important aspects of how growth actually happened. Instead, we need to begin with actual institutions, trace their dynamics, and understand the sometimes surprising ways in which they could stimulate better or worse outcomes. The result of Epstein’s insistence on a firmly historical approach was often to show how relatively subtle distinctions—for instance the dif- ferences between state/guild relations in different parts of Northern Italy— might set in motion very different long-run dynamics; consequently his work shows us that it is at least as important for us to understand these differences as to emphasise how all these state-society relationships shared fundamental features which made them quite different from a theoretical optimum. One part of what Epstein emphasised in his account of late medieval and early modern Europe was the rise of unified jurisdiction through the growth of larger, territorially-defined states. This process eliminated many local monopolies, internal tolls, and other privileges that had hampered market integration and reduced potential incentives to produce more ef- ficiently. This type of change is much less of a story in post-1000 East Asia, especially China, since these kinds of barriers to supra-local trade had been largely abolished centuries earlier (or never established in the first place). Indeed, the latest big book on China/Europe comparisons—Rosenthal and Wong’s Before and Beyond Divergence—takes the absence of significant jurisdictional divisions within the Chinese empire as one of its major points of departure.1 Even in Japan, which was politically much more fragmented than China, such divisions do not seem to have been nearly as big a prob- lem as in much of medieval and even early modern Europe. Another argument Epstein emphasised—that although guild privileges did sometimes interfere with market integration, guilds in many places were comparatively open and flexible institutions, and played a significant role in creating and diffusing new technologies—travels somewhat better to East Asia, but still arrives looking significantly altered. China’s urban guilds generally lacked legal authority over economic matters, and uni- formly lacked legal authority to exclude non-members from their trades; they thus appear comparatively weak. Insofar as Epstein was right that most European guild ‘monopolies’ were also quite porous and flexible in fact, the gap between European and Chinese guilds would become nar- rower in reality than it appears to be based on older literatures; nonethe- less, the difference in formal legal arrangements could hardly help having some significance.

1 Jean-Laurent Rosenthal and R. Bin Wong, Before and beyond divergence: The politics of economic change in China and Europe (Cambridge: Harvard University Press, 2011). skills, ‘guilds’, and development 95 Knowing how important these formal legal differences were is impor- tant in its own right; but it also matters because their shadow falls across our analyses of other issues. To the extent that either Chinese or European guilds lacked effective power to monopolise their trades—helping their members at the expense of the broader economy—they needed to attract members by offering other sorts of benefits; and some of these benefits, such as helping people learn their trades, would have benefited the econ- omy as a whole. In both China and Europe, people were apprenticed to individual mas- ters, not to the guild per se; the question then becomes how much of a role guild standards, oversight and organisational resources played in making these apprenticeships happen and setting expectations for them. It has generally been assumed that in most European cases, that role was sub- stantial; the controversy has been more about whether the influence of guilds was primarily facilitative (Epstein’s position) or restrictive. In the literature on China, there has been more doubt about whether the guilds mattered at all to technological change. The restrictionist argument has been made, but there is relatively little evidence that guilds had a powerful influence in that direction, as we shall see later. Meanwhile, those who have looked for the positive influence of Chinese guilds have tended to focus on the social services they provided to members. While one can certainly ar- gue that these services made it much easier for people to go to the places in which they learned their trades, and thus indirectly facilitated skill ac- quisition—or that, by making it easier to take your skills to a new places, they increased the incentives to acquire skills in the first place—there has been very little discussion of Chinese guild structures in those terms, and even less of any direct role that guilds may have played in promoting skill diffusion and/or technological progress. Guilds may nonetheless have sometimes played significant roles in skill formation, but if so they have left little record of it; and because they were structured so differently from European guilds, it is hard to analyse them the same way. Indeed, it is doubtful whether ‘guild’ is even a good transla- tion for the various Chinese terms (most notably hang) that are conven- tionally translated that way when they refer to occupational organisations, though I will follow convention and continue to do so.2

2 Peter Golas, ‘Early Ch’ing guilds’, in: G. William Skinner (ed.), The city in Late Imperial China (Stanford: Stanford UP, 1977), 558-59 briefly discusses some of the pros and cons of such a choice. 96 kenneth pomeranz Meanwhile in the countryside—which probably housed a considerably larger share of Chinese manufactures than in medieval or early modern Europe—guilds seem to have been much less important than either ex- tended families or villages as foci for skill transmission, marketing, dispute resolution, and so on. Chinese guilds’ near-total lack of de jure power means that the analogy that Epstein invoked between certain privileges in late medieval Europe and modern ‘infant industry’ protection would be much weaker in China. The situation is more complicated for Japan, and Euro- pean analogies may fit better there, but still not that closely: and as we will see, it is not clear that employing these analogies in East Asia ‘predicts’ the right outcomes. Nonetheless, the very starkness of the differences between the ends of Eurasia on these issues gives us an unusual perspective from which to re- flect on some of the issues Epstein worked on. After all, market integration, skill formation, innovation and state power surely mattered in both places, however differently they were configured. Furthermore, a point which Epstein particularly emphasised—that lacking states strong enough to enforce reasonably uniform rules and prevent private interference with markets probably inhibited pre-1800 growth at least as much as overly strong states that allegedly crushed private initiative– is an important cor- rective to ‘Whiggish’ narratives in settings besides the European ones that he wrote about. Analysing these issues in East Asia is further complicated by a problem that is both historical and historiographical: that we are not always sure whether what we are supposed to be explaining is ‘failure’ or ‘success.’ Clearly, East Asia did not pioneer the major technologies of the Industrial Revolution. And since half a millennium earlier, at the outset of the period Epstein studied, China may well have been the world’s technological lead- er, one way of framing the issue is to look at how that lead was lost: the famous ‘Needham puzzle.’ It is worth remembering that both China and Japan had very technologically sophisticated agricultures, and were very good at diffusing agricultural best practices; thus they were not uniformly ‘technologically backward,’ even in the eighteenth century. Nonetheless, for current purposes, it makes sense to leave that aside and focus on (proto)- industry, where the East Asian record of technical innovation after ca. 1300 certainly lags that in Europe. With the story of Chinese technology in the early modern period generally seen as one of ‘failure,’ there has been little impetus for people to ask the kind of questions that Epstein asked about what positive contributions its guilds may have made to technological and economic progress. skills, ‘guilds’, and development 97 At the same time, however, we should also remember that even in 1850 much of Japan and at least parts of China had reasonably technologically advanced industries compared to much of the rest of the world. Moreover, those same parts of East Asia (plus a few others) have been unusually suc- cessful in adopting, adapting, creating and diffusing new industrial tech- nologies since the mid-late nineteenth century (especially in the last several decades); and that many scholars (notably Kaoru Sugihara) have attributed this success to institutions that successfully nurtured and utilised a rich store of human capital.3 A recent study of early modern skill premiums—the difference in earn- ing power between skilled and unskilled workers in the same field (in this case construction)—by Jan Luiten van Zanden provides another reason to look closely at Chinese institutions for creating and diffusing useful knowl- edge. Lower skill premiums are strongly correlated with more successful economic development among modern economies; this is quite logical, since a high skill premium suggests that needed skills are in short supply. Van Zanden shows that in early modern times, skill premiums in Northwest European cities—London, Amsterdam, Paris, Ghent, Antwerp, Oxford— were much lower than in the rest of Europe, and also much lower than those derived from the (often scanty) data for cities in Russia, the Middle East, India, Southeast Asia, or Korea. One Japanese data set also shows very high skill premiums, while another suggests a figure lower than these oth- er places, but still much higher than Northwest Europe. In China, however, or at least in South China, the skill premiums derived from a relatively good (though still far from perfect) eighteenth century source generate wage premiums as low as those for Northwest Europe.4 Some Chinese institution or institutions thus seem to have done rather well at meeting the demand for skilled labour. In short, one can find a positive story that needs explain- ing, along with the more familiar ‘what went wrong’ question. Human capital is not the same as skills. It includes, for instance, more general dispositions, such as punctuality, habituation to taking orders, and

3 Kaoru Sugihara, ‘The East Asian path of economic development: A long-term perspec- tive’, in: Giovanni Arrighi, Takeshi Hamashita, and Mark Selden (eds), The resurgence of East Asia: 500, 150 and 50 year perspectives (London: Routledge, 2003), 78-123. 4 Jan Luiten van Zanden, ‘The skill premium and the “Great Divergence”’, European Review of Economic History 13 (2009), 127, 131-33. These results are particularly striking since at least one explanation of Northwest Europe’s low premia—low interest rates, which should decrease the amount of compensation a skilled worker would need to make up for having delayed making money to undergo training—would not apply to South China; Chinese interest rates were certainly higher than those in Northwest Europe, and probably higher than in several of the other places mentioned here. See Ibid. 133-39 for a discussion. 98 kenneth pomeranz so on. Partly for that reason, human capital is also notoriously hard to document and measure; proxies often used for it in the contemporary world, such as years of formal education, have serious limitations even for today’s conditions, and are almost certainly inadequate for earlier periods. Moreover, the institutions responsible for promoting (or not promoting) these varied components of productive labour are likely to be quite varied, even within a single country. In the East Asian case, the so-called ‘pluri- active’ household—one switching its members back and forth among many market-oriented activities—may have been the crucial one, rather than extra-familial guilds and apprenticeships. Tanimoto Masayuki has recently made an interesting case for the relevance of dispositions acquired in rural families with multiple economic activities to the small sub-con- tracting firms so important to the rise of Japanese industry to global com- petitiveness in the early twentieth century; a similar argument could probably be made for developments in Taiwan and in coastal China.5 But again, I hope that the combination of some sharp East-West differences and a few similarities are sufficient to make the juxtaposition of different development paths interesting.

The Rise of Guilds in Late Imperial China: Trade-based and Native Place-based Organisations

The Ming (1368-1644) founder wished to freeze China’s social structure, and make most craft occupations hereditary. His officials therefore sought to register people with industrial skills, and to impose heavy obligations of rotating service in state manufactories. The system did not last long, how- ever. Passive resistance and strong incentives for officials to look the other way made it largely a dead letter by the middle of the dynasty (ca. 1450); state manufacturing increasingly relied on hired labour and outsourcing at market rates.6

5 Masayuki Tanimoto, ‘Labour intensive industries in industrial districts: Another phase of Japan’s industrialization’, in: Gareth Austin and Kaoru Sugihara (eds), Labour-intensive industrialisation in Global History (London: Routledge, forthcoming); see also Penelope Francks, ‘Rural industry, growth linkages, and economic development in nineteenth-century Japan’, Journal of Asian Studies 61 (2002), 33-55; Harry Oshima, ‘The Transition from an agricultural to an industrial economy in East Asia’, Economic Development and Cultural Change 34 (1986), 783-809; Osamu Saito, ‘Land, labour, and market forces in Tokugawa Japan’, Continuity and Change 24 (2009), 169-96. 6 Timothy Brook, The confusions of pleasure: Commerce and culture in Ming China (Berkeley: University of California Press, 1999) gives a good overview of this attempt and skills, ‘guilds’, and development 99 Meanwhile growing luxury markets provided much more rewarding outlets for the labour of craftsmen; the increasing irrelevance of sumptuary laws by the late Ming (documented at length by Wu Renshu)7 was the flip side of the failure of the Ming attempt to control production in major crafts. In some places, the combination of burgeoning markets for craftsmen and craft products with a moribund but still officially extant set of state of la- bour requisitions created repeated conflicts. Numerous new hang—a term which sometimes, but not always, refers to an association of people in the same trade—appear in the record starting in the late Ming, and even more in the Qing; many seem to have emerged out of efforts to limit the impact of corvée demands on any particular tradesman, while providing enough cut-rate labour on an orderly basis to meet the essential needs of cash-poor local governments.8 These were in large part new organisations, though some shared the names of earlier organisations of craft producers and merchants. These guilds also acquired other functions, including organising shared worship, providing social services for poor members, and sometimes regulating the trade’s internal affairs (wages and hours, quality standards, and so on). As population and commercialisation soared in Qing cities, the dockets and backlogs of magistrates’ courts exploded; many officials were therefore happy to effectively delegate a good deal of regulation and dispute settle- ment to guilds, even though the state remained suspicious of private or- ganisations.9 Many guilds also assured the state of their usefulness by contributing to firefighting, poor relief, and other important local func- tions; a few guild leaders (presumably merchants rather than artisans) also reinforced their respectability by purchasing official degrees.10 All of this gave guilds a fairly secure position, but not one that was en- shrined in law; and this semi-formal position probably did much to skew the public self-presentation of the guilds’ functions (which are most of what has survived) towards their least controversial functions. It is thus its failure. 7 WU Renshu, Pinwei shuhua: Wan Ming de xiaofei shehui yu shidafu (Elegant tastes: Literati and the consumer society of the Late Ming) (Taiwan: Zhongyang yanjiuyuan and Lianjing chuban gongsi, 2007). 8 QIU Pengsheng, Shiba, shijiu shijiji Suzhou cheng di xin xing gongye tuanti (Newly emerging industrial organizations in 18th and 19th century Suzhou city) (Taibei: Guoli Taiwan daxue chuban weiyuanhui, 1990), 43, 92-99. 9 Ibid., 44; Timothy Bradstock, Craft guilds in Ch’ing Dynasty China, Unpublished Ph.D.dissertation, Harvard University, 1984, 63-66. 10 William Rowe, Hankow: Conflict and community in a Chinese city, 1796–1895 (Stanford: Stanford University Press, 1989), 94, 97-103, 112-13, 155-57, 164-66. 100 kenneth pomeranz quite hard to know how effective they were in protecting members’ eco- nomic interests—or, to be more precise, some interests of some mem- bers—or even how central that was to their goals. Knowing how they did so is even harder; we have, for instance, a few accounts suggesting that urban guilds wielded considerable direct, coercive power (while officials looked the other way), while other accounts suggest that guilds were all but powerless in fundamental matters, such as excluding interlopers from the nearby countryside.11 The Ming had been largely concerned with regulating those artisans who produced goods that the state needed and/or conferred high prestige on those who acquired them. Consequently, the huge numbers of people producing relatively simple consumer goods for largely rural markets or producing basic intermediate goods (e.g. barrels for shipping oils) had gen- erally remained outside the Ming regulatory and corvée system. Their num- bers expanded considerably over the course of the Ming and even more in the Qing, often without any guild system at all that is visible to us now. Meanwhile, rural demand for manufactures—aside from what farm families made for themselves—appears to have been met largely by two groups of people, who were also based in the countryside. On the one hand, villages generally had a small number of resident artisans with permanent shops: a few shoemakers, carpenters, bamboo workers, followers of the needle trades, barbers, etc. Such people often also had some farmland, but largely lived from doing everyday artisanal work for their fellow villagers.12 The nearby market town would have another set of resident craftspeople, perhaps (though not necessarily) more specialised and fully detached from farming. The crucial difference, however, seems not to have been between those resident in villages and market towns: i.e. between artisans in the country- side and those at the lowest level of the ‘urban’ world. Instead it was be- tween these resident artisans and another group of mobile artisans. The latter would travel established circuits—either singly or in small groups— carrying at least some tools and providing a very particular good or service in which they were specialised. Depending on precisely what they did, such

11 Bradstock, Craft guilds, 152, 162-71; Golas, ‘Early Ch’ing guilds’, 579-80; Christine Moll-Murata, ‘Chinese guilds from the seventeenth to the twentieth centuries: An overview’, International Review of Social History 53, Supplement (2008) , 226-27, 244. 12 PENG Nansheng, Ban gongyehua: Jindai Zhongguo xiangcun shougongye de fazhan yu shehui bianqian (Semi-industrialization: The Development of rural village handicraft industries and social transformation in modern China.) (Beijing: Zhonghua shuju, 2007), 301-02. skills, ‘guilds’, and development 101 workers might either arrive in a location and work for a time in the shop of a resident artisan or set up independently, using only their own equip- ment. Some might travel year-round, but a more typical pattern was for these men to combine farming with seasonal work on the road. Particular villages quite often specialised in a particular trade and dominated it over some specific region; the requisite skills were generally passed down through families, but were in some sense also the joint patrimony of the village.13 Some of these skills might be quite specific, though that did not necessarily mean that they catered to rarefied needs: rather, their niches were defined by needs that could be serviced on a seasonal basis, rather than needing somebody who was present every day. On the one hand, these kinds of workers seem to have made and repaired many of the highest- quality icons in rural temples; on the other, men from one semi-arid North China village supplemented their farming incomes for generations by trav- eling the region as itinerant barbers, pig castrators, veterinarians, rope- makers and coffin-makers.14 In some ways, then, the institutional configuration of craft production in the more advanced regions of China bears a superficial resemblance to that in late medieval Europe. Much high-end production was done in the cities, often in large workshops, by craftsmen who were members of or- ganisations which linked workers across many workshops and who often came from someplace else. Meanwhile, middle and lower end production was done by suburban and rural people (many of them women, especially in the textile trades), usually in homes or small workshops. Some of that work was sub-contracted from urban producers, and some was autono- mous from and at least potentially competitive with urban output. The rural workers often had no formally institutionalised horizontal ties— though the merchants who bought from them were more likely to have some.15 But these resemblances to European circumstances were probably more apparent than real.

13 Ibid., 302-04. 14 Edward Friedman, Paul Pickowicz, and Mark Selden, Chinese village, socialist state (New haven: Yale UP, 1991), 3. We have almost no information on the living arrangements of rural itinerants, who would not have found native-place associations that could house them; presumably most stayed either in their workplaces or in cheap inns, and many (though presumably not all) must have travelled geographically narrow enough circuits that unfa- miliar dialects, food, etc., were not a major problem. 15 See Gary Hamilton, Commerce and capitalism in Chinese societies (London: Rout- ledge, 2006), 6, 93-125, 184-99, for discussion of this as characteristic of Chinese capitalism. 102 kenneth pomeranz On the one hand, the urban workers’ organisations generally had rela- tively little power (and certainly no de jure power). Many did not even see economic regulation as their most important task. Some, as we have seen, had emerged in response to Ming efforts to organise trades for purposes of meeting corvée obligations. Many others evolved out of organisations that were originally based on a shared place of origin rather than a shared trade: for instance the organisation of people from Shaoxing currently residing in Beijing. Such groups might include merchants, officials, artisans, and so on. In many cases, officials and aspiring officials (students preparing for civil service exams, literati traveling in search of connections) were the founders of these organisations, and dominated them for quite some time.16 Even when commercial/artisanal groups seemed to be the mainstay of a native-place organisation, their control was often insecure. In one particularly notorious eighteenth century case, a traveling official moved into the guild hall for people from the Hangzhou area living in Suzhou, taking over 30 rooms customarily used by merchants to store their goods; it took a lawsuit to get him out.17 In the short run, it was not easy for merchants and artisans to create and gain recognition for a separate organisation, and sometimes remaining together with gentry sojourners served their needs well enough. But in the longer run, it was fairly common for such groups to eventually subdivide; the initiative to do so might come either from elites who wanted a more exclusive organisation or from more plebeian members. When such a divi- sion did occur, the non-elite group then sometimes re-organised itself based on trades. Thus if, say, people from Guangzhou dominated boat- building in a particular city, a boat-builders organisation would eventually emerge out of what had been an association for merchants and artisans from Guangdong, and prior to that had been an association for sojourners from Guangdong dominated by its degree-holding members. In some cases, a still further division occurred later, with merchants and artisans in the same trade dividing from each other. But this was much less common, in part because officials generally frowned on worker-only organisations, believing them to be likely sources of disorder or even rebellion.18

16 Bradstock, Craft guilds, 43; Qiu, Shiba, 33; Moll-Murata, ‘Chinese guilds’, 215-17; HE Bingdi, Zhongguo huiguan zhi lun (An Historical Survey of Chinese Landsmannschaften.) (Taibei: xuesheng shuju, 1966) remains the classic account, but is primarily focused on native place as a principle of organization, rather than occupational organization and its implications for the economy. 17 Qiu, Shiba, 34; QU Yanbin, Hang hui shi (A History of Guild Associations) (Shanghai: Wenyi chubanshe, 1999), 74-75. 18 Qiu, Shiba, 67, 75; Bradstock, Craft guilds, 58-62. skills, ‘guilds’, and development 103 The scale of these organisations could vary wildly. Early eighteenth- century Suzhou, for instance, had over 400 calendering firms, employing at least 20,000 people;19 while they were generally divided among two or three guilds, these were nonetheless large organisations. By contrast, a trade that served a local clientele in a county seat, with fewer than a dozen firms and perhaps 50 workers, might nonetheless be organised as a guild in some cases; as we shall see later, this became increasingly common in the late nineteenth century, when the management of proliferating com- mercial levies often required that a formal organisation be created through which even small trades could interact with local government. For organi- sations based on ‘native place’, that unit was most frequently a province, but huiguan which grouped only people from a single prefecture (out of perhaps a dozen in an ‘average’ province) were also common, and there were a few notable examples of groups which included people from two provinces (e.g. Hunan and Hubei or Shanxi and Shaanxi).20 Not surpris- ingly, organisations based on smaller geographic units were most common where natives of one prefecture dominated a particular trade (e.g. law clerks from Shaoxing, who had an organisation in many administrative centers) or where two prefectures from the same province both sent a lot of migrants to some particular place (e.g. various Fujian and Guangdong prefectures on Taiwan). In some cases, these names of relatively large units may have hidden a more intimate reality: since skills were often transferred along lines of pre-existing personal relationships, what appears in the re- cords as, say, a group of ‘boat builders from Chaozhou prefecture’ may have sometimes been controlled by people from a few villages in that prefecture. But barring the appearance of new records, any assertions about how com- mon this was must remain speculative. Regardless of whether workers and employers separated, geographic issues often remained even once an ostensibly trade-based organisation came into being. Since a given trade in a given city might well be practiced by people of various local origins, this could result in a single city having three or four ‘guilds’ for a single trade, making each of them that much less able to control their economic environment. Sometimes hang addressed this problem by dividing up the city, giving each native place group within

19 Paolo Santangelo, ‘Urban Society in Late Imperial Suzhou’, in: Linda Cooke Johnson (ed.), Cities of Jiangnan in Late Imperial China (Albany: State University of New York Press, 1993), 108, 110. 20 By mid-Qing times, some densely-populated prefectures had well over 3 million people, though a bit under 2 million would be more typical; an average province would have had over 20 million people. 104 kenneth pomeranz it a territory; but since buyers were free to move around town, the effective- ness of such arrangements depended on strong and continuing coordina- tion, and we lack sufficient evidence to say how common that was. In other cases, workers divided the production process, granting different native place groups control over different parts of the work, even if the skills required for these different steps were essentially the same.21 In what was probably a smaller number of cases, this was addressed by groups originally based on both trade and native place formally merging, with shared occupation trumping diverse geographic origins. This happened, for instance, when Suzhou printers, who had previously been organised into three separate groups based on place of origin, created one printers’ guild in the seventeenth century.22 But this was far from being an inevi- table outcome; and even when a hang organised on occupational lines managed to ‘unite’ people from multiple native places, they often main- tained distinct sub-groups within the organisation, while guild records listed donors to its projects in separate native-place lists.23 Nor is it always obvious that disambiguation always made organisations more useful. While neither occupational nor native-place organisations created formal federations across places prior to the twentieth century—there was no empire-wide institution linking either silversmiths or Sichuanese in Hang- zhou to their counterparts in Changsha—migrants frequently took advan- tage of the existence of both kinds of organisations. Knowing that they might be able to get help from either a native-place or an occupational group in their destination city (which might be one where a different pop- ulation group dominated the migrant’s trade) presumably suited these migrants better than having to rely on occupation alone. In short, we need to be skeptical of any story—particularly common in mainland historiog- raphy which seeks to conform to a narrative about the rise of indigenous capitalism—in which functional organisation inevitably supplants that by native place as the economy grows.24 Even in the mid-twentieth century, with modern-style chambers of commerce and labour unions very much in evidence, native place organisations still held on to some of the ‘turf’ we might expect these institutions to occupy.25

21 Bradstock, Craft guilds, 139. 22 Golas, ‘Early Ch’ing guilds’, 565; He, Zhongguo huiguan zhi lun, 102-03. 23 Bradstock, Craft guilds, 137. 24 E.g. Liu Yongcheng, “Lun Qingdai guyong laodong,” (On Hired Labor in the Qing) Lishi yanjiu (1962, no. 4), 104-28; also Golas, ‘Early Ch’ing guilds’ (though without Liu’s Marx- ist assumptions). 25 Bryna Goodman, Native place, city and nation: Regional identities and networks in Shanghai, 1853-1937 (Berkeley: University of California Press, 1993). skills, ‘guilds’, and development 105 Guild Functions, Economic and Otherwise

Urban guilds seem to have focused most of their energies on providing social services for their members, at least if their written rules are any guide: provisions about providing for burials, assisting the sick and injured, orga- nising shared worship of patron deities, and so on predominate in those texts.26 They did promote standardised weights, measures, quality stan- dards, and, in some cases, prices, and regulate the length of apprentice- ships.27 Many also claimed the power to grant or refuse permits for visitors to work in the trade, though interlopers seem to have frequently gotten away with flouting such strictures. In general guilds seem to have tried to make such people join the guild rather than trying to stop them from prac- ticing their trade;28 this seems unlikely to have deterred people from trying to trespass on the guild’s turf, but probably offered a resolution of particu- lar cases relatively unlikely to generate complaints to officials. We have very few documented cases of truly effective exclusion of outsiders before the late nineteenth century. This may reflect problems of documentation. After all, the most effective exclusion would have discouraged outsiders from even trying to compete, and thus leave no records, while guilds would have been foolish to boast about succeeding at something which they were not legally empowered to do. But it may also be the case that efforts at exclusion became more com- mon in the late nineteenth century. Tax-farming arrangements became much more common after 1853, and often involved substantial grants of power over a trade to the guild which farmed some new commercial tax.29 At the same time, some guilds no doubt felt increasingly threatened by new technologies. Relations between some guilds and local government grew closer in the very late Qing; while this sometimes meant greater govern- ment control over the guild, it sometimes gave guilds greater power to coerce third parties.30 By the twentieth century, a significant number of Chinese reformers, implicitly or explicitly invoking medieval European parallels, were referring to guilds as ‘feudal’ impediments to a free exchange of goods, labour, and knowledge that China needed, and calling on the state

26 Moll-Murata, ‘Chinese guilds’, 223; Golas, ‘Early Ch’ing guilds’, 568-9, 574-5 27 Golas, ‘Early Ch’ing guilds’. 28 Ibid., 562; Moll-Murata, ‘Chinese guilds’, 223. 29 Susan Mann, Local merchants and the Chinese bureaucracy, 1750 -1950 (Stanford: Stanford UP, 1987), 128-44. 30 Moll-Murata, ‘Chinese guilds’, 226, 106 kenneth pomeranz to break up the guilds.31 Ironically, it seems likely that many of the more striking examples of restrictive behaviour that these reformers complained about were relatively recent developments, enabled by the rise of a modern state: one which was willing to give guilds much more power over their trades than they had had before in return for an orderly flow of revenue. It is worth noting that in those towns which became commercially im- portant without ever becoming major administrative centers, Qing guilds seem to have been more likely to form confederations than they were else- where, and those confederations were more likely to take over various significant functions for the town as a whole: arranging for fire-fighting, street maintenance, poor relief for the whole town (not just for members) and so on.32 One might suspect that where guilds did play this kind of expanded role, and thereby became more influential—while also having relatively little direct government oversight—they would take advantage of the situation to exercise more control of markets. However, the limited evidence we have so far suggests no such pattern. What we do find is a chronological pattern, rather than a geographic one: various merchant and artisan guilds making stronger efforts to create and enforce local monopo- lies after 1853 as they became more strongly connected to the state.33 Prior to that there is not much evidence of strong local monopolies or oligopolies, except in certain service occupations.34

City and Countryside

Perhaps unsurprisingly, given that urban guilds in China lacked coercive power, rural producers were not systematically excluded from high value- added proto-industrial production. True, urban workshops dominated production of the very highest-end goods—top quality silks, porcelain, things made with gold, and so on—while rural producers made most of

31 See the discussion in Jacob Eyferth, Eating rice from bamboo roots: The social history of a community of handicraft papermakers in rural Sichuan, 1920–2000 (Cambridge Mass.: Harvard UP, 2009), 94 32 Rowe, Hankow, describing Hankou, is the classic example; the same was true, in many ways of the Chinese-ruled parts of late Qing Shanghai. But in Yangzhou, where com- merce was also more important than field administration, guilds barely existed: perhaps as Moll-Murata (‘Chinese guilds’, 243-244) suggests, because the dominance of Huizhou merchants was so complete that no other commercially-oriented associations were needed. 33 Mann, Local merchants, 128-32, 155-62, 181-82. 34 The difficulty of re-selling many services—I cannot, for instance, pass on my haircut to somebody else, as I can pass on a piece of cloth I have purchased—makes enforcement of monopolies potentially easier. skills, ‘guilds’, and development 107 the goods destined for low end markets, such as medium and low-grade cotton cloth. But some relatively expensive goods were also made outside of cities; so were subcontracted parts of some supposedly urban luxury goods. Indeed, rural handicraft production appears to have been more specialised, more dependent on long distance trade for its markets, and perhaps more skilled in Ming China than most rural European proto-in- dustrial production of the late medieval period. In these ways it may have been more reminiscent of Europe’s next round of proto-industrialisation in the late seventeenth – early nineteenth centuries. No special permissions from urban authorities were needed for rural people to enter artisanal production, and the capital needed for most kinds of production aimed at low and medium-level markets was inexpensive. Even producers of many high quality artisanal products located in the countryside near raw materi- als; this minimised transport costs, since the materials were bulkier than the finished goods, and may have meant lower labour costs as well. Such locational choices were not unknown in Europe, either, of course; stone was often dressed at the quarry, for instance, even when destined for urban buildings.35 But they may have been particularly common in China.36 As Rosenthal and Wong suggest, Europe’s political fragmentation, which meant frequent war, created incentives to locate fixed capital behind city walls;37 this reinforced other incentives, stressed by Epstein, that this same fragmentation created to locate near one’s (mostly urban) markets.38 The merchants who sold China’s rural-made artisan goods, often in dis- tant markets, enforced quality standards, and many engaged in a kind of trademarking,39 but there was no legal authority behind this. In general, neither merchant nor artisan guilds could act as price-setting cartels, either

35 Prak, in this volume, 139. 36 Note, for instance, that the idea that labour costs were lower in the countryside requires at least two important conditions. First, the local food supply has to be sufficiently elastic (either through local production or trade) that a concentration of non-agricultural workers does not result in food prices rising to urban levels; second, there must be sufficient reservoirs of human capital available that productivity-adjusted labour costs remain below urban ones. The political unity of China presumably increased the number of rural locations where these conditions could be met, but they were still far from universal, especially before the arrival of New World food crops made it easier to feed people in certain highland loca- tions 37 Rosenthal and Wong, Before and beyond divergence. 38 S.R. Epstein, Freedom and growth: The rise of states and markets in Europe 1300-1750 (London: Routledge, 2000), 109-10, 122, 147-55. 39 For examples see Gary Hamilton, and Lai Chi-kong, ‘Consumerism without capital- ism: Consumption and brand names in Late Imperial China’, in: Henry J. Ruiz and Benjamin 108 kenneth pomeranz legally or informally, (the latter situation did change in some cases in the late nineteenth century, as we shall see); nor did they significantly restrict the supply of whatever they provided, since they aimed at bringing poten- tial competitors into the guild rather than preventing them from practicing the same trade. Thus rent-seeking by local organisations was probably quite limited, in part because China had long had, on a grand scale, the kind of state that Epstein saw emerging on medium scales in late medieval/early modern Europe: absolutely sovereign but generally supportive of secure private property, rather than a threat to it.40

Skill Formation and Technological Creativity

In Epstein’s relatively positive view of guilds, their most important features included their roles in teaching skills and in creating and diffusing innova- tions. His argument that certain local monopoly privileges constituted a kind of infant industry protection which actually facilitated technical in- novation was particularly interesting (and controversial).41 Essentially he argued that European guilds: a. facilitated training by enforcing contracts (that part is not controversial, though how important it was, relative to rent-seeking, has been disputed); b. created a structure that facilitated the movement of workers, and thus the sharing of skills, and c. provided a form of temporary monopoly rents which (somewhat like modern patents) pro- vided incentives for innovators. He also effectively called into question many of the standard bases for claiming that guilds stifled innovation, without denying that this sometimes happened, or that some guilds—par- ticularly those which had close relations with governments that also ruled the surrounding countryside—were fairly consistently conservative, or that the ways in which guilds did promote innovation were often inadvertent. Even with these nuances, his view differed quite sharply from traditional views in which guilds were seen as almost always standing athwart both

S. Orlove (eds), The social economy of consumption (Lanham, MD: University Press of America, 1989), 253-79; Eyferth, Eating rice, 75 40 Sporadic rent-seeking by greedy officials and other individuals is of course a differ- ent matter, and most of it is probably invisible to us, whether in China or elsewhere. 41 For a heated controversy about Epstein’s ‘rehabilitation’ of the guilds, still on-going at the time of his death, see Sheilagh Ogilvie, ‘“Whatever is, is right”? Economic institutions in pre-industrial Europe’, Economic History Review 60 (2007), 649-84; S.R. Epstein, ‘Craft Guilds in the pre-modern economy: A discussion’, The Economic History Review 61 (2008), 155-74; Sheilagh Ogilvie, ‘Rehabilitating the guilds: A reply,’ Economic History Review 61 (2008), 175-82. skills, ‘guilds’, and development 109 technological change and the making of more competitive markets. One sub-point in his argument that rings especially true for an East Asianist is the reminder that pre-modern technological change was usually not la- bour-saving or de-skilling—and thus often posed no particular threat to guilds. What, then, can we say about guilds, learning, and technical change in East Asian contexts? How were skills learned within these kinds of guild structures? We know very little. The guild rules that have survived from late imperial and Republican China almost all specify the same three-year ap- prenticeship.42 Since not every craft takes the same amount of time to learn, this suggests that other goals—such as regulating the number of masters and apprentices, and balancing the interests of employers in ex- tending a process which gave them access to very cheap labour with those of the apprentices themselves—may have been more important to guild leaders than insuring proper training. Moreover, guilds do not appear to have formally tested masters before allowing them to set up shop as long as they paid the required fees and met the other requirements of member- ship and experience. This may well indicate that ‘the guilds played little role in assuring that people had proper skills’,43 as Moll-Murata concludes, but we cannot be certain; our records are scanty, and since guilds did en- force product quality standards, it seems unlikely that they did not at least do some informal monitoring to weed out egregious cases. It is also note- worthy that apprenticeships in Europe tended to be longer than in China— typically seven years in early modern England, three to five in various parts of continental Europe44—and in Europe, too, the official length of ap- prenticeships varied more with local political configurations than with the nature of the craft being learned.45 So if the uniform length of Chinese apprenticeships indicates the importance of considerations other than creating conditions for learning a trade, then there is no reason to believe that this was less true in Europe; and, conversely, the alternate explanations one can imagine for European cases might apply to Chinese ones as well.46

42 Bradstock, Craft guilds, 172. 43 Moll-Murata, ‘Chinese guilds’, 240. 44 Patrick Wallis, ‘Apprenticeship and training in premodern England’, Journal of Economic History 68 (2008), 834-35. 45 Ibid., 852-53. 46 For instance, apprenticeships of different lengths for the same trade in different parts of Europe (e.g. for tailors in Britain and Holland) could reflect different levels of skills attained, always remembering that the end of an apprenticeship did not necessarily rep- resent the end of improving one’s skills: S.R. Epstein, and Maarten Prak, ‘Introduction: 110 kenneth pomeranz At least four points are worth considering here. First, though both Eu- ropean and Chinese guilds set rules for apprenticeships—in the European case, usually under the eye of the state—the actual implementation of apprenticeships was carried out in individual shops, and must have often differed from the theoretical norms. (That so many European apprentice- ships were never completed47 makes it all the more likely that private agreements and disagreements shaped these relationships at least as much as formal rules.) Second, whatever conclusions we reach about one of these institutions therefore have only limited significance for the other. Particu- larly in cases where much teaching of skills went on within extended families, we must be also distinguish between ‘apprenticeship’ as a formal institution enshrined in a contract or other formal document, and ‘ap- prenticeship’ in the loose sense of an arrangement involving teaching on the job; the latter might be quite important in cases where the former is not. Third, rent-seeking—whether of masters hoping to extend the period in which they received cheap labour from apprentices, or of the whole trade in limiting the supply of their good by slowing the entry of new pro- ducers—may well have been among the non-technical considerations that determined the lengths of apprenticeships, but we cannot assume they were dominant. When dealing with the tacit knowledge component of skills, we cannot even assume that the parties themselves knew how long it would take to learn alone what they needed to learn. Lastly, that appren- ticeships may have been shaped by considerations other than the maxi- mally efficient transfer of skills need not mean that they were not also an important means by which skills were transferred. Indeed, evidence of other avenues by which artisanal skills were trans- ferred in urban China is also very scant. While there are some hints of very high literacy rates in certain towns (such as the iron-making center of Fos- han), printed texts explaining artisanal practices do not seem to have been common; most that do survive, such as the richly detailed Tiangong kaiwu, were clearly intended more for the curious gentleman than the actual craftsman.48 By contrast, significant numbers of do-it-yourself medical

Guilds, innovation, and the European economy, 1400—1800’, in: Epstein and Prak (eds), Guilds, innovation, and the European economy, 1400-1800 (Cambridge: Cambridge UP, 2008), 8. 47 Wallis, ‘Apprenticeship and training’; Chris Minns and Patrick Wallis, ‘Rules and reality: Quantifying the practice of apprenticeship in Early Modern England’, Economic History Review 64 (2011), 1-24. 48 See Dagmar Schäfer, The crafting of the 10,000 things: Knowledge and technology in seventeenth-century China (Chicago: University of Chicago Press, 2011). skills, ‘guilds’, and development 111 manuals with simple instructions published in relatively cheap formats survive, as well as manuals for householders which give very basic instruc- tions for a number of practical skills, but not much detail on any one of them. At the opposite end of the scale of formality, many artisanal skills were presumably learned within families, but we have very little information about how prevalent this was, much less how it worked. We know that a few high status trades, such as medicine and pharmacy, were mostly passed from fathers to sons in China, but for a much larger array of skilled occupa- tions, we have only a vague sense that father-son transmission was com- mon, without any detailed knowledge of how common. Such transmission was common in Europe, too, but—contrary to stereotypes of pre-modern Europe as a world of inherited statuses and occupations—this does not seem to have been the most common way to acquire an artisanal niche.49 We have some records of skilled Chinese metalworkers which suggest transmission of skills through other kinds of kinship—i.e. through marital networks50—but again, this evidence is far too fragmentary to support any precise statements about frequency. Knowing more about family-based training is clearly important for placing the learning that occurred via for- mal apprenticeships in its proper context, but it seems unlikely that we will have the data we want any time soon. Because oral transmission of skills often leaves no trace in the record, we cannot say very much about these processes, but it seems likely that, given an absence of de jure power over markets, the benefits of technique- sharing probably played a significant role in sustaining guild membership, along with the social and religious services discussed earlier. Since a great many skilled workers seem to have moved to another city to learn their trade—and then sometimes to still another city to practice it—guilds that provided important social supports, comrades who spoke one’s home dia- lect, and so on, were presumably quite important to making the acquisition and diffusion of skills possible. Indeed, the support that they provided for geographic mobility was the principal raison d’être of many Chinese guilds, and might be one area in which they were more important than their Eu- ropean counterparts. While the latter also facilitated mobility (and thus exchange of information), they faced a far more fragmented jurisdictional landscape on the one hand, and a less fragmented religious one on the

49 Epstein and Prak, ‘Introduction’, 10. 50 I draw here on unpublished research by Susan Naquin, shared via personal com- munication, December 2011. 112 kenneth pomeranz other (meaning that a sojourner in almost any European town could likely find one institution that was relatively familiar to him, and that might provide emergency aid and spiritual solace). Furthermore, a larger percent- age of craftsmen in most European cities than in most Chinese ones would probably have been local natives,51 and thus not needed a guild to facilitate their adjustment to the town where they plied their trade.

Did Guilds Obstruct Technical Learning?

It is less clear to what extent guilds may have also prevented some people from acquiring skills. One large group—women—seems to have been ex- cluded by Chinese even more than by European guilds; but it is unclear to what extent guilds were important movers in these restrictions, rather than simply reflective of broader forces. And in both places—but especially in China, as I have emphasised here—rural craft production, even of some relatively high-end goods, offered a capacious route for circumventing guilds’ gender exclusivity; to the extent that rural women were also di- rected away from many kinds of skilled craft work, forces other than guilds were clearly responsible. How much evidence do we have for guilds restricting access to skill for groups other than women (and, in the European case, Jews)? Epstein, of course, argued forcefully that such complaints were at least over-stated in the case of European guilds, and often completely without foundation. We know much less in the Chinese case, but at least at this point, it seems that a broadly similar revisionism may be appropriate there. As we have seen, many Chinese craft groups were organised by native place, though the exclusions were generally not completely rigid; to the extent that a given set of skilled workers managed to colonise every place where there was demand for their services, their common origin (and dialect) would not have inhibited transmission of skills. Conversely, to the extent that, say, iron workers from Zhili did not communicate and share techniques with

51 Though here, too, any conclusions must be tentative. We think that the percentage of native-born craftsmen in Chinese cities was very low, but it was often quite low in Euro- pean cities, too: see for instance, Epstein and Prak, ‘Introduction’, 16. Moreover, the com- plexities of Chinese ‘native place’ designations, in which a person could be considered a native of one’s father’s place of origin even if they had never lived there, complicate the analysis of even the very limited data we have. Nonetheless, we know that Chinese artisans frequently left their wives and young children behind when they relocated to new cities; this must have reduced considerably the number of sons who would have spent their early youth, the period of learning their craft (from their father or somebody else) and their subsequent career all in one city. skills, ‘guilds’, and development 113 ironworkers from Jiangxi, any resulting problems would seem more at- tributable to their underlying linguistic and cultural diversity than to the structure of the guilds themselves. In some cases, guilds seem to have subdivided very minutely, often along native-place lines. This could make it (in theory) impossible for some work- ers to do jobs of which they were perfectly capable, as when workers who plated copper with gold and those who plated copper with silver belonged to different guilds; if these subdivisions separated a single production pro- cess into tasks reserved for different sets of workers, it might effectively de-skill these workers in order to preserve narrow niches for all.52 It is unclear how common this was. Some of the very limited direct evidence of such phenomena that we have comes from workers pleading before magistrates that they could not mitigate some damage done to their trade by taking up any other work; in that context, the workers had an obvious incentive to make their options seem as circumscribed as possible. Nor is it always clear when a narrow range of activity on the part of an artisan reflects protectionist barriers and when it reflects perfectly rational spe- cialisation. The family of Suzhou apothecaries who made only one medi- cine, the formula of which they kept secret from the rest of their guild, might have had extremely narrow skills, as Bradstock suggests,53 or they might have been focusing all their efforts on the most profitable use of their time. We do not know enough to distinguish. By contrast, scholars looking at Japanese materials have argued that the proliferation of more and more specialised guilds during the Tokugawa—e.g. the subdivision of the car- penters’ guild into separate guilds for people making different wooden products—is indicative of on-going economic and technological creativ- ity, not of protectionism run amok.54 This could indeed be the case, but one wonders how much the different significances attributed to specialisa- tion in what we might call ‘late pre-industrial’ China and Japan stem from reading backwards differences in economic performance that became ap- parent after 1860: differences quite likely unrelated to previous guild struc- tures. Qiu Pengsheng has proposed a distinction that might help us think through this issue—though it does not seem to have been the problem he had in mind. He argues that Suzhou craft guilds behaved very differently

52 Bradstock, Craft guilds, 130, 139, 144-45, 147. 53 Ibid., 145. 54 Tessa Morris-Suzuki, The technological transformation of Japan: From the seventeenth to the twenty‑first century (New York: Cambridge University Press, 1994), 51-52. 114 kenneth pomeranz depending on the external market conditions they faced. In trades with stagnant or shrinking markets, he suggests, the paramount goal, shared by employers and employees alike, was to restrict, so as to protect their threat- ened livelihoods: for such groups efforts to define trades very narrowly might have made good sense, at least in the short run.55 In sectors with more buoyant demand—Qiu lists cotton, silk, and paper as examples— maintaining barriers to entry was relatively unimportant, and guilds fo- cused more on managing employer/employee relations, usually to the advantage of the employers.56 If this distinction holds—and Qiu, too, is working from limited evidence—it suggests that any guild-based obstacles to the acquisition or improvement of skills did not matter much; there was, after all, limited incentive to invest in developing skills and techniques for declining sectors, and the reach of these guilds was limited, anyway. Still, one cannot rule out some negative effects of urban guilds on innovation and skill development, to be balanced against whatever positive effects they had. Finally, Timothy Bradstock, who takes a particularly negative view of Chinese guilds, has argued that the strong pressures on guild members to share any new knowledge that they acquired with all other members in- hibited individual efforts to improve production techniques.57 To put the argument in Epstein’s terms, Bradstock’s claim would be that it is indi- viduals who needed temporary monopoly rents to initiate innovation, and the guild system denied this to them. It is unclear how widespread or ef- fective such pressures to share knowledge actually were—indeed we have very little direct evidence of them—or how unique to China they were. Nor is it clear how to reconcile this claim with counter-examples: including some that Bradstock himself mentions, such as the family of apothecaries that kept its special recipe for a medicine secret from other apothecaries. Overall, I suspect this argument, like other complaints about ‘backward’ guilds that we find reported by modern reformers, are exaggerated—much as Epstein argued that modern liberals’ reflexive dislike of European guilds led them to exaggerate their drawbacks and overlook their virtues. But the claim cannot be dismissed out of hand. The role of ascribed characteristics such as native place in providing access to skills may, however, have slowed the development of innovations requiring people with different skills (and thus, in many cases, different

55 Qiu, Shiba, 99, 104. 56 Ibid., 99-102. 57 Bradstock, Craft guilds, 194 skills, ‘guilds’, and development 115 dialects) from collaborating. If, for instance, an improved loom required inputs from a highly skilled weaver, carpenter and ironworker, it is possible that the strength of native-place traditions could have been a barrier. But this, too, is speculation.

Migration and Skill Diffusion in More Recent Times

We know a little bit more about how migration, diffusion of skills, and in- novation worked in the nineteenth and twentieth centuries. It is interesting that when Shanghai began its rapid development, the vast majority of skilled workers came from very specific areas, often in the nearby Yangzi Delta, and continued to pursue occupations with which their home terri- tories were associated. The pattern for unskilled workers was quite differ- ent—they mostly came from poorer areas further away.58 Delta residents who did move to Shanghai were primarily artisans, organised in guilds with a strong native-place component: carpenters, printers, various sorts of met- alworkers, painters, masons, and so on.59 They also dominated the limited ranks of skilled workers in cigarette factories, on the docks, and else- where.60 Perhaps aided by the fact that technological change tended to be slower in Shanghai industry than it was in the corresponding Japanese industries—so that their skills remained relevant for longer—these arti- sanal groups were also successful in colonising many of the new skilled occupations that sprung up in Shanghai: virtually all boilermakers in Re- publican Shanghai came from Wuxi, for instance, and most machinists from Ningbo and Shaoxing.61 A less positive (from the Chinese perspective) way of looking at the same pattern might be to say that while Japan’s more rapid creation of a cohort of engineers and artisans with modern skills enabled them to move much more quickly beyond reliance on partially re-trained ‘traditional’ artisans, who had proved themselves inadequate to

58 Elizabeth Perry, Shanghai on strike: The Politics of Chinese labor (Stanford: Stanford University Press, 1993), 48-57. 59 Ibid., 32-47. See also MA Junya, Hunhe yu fazhan: Jiangnan diqu chuantong shehui jingji de xiandai yanbian (1900-1950) (Mixture and Development: The Modern Evolution of Traditional Society and Economy in the Jiangnan Region, 1900-1950) (Beijing: Shehui kexue wenxian chubanshe, 2005), 98-99, 105. The same was true of another stream of relatively skilled immigrants from the relatively prosperous Canton area, who moved to Shanghai when it supplanted Canton as the primary center of foreign trade. 60 Ma Junya, , Hunhe yu fazhan, 116-7. 61 Shanghai shehui kexueyuan, Jiangnan zaochuanchang changshi, 1865-1949 (History of shipyards in the Yangzi Delta Region, 1865-1949) (Nanjing: Jiangsu Xinhua shudian faxing, 1983), 154-55; Perry, Shanghai on strike, 36. 116 kenneth pomeranz many of the demands of modern industry, Shanghai relied on such people for a longer time because its technological progress was slower than Osa- ka’s.62 However one understands the cause and effect relationships, what is clear is that the kind of relatively skilled labour-intensive industry that characterised both urban and rural settings in the Yangzi Delta before 1850 remained important in twentieth century Shanghai, even as those indus- tries and workers were increasingly surrounded by less skilled factory op- eratives from elsewhere. At least as a working hypothesis, I would suggest that the continuities between old kinds of skilled work and new industries were even more important in Shanghai than in pre-war Japanese cities, where they were far from trivial;63 and this region’s growth, if not as rapid as in Japan (and, of course, buried in a huge Chinese economy which included a number of stagnant or declining areas)64 was nonetheless impressive.65 Moreover, at least until 1949, neither the skilled nor the unskilled im- migrants to Shanghai cut their ties to their native places: this appears to be true both for those who relocated from other towns and cities (e.g. ship- building carpenters who had previously worked in Ningbo or Canton) and for those who came from rural villages. Thus not only did rural industry remain very important, as we shall see shortly, but even in the country’s major industrial metropolis, urban and rural labour markets remained

62 See Erich Pauer, ‘Traditional technology and its impact on Japan’s industry during the early period of the Industrial Revolution’, The Economic Studies Quarterly 38 (1987), 354-71 for a view of Japanese industrialization which minimizes the contribution of ‘tradi- tional’ skills; Kaoru Sugihara, ‘The European miracle and the East Asian miracle: Towards a new global economic history’, Sangyô to Keizai 11 (1996), 27-48, Sugihara, ‘Agriculture and industrialization: the Japanese experience’, in: Peter Mathias and John Davis (eds), Agri- culture and Economic Growth (Oxford: Blackwell, 1997), 148-66 and Morris-Suzuki, The Technological Transformation are important examples of scholarship that places greater emphasis on the contribution of this traditional accumulation of skills and discipline. The present comparative point, I think, would hold even if Pauer is largely right about Meiji Japan, but is that much stronger if we accept Morris-Suzuki, Tanimoto, and Sugihara. 63 Tanimoto, ‘Labour intensive industries’. 64 For a brief discussion of regional trends in Chinese development in the century preceding 1949 see Kenneth Pomeranz, ‘Re-thinking the Late Imperial Chinese economy: Development, disaggregation and decline, 1730-1930’, Itinerario 24 (2000), 47-67. 65 For estimates of the growth rate in Shanghai and the Lower Yangzi more generally in the early 20th century, see Debin Ma, ‘Modern economic growth in the Lower Yangzi in 1911-1937: A quantitative, historical, and institutional analysis’, Discussion paper 2004-06- 002, Foundation for Advanced Studies on International Development, Tokyo. See also Thomas Rawski, Economic growth in prewar China (Berkeley: University of California Press, 1989); Ma, Hunhe yu fazhan. skills, ‘guilds’, and development 117 closely intertwined. This has persisted even into the present, despite post- 1949 restrictions on urban/rural migration and the imposition by law of two very different institutional structures for urban and rural production. As is well known, a remarkable amount of China’s recent industrial boom is rural and/or suburban, especially in the Lower Yangzi region, which still has a very disproportionate share of China’s most skill-intensive, high value-added sectors. Quite a bit of the training of skilled workers in these plants has been implemented by luring Shanghai workers originally from these areas back to their homes, often after they retired or were laid off from state-owned factories in Shanghai.66

Skill Acquisition in Rural Industries

A still larger share of manufacturing remained rural and artisanal through the first half of the twentieth century. Even in Jiangsu province, which included Shanghai and Wuxi, half of industrial output was still rural in 1937.67 The share of handicraft production (some of which was urban) is equally striking, accounting for close to 75% of industrial output nation- wide in 1933.68 Since the formal state, with a few exceptions, did not reach down to the village level in a systematic fashion until after 1949, economic organisation at the grass roots, including that of craft production, was relatively ‘spontaneous,’ and written documentation is correspondingly limited. We do, however, know a bit about how some skills were learned and transmitted. Papermaking was a skilled rural craft which was generally concentrated in the countryside: often (though not always) in hilly areas that provided ample raw materials and relatively poor prospects for supporting oneself

66 There was also a more circuitous version of this path that was probably quite important for skill diffusion in certain areas and industries. During the Second World War retreat to Sichuan, and then, more lastingly, during the extended Maoist campaign to build heavy industries in China’s interior (where they would supposedly be safe from air attack in the event of war with the USA or USSR), large numbers of skilled ‘Shanghai’ workers (many of whom actually saw other Yangzi Delta towns as their ‘native place’) were relocated to interior provinces. When these people were allowed to return to Shanghai in the 1980s, many of them then took the further step of moving from Shanghai to Changzhou, Wuxi, etc. See Chris Bramall, The industrialization of rural China (Oxford: Oxford UP, 2007), 147-49 on the transfer of expertise to ‘third front’ zones. He places more emphasis, however, on the role of ‘sent-down youth’ during the Maoist era. 67 Ma, Hunhe yu fazhan, 280-90 68 LIU Ta-chung and Kung-Chia YEH, The economy of the Chinese mainland: National Income and economic development, 1933-1959 (Princeton: Princeton UP, 1965), 66. 118 kenneth pomeranz through agriculture alone. Geographically scattered concentrations of skill could prove very durable; Jing county, Anhui, remained the leading pro- ducer of high quality xuan paper (widely sought by elite calligraphers) for over 1,000 years. There and elsewhere, techniques appear to have mostly been learned on the job, in family-based firms; though the quantity of pa- per production in some locations was quite large, artisanal paper produc- tion has few economies of scale, so that growth generally meant a proliferation of small workshops. 69 Personal transmission of technical knowledge in this, as in many other rural crafts, was dominant—so much so that major demographic events could significantly change the local stock of knowledge. In parts of Shanxi, for instance, the great famine of 1876-79 seems to have resulted in a com- plete loss of the knowledge needed for certain industries; meanwhile an influx of refugees from Henan during the same famine led to significant changes in paper-making techniques.70 In Jiajing, Sichuan—an important regional paper-making center for which we have an unusually good study— three-year long formal apprenticeships did exist, but were used only in cases where a boy needed to learn a skill that no family member had.71 Formal associations of papermakers, recognised by the local government, existed in Jiajing by the early nineteenth century (and quite likely earlier), but as far as we know their principal activities were religious.72 New technical knowledge was often developed within particular work- shops, but generally spread rapidly within particular paper-making dis- tricts. Until quite recently, few Jiajing workshops had walls which would prevent somebody from looking into them; moreover, tools, labour, and the use of fixed facilities were frequently shared in ways that would have made it very difficult to keep secrets, though attempts to do so were not unknown.73 However women, who generally married out of their natal village, were frequently excluded from learning the most valued technical knowledge; this is a pattern that persists in village-owned industries in various parts of China today.74 Eyferth emphasises that in most cases, the

69 Eyferth, Eating rice, 29, 78-83, 181-89. Urban paper-making, on the other hand, was sometimes done in much larger factories. In some cases, marketing was more concentrated. 70 Henrietta Harrison, ‘Village industries and the making of rural-urban difference in early twentieth century Shanxi,’ in Jacob Eyferth, (ed.), How China works: Perspectives on the twentieth century industrial workplace (London: Routledge, 2006), 28. 71 Eyferth, Eating rice, 36. 72 Ibid., 98-103. In the 20th century, when the state sought to levy taxes specifically on paper-making, these organizations took on a more political role. 73 Ibid., 39-41. 74 Ibid., 41-42; Ellen Judd, Gender and power in rural North China (Stanford: Stanford UP, 1994), 86-107. skills, ‘guilds’, and development 119 keeping of technical knowledge within the existing paper-making zone was not so much the result of deliberate concealment as of the difficulty of learning many of the skills without actually being involved in the work, and the cooperative efforts it requires:75 the requisite skills do not fully exist until the moment when .. partners face and give a visual cue—a look or nod—to begin. In papermaking, such dis- tributed skills are most evident in the production of large six- or eight chi paper (97 × 180 or 124 × 248 cm), molded by teams of two to four workers who synchronise their movements in a slow, rhythmic dance. Socially dis- tributed skills also exist in the continuous operations around the vat, where teams of pulpers, vatmen, and brushers work in close cooperation, and in seasonal steaming, where groups of labor-exchanging neighbors constitute a second site of skill reproduction. Clearly, this was not true of all paper-making skills, much less of all artisanal skills more generally. A brief introduction to the history of xuan paper provided by the Anhui Provincial Library refers to parts of the production process being deliberately kept secret,76 as do some popular accounts of encounters with Japanese and English industrial spies early in the twenti- eth century. Yet the point that much artisanal knowledge is not only ‘tacit,’ or ‘embodied,’ but may be embodied in a group rather than in individuals, is an important one. Among other things, it suggests that the implicit counter-factual behind some of our assessments of how institutions af- fected the development and dissemination of technical knowledge—‘how would the knowledge have spread if the individuals who possessed it were on their own in the labour market?’—may sometimes be the wrong one. It is also striking that what observers described as a highly ‘traditional’ organisation for production and technical training in Jiajing paper-making was not inconsistent with considerable dynamism and technical change. In the 1920s, Jiajing papermakers learned to make newsprint, and soon dominated the provincial market (helped by political opposition to Japa- nese imports). A decade later, when the national government moved to Sichuan during World War Two, Jiajing paper-makers adapted successfully to the huge jump in demand that resulted; this required learning to make paper that would work in rotating printing presses and needed a different chemical composition as well as a different format. Lesser but still signifi- cant changes (such as watermarking, and using chlorine to bleach paper)

75 Eyferth, Eating rice, 42-44; quotation from p. 44. 76 http://www.ahlib.com/ahlib/tszy/gongyimeishu/wenfangsibao/xuanzhi.htm?sub fcode=007090110 120 kenneth pomeranz were also introduced during these years and some spread rapidly through the industry.77 It does not appear, then, that being unable to prevent one’s innovations from spreading to other producers within the local network necessarily led to technical stagnation. Meanwhile it appears quite possible that the abil- ity to keep innovations within the regional network may indeed have pro- vided—like the guilds that Epstein described—something akin in its effects to infant industry protection. The Gaoyang textile industry gives us another Republican-era case in which the sharing of technical knowledge across a (more formally-structured) rural industrial district proved quite compatible with on-going technological progress.78 But this is still a very limited sample, and we do not know at what rate (if any) such knowledge eventually leaked beyond the locale; ‘infant industry protection,’ after all, becomes more problematic if it lasts far beyond infancy (as it generally did not in Epstein’s cases). One might nonetheless tentatively suggest that within China’s very different institutional setting, rural kinship and local- ity networks may have performed some of the same functions that Epstein saw some European guilds performing, and provided some of the same (largely unintended) benefits to overall technical development. If this is indeed the case, we would have a very interesting example in which mem- bership in communities which had no formal legal status did the same sort of work—both for individuals and for the society—that the formal organisations and legal rights emphasised by Epstein did. We might then be in a position to go on to more nuanced discussions of how particular similarities and differences mattered. With the present state of our knowledge, however, this is piling conjec- ture upon conjecture. It is safer to say that, in rural China as in urban Eu- rope, organisational forms often accused by modernisers of being inflexible guardians of unchanging ways, turn out to have been a good deal more flexible and innovative in practice.79 If we were to attempt a very tentative overall comparison between China and Europe, we might suggest that China actually had less overall interference with markets, and thus very good prospects for Smithian growth; but that those same institutions, while they did not block technological change, may have been less conducive to

77 Eyferth, Eating rice, 40-41, 109. 78 Linda Grove, A Chinese economic revolution: Rural entrepreneurship in the twentieth century (Lanham, MD: Rowman and Littlefield, 2006). 79 For a Chinese example, see Eyferth, Eating rice, 110. This is, in fact, rather close to Bradstock’s basic view of Chinese guilds. skills, ‘guilds’, and development 121 it than Europe’s. (A bolder alternate hypothesis would be that they were only less conducive to capital- and energy-intensive kinds of technological change.) Such an argument, tentative though it is, would be broadly con- sistent both with Epstein’s arguments about Europe and with the main outlines of so-called ‘California school’ arguments about Chinese and com- parative development.

Japan

In Tokugawa (1603-1868) Japan, guilds were considerably stronger than in China, in large part because they often cooperated closely with government authorities. There has been an increasing tendency to question how much various rules that authorised monopolies or imposed strict standardisation were enforced in practice.80 This reinterpretation parallels what Epstein and others have done with respect to European guilds, but it is thus far more limited in chronological scope, focusing mostly on the period after about 1730. However far this revisionist literature on Japan ultimately winds up going, it is worth noting for our present purposes—as is also the case for the new view of European guilds—that the huge difference between statutory schemes in both Japan and Europe on the one hand and China on the other must have had some impact. Consequently, direct comparisons between Japanese dynamics and those which Epstein described for Europe are probably closer than those we have just discussed for China. In fact, one could argue that the combina- tion of fragmentation and unity in the Tokugawa system had very signifi- cant resemblances to what Epstein described for late medieval Europe. There were large numbers of jurisdictions in both places, many of which granted theoretical monopolies of various sorts to particular groups of producers; and for both places, enforcement of such monopolies seems to have varied greatly in its effectiveness, especially vis-à-vis rural produc- ers.81 European and Japanese polities also shared the experience of chron- ic fiscal pressures, though in the Japanese case these came largely from the costs of attendance at court and samurai stipends rather than warfare. These pressures created common incentives for domains to improve their ‘balance of payments,’ both by exporting high end products and by making do with whatever version of cheaper goods could be produced locally—

80 See e.g. Saito, ‘Land, labour, and market forces’. 81 Ibid., 183. 122 kenneth pomeranz whether local producers had a comparative advantage or not. And in Japan, as in early modern ‘proto-industrial’ Europe, we see the gradual decay in many domains of whatever monopolies had existed as rural industries developed—though in the Japanese case, the industries were more often developed with an eye towards import substitution, while European proto- industry was more likely to be export-oriented.82 Whether that difference affected the degree of skill development and innovation in these different settings is unclear. But here, too, the differences were as important as any similarities. In many Japanese cases, the connection between the domainal regime and the monopolistic guild was very strong, so that enforcement of restrictions could be very firm: at the famous Arita porcelain works, guards were post- ed around the village where the key artisans lived to prevent any secrets from leaking, and special care was taken to sub-divide the labour process to ensure that no one person knew all the relevant secrets of production. The key enforcers appear to have been the domainal authorities, inciden- tally, not the guilds.83 Likewise, the migration of Japanese skilled workers often seems to have been initiated from the top—by local authorities invit- ing skilled workers to visit or immigrate, rather than through worker-con- trolled patterns of circulation that connected different domains. There was certainly some flow of skilled workers towards the growing cities of Edo, Osaka, and Kyoto, with their huge luxury markets, but I am unaware of any worker-initiated flow from these major cities (which were probably on the technological cutting edge) towards second-tier cities, or among second- tier cities. There were, however, a significant number of artisans who left the cities entirely in the second half of the Tokugawa, as rural industry grew and urban/rural wage gaps narrowed. Overall, then, Tokugawa restrictions on migration were far from perfectly enforced; but they were nonetheless significantly stronger than in China after the early Ming, and probably stronger than those in much of Europe.84 Based on these indications of restrictionist measures, one might expect to see a relatively poor technological performance in Tokugawa Japan. But

82 The classic account for Europe is Peter Kriedte, Hans Medick, and Jürgen Schlum- bohm, Industrialization Before Industrialization (Cambridge: Cambridge UP, 1981). 83 Morris-Suzuki, The technological transformation, 30-31. 84 Saito, ‘Land, labour, and market forces’, 184-85 briefly describes these regulations and alludes to their highly variable enforcement. The unusually rapid growth of Western Europe’s urban population, at a time when cities remained demographic sinks (i.e. places with higher death rates than birth rates), suggests how ineffective restrictions on rural-urban migration had become in early modern Western Europe as a whole, skills, ‘guilds’, and development 123 there were also strong countervailing forces. These included a relatively literate population and widespread printing of technical manuals, plus considerable importation of technology—first mostly from China, later mostly from Europe—which was often encouraged by domainal govern- ments.85 And in many ways, the creation and especially diffusion of tech- nology in Tokugawa Japan looks quite impressive, particularly when compared to much more ‘liberal’ China. There seem, for instance, to have been far more changes in tools and tool-making, evident not only in the pages of those technical manuals, but in the seemingly endless sub-division of the carpenters’ guild into loom-makers, cart-makers, etc.86 The cross- fertilisation of techniques in different realms of endeavour seems to have been more vigorous than in China: the quick adoption of the thermometer (originally introduced by doctors) for use in monitoring conditions in silk- worm sheds is just one striking example87 that took much longer on the other side of the Yellow Sea. A couple of decades ago, this might have seemed to add up to a neat synthesis. Japanese artisanal organisation, one might suggest, was much more ‘feudal’ than China’s; now that we have seen that ‘feudal’ institutions like guilds had their technologically progressive side, we know why Japan did particularly well. But we are nowhere near a point where we can jump to such a conclusion. As Epstein was also intent on reminding us, there was no single model of how European guilds or city-states worked, even if we can recognise family resemblances: and the Japanese institutions in many ways seem more like those which he blamed for technological conservatism in Florence (with the urban guilds and the rulers of the surrounding coun- tryside very closely aligned) than like those of Lombardy, which he thought created more positive dynamics.88

85 Kawakatsu Heita, Nihon Bunmei to Kindai Seiyō: ‘Sakoku’ Saikō (Japanese Civilization and the Modern West: ‘The Closed Door Policy’ Reconsidered) (Tokyo: Nippon Hōsō, 1991); Kawakatsu Heita, ‘Nihon no kōgyōka o meguru gaiatsu to Ajia kan kyōsō’ (Outside Pressures Surrounding Japanese Industrialization and Intra-Asian Competition), in: Hamashita and Kawakatsu (eds), Ajia kōekiken to Nihon kōgyōka 1500-1900 (The Asian Trading Sphere and Japanese Industrialization) (Tokyo: Riburopōto, 1991); Christian Daniels, ‘Jushichi, hachi seiki Higashi, Tōnan Ajia ikinai boeki to seisan gijutsu iten: seitō gijutsu o rei to shite’ (The Intra-regional Trade in Seventeenth and Eighteenth Century East and Southeast Asia and the Transfer of Productive Technologies: the Case of Sugar Refining), in: Takeshi and Heita (eds), Ajia kōekiken to Nihon kōgyōka 1500-1900. 86 Morris-Suzuki, The technological transformation, 51-52. There was less proliferation of metal tools than in Europe, but that is hardly surprising given the costs of fuel—something that was an even bigger problem in key regions of China. 87 Ibid., 40. 88 Morris-Suzuki, The technological transformation, 28-32; Epstein, Freedom and growth, 124-38, 161-67. 124 kenneth pomeranz And when one gets to the nineteenth-century adoption of modern tech- nologies, the relationships among skill levels, organisation, and techno- logical progress seem to flip again: Japanese industries seem to have depended more on new institutions and formal, book-based training to create a labour force for modern industries than Chinese industries did, and to have had less continuity of personnel between old and new ship- builders, iron-workers, and so on.89 Of course, as noted before, the rate of technological change was faster in Japan, and the state moved more vigor- ously to create and sustain arsenals, technical schools, and other centers for the acquisition of new skills.90 Thus some of this difference may be less a matter of the adaptability of Chinese artisans than of the lesser adequa- cy of any artisanal tradition amidst the especially rapid technological change in Meiji and Taishō Japan; and, conversely, of the slower rates at which modern channels of skill formation were developed in China. But while those are certainly plausible explanations of the persistence of arti- sanal skills within Chinese factories, they may not be the whole story: the adaptability of Yangzi Delta craftsmen to twentieth-century industrial roles is very impressive. Moreover, the role of pluri-active rural families in shap- ing what Sugihara calls the ‘quality of labour’ may be a similarity across East Asian cores as important as any differences among them.

Conclusion

I would like to end with hypotheses that are much more modest than many of the quite speculative ideas which have been suggested in the rest of this chapter. These final propositions consist mostly of negative statements about conclusions we should not jump to and guesses about issues that would be worth exploring, rather than declarative statements about the nature of particular institutions (much less whole societies). However, I think they take us back to some of the points that Epstein’s work helped bring out within a European context, and suggest additional ways to make use of them in East Asian and comparative contexts. One such point is that there was no necessary reason that the same institutions that were good for encouraging market integration were good

89 Pauer, ‘Traditional technology’. 90 Benjamin Elman, On their own terms: Science in China, 1550-1900 (Cambridge Mass.: Harvard UP, 2005); Wang hsien-chun, ‘Discovering steam power in China, 1840s–1860s’, Technology and Culture 51 (2009), 31-54.; for Japan Kozo Yamamura, ‘Success ill-gotten? The role of Meiji militarism in Japan’s technological process’, Journal of Economic History 37 (1977), 113-35. skills, ‘guilds’, and development 125 for encouraging technological innovation. Or, to put in another way, insti- tutions which helped an economy move close to its production possibility frontier didn’t necessarily have much in common with those that would shift that frontier outward. Epstein made this point, in a more general form, in the introduction to a posthumously published volume he co-edited with Maarten Prak, that trying to label institutions as either good or bad for development is often not very helpful in helping us see how they actually worked.91 While this may seem obvious, this is so only once the existence of features inconsistent with the old consensus have been acknowledged. Indeed, some participants in the discussion of European guilds would strongly disagree that it applies here.92 A second, and related point, is that changes in the volume of industrial production might not greatly affect the likelihood of changes in the tech- nology of production. From the point of view of employment, output, in- come distribution and living standards at a given moment—all important issues—rural proto-industrialisation, in all its many forms and locations, remains a huge story; but that did not necessarily make it a promising source of major technological changes, and especially not of the capital- and energy-intensive kinds of technological change that ultimately created a very different economic world. Several of the possible reasons for this that Epstein mentioned are well worth exploring further: particularly, I think, the high monitoring costs for anyone who has wanted to introduce new techniques. In this, process in- novations often differed from the introduction of new products, where the compliance of workers could be easily monitored—and which happened all the time, in both urban and rural settings. Product innovations within a roughly constant technology of production was, for instance, a very sub- stantial part of the way that proto-industrial producers in the leading re- gions of East Asia coped with competition from new low-wage entrants. (The development of new cloth patterns in Jiangnan when other people in other areas became able to match their plainer weaves is a good example.)93 In many rural-based Chinese industries, the predominance of

91 Epstein and Prak, ‘Introduction’, especially 23-24. 92 Ogilvie, ‘“Whatever is, is right”?’, and ‘Rehabilitating the guilds’. 93 A number of examples are given in Li Bozhong, Jiangnan de zaoqi gongyehua 1550- 1850 (Early industrialization in Jiangnan)(Shanghai: Shehui kexue wenxian chubanshe, 2000). Product innovations in the early modern textile industries of India seem to have been even more numerous and important; see Prasannan Parthasarathi, Why Europe grew rich and Asia did not: Global economic divergence, 1600-1850 (Cambridge: Cambridge Uni- versity Press, 2011). 126 kenneth pomeranz small family-based production units kept monitoring costs low, and cer- tainly did not rule out significant process innovations; as we saw in the case of Jiajing paper-making, such innovations could spread rapidly through an industrial district, even when production was scattered among many small workshops. But—even leaving aside important questions about diffusion beyond one such district—such firms were unlikely to initiate process in- novations that would involve significant economies of scale, or greatly increase capital/labour ratios, and thus labour productivity. Conversely, this suggests that in looking for institutional explanations of rates of technological change in pre-modern industry, we may want to pay particular attention to spillovers from a relatively small group of high- end industries producing for relatively exacting consumers (e.g. instru- ment-production, clocks, weapons, etc.) who were not particularly price-sensitive—and at the opportunities for spillovers created when work- ers in those industries came into contact with new people who had differ- ent problems. In the early modern European case, a renewed interest in the role of weapons productions and spillovers from it already seems to be underway.94 For modern East Asian cases, this adds one more reason for going back to look again at the role of arsenals. These institutions figured prominently in early accounts of Japanese and Chinese industrialisation,95 then largely got dropped from the story as economic historians switched from narrative to econometric approaches, and became more skeptical of the contributions of firms that were not themselves profitable without state subsidies. They are now attracting some interest again, though thus far more from historians of science, technology and culture than from eco- nomic historians.96 Finally, all these issues suggest one last point that I think Epstein himself would have emphasised. We need to look at the

94 Philip Hoffman, ‘Prices, the Military Revolution, and Europe’s comparative advantage in violence’, The Economic History Review 64 (2011 issue supplement), 41-51, 57. Certain important spin-offs—e.g. from metal-cutting and precise boring for weapons production to the making of better-fitted pistons for steam engines—have long been known. But see Joel Mokyr, The lever of riches: Technological creativity and economic progress (Oxford: Oxford University Press, 1990), 183-86 for a view that warfare did not have positive net effects on the development of civilian technology. 95 Yamamura, ‘Success ill-gotten?’. 96 Elman, On their own terms; Meng Yue, ‘Hybrid science versus modernity: The prac- tice of the Jiangnan Arsenal, 1867-1904’, East Asian Science, Technology, and Medicine 16 (1999), 13-52; WANG Hsien-chun, ‘Zhengqi tuidong de lishi: zhengqi jishu yu wan Qing Zhongguo shehui bianqian 1840-1890’ (History driven by steam: Steam technology and social transformation in Late Qing China, 1840-1890), Zhongyang yanjiuyuan jindai shi yanjiusuo jikan vol? (2009), 41-85. skills, ‘guilds’, and development 127 rights, bargaining positions and freedoms (in the plural) of rather specific, strategic, sub-groups of people—including some that may have infringed on the rights of other groups—rather than at a more abstract summary of how much ‘freedom’ (in the singular) characterised a society as a whole. 128 kenneth pomeranz skills, ‘guilds’, and development 129

PART TWO

INDUSTRY PERSPECTIVES 130 kenneth pomeranz mega-structures in the middle ages 131

CHAPTER FOUR

MEGA-STRUCTURES OF THE MIDDLE AGES: THE CONSTRUCTION OF RELIGIOUS BUILDINGS IN EUROPE AND ASIA, C.1000-1500*

Maarten Prak

Introduction

When building began in 1136 on the new choir for the church of Saint De- nis, north of Paris, the Latin West started a building boom that was to last some three hundred years.1 The Gothic building style, of which the Saint Denis choir was the first example, originated as the visual expression of a bid for power by the French kings, and early projects were undertaken in the region around their capital, Paris. Their size and symbolism were trib- utes in equal measure to the power of God, and that of the sovereign as his worldly representative.2 The style became popular throughout Europe. Medieval cathedrals are still among the tallest structures in many Euro- pean towns. They are triumphs of human ingenuity, culturally as well as technologically.

* This is a slightly modified version of an article that was first published in the Journal of Global History 6 (2011), 381-406. It is reprinted here with permission from Cambridge University Press and the London School of Economics. Helpful suggestions came from Lex Bosman (Amsterdam), Qinghua Guo (Melbourne), and Aart Mekking (em. Leiden), from audiences at Universität Konstanz, Cambridge University, the London School of Econom- ics, and Utrecht University, from five anonymous referees and from the editors of the JGH, especially William Gervase Clarence-Smith, who was very supportive of the project. Throughout the work on this article, the enthusiasm of Merlijn Hurx (Utrecht) and Jan Luiten van Zanden (Utrecht) has been most encouraging. Essential assistance was provided by Koen Ottenheym (Utrecht). Daniel Curtis (Utrecht) improved my English. The original idea of the article was inspired by materials found in S.R. (Larry) Epstein’s study after his death in 2007, and by his paper published in this volume. All errors are mine alone. 1 Robert A. Scott, The Gothic enterprise: A guide to understanding the Medieval cathedral (Berkeley etc.: University of California Press, 2003), ch. 1. The term ‘Latin West’ is used because later the Byzantine Empire is also discussed, which was partly located in Europe. To avoid cumbersome description, I use the word ‘Europe’ when in fact I mean the Latin West. 2 Dieter Kimpel and Robert Suckale, Die gotische Architektur in Frankreich 1130-1270, Munich: Hirmer Verlag, 1985, 67-76. Compare Elena Paskaleva, ‘The architecture of the Four-Īwān building tradition as a representation of paradise and dynastic power aspirations,’ PhD-dissertation Leiden University 2010. 132 maarten prak The Latin West however, was not the only area of Eurasia where major building projects were undertaken for religious purposes.3 In China during the Song (960-1279), many pagodas and pagoda towers were built, of heights not previously attained.4 In Cambodia from the eighth to the thirteenth centuries, ‘admirable structures of exceptional scope’ were built, most no- tably the Angkor Wat temple complex.5 In Asia Minor, spectacular works had been completed earlier, like Hagia Sophia (Aya Sofya) in Istanbul, and during the Byzantine era routine church construction continued apace.6 In the Islamic world, a great many architectural monuments date from the first half of the previous millennium, and many of these were built with a religious purpose. These include several mosques and madrassas in Cairo, the Great Mosque in Bursa, Turkey, and the 72.5 metres high Qutb minaret in Delhi, as well as non-religious buildings like the Alhambra in Granada, Spain.7 Much as we can admire the artistic achievements of these works, they also raise questions of a more practical nature. How did builders mas- ter the techniques necessary to erect these large structures and ensure their technical integrity? Because of the complex challenges associated with tall constructions, and their presence throughout Eurasia, the building industry provides us with an opportunity to compare the application of cutting-edge technol- ogy under more or less identical circumstances across many different regions and cultures. It allows us in other words, to identify the character- istics of the technological platforms for economic development available to the societies that are central to the debate on the ‘Great Divergence’.8 By focusing on a single industry, this chapter may offer a form of comparison

3 James W.P. Campbell, Brick: A world history (London: Thames & Hudson, 2003), ch. 3. 4 Ibid., 92. 5 Jacques Dumarçay and Pascal Royère, Cambodian architecture, eighth to thirteenth centuries, transl. and ed. by Michael Smithies, Handbook of Oriental Studies, section 3: South-East Asia, vol. 12 (Leiden: Brill, 2001), 109. 6 Robert Mark and Ahmet Ş. Çakmak (eds), Hagia Sophia from the age of Justinian to the present (Cambridge: Cambridge University Press, 1992); Robert Ousterhout, Master builders of Byzantium (Princeton: Princeton University Press, 1999). 7 George Michell (ed.), Architecture of the Islamic world: its history and social meaning, London: Thames & Hudson, 1984: 209-80. Other surveys include K.A.C. Creswell, A short account of early Muslim Architecture revised and supplemented by James W. Allan (Alder- shot: Scolar Press, 1989); Robert Hillenbrand, Islamic architecture: form, function and mean- ing (Edinburgh: Edinburgh University Press, 1994), 66-128; M. Shokooy, Muslim architecture of South India (London: Routledge, 2002). 8 Referring to Kenneth Pomeranz, The great divergence: China, Europe and the making of the modern world economy (Princeton: Princeton University Press, 2000), and the subse- quent debate about the argument. mega-structures in the middle ages 133 that has so far been under-utilised in global economic history, where schol- ars instead have followed in the footsteps of Pomeranz’ book, by preferring to compare geographical regions, in spite of the problems created by such heterogeneous units of analysis.9 As in other historical disciplines, the history of technology is usually subdivided into two clearly distinguishable periods: before and after the eighteenth century. In earlier ages, technological improvements were not entirely absent, but they were haphazard, the result of trial and error.10 Once scientists were able to explain the underlying principles, that is when the Scientific Revolution kicked in, inventors could develop new technolo- gies much more systematically. As a result, the pace of technological improvements increased dramatically, leading in turn to long-term eco- nomic growth. There is no reason to quarrel with the general outline of the story. It does, however, offer some intriguing questions about technology and technological change in the pre-modern era. To address these it is useful to start with a distinction that Joel Mokyr made between propositional or ‘what’ knowledge, and prescriptive or ‘how’ knowledge, which he also called Ω or epistemic, and λ or technical types of knowledge.11 For Mokyr, the two are intimately related: ‘for a technique to exist, it must have an epistemic base in Ω’.12 It is immediately clear that the weakness of Ω before the Scientific Revolution of the seventeenth cen- tury, must have fatally held back the development of λ, or technology, and thus by implication economic progress. However, the building industry seems to suggest a different relation between Ω and λ types of knowledge. The industry achieved impressive accomplishments in practice, without much change in the theoretical foundations of the building craft. This raises three specific questions: first how could the building industry progress along the scales of size and com- plexity, second what was the nature of the knowledge utilised by builders, and third how was this knowledge transmitted from one generation to the next?

9 Other examples include Karel Davids, ‘River control and the evolution of knowledge: a comparison between regions in China and Europe, c. 1400-1800’, Journal of Global History 1 (2006), 59-80, and Robert Finlay, The pilgrim art: cultures of porcelain in world history (Berkeley: University of California Press, 2010), as well as the papers in this part of the pres- ent volume. 10 Steven A. Epstein, An economic and social history of later medieval Europe, 1000-1500 (Cambridge: Cambridge University Press, 2009), 190-215. 11 Joel Mokyr, The gifts of Athena: Historical origins of the knowledge economy (Prince­ton: Princeton University Press, 2002), ch. 1. 12 Ibid., 13. 134 maarten prak The building industry is an appropriate laboratory to investigate issues of knowledge and human capital formation, for several reasons. First and foremost, the construction of large religious buildings far outstripped the scientific understanding of the mechanics working on such large struc- tures. Secondly, the industry experienced a string of innovations necessary to realize the artistic ambitions of those who commissioned Gothic church- es and similar structures. Thirdly, the industry was a truly global one. Fi- nally, thanks to the record keeping of church institutions, and the hard work of generations of art historians, the source material for this industry is relatively abundant, especially for Europe. The more complete European picture is used here to flesh out the fragmented evidence from the Middle East and South and East Asia. This comparison allows us to gauge whether Europe’s production and reproduction of λ-knowledge was unique, or instead fitted a wider pattern. Unfortunately medieval builders have left preciously few written docu- ments about the theories underpinning their work. Where available, those documents articulating construction theories are used here, but we also have to rely on the material record, that is the church buildings themselves, and what these can tell us about the ways in which they were constructed.13 In this chapter, data from the art-historical literature is applied to questions raised by global historians and the history of technology.

Church Building

Between roughly 1150 and 1250, the new Gothic style transformed both the outlook and the construction of European religious buildings in funda- mental ways. The stylistic innovations were first and foremost a radical shift in the balance between walls and windows. Romanesque buildings had thick walls and relatively few windows, whereas Gothic churches con- sisted almost entirely of glass, held together by slim stone pillars. Other innovations included the placing of towers, which proliferated in Gothic buildings, the development of sculpted porches placed in deep recesses,

13 This approach has been inspired especially by the works of John James and Robert Mark cited in footnotes 6, 15, 18, 21, 49, 65, 94, as well as Robert Mark (ed.), Architectural technology up to the Scientific Revolution: The art and structure of large-scale buildings (Cambridge Ma.: MIT Press, 1993). See also Salvatore D’Agostino, ‘Historical buildings as an archive of the material history of construction’, in: Antonio Becchi, Massimo Corradi, Federico Foce, and Orietta Pedemonte (eds), Towards a history of construction: Dedicated to Edoardo Benvenuto (Basel: Birkhäuser, 2002), 369-84. mega-structures in the middle ages 135 the creation of round or ‘rose’ windows, the ribbed vault, and the flying buttress.14 The Gothic design was first developed in the region around Paris, and many innovations also originated there, although not always in connection with the famous cathedrals.15 From the Île-de-France, the style spread to other parts of Europe, being adapted to local tastes and traditions. In England, for instance, flying buttresses were rarely found, and builders continued the tradition of constructing thicker walls.16 In Northern Ger- many, where natural stone was in short supply, builders had to construct Gothic designs in brick, creating the distinctive Backsteingotik, or brick Gothic.17 Construction challenges for tall buildings include the problems created by the structure’s own weight, the resistance to pressures from the natural elements, especially wind, and the necessity of natural light to illuminate the interior of the building.18 Gothic design was a huge improvement over its predecessor (Romanesque) in terms of lighting, because it allowed far larger surfaces of glass. At the same time, the height of the buildings made them more susceptible to wind pressures. The less solid construction of the vertical elements only increased the problems that builders had to overcome.19 An important aid was the ‘invention’ of the flying buttress.20 Through a complex combination of horizontal extension and vertical loading, build- ers managed to prevent the vaults from pushing the walls of the church outwards, and the vaults themselves from collapsing as a result.21 ­Another

14 Christopher Wilson, The Gothic cathedral: the architecture of the great church 1130- 1530 (London: Thames & Hudson, 2000), 69-70; Bill Addis, Building: 3000 Years of design engineering and construction (London: Phaidon, 2007), 95-99. 15 John James, The template-makers of the Paris Basin (Leura: West Grinstead Nominees, 1989), ch. 9. 16 Wilson, Gothic cathedral, 74. 17 Rudolf Pfefferkorn, Norddeutsche Backsteingotik (Hamburg: Christians, 1984). 18 Robert Mark, Light, wind, and structure: the mystery of the master builders (Cambridge Ma.: MIT Press), 1990, ch. 2. 19 Philip Ball, Universe of stone: a biography of Chartres cathedral (New York: Harper Collins, 2008), ch. 8; also Jacques Heyman, The stone skeleton: Structural engineering of stone masonry (Cambridge: Cambridge University Press, 1995). 20 Jean Gimpel, The medieval machine: The Industrial Revolution of the Middle Ages (London: Victor Gollancz, 1977; orig. La revolution industrielle du Moyen Âge), 121. 21 Alan Borg and Robert Mark, ‘Chartres cathedral: a reinterpretation of its structure’, The Art Bulletin 55 (1973), 367-72; M. Wolfe and R. Mark, ‘Gothic cathedral buttressing: the experiment at Bourges and its influence’, Journal of the Society of Architectural Historians 33 (1974), 17-26; William Clark and Robert Mark, ‘The first flying buttresses: a new recon- struction of the nave of Notre-Dame de Paris’, The Art Bulletin 66 (1984), 47-65; Mark, Light, ch. 1, and 108-22. 136 maarten prak was the development of the pointed arch, which through the double arc of its construction created downward rather than outward pressure; this they probably copied from Arabic examples.22 Builders had to intuitively understand these problems, and find solutions for them, as they had no laboratories to test the various pressures to which their constructions were exposed.23 In the Byzantine Empire the heyday of church building was long over by AD 1000. Hagia Sophia dated from the sixth century, and no equivalent church buildings were constructed during the subsequent millennium. However, a great many new churches were created in the revival following the Iconoclast Controversy, which ended in 843. Moreover, this period saw the development of a new church lay-out, the so-called cross-in-square. Contrary to the Latin churches, where the cross shape determined the ground plan of the building, this Byzantine type was almost square, with a cross shape inserted into the interior lay-out.24 Brick was the dominant building material in South Eastern Europe and Asia Minor. Such was also the situation in much of the Islamic world, where building stone was in short supply. Many early mosques were relatively simple con- structions, consisting of a flat roof supported by rows of pillars. Their de- signers, descendants of nomads, had no great building tradition to help them.25 Early mosques incorporated Byzantine examples, building knowl- edge and building materials.26 Ottoman builders later also used Byzantine examples as a source inspiration, but even they were intimidated by the constructional challenge of Hagia Sophia, as they found it difficult to con- struct a dome of equal size. This, incidentally, was used by Christian critics as proof that Christianity was superior to Islam.27 In the sixteenth cen- tury, however, Ottoman architecture blossomed again, under the inspiring leadership of Sinan ibn Abdülmennan, who was appointed Chief Court Architect in 1538. Sinan single-handedly designed almost 500 mosques, and

22 Donald Hill, A history of engineering in Classical and Medieval times (London: Croom Helm, 1984), 99. 23 Cf. Nicola Coldstream, Masons and sculptors: medieval craftsmen (London: British Museum Press, 1991), 60. For modern laboratory experiments on Gothic constructions, see the works of Robert Mark, esp. Light. 24 Ousterhout, Byzantium, 12. 25 Ulya Vogt-Göknil, Die Moschee: Grundformen sakraler Baukunst (Zürich: Verlag für Architektur, 1978), 22, 26; Cresswell, Muslim architecture, 16. 26 Henri Stierlin, Islam von Bagdad bis Córdoba: Frühe Bauwerke vom 7. bis 13. Jahrhun- dert (Hong Kong etc.: Taschen, 2009), 28, 32, 48, 52. 27 Metin and Zeynep Ahunbay, ‘Structural influence of Hagia Sophia on Ottoman mosque architecture’, in: Mark and Çakmak (eds), Hagia Sophia, 180. mega-structures in the middle ages 137 was the main source of inspiration for the best-known building from the era, the Sultan Ahmed Mosque, or Blue Mosque, in Istanbul.28 Most of the building in Song China (960-1279) remained of the tradi- tional wooden post and beam type. An important innovation during this period, however, was the construction of brick towers, of which quite a few survive. They seem to have been introduced by the Liao dynasty (907-1125), who were the first non-Chinese rulers of the empire.29 Their Buddhist tow- ers often achieved impressive heights. The ‘Iron Pagoda’ of Kaifeng, which in fact was made of iron-grey glazed bricks, reached 57 metres, the wooden pagoda of the Fogong temple in Shanxi came to 67 metres, and the Liuhe, or Six Harmonies Pagoda in Hangzhou, also built in brick, stood 60 metres tall. The tallest still in existence is the Liaodi Pagoda, built in 1055, which comes to 84 metres.30 Unlike most European towers, the Buddhist pagodas in China were free-standing buildings, without any support from an ap- pended church building.31 The buildings therefore had wide-spreading eaves, which helped to stabilise the construction.32 Aside from the usual dangers of fire (the pagoda in Kaifeng was struck by lightning and burned down in 1044), they also had to cope with problems of high winds and earthquakes.33 For this reason, the wooden posts and beams method of building was often preferred, set on top of a concrete platform, but not inserted into it. The curved shape of the roofs helped the wind to skid over the structure, while the extension of the roofs also protected the wooden frame from the rain. Chinese builders used roof tiles up to four times as heavy as those utilised in Europe, again to stabilise the constructions. Com- pared to European cathedrals, pagodas were built quickly, in less than ten years.34

28 Aptullah Kuran, Sinan: The Grand Old Master of Ottoman architecture (Istanbul: Ada Press, 1987); Robert Hillenbrand, Islamic art and architecture (London: Thames & Hudson, 1999), 264. 29 William Watson, The arts of China 900-1620 (New Haven: Yale University Press, 2000), 67-72. 30 Watson , China, ch. 5; also Ernst Börschmann, Die Baukunst und religiöse Kultur der Chinesen vol. 3: Pagoden (Berlin: Walter de Gruyter, 1931). 31 Qinghua Guo, Chinese architecture and planning: ideas, methods, techniques (Stutt- gart: Edition Axel Menges, 2005), ch. 5. 32 Watson, China, 78. 33 Else Glahn, ‘Chinese building standards in the 12th century’, Scientific American 244/5 (1981), 132; Thomas Ledderose, Ten thousand things: Module and mass production in Chinese art (Princeton NJ: Princeton University Press, 2000), 106, 110. 34 Guo, Chinese architecture, 64. 138 maarten prak The Workforce

Major European building projects employed highly variable workforces. During the summer months of 1253, up to 435 people were working on Westminster Abbey, of whom 130 were stone masons, and 220 were assis- tants employed on various tasks. In September the number was halved, mainly because the assistants left the site, presumably to participate in the harvest. For the building of the cathedral in Regensburg, in 1459 master Conrad Roriczer employed between eight and twelve stone masons and a lodge-assistant, an apprentice for three weeks, as well as two to four car- penters.35 The employment of a few dozen workers seems to have been the normal situation. ‘Routine’ activities could be interspersed with short cam- paigns, when much larger numbers were employed. The building of Can- terbury cathedral, which lasted from 1175 till 1517, consisted of 161 years of high activity and 182 years of low activity.36 The specialist workforce was recruited from a wide area and was there- fore highly mobile.37 On the building site of Prague Cathedral in 1372-78, next to the local workers, Germans from twenty different towns and regions were found, as well as workers originating from Brabant in the Low Coun- tries, Hungary and Poland. Of 131 stone masons employed at Vale Royal Abbey between 1278 and 1280, at most 10 percent were recruited locally; the rest came from a variety of English counties. On the other hand, at least half the carpenters and blacksmiths were local people, and of the semi- skilled or unskilled assistants, around ninety percent came from the area itself.38 Similar mobility among stone masons was found in the town of Kampen, in the Netherlands. During the fifteenth and sixteenth centuries, the town’s stone masons, mainly charged with the building of the town’s major churches, came from a wide area in the Low Countries and Western Germany. The town’s official carpenters, however, seem to have been re- cruited mostly (although not exclusively) from the ranks of local masters.39 Both the organisation of the building process and the distribution of tech-

35 Günther Binding, in collaboration with Gabriele Annas, Bettina Jost and Anne Schunicht, Baubetrieb im Mittelalter (Darmstadt: Wissenschaftliche Buchgesellschaft, 1993), 272-76. 36 Scott, Gothic enterprise, 40. 37 Spiro L. Kostof, ‘The architect in the Middle Ages, East and West’, in: eod (ed.), The architect: chapters in the history of the profession (Berkeley: University of California Press, 2000), 83. 38 Ibid., 284-85. 39 Chr. J. Kolman, Naer de eisch van ’t werck: De organisatie van het bouwen te Kampen 1450-1650 (Utrecht: Matrijs, 1993), ch’s 7 and 9. mega-structures in the middle ages 139 nical knowledge required a collaboration of local and itinerant crafts- men.40 Transportation costs were such that stone was preferably quarried on the site of the building project. Stone used in Marlborough in 1237 cost 3 shilling per unit at the quarry, but another 22 shillings for carriage. In the early fourteenth century, land transport of stone across five or six miles cost the equivalent of the stone itself, and it was much cheaper to carry it across water. Stone from Normandy was shipped through Caen to England, and in 1287 the transport costs of a shipload ordered for the building of Norwich Cathedral was only twice as expensive as the stone itself.41 A first implication was that, whenever the two were situated far apart, stone was preferably prepared at the quarry, rather than at the building site, otherwise expensive transport would be wasted on excess material. Moreover this allowed the identification of low-quality pieces before shipping.42 In the Low Countries, stone was found in a crescent-shaped region running roughly from Ghent, across Brussels, to Maastricht. This stone supplied the northern territories, which were deprived of their own stock of raw mate- rials.43 The system eventually became so sophisticated that, in the course of the Middle Ages, complete pre-fabricated churches were prepared at the quarry to be shipped as parts to the building site, where they were assembled by local workmen.44 The other implication was that crucial knowledge tended to be con- nected to the building material. Again the Low Countries provide intrigu- ing examples.45 Between the late fourteenth and mid-sixteenth century, successive generations from the Keldermans family, who originated from Brussels but later moved to Antwerp, were contracted to supply stone, as well as design and execute numerous public buildings, both religious and civic, in the Northern Netherlands. They prepared complete plans (‘pattern’

40 S.R. Epstein, ‘Transferring technical knowledge and innovation in Europe, c.1200- c.1800’, unpublished paper, presented at the Economic History Association Annual Confer- ence, San Jose, September 2004. 41 L.F. Salzman, Building in England down to 1540: a documentary history (Oxford: Clarendon, 1967), 119, 132-33. 42 H. Janse and D.J. de Vries, Werk en merk van de steenhouwer: Het steenhouwersam- bacht in de Nederlanden voor 1800 (Zwolle: Waanders, 1991), 10. 43 Ibid., 10-19. 44 R. Meischke, De gothische bouwtraditie (Amersfoort: Bekking, 1988), 79-84; Merlijn Hurx, ‘Middeleeuwse “pre-fab” in de Nederlanden? De Hollandse kerken van de Antwerpse loodsmeester Evert Spoorwater’, KNOB-Bulletin 106 (2007), 112-34, and Hurx, Architect en aannemer: De opkomst van de bouwmarkt in de Nederlanden 1350-1530 (Nijmegen: Vantilt, 2012). 45 Janse and de Vries, Werk en merk, ch. 5. 140 maarten prak in the sources), as well as specific features, and provided the templates for the various stone parts necessary to erect the structure. The Keldermans firm also supplied their customers with sculptures, and were charged with working the stone from quarry to finished building.46 Before a church was built in stone, it had to be constructed in wood. Because these wooden structures were taken down after the building was finished, one often overlooks the contribution of carpenters, who were vital to the whole construction process.47 First, carpenters set up the scaf- folding that allowed builders to move around. For the building of Westmin- ster cathedral, the accounts of 1324 refer to 400 pieces of alder of 38 feet, 25 pieces of 20 feet, and 61 pieces of ash of 42 feet; all this wood was used for scaffolding.48 Secondly, complex parts like stone vaults were set in wooden frames until the mortar hardened, a process that could take six months and longer.49 Thirdly, some important parts of the church were built of wood, notably the roofs. The vaults were usually made of stone, but these were covered by roofs made entirely of timber and covered in slate, tile or lead to keep the water out. These timber roof structures were often very elaborate. It is estimated that the roof of Our Lady’s Church (Lieb- fraukirche) in Ingolstadt, Germany, contained the wood of some 3,800 trees.50 Unfortunately very little is known about the carpenters who undertook this work. We do know that carpenters had to be able to use a variety of tools to exercise their trade. The largest number of different items listed in a study of medieval building tools concerned carpentry. These included various types of axes, saws, drills, hammers, and chisels, as well as a com- pass, square, plane, file, and rasp.51 Carpenters seem to have been mostly local craftsmen, but this is inferred from the absence of references to itin- erant carpenters, rather than from any positive evidence.

46 J.H. van Mosselveld, Keldermans: Een architectonisch netwerk in de Nederlanden (The Hague: Staatsuitgeverij, 1987), ch. 1. 47 John Fitchen, The construction of Gothic cathedrals: A study of medieval vault erection (Oxford: Clarendon Press, 1961), ch. 2; James W.P. Campbell ‘The carpentry trade in seven- teenth-century England’, The Georgian Group Journal 12 (2002), 216-17. 48 Salzman Building in England, 318-19. 49 John James, Chartres: The masons who built a legend (London: Routledge & Kegan Paul, 1982), 41. 50 Barbara Schock-Werner, ‘Bauhütten und Baubetrieb der Spätgotik’, in: Anton ­Legner, (ed.), Die Parler und der Schöne Stil 1350-1400: Europäische Kunst unter den Luxemburgern vol. 3, exhib. catalogue (Cologne: Kunsthalle, 1978), 58. 51 Frieda Van Tygem, Op en om de middeleeuwse bouwwerf: De gereedschappen en toestellen gebruikt bij het bouwen van de vroege middeleeuwen tot omstreeks 1600 (Brussels: Paleis der Academiën, 1966), 1-70. mega-structures in the middle ages 141 One source to have survived on the building workers who constructed the Timurid sacred buildings, in what is today northern Iran, Afghanistan, Uzbekistan, and Turkmenistan, is a picture copied in 872/1467, which is now in the Johns Hopkins University library. It shows a variety of artisans, including stone cutters, tile makers, brick masons, and mortar makers. There is also scaffolding in evidence, suggesting a contribution by carpen- ters. Equally interesting is the variety of ethnic backgrounds of the work- men pictured here, which include people from the region, as well as people seemingly of Mongol and African descent, suggesting mobility among building workers.52 The variety of crafts is confirmed by a compilation of the names of craftsmen found on a range of Timurid buildings which cov- ered over twenty different crafts; carpenters were among the most numer- ous with 15 references, out of a total of 107 craftsmen.53 This seems typical of the situation in other parts of the Islamic world, where we find the same types of crafts on building sites.54 These must have been highly skilled men. One of their great feats was the building of large domes without centering, that is without the help of a wooden sup- port structure. A striking difference with European building practices was the speed of construction; in the Islamic areas large structures were usu- ally erected within a decade.55 Although some master-builders were iden- tified by name, the great majority of them worked anonymously, as they did in the Latin West.56 This underlines the dominant role of the com- munity of builders, as opposed to the single creator. The Islamic building community was highly mobile, possibly due to conscription or slavery, rather than voluntary migration.57 In Egypt occupational terms in the building industry closely followed those of raw materials.58

52 Lisa Golombek and Donald Wilber, et al. (1988), The Timurid architecture of Iran and Turan, 2 vols (Princeton: Princeton University Press, 1988), vol. I, 91-92, and vol. II: pl. 481. 53 Ibid. vol. I, 65-66; Oleg Grabar, Renata Holod, ‘A tenth-century source for architec- ture’, in: Ihor Sevchenko, Frank E. Sysyn (eds), Eucharisterion: Essays presented to Omeljan Pritsak on his sixtieth birthday by his colleagues and students, special issue of Harvard Ukra- nian Studies 3-4 (1979-80), 316. 54 Hans E. Wulff, The traditional crafts of Iran: their development, technology, and influ- ence on Eastern and Western Civilizations (Cambridge Ma.: MIT Press, 1966), 102-35. 55 L.A. Mayer, Islamic architects and their works (Geneva: Albert Kundig, 1956), 23. 56 Ibid., 21, and Kostof, ‘Architect’, 65; but see also Donald L. Wilber, ‘Builders and craftsmen of Islamic Iran: the earlier periods’, Art and Archaeology Research Papers (AARP) 10 (1976), 31-40. 57 Ronald Lewcock, ‘Architects, craftsmen and builders: materials and techniques’, in: Michell (ed.), Architecture of the Islamic world, 131, 134; Mayer, Islamic architects, 28. 58 Maya Shatzmiller, Labour in the medieval Islamic world (Leiden: Brill, 1994), 211. 142 maarten prak In India, architects and sculptors were indicated with the same word rūvār, suggesting a connection between knowledge of stone and construc- tion expertise that was similar to what we observed in Europe.59 Indian builders were initially seen as being of inferior social status, but this had improved markedly by the twelfth century.60 Building sites used both local and migrant artisans.61 This is very clear in Gujarāt, where a new style developed in the late tenth and early twelfth centuries. The rapid distribu- tion of this Maru-Gurjara style is difficult to explain without artisan migra- tion. At the same time, local variations suggest the involvement of local workers.62 Cambodian records do not mention wages, and we must thus assume that their remarkable temple complexes were constructed with the help of forced labour.63 Chinese building workers, too, were required to provide corvée labour. A register of 129,977 provincial (lunban) artisans from 1393, divided over 62 different trades, showed half of them to be builders. Not surprisingly, given the prominence of wooden constructions in China, a great many were carpenters and sawyers, but the register also mentioned thousands of tilers, bricklayers and masons. All these people were commanded to come and work for a specified time in the capital. By implication they travelled long distances and therefore had an opportunity to meet colleagues from other regions.64

Knowledge

In 1516, Lorenz Lechler, a mason from the Neckar area in Germany, wrote down a number of ‘Instructions’ on the building trade for his son. Although he did not set this out in a particularly systematic way, one can identify four major problems facing the builder, once the decision to launch a ma-

59 S. Settar, The Hoysala temples vol. 1: text (Bangalore: Kalā Yātra Publications, 1992), 84; Sudhākara Nāth Miśhra, Gupta art and architecture (with special reference to Madhya Pradesh) (Delhi: Agamkala Prakashan, 1992), 231-36; Gerard Foekema, Architecture decorated with architecture: Later medieval temples of Karnātaka, AD 1000-1300 (New Delhi: Munshiram Manoharlal, 2003), 79. 60 Settar, Hoysala temples, 89. 61 Ahsan Jan Qaisar, Building construction in Mughal India: the evidence from painting (Delhi: Oxford University Press, 1988), 11-12. 62 Alka Patel, Building communities in Gujarāt: Architecture and society during the twelfth through fourteenth centuries (Leiden: Brill, 2004), 79. 63 Dumarçay and Royère, Cambodian architecture, 7. 64 Klaas Ruitenbeek, Carpentry and building in Late Imperial China: A study of the fif- teenth-century manual Lu Ban jing (Leiden: Brill, 1996, 2nd rev. ed.), 16-17. mega-structures in the middle ages 143 jor project had been taken.65 First, he had to make sure he obtained the right quality of stone. We have already discussed in the previous section how a vertical integration of the industry could take care of this problem. Second, a decision was required about the scale of the building and the type of measurements to be used. Third, the builder had to apply this ac- cumulated knowledge to overcome pressing problems that could arise during the building process. At several points Lechler emphasised that a builder could not rely on rules and principles alone, but had to apply judge- mental decisions. Finally and most importantly, the builder had to ensure the structural integrity of the building: ‘for an honourable work glorifies its master, if it stands up’.66 To achieve this masons used a combination of experience and practical geometry. Geometry was taught in universities as a subject, but there is no reason to assume that masons took academic degrees.67 Therefore, what they required was a practical form of geometry that did not entail compli- cated calculations, which was exactly what Lechler was providing for his son. The underlying principle was that a small number of dimensions al- lowed the builder, through a few intermediate steps, to arrive at all the dimensions of the large structure that he was commissioned to erect. In other words, one ‘macro-module’68 determined most of the other crucial variables of the church; it was the cathedral’s ‘genetic code’.69 In Lechler’s instructions, this macro-module was the width of the choir. The choir had to be twice as long as it was wide, while the nave of the church was to be twice as long as the choir. The width of the nave, Lechler suggested, was to be the same as that of the choir, and the aisles of the nave half the width of the choir and nave. The height of the building was equally determined by the basic module.70 Other structural elements were derived from the same fundamental measurement. The walls of the choir, Lechler recom- mended, should be a tenth of the width of the choir. The precise ratio, however, also depended on the quality of the stone. With good stone, three

65 Lon Shelby and Robert Mark, ‘Late Gothic structural design in the “Instructions” of Lorenz Lechler’, Architectura 9 (1979), 115; also Pamela O. Long, Openness, secrecy, author- ship (Baltimore: Johns Hopkins University Press, 2001), 211-15. 66 Shelby and Mark, ‘Structural design’. 67 Lon Shelby, ‘The geometrical knowledge of mediaeval master masons’, Speculum 47 (1972), 397. 68 Shelby and Mark, ‘Structural design’, 117; Paul Frankl, ‘The secret of the mediaeval masons’, The Art Bulletin 27 (1945), 49. 69 Scott, Gothic enterprise, 106. 70 Shelby and Mark, ‘Structural design’, 118 144 maarten prak inches could be subtracted, but when the stone was poor, three inches had to be added.71 This principle was thus one of modularity, although the term ‘propor- tional’ knowledge has also been used.72 The same underlying principle had already applied to Romanesque building.73 It has been argued that ‘modu- larity’ was necessary to organise work across the long periods of time re- quired to build a church.74 The ‘modular system’ has also been interpreted as the grid for the symbolic programme of the great churches.75 Similar principles governed the design of the smaller elements of a building.76 A booklet by Mathes Roriczer, another German builder active in the second half of the fifteenth century, described the design of pinnacles in terms similar to those used for the ground-plan of the church as a whole. By drawing squares within squares, according to fixed patterns, the mason could derive a pinnacle from a basic measurement without any under- standing whatsoever of mathematical principles.77 For the design of vaults, even where their patterning seemed very com- plicated at first sight, the same modular principles applied. The ground- plan of the nave and choir were sub-divided into a series of rectangles, of which the central point was then easy to establish. Having set out these measurements in real-life size, the builder could establish with a compass the curvature of the ‘principal arch’ (Prinzipalbogen) connecting the four corners of the rectangle and the central point. From this principal arch, the smaller arches making up the vault were then derived, again using proportional measurements.78 Perhaps predictably, there was more than one system in use at the time. This is brought into sharp focus by two debates, based around the commis- sioners of the cathedral in Milan, who were insecure about both the aes-

71 Ibid., 118-20. 72 John Harvey, Mediaeval Craftsmen (London: B.T. Batsford, 1975), ch. 5. 73 Charles M. Radding and William W. Clark, Medieval architecture, medieval learning: builders and masters in the age of Romanesque and Gothic (New Haven: Yale University Press, 1992), 37, 48. 74 Scott 2003: 141-42. 75 Aart J.J. Mekking, De dom van bisschop Adalbold II te Utrecht: De architectuur en de rol van een elfde-eeuwse bouwheer (Utrecht: Clavis, 1997), 8. 76 François Bucher, ‘Design in Gothic architecture: a preliminary assessment’, Journal of the Society of Architectural Historians 27 (1968), 51. 77 Gothic design techniques: The fifteenth-century design booklets of Mathes Roriczer and Hanns Schuttermayer, ed. by Lon Shelby (Carbondale: Southern Illinois University Press, 1977), 84-85. 78 Shelby and Mark, ‘Structural design’, 124-26. mega-structures in the middle ages 145 thetics and the construction of their expensive project. In May 1392, experts from Lombardy met Heinrich Parler, from Gmünd in Germany, to discuss the issues, but failed to agree on the best way forward. A similar debate in January 1400, between builders from Lombardy and Jean Mignot, a French architect brought in by the Duke, likewise failed to reach a compromise. Their exchanges, faithfully recorded by the cathedral administrators, point to the clash of two different sets of design principles: both modular, but with different proportions. Mignot was convinced that the Italian system was aesthetically imperfect and structurally unsound, but his opponents were unwilling to accept his opinions, and in the end had their way.79 It is also worth emphasising that these sources hardly ever discussed religious or cosmic principles behind procedures, although these were evidently important to their patrons.80 They did however, keep on referring to building experience as the guiding light. As Lechler wrote to his son: ‘Give to this writing careful attention, just as I have written it for you. How- ever, it is not written in such a way that you should follow it in all things. For [in] whatever seems to you that it can be better, then it is better, according to your own good thinking’.81 Architectural drawings as we now understand them have rarely survived prior to the fourteenth century, and when they do, they tend to show de- tails, like towers or a chapel, rather than the building as a whole.82 When the chapel for King’s College, Cambridge, was commissioned by Henry VI in 1448, its principal dimensions were simply outlined on the ground; there was no mention of any plan.83 Wooden models likewise seem to appear only in the 16th century.84 It is possible, even though precious few survive,

79 James S. Ackerman, ‘“Ars sine scientia nihil est”: Gothic theory of architecture at the cathedral of Milan’, The Art Bulletin 31 (1949), 84-111. 80 Kimpel and Suckale, Gothische Architektur, 45-54; Nigel Hiscock, The symbol at your door: Number and geometry in religious architecture of the Greek and Latin Middle Ages (Aldershot: Ashgate, 2007); Ball, Universe of stone, ch. 5. 81 Shelby and Mark, ‘Structural design, 115. 82 Arnold Pacey, Medieval architectural drawing: English craftsmen’s methods and their later persistence (c.1200-1700) (Stroud: Tempus, 2007), 59, 139; Anthony Gerbino and Stephen Johnston, Compass and rule: Architecture as mathematical practice in England (New Haven: Yale University Press, 2009), ch. 1; James S. Ackerman, Origins, imitation, conventions: ­Representation in the visual arts (Cambridge Ma.: MIT Press, 2002), 31. Bucher, ‘Design’ disagrees, but his examples refer to late Gothic. 83 Salzman, Building in England, 520. Radding and Clark, Medieval architecture, ch. 7, argue that there was a single individual who designed the whole building, but fail to address the implications of the fact that the execution of that design usually took many generations, and was thus subject to numerous personnel changes among the builders in charge of the project. 84 Meischke, Bouwtraditie, 163. 146 maarten prak that the general plan of the building was sketched on a piece of paper or parchment, but they were no more than a rough guide to the building that was eventually to emerge. The lay-out of Soissons cathedral, for example, underwent at least six major revisions during a fifty-year period.85 The basic dimensions of the ground-plan were instead set out on the building site, while the various elements of the elevation were drawn in real-life size on the floor of the building lodge.86 The main instruments used by the master mason for his design work were a ruler, square, and compass.87 Given the patchiness of the general design, much depended on the spe- cific interventions of the builders during the building process itself. As we have seen, this often was a discontinuous process, in which low levels of activity were punctuated by short, intense campaigns. Churches, which at first sight suggest unity of design, in fact displayed all kinds of ‘junctions’, marking the end of one campaign and the start of another.88 The reason for these junctures seems to have been that each group of builders used their own template.89 Templates guided the stone-cutters’ work. There must have been many identical templates to allow groups of stone-cutters to work simultaneously on similar elements of the building. Because of the huge weights involved, and given the characteristics of the lime mortar that kept the stones in place, the various elements, particularly of the vault structures, had to be executed with great precision. Small margins of error had fatal consequences for the stability of the vault’s construction.90 Rigorous quality controls of stone-cutter’s work were imposed by insisting that they mark individual pieces with their own distinctive signature.91 Templates were therefore a repository of the most significant information about the vital elements of the building.92 It was particularly the building of large churches that was undertaken in short campaigns, separated by intervals that could be lengthy. These intervals were necessary to re-stock the treasuries of the commissioning institutions, and also to allow the mortar of the construction to set and

85 James, Template-makers, 119, 198. 86 Pacey, Medieval architectural drawing, esp. ch. 2. 87 Lon Shelby, ‘Medieval mason’s tools: the level and the plumb rule’, Technology and Culture 2 (1961), 127-30 88 James, Template-makers, chs. 2 and 7 ; Marvin Trachtenberg, Building-in-time: From Giotto to Alberti and modern oblivion (New Haven: Yale University Press, 2010), ch. 5. 89 Ibid. 90 Ibid.: 86. 91 Harvey, Mediaeval craftsmen, 126; Janse and de Vries, Werk en merk, 50. 92 James, Template-makers, 34. mega-structures in the middle ages 147 create stability. During the setting, which might take more than a year, signs of pressure could appear as cracks in the newly finished work. Cracks led to adaptations on the construction, to reduce pressure on vulnerable points.93 Although this method produced satisfactory results in most building projects, the experiment at times went horribly wrong. The best-known example in Western Europe was the collapse of the vaults in Beauvais ca- thedral in 1284, and the collapse of its tower in 1573. The present building shows clear signs of reinforcement of precisely those areas that laboratory tests have shown to be most at risk.94 In Troyes, one of the towers collapsed in 1362, whilst the upper nave came crashing down in 1395.95 Both the organisation of the workforce and the evidence related to the organisation of knowledge point in the same direction: the building of churches was a ‘messy’ process.96 It was conducted as a full-scale experi- ment.97 The building site was a ‘laboratory’, where builders applied their ‘local and tacit knowledge’.98 This practice was also common in the building of Hindu temples. An eleventh century treatise required the architect (sthapati) to be knowledge- able in both the ‘science of architecture’ and ‘skilled in the work’. Yet that was not enough: intuition, ‘readiness of judgement in contingencies’, and the ability to combine theoretical and practical skills were the hallmark of the all-round builder.99 The building industry in Byzantium had characteristics reminiscent of those of the Latin West. The people in charge of design and construction were engineers rather than artists. No architectural plans have survived, and traces in extant buildings suggest that whatever drawing was needed was done full-scale and on location, with the help of practical geometry.100

93 Clark and Mark, ‘Flying buttresses’; Mark, Light, 105. 94 M. Wolfe and R. Mark , ‘The collapse of the vaults of Beauvais cathedral in 1284’, Speculum 51 (1976), 462-76 95 Stephen Murray, Building Troyes cathedral: The Late Gothic campaigns (Blooming- ton: Indiana University Press, 1987), 36. 96 James, Chartres, ch. 2. 97 Stephen Murray, Beauvais cathedral: Architecture of transcendence (Princeton N.J.: Princeton University Press, 1989), 58. 98 David Turnbull, ‘The Ad-hoc collective work of building Gothic cathedrals with templates, string, and geometry’, Science, Technology, & Human Values 18 (1993), 316-17. 99 Stella Kramrisch, The Hindu temple 2 vols (Delhi: Motilal Banarsidass, 1976), vol. I, 8. 100 Charalambos Bouras, ‘Master craftsmen, craftsmen, and building activities in ­Byzantium’, in: Angeliki E. Laiou (ed.), The economic history of Byzantium: From the seventh 148 maarten prak Churches were designed and built as ‘modular units’, and designs were changed repeatedly in the course of the construction process.101 The churches were remarkably uniform in general design, suggesting an inten- sive exchange of information throughout the empire, but at the same time they displayed local variations in detailing, which suggests that labour markets for building workers were regional rather than national.102 Pro- portional geometry was key to the designing of a church. Whereas in the Latin West the choir was the most important element of the building, and hence determined the proportions of its other features, the central ‘module’ in Byzantine churches was the dome. Its height therefore acted as the ‘con- trolling element’ of the design.103 Islamic architects were known for the intricate detail of their designs, and there was immense variation. However, the proportions of their build- ings were extremely regular, and utilised a severe logic of ‘modules’, which determined the size of all crucial elements.104 It has been suggested that both general and detailed drawings were made, as well as scale models,105 but none have survived. The first treatise on building in the Islamic world was written at the end of the thirteenth century by the prime minister of the Il-Khanid rulers in Iran, presumably an administrator and not a build- er; unfortunately no copies are left.106 Timurid builders used sophisticated geometry, again based on ‘modular’ principles.107 A series of drawings by a sixteenth-century architect from Bukhara in Uzbekistan showed various complex buildings set out on modular grids. No single element was con- sistently used as the basic module, and the most prominent element was normally picked for this purpose. In the case of religious buildings this was most obviously the dome.108 It has been suggested that one reason why through the fifteenth century (Washington, D.C.: Dumbarton Oaks Research Library and Collection, 2002), 546-47. 101 Ousterhout, Byzantium, 44, 58 (quote), and 86. 102 Ibid., 26, 56-57, 116. 103 Ibid., 72, 80-81. 104 Lewcock, ‘Architects’, 130, 132. 105 Ahmad Y al-Hassan and Donald R. Hill, Islamic technology: An illustrated history (Cambridge: Cambridge University Press, 1986), 265. 106 Lewcock, ‘Architects’, 133. 107 Golombek and Wilber, Timurid architecture, vol. I, 139; Gūlru Necipoğlu, The Topkapi scroll: geometry and ornament in Islamic architecture (Santa Monica: Getty Center for the History of Art and the Humanities, 1995), 44. 108 Golombek and Wilber, Timurid architecture, vol. I, 139-40; Yvonne Dold-Samplonius, ‘The volume of domes in Arabic mathematics’, in: M. Folkerts, J.P. Hogendijk (eds), Vestigia mathematica: Studies in medieval and early modern mathematics in honour of H.L.L. Busard (Amsterdam: Rodopi, 1993), 93-106. mega-structures in the middle ages 149 both drawings and manuals appeared relatively late in Central Asia and Europe was because the medieval architect’s practical geometry provided its own design dynamic, once the outline of the ground-plan had been established.109 In Hindu architecture the square determined both the overall plan and the details of architectural design. Temples had a square ground-plan, usu- ally sub-divided into 64 smaller squares, each with their specific religious identity.110 In Kerala, in southwestern India, a variety of basic forms were used for the ground-plan, but by far the most common was the square.111 The Maru-Gurjara style of Gujarāt consisted of a series of ‘repeatable mod- ules’.112 Building treatises were rare and described buildings, rather than providing prescriptions as to how to build them. It is possible that builders had access to these texts, but they were always applied in combination with locally developed praxis. As the volume of temples increased in the course of the fourteenth century, builders had to rely increasingly on their ‘expe- rientially gained knowledge’.113 Chinese architecture was likewise based on modular principles, which came out very clearly in a treatise completed by Li Jie around 1100, at the end of the reign of Emperor Zhe Zong. The Yingzao fashi, or State Building Standards, was an administrative document meant to regulate wooden post and beam building.114 At 1,078 pages long, and first printed in 1103, it pro- vided guidelines for the construction of a variety of public buildings and private homes. Li Jie was not a builder but a bureaucrat, as well as a paint- er, and the author of other books, on geography, history and philology. The Yingzao fashi was compiled by the government to regulate a number of its own activities (the so-called Wang Anshi Reform), including the construc- tion of public buildings.115 Li worked as Superintendent for State Buildings in the Ministry of Works, and as such had first-hand experience of the

109 Necipoğlu, Topkapi scroll, ch. 3. 110 Kramrisch, Hindu temple, vol. I, parts ii and iii. 111 H. Sarkar, An architectural survey of temples of Kerala (New Delhi: Archeological Survey of India, 1978), 63; also Michael W. Meister, ‘Analysis of temple plans: Indor’, Artibus Asiae 43 (1981-82), 302-20; Meister, ‘Geometry and measure in Indian temple plans: rectan- gular temples’, Artibus Asiae 44 (1983), 266-96; Meister, ‘De- and reconstructing the Indian temple’, Art Journal 49 (1990), 395-400. 112 Patel, Building communities, 85, 87. 113 Ibid., 163 114 Ledderose, Ten thousand things, ch. 5; also Yuhuan Zhang (ed.), History and develop- ment of Ancient Chinese architecture (Beijing: Science Press, 1986), 471; Guo, Chinese archi- tecture, 93; Watson, Arts of China, 85. 115 Guo, Chinese architecture, ch. 7. For other Chinese building manuals Ruitenbeek, Carpentry, pp.24-31. 150 maarten prak building trade. The second part of the Yingzao fashi discussed standards and regulations for design and construction, as well as guidelines for the production of bricks and tiles.116 As in Europe, the mathematics was mostly based on proportions. The Yingzao fashi identified eight building sizes, and recommended a standard size of beam for each of them. That beam size, the cai, determined the proportions for all the elements used for complex construction of roofs and their support structures, as well as other aspects of the building. Stan- dard measurements, incidentally, also allowed the readers of the Yingzao fashi, especially government officials, to calculate the number of workdays involved in the construction process, and hence the cost of projects.117 As in Europe, this combined prescriptive knowledge of the module’s mea- surement and a tacit understanding of how to apply it in practice.118 Chinese builders also used architectural drawings. The Yingzao fashi provided various types of working drawings, although no others have sur- vived from the medieval period.119 Perhaps the standardised form of build- ing made these superfluous. Interestingly, many architectural clay models have survived in China, although the majority date from the earlier Han Period (206 BCE—220 CE). Most of them are so unsophisticated that we must assume that they served decorative purposes.120 Some clearly had a function similar to European models, that is to provide a small-scale im- pression of the building for both patrons and builders.121

Apprenticeship and Other Sources of Information

Like many other pre-industrial products, buildings made huge ­demands on the skills of their producers. Cognitive psychologists have discovered that the time of training required to master complicated skills is remark-

116 Ibid., ch. 7. 117 Ibid., 93-96; Zhang (ed.), Chinese architecture, 471-72; Ledderose Ten thousand things, 137. 118 Cf. S.R. Epstein and Maarten Prak, ‘Introduction: guilds, innovation and the Euro- pean economy, 1400-1800’, in: ead. (eds), Guilds, innovation and the European economy, 1400-1800 (Cambridge: Cambridge University Press, 2008), 14. 119 Guo Chinese architecture, ch. 10. The drawings discussed in Zhang, Chinese archi- tecture, 473-74 seem to relate either to urban planning, or date from a later era. Those pictured in Ruitenbeek, Carpentry, 53 are highly stylised, suggesting variations of standard models. 120 Nicole De Bisscop, Onder dak in China: Oude architectuurmodellen uit het Henan museum, exhib. cat. Koninklijke Musea voor Kunst en Geschiedenis (Brussels: Mercator- fonds, 2007), 53. 121 Zhang Chinese architecture, 474-75; Guo, Chinese architecture, 123-24. mega-structures in the middle ages 151 ably similar across a wide variety of tasks: it takes roughly ten years to become a top-level expert in any kind of skill-based task.122 One reason why it took, and still takes, so long for adolescents and young adults to become fully trained, is that crafts typically combine propositional and tacit types of knowledge.123 Tacit knowledge cannot be articulated; ‘we can know more than we can tell’, as one scholar puts it, and it needs to be transferred from person to person.124 This happens most effectively in ‘communities of practice’.125 It is therefore important to not only pay at- tention to the contents of skills education, but also to its social organisa- tion. Before we look at apprenticeship however, we have to discuss briefly two other potential sources of information on building: the clergy in their capacity as patrons of buildings, and written documentation. For a long time it has been thought that the original impulse for the Gothic style in Europe, its innovative design, came from the church itself. There are two good reasons for this assumption. The clergy were the learned class in medieval society, and surely the beauty of the Gothic style must have been inspired by faith and knowledge, two forms of human capital that were uniquely concentrated in the hands of clergymen. It is an assumption confirmed in one of the great architectural documents of the period, the writings of Suger, abbot of Saint-Denis at a time when the mon- astery obtained a new choir, often identified as the start of the Gothic style. However, it is not evident what Suger´s precise role was in the design pro- cess, and whether he dictated the general direction at all. There is no evi- dence that he had any specific ideas on how to execute plans, something generally true of the contribution of the clergy.126 Written documentation about building and related crafts is rare before the sixteenth century. The famous Villard de Honnecourt, author, or rather

122 K.A. Ericsson, ‘The acquisition of expert performance: an introduction to some of the issues’, in: Ericsson (ed.), The road to excellence: The acquisition of expert performance in the arts and sciences, sports and games (Mahwah NJ: Lawrence Erlbaum Associates, 1996), 10-11. 123 S.R. Epstein, ‘Property rights to technical knowledge in premodern Europe, 1300- 1800’, American Economic Review 94 (2004), 383. 124 M. Polanyi, The tacit dimension (Gloucester: Peter Smith, 1966), 4. 125 A.C. Cianciolo, C. Matthew, R.J. Sternberg, and R.K. Wagner, ‘Tacit knowledge, practical intelligence, and expertise’, in: K.A. Ericsson, N. Charness, P.J. Feltovich, R.R. ­Hoffman (eds), Cambridge handbook of expertise and expert performance (Cambridge: Cambridge University Press, 2006), 623-24. 126 Radding and Clark, Medieval architecture, proposes a more elegant solution, point- ing to parallel changes in philosophy and building style in the Paris basin in the first half of the twelfth century as an expression of similar intellectual developments, without claim- ing a direct impact from one upon the other. 152 maarten prak compiler, of a famous manuscript on French Gothic cathedrals dating from around 1230, was possibly not a builder himself. His drawings seem to have been made either from the building as it was constructed, or from plans that he observed on site.127 It was only in the later decades of the fifteenth century that the first building ‘manuals’ started to appear. Several small books from Germany survive, with detailed instructions on the design and execution of major building works; we have already discussed their con- tents in the previous section.128 During the sixteenth century, similar guides for carpenters were published, containing the same practical types of calculations found in stone masons’ manuals. This was mathematics without the theory, because carpenters could not be assumed to know how to multiply.129 If neither learned institutions like the church, nor written documenta- tion, were much of a help to builders, we must look for alternative sources of information that they could tap into. Apprenticeship was (and still is) found throughout the world as the main institutional context for the ac- quisition of craft skills.130 Apprenticeship was however, embedded in other institutions, notably the family, the guild and the state. In Europe, there were several dynasties of architects, the best-known of whom were perhaps the Parler, who were active in Central Europe in the fourteenth and fifteenth centuries. Their last name was derived from the word for building lodge foreman: parlier. Their activities were first recorded in Cologne and Gmünd in Central and Southern Germany, where Heinrich Parler was active around the middle of the fourteenth century. The portrait of his son Peter Parler, the most famous of the family, was incorporated into the structure of Prague cathedral, where he was in charge of construc- tion work in the second half of the fourteenth century. Two of his brothers were working as stone masons in Prague and Freiburg, and in Basel. Peter’s sons Wenzel and Johann succeeded him as Master of the cathedral works in Prague, and Wenzel also worked in Vienna. Their cousin Michael was

127 Robert Branner, ‘Villard de Honnecourt, Reims and the origin of Gothic architectural drawing’, Gazette des Beaux-Arts 61 (1963), 137-38; Kimpel and Suckale, Gothische Architek- tur, 44; Wilson, Gothic cathedral, 141. 128 Shelby (ed.), Gothic Design Techniques; Shelby and Mark, ‘Lorenz Lechler’. 129 David T. Yeomans, ‘Early carpenters’ manuals 1592-1820’, Construction History 2, 1986, 14. 130 Bert De Munck, Steven L. Kaplan, Hugo Soly (eds), Learning on the shop floor: His- torical perspectives on apprenticeship (New York: Berghahn, 2007); Michael W. Coy (ed.), Apprenticeship: from theory to method and back again (Albany: State University of New York Press, 1989); Trevor H.J. Marchand, The masons of Djenné (Bloomington: Indiana University Press, 2009), ch. 5. mega-structures in the middle ages 153 Master in Strasbourg, while Johann’s son, another Johann, worked as stone mason in Prague.131 Four members of the Roriczer family in three succes- sive generations held the office of cathedral architect at the building lodge in Regensburg.132 The family was by no means the only environment in which building knowledge was transferred in Europe, and data from a later period point in exactly the opposite direction. Of 80 apprentice bricklayers in York be- tween 1654 and 1752 whose fathers’ occupation was known, only 21 were also bricklayers (and 11 of them apprenticed their son). In other words, three-quarters came from outside the trade. If we take into account the likelihood that there were very few bricklayers’ sons among the 200-plus apprentice bricklayers during that same period whose fathers’ occupations were unknown, it could be that less than ten percent were actually continu- ing their fathers’ trade.133 Possibly the great majority of builders learned the tricks of their trade outside the family environment. Three settings have already been suggested in the course of the previous discussion: the stone quarry, the building lodge, and the local guild. The actual work in the quarries has thus far not been properly investi- gated, and we can therefore say next to nothing about its contribution to the training of the skilled workforce. However, the names of stonemasons in the Northern Low Countries very often referred to places of origin in the areas where the stone was found, strongly suggesting a close connection between quarrying and the acquisition of skills necessary to work the stone.134 The building lodge was both a concrete location and an organisational form. Like families and guilds, the building lodge was a hierarchical and regulated institution, headed by one or more directores fabricae.135 The lodge was financed by its patrons, but some had more or less indepen- dent sources of income, as in Strasbourg, where lodge property secured a steady and substantial income stream.136 Under the Master and his

131 Barbara Schock-Werner, ‘Die Parler’, in: Legner (ed.), Die Parler, 7-11 132 Shelby (ed.), Gothic design techniques, 7-28. 133 Woodward Men at work: labourers and building craftsmen in the towns of northern England, 1450-1750 (Cambridge: Cambridge University Press, 1995), 54-55. This was true in many other trades as well: Epstein and Prak, ‘Introduction’, 9-10. 134 Janse and De Vries, Werk en merk, 25. 135 Schock-Werner, ‘Bauhütten’. 136 On the financing of cathedral building: Henry Kraus, Gold was the mortar: The economics of cathedral building (London: Routledge & Kegan Paul, 1979); Wim Vroom, Financing cathedral building in the Middle Ages: The generosity of the faithful (Amsterdam: Amsterdam University Press, 2010). 154 maarten prak substitute, the parlier, were the journeymen, as well as the apprentices.137 The masons’ guilds had an ambiguous role in the history of European church building, as the earliest craft guilds were established halfway through the twelfth century.138 This was well into the Gothic building boom. Moreover, guilds in the building industry were not among the early foundations. The masons’ guild established in Lincoln in 1313 was a religious confraternity, not a craft guild. In London, the first sign of an organisation of masons dates from 1376, although regulations for the trade had already been introduced a little earlier, in 1356.139 Clearly guilds were not a prereq- uisite for the construction of Gothic churches. Still, they became more important as time went on, and building crafts in Italian towns obtained statutes in the course of the fourteenth century.140 Guild statutes regulated the training of aspiring workers in the trade, even though the training itself was left to individual masters. The minimum training period varied significantly, even in relatively small areas. In Verona the minimum was just one year, in Piacenza four, in Bologna five, in Genoa and Savona six, and in Venice six to seven. The statutes in Padua distinguished between stonecutters, who were required to learn for six years, and the wallers, whose apprenticeship took up to eight years.141 In Germany it was the other way around: wallers had to be apprenticed for four years, stonemasons for six.142 Clearly these regulations were referring to very different levels of accomplishment, suggesting that five to six years were required to become fully competent. It was very likely that addition- al experience was necessary before admission as an independent master.143 The regulations for Parisian masons from the mid-thirteenth century stip- ulated the following:144

137 Volker Segers, ‘Studien zur Geschichte der Deutschen Steinmetzbruderschaft: Mit besondere Berücksichtigung der für das Strassburger Gebiet geltenden Ordnungen und Bestätigungsurkunden (15. bis 17. Jahrhundert)’, unpubl. PhD-dissertation, Freie Universität Berlin, 1980. 138 Steven A. Epstein, Wage labor and guilds in medieval Europe (Chapel Hill: University of North Carolina Press, 1991), ch. 2. 139 Douglas Knoop and G.P. Jones, The mediaeval mason: an economic history of English stone building in the later Middle Ages and early modern times (Manchester: Manchester University Press, 1967), 135-36. 140 Richard A. Goldthwaite, The building of Renaissance Florence: An economic and social history (Baltimore: The Johns Hopkins University Press, 1980), 431-34. 141 Ibid., 260. 142 Segers, ‘Studien’, 179-80. 143 Epstein and Prak, ‘Introduction’, 8. 144 Quoted from Binding, Baubetrieb, 104-05. mega-structures in the middle ages 155 In Paris everyone can be a mason, provided he knows the trade (le mestier) and works according to the customs and practices of the trade, which are as follows: Only those who have been apprenticed can work in a workshop, and after his apprenticeship he can exercise the trade independently only after six years of experience. Rules drafted by the meeting of master stonemasons from the in Regensburg in 1459 included a special section on the dienner, or apprentices.145 The most important aim of these regulations seems to have been the continuous supply of qualified labour. Thus there existed in Europe a range of institutions offering opportuni- ties for the acquisition of skills. The available evidence suggests that these institutions existed side by side, working out their relations in local settings according to contextual peculiarities.146 Neither the family, nor the quarry or the building lodge, and not even the guild, provided a fully-fledged train- ing programme. What these institutions did provide was an environment with educational potential. Among these, the guilds were most focused towards the training of the workforce. As the guilds seem to have become more significant in the regulation of the building trade over time, training became a more important concern for the industry as a whole. Because of the relative unimportance of stone in Byzantine building, quarries were insignificant as locations for the training of its workforce. The two other environments we identified in the Latin West were, how- ever, equally important for the Byzantine building industry: the workshop and the guild.147 Unfortunately very little is known about the early history of Byzantine guilds, although we do know that they emerged earlier than in the Latin West, and that they continued to function throughout the Middle and Late Byzantine periods. The guilds enforced quality controls by holding the master accountable (for brick buildings the warranty ex- tended to a ten year period) and prohibiting him from taking on a new project before he had finished the current one. They also insisted on prop- er skills: ‘Those who build walls and domes or vaults must possess great exactitude and experience lest the foundation prove unsound and the building crooked or uneven’, reads the Book of the Eparch, a source from the tenth century listing some craft regulations.148 Whereas guilds were permanent institutions, workshops were temporary arrangements,

145 Segers, ‘Studien’, 179-80. 146 Schock-Werner, ‘Bauhütte und Zunft’, in: Legner (ed.), Die Parler, 64-65. 147 Ousterhout, Byzantium, 49-57. 148 Ibid., 50. 156 maarten prak attached to the building project itself. They were headed by a master build- er or master mason, in charge of the work force. Apprentices must have been trained on the job, but the sources are silent on this aspect. In the Timurid Empire, craft skills seem to have been acquired princi- pally through the family system.149 In some parts of the Islamic world, notably Iran, Iraq, Egypt and Turkey, craft guilds emerged, albeit somewhat later than in the Latin West. Guilds were closely tied to religious functions, as in Europe.150 These organisations, called futuwwa in places, had a patron saint, organised examinations, and provided the title of ‘master’.151 But the comparison only takes us so far. A French traveller from the seventeenth century was struck in Isfahan by how different local craft organisations looked from what he knew about the corporations back home. They united men of the same profession, and there was a chief officer, but there were no rules of membership, apprenticeship schemes, or quality controls im- posed on the members.152 There are hints, from the South-Indian Hoysala Kingdom (eleventh to fourteenth centuries) of builders’ ‘assemblies’, which suggest they were organised, either as professionals or in castes, two types of organisation that overlapped. Like European guilds, these organisations also had a reli- gious component, and one of them was described as ‘the servants of God- dess Saravastī’. However these Indian crafts were not as regulated as guilds.153 Chinese builders were definitely organised in family guilds, which were registered, regulated and supervised by the government.154 However, craft guilds somewhat similar to the European guild system only appeared in the seventeenth century.155 This was a time when compulsory labour for the government declined.156 Until then, the craft, and therefore its skills,

149 Golombek and Wilber, Timurid architecture, vol. I, 67. 150 Lewcock, ‘Architects’, 133. 151 al-Hassan and Hill, Islamic technology, 267-69. 152 Ronald W. Ferrier (transl. and ed.), A journey to Persia: Jean Chardin’s portrait of a seventeenth-century empire (London: I.B. Taurus, 1996), 167; also Lewcock, ‘Architects’, 133. 153 Settar, Hoysale temples, 86-88. 154 Guo, Chinese architecture, 90; Dagmar Schäfer, The crafting of the 10,000 things: Knowledge and technology in seventeenth-century China (Chicago: University of Chicago Press, 2011), 94-108; also the contributions by Pomeranz and Moll-Murata in this volume. 155 Peter J. Golas, ‘Early Ch’ing guilds’, in: G. William Skinner (ed.), The city in Late Imperial China (Stanford: Stanford University Press, 1977), 555-80. 156 Christine Moll-Murata, ‘Chinese guilds from the seventeenth to the twentieth centuries: an overview’, in: Jan Lucassen, Tine de Moor, Jan Luiten van Zanden (eds), The return of the guilds, International Review of Social History Supplement (Cambridge: Cam- bridge University Press, 2008), 219. mega-structures in the middle ages 157 had to be transferred to at least one member of the next generation in each craft family.157 The Yingzao fashi and similar manuals also testify to the importance of the Chinese state in the transfer of knowledge; under the Qing, this role was taken over by the guilds. Another Chinese peculiarity was that building knowledge was transferred orally in verse form, presum- ably because that made it easier to memorise.158 As in Europe this knowl- edge was embedded in metaphysical beliefs and rituals.159

Conclusions

In his 1998 Mellon Lectures in the Fine Arts, the German sinologist, Lothar Ledderose, argued that Chinese art, from calligraphy to images to architec- ture, was essentially modular. Aside from a general point about Chinese culture, Ledderose saw this as an explanation for the emergence of artistic mass production in China, which nonetheless managed to maintain dis- tinctive characteristics in individual pieces, simply by recombining existing elements.160 Likewise in this article it has been argued that the builders of some of the greatest works of medieval architecture throughout Eurasia used modules to design their projects.161 Where we differ from Ledderose is in the intention to explore how pre-modern builders were able to over- come the challenges of projects whose mechanics they were unable to fully understand. Three points stand out. Firstly, builders’ knowledge was practical, and related to their working experience, rather than theoretical. This is probably an unsurprising con- clusion, but it is a point worth making nonetheless,162 because the other two conclusions follow on from it. It is also important because it under- scores a point made by S.R. Epstein, namely that it was not technological constraints as such that hampered the pre-modern economy, but the

157 Guo Chinese architecture, 90. 158 Ibid. 159 Ruitenbeek, Carpentry, 82-89. 160 Ledderose, Ten thousand things. 161 More specific parallels between the Latin West and Central Asia are highlighted in Gūlru Necipoğlu, ‘Geometric design in Timurid/Turkmen architectural practice: thoughts on a recently discovered scroll and its late Gothic parallels’, in: Lisa Golombek, Maria Sub- telny (eds), Timurid art and culture: Iran and Central Asia in the fifteenth century (Leiden: Brill, 1992), 48-66. 162 See also Tim Ingold, The perception of the environment: Essays on livelihood, dwelling and skill (London: Routledge, 2000), pt. iii; Pamela H. Smith, The body of the artisan: Art and experience in the Scientific Revolution (Chicago: University of Chicago Press, 2004), ch. 3; Richard Sennett, The craftsman (New Haven: Yale University Press, 2008). 158 maarten prak under-utilisation of available technologies.163 The knowledge that we de- scribed in this paper was already available to the Romans. It was insufficient to build a steam engine, but the accomplishments of the constructors of medieval religious buildings across Eurasia demonstrate that, despite the limitations of their practical knowledge, they were capable of remarkable progress in the construction of other quite complex ‘machines’.164 The second point to emerge from our investigation of the medieval building industry is that modular knowledge could indeed substitute for theoretical knowledge. It is not so difficult to see why, across all the major civilisations of Eurasia, builders used modules to create large structures. On the basis of practical experience, builders had worked out how the various modules of a complex structure like a church building, or the tall towers attached to these buildings, could be reduced to a coherent set of proportional dimensions. These dimensions were often embedded in com- plex religious, cosmological and scientific theories.165 However, it is argued in this article that an important function, perhaps even the primary func- tion, of modular design principles was their easy transmission between builders.166 Modular dimensions guided the general patterns of builders’ work. At the same time, the details of that work had to be adapted to local circumstances, such as the quality of the surface on which the building was to be constructed, the type of building material available for the construc- tion, and the financial constraints of the project. Therefore the application of this knowledge was always embedded in the practice of the building process itself. The construction of a large church building was an ‘experi- ment’, and the building site was a ‘laboratory’.167 The propositional knowl- edge used in this laboratory was relatively simple, while the prescriptive knowledge was embodied in the builders’ tacit skills. Work in this laboratory, and this is our third conclusion, was almost by definition, collective, and so was the process of acquiring the knowledge of how to build properly.168 Building a complex structure like a church was

163 S.R. Epstein, Freedom and growth: The rise of states and markets in Europe, 1300-1750 (London: Routledge, 2000), 7. 164 Gimpel, Medieval machine. 165 Hiscock, Symbol; Richard Padovan, Proportion: Science, philosophy, architecture (London: Routledge, 1999). 166 Rowland J. Maidstone, ‘Structural theory and design before 1742’, Architectural Review 143 (1968), 304. 167 Turnbull, ‘Collective work’. 168 Howard S. Becker, Art worlds (Berkeley: University of California Press, 1984); Pamela H. Smith, ‘In a sixteenth-century goldsmith’s workshop’, in: Lissa Roberts, Simon Schaffer, mega-structures in the middle ages 159 therefore a social, as much as a technological challenge. We have seen how the medieval building industry utilised a number of different social institu- tions to accomplish its remarkable feats: the family, the building lodge, and the guild. These three institutions were found throughout Eurasia as the appropriate environments for construction work. Religion (everywhere) and the state (in China) also provided important contexts for the consoli- dation of building technology. The specific mixtures of these institutions are, however, still something of a mystery, as is the possible implication of varying mixtures for the development of the industry. Each area of Eurasia reached its peak level of building achievements at a different point in time, but those peak levels seem to have been functions of general socio-eco- nomic and cultural trends, rather than autonomous developments. Fundamentally new directions were only explored in European archi- tecture after our period, with the application of mathematics to construc- tion challenges in the eighteenth and nineteenth centuries.169 Although it is tempting to argue that this development was an expansion of earlier connections between geometry and building,170 it is worth remembering that such connections existed throughout Eurasia during the Middle Ages, and that everywhere academic geometry during the Middle Ages served aesthetic and religious, rather than constructional purposes. St. Paul’s ca- thedral in London, designed in the late 1660s by mathematician Christo- pher Wren, professor of astronomy at Oxford, was still built as an experiment, complete with collapsing vaults and major design changes during the building process.171

Peter Dear (eds), The mindful hand: Inquiry and invention from the late Renaissance to early Industrialisation (Amsterdam: KNAW, 2007), 35, 39. 169 Gerbino and Johnston, Compass, 98; Addis, Building, chs. 5-6; A. Kahlow, ‘Between mechanics and architecture: The foundation of the Berlin “Bauakademie” in 1799’, in: ­Becchi et al. (eds), Towards a history of construction, 385-92. 170 Pamela O. Long, ‘The contribution of architectural writers to a “scientific” outlook in the fifteenth and sixteenth centuries’, Journal of Medieval and Renaissance Studies 15 (1985), 265-69. 171 James W. Campbell, Building St Paul’s (London: Thames & Hudson, 2007), ch. 12; Gerbino and Johnston, Compass, 100, 164. 160 maarten prak the technology and teaching of shipbuilding 161

CHAPTER FIVE

THE TECHNOLOGY AND TEACHING OF SHIPBUILDING 1300-1800

Richard W. Unger

Introduction

Economic historians are students of how people in the past solved or did not solve the problem of scarcity. It is the roots of economic success, and by implication the roots of economic failure, which in the end worry all economic historians. That, by definition, applied to Larry Epstein. Two large issues drew his attention, especially over recent years. He was concerned about the incongruities of facts and theoretical predictions about the post Black Death European economy. The standard view seemed less satisfac- tory in explaining the recovery or lack of it from the demographic disaster than did the actions of some governments in regulating and taxing trade. His earlier work on Italian towns suggested that part of the answer lay with the lack of market integration and that sent him to explore the role of trade in what economic success there was in early modern Europe.1 A second issue was the place of guilds in the European economy over the long term, that is from the High Middle Ages when they first emerged down to the Industrial Revolution, and the attacks on them by Enlightenment thinkers, attacks which led to their abolition. It seemed to him strange that institu- tions which many eighteenth century writers and policy makers thought of as drags on the economy should have survived and thrived for so long.2 The concerns about market integration and about the effects of guilds on the economy converged in his growing interest in shipping and ship- building. The efficiency and effectiveness of waterborne transportation in early modern Europe was one source of expanding trade and so of recovery from the Black Death. Shipbuilding, the source of capital goods for ship-

1 S.R. Epstein, ‘Town and country: economy and institutions in late medieval Italy’, Economic History Review 46 (1993), 453-77. 2 S.R. Epstein, ‘Craft Guilds, Apprenticeship and Technological Change in Pre-indus- trial Europe’, in: S.R. Epstein and Maarten Prak (eds), Guilds, innovation, and the European economy, 1400-1800 (Cambridge: Cambridge University Press, 2008), 52-80. [originally in Journal of Economic History 58 (1998), 684-713]. 162 richard w. unger ping, depended on skill and knowledge about how to build ships and the effective transfer of that skill and knowledge. Guilds were, as argued for other industries, a potential vehicle for training and innovation in ship- building. There was considerable advance in ship design and construction both in Europe and Asia in the ten centuries before the Industrial Revolu- tion. The pattern of those developments indicates not only the potential influence on commerce but also the ways in which new knowledge was developed and transferred among shipbuilders both near home and at some distance. Practitioners passed on information in person but in the fifteenth and sixteenth centuries written works on shipbuilding offered another vehicle. It would not be until the late eighteenth century that trea- tises on the topic proved effective in disseminating know-how. Even the presence of knowledge, however exchanged, was not enough to guarantee advances in practice on the shipbuilding wharf in Europe or Asia or the New World. There were strong disincentives to innovate created by institu- tions, including guilds, as well as by market structures. There were also incentives to innovate created by profit as well as by training and the insti- tutions responsible for that training in some parts of the world: guilds. Shipbuilding was an enterprise, albeit on varying scales and for varying purposes, which was common almost everywhere and so the study of ves- sel construction, what was built and how, is eminently suited to global comparisons. Constraints on delving into relationships in shipbuilding technology are many and inflexible. Ignorance—that of specific econom- ic historians and of historians in general—is an almost insurmountable barrier. The potential rewards from the comparison in indicating the com- parative role of shipbuilding in market development around the world, however, make it worth the risk of missing the mark. One central question in examining that role is the place of guilds, or rather of institutions in general, in generating technical change. Better ways of doing things and better products have been a major source of econom- ic growth in the nineteenth century and even more so in the twentieth and perhaps yet more in the twenty-first. It is no longer possible to sustain the presumption that technical advance, a major contributor to the Industrial Revolution, only emerged in the eighteenth century and only in Western Europe.3 A mass of work by historians of medieval and early modern Eu- rope as well as of Asia has demonstrated that instead of being technically

3 S.R. Epstein, ‘Property Rights to Technical Knowledge in Premodern Europe, 1300- 1800’, American Economic Review 94 (2004), 382. the technology and teaching of shipbuilding 163 dormant, the period up to industrialization was marked by ongoing and continued introduction of new methods of doing and in many sectors across economies. If the pre-modern world was alive with technical ad- vance then an obvious question is what were the roots, the sources, of that advance. The creation of knowledge, the nurturing of that knowledge, its adaptation to practical use, the implementation of new methods to help in solving the problem of scarcity, and the transmission of knowledge about new products and new ways of doing things were all attributes, more or less, of pre modern societies. Explanations for the presence of those phe- nomena should exist. Epstein thought that one reason for the observed technical advance was institutional, and more specifically guilds. They were for him sources of education and innovation. The primary purpose of guilds was training skilled workers and invention and innovation were unintended but wel- come by-products of the purpose.4 It was the training function which explained why they existed for so long and why they were so slow to be dismantled, even if eighteenth century thinkers declared them bad. The Scottish professor of moral philosophy, Adam Smith, was by no means alone in his dislike of guilds. His fears of the actions of businessmen, their tendency to collude to gain collective advantage, were of course well found- ed. Not only was there that danger with having organizations of producers, but also guilds were by definition restrictive since they enjoyed monopolies over production, sale and distribution and so prevented individuals from pursuing their natural tendencies to truck, barter and exchange.5 Guilds were detrimental because they prevented economic growth and they pre- vented technical change or at least the adoption of new techniques. ­Epstein, with Maarten Prak, argued the contrary, saying that craft guilds, ‘created an environment that was conducive to the type of tacit, embodied, and incremental innovation typical of most industrial development before the Industrial Revolution, and much of it even after that momentous event’.6 Because shipbuilding was a source of technical advance, because ship- building supplied tools for the transformation of the relative place of Eu- rope in the world and because there were guilds of shipbuilders in Europe,

4 Epstein, ‘Craft guilds’, 52, 56. 5 Adam Smith, An inquiry into the nature and causes of the wealth of nations, ed. by Edwin Cannan, and Max Lerner (New York: The Modern Library, 1937), 118-24. 6 S.R. Epstein and Maarten Prak, ‘Introduction: Guilds, Innovation, and the European Economy, 1400-1800’, in: Epstein and Prak (eds), Guilds, innovation, and the European economy, 23. 164 richard w. unger the industry offers fertile ground for examining the influence of guilds on the economy through the agency of generating new methods and transmit- ting them. The types of knowledge created and passed on, the actions of guilds to promote or limit the development and transmission of knowledge, and the ways in which other public institutions affected shipbuilding di- rectly or indirectly through guilds, are essential elements in assessing the contribution of guilds to shipbuilding and so to shipping and so, finally, to solving the problem of scarcity.

Developments in Ship Design

Whatever the role of institutions, there is no doubt that technical advance was a feature of European and, perhaps to a lesser degree, Asian shipbuild- ing between 1300 and 1800. The improvements in technology were impres- sive because of the effects on politics as well as the economy. That impact was more obvious for Europe, though that may be in part a product of the concentration on the study of the European past in the period. The techni- cal change was impressive also because of the variety of technologies ship- builders employed, the uniqueness and novelty of designs and methods, and because of the integration of their efforts with other technologies. The changes and developments are elusive in part because of the nature of surviving information which, though varied in type, still leaves out much of the process of technical advance, and in part because technological change took different paths in different places. There was a rich pre-history to the developments in shipbuilding be- tween 1300 and 1800. That varied background set the stage for two periods of advance, loosely the Renaissance and early modern transformation dat- ing from the fourteenth through the middle of the seventeenth century, and the years from around 1650 through to about 1820, the later period being marked by stability and refinement of earlier developments. The history of shipbuilding took a sharp turn after 1820 as steam power started to have an effect on shipping. The periodisation is based principally on the pattern in Europe because of the faster pace of change there and, equally important, because the tendency over time was for European techniques to spread through much of the rest of the world. Major advances in ship design came first in China but they were soon followed by essential chang- es in Europe which translated over the long run into superior types of craft, the ones that were responsible for the stability in the almost two centuries before the effects of steam came to be felt in shipping. the technology and teaching of shipbuilding 165 The dating of the first construction of junks, the highly versatile sea- going vessels of China, is difficult to establish. By the eighth and ninth centuries varied influences of domestic construction of river craft came together with knowledge of high seas ships from Southeast Asia to create a unique and highly effective vessel.7 That type made it possible for Chinese shipping to dominate the East and South China Seas and to some extent the Indian Ocean by the fourteenth century. The number and size of Chi- nese ships at the time left a deep impression on foreign visitors such as Marco Polo and Ibn Battuta.8 Writers even in the seventeenth century still remarked on the number of ships in China.9 In the period when the Roman Empire flourished Chinese ships were already numerous and maintained trading relations with Southeast Asia and even Africa.10 By 1300 the junk had evolved into a cargo carrier of great flexibility. It could be built in vary- ing sizes and, in larger versions, could carry significant payloads by river or across the open sea. Archeological investigations of wrecks indicate a sophisticated and varied building tradition.11 The rig was made up of bal- anced lug-sails, with battens, on anywhere from one up to seven or even more masts on the largest of junks. The masts were staggered to the port and starboard of the centre line of the ship.12 There was no keel and no pointed bow so the junk looked very different from European types. There were bulkheads spaced through the hull giving internal strength and creat- ing watertight compartments so that even if one flooded the integrity of the vessel was not compromised. There was an axial rudder which added to the already considerable manoeuvrability thanks to the configuration of the rig. Junks were also durable, apparently having typically long life spans. Merchants travelled on board, renting a section or more of the junk and then the vessel worked its way along the coast, stopping at various ports for the merchants to trade. That tramping was the standard

7 Pierre-Yves Manguin, ‘The Southeast Asian Ship: An Historical Approach’, Journal of Southeast Asian Studies 11 (1980), 266-76 8 L. Audemard, Les Jonques Chinoises. (Rotterdam: Museum voor Land- en volken- kunde and Maritime Museum Prins Hendrik, 1957-1963), 25-30; Sean McGrail, Boats of the world: From the Stone Age to Medieval times (Oxford: Oxford University Press, 2001), 377-78; Joseph Needham, with the collaboration of Wang Ling and Lu Gwei-Djen, Science and civilisation in China. Vol. 4, III: Civil Engineering and Nautics (Cambridge: Cambridge Uni- versity Press, 1971), 465-70. 9 Needham, Civil engineering, 423. 10 Needham, Civil engineering, 440-46. 11 William Wayne Farris, ‘Shipbuilding and Nautical Technology in Japanese Maritime History’, The Mariner’s Mirror 95 (2009), 265; McGrail, Boats of the world, 360-76. 12 Farris, ‘Shipbuilding’, 265. 166 richard w. unger

Illustration 5.1. The sea-going junk in the Liu-Chhiu Kuo Chih Lüeh of about 1757. Joseph Needham, Wang Ling and Lu Gwei-Djen. Science and civilisation in China. Volume 4, Part III, Civili Engineering and Nautics. Cambridge: Cambridge University Press, 1971, p. 400. pattern and one which decreased dangers for junks travelling the high seas. In the late Middle Ages the products of Chinese shipyards made regular trading voyages to Java and other parts of Southeast Asia. The much dis- cussed expedition of Zheng He, reaching as far as East Africa in large ships built at yards on the Yangtze near Nanking, demonstrated the potential of Chinese shipbuilding in the first two decades of the fifteenth century.13 The largest of those ships may have been about 70 metres long and reached 2,000 tons in capacity.14 That was a case of state-sponsored and managed shipyards producing ships that probably pressed technical limits. The pro- ducers of those ships were not the first yards created by imperial authority. The emperor Liu Yen (917-942 AD) had already set up shipbuilding work- places in the tenth century.15 They were not the last. The Dragon River Shipyard at Nanjing at about 500,000 square metres was the largest shipyard

13 Needham, Civil engineering, 479. 14 McGrail, Boats of the world, 379-81. 15 Needham, Civil engineering, 460. the technology and teaching of shipbuilding 167

Illustration 5.2. A Malaysian jong and other Malaysian craft from Willem Lodewycksz., D'eerste boeck, Historie van Indien, waer inne verhaelt is de avontueren die de Hollandtsche schepen bejeghent zijn, Amsterdam, 1598. Reprinted in Eerste schipvaart der Nederlanders naar Oost-Indië onder Cornelis de Houtman, 1595-1597; journalen, documenten en andere bescheiden, ed. G.P. Rouffaer en J.W. IJzerman. ’s-Gravenhage: M. Nijhoff, 1915, page 132, plate 28. in the world in the late seventeenth century.16 Though Chinese supreme rulers did, on occasion, create shipyards it appears that the overwhelming majority of building always took place in private yards not only beyond the reach of decisions of the emperors but with limited regulation from any source, either governmental or generated by the builders themselves. Large vessels from China visited Japan as well but the influence they had on shipbuilding there before the fourteenth century was limited. Typ- ically high seas trade was in the hands of foreigners and Japanese builders concentrated their efforts on small coastal vessels, many of them powered by oars. Even with increased commercial contact it appears that the ships that came from Japanese yards remained distinctive with little transfer of Chinese techniques. Planks were fitted edge-to-edge but otherwise the vessels were different from Chinese, Indian Ocean or European ships. There were no bulkheads to create separate compartments in the hold, no keels,

16 Christine Moll-Murata, State and Crafts in the Qing Dynasty (1644-1911). Habilitations- schrift, Eberhard Karls Universität Tübingen, April, 2008, 182-83. 168 richard w. unger and no stem- or sternposts. The size of Japanese ships grew but within the established tradition with more masts and larger sails. That was true of all types including warships. Toyotomi Hideyoshi (1537-98), the first daimyō to rule all of Japan, even went so far as to establish standard specifications for warships throughout the country in the late sixteenth century, some- thing not tried in Europe for another hundred years. Japanese ships of unique design visited ports in East and Southeast Asia through the seven- teenth and eighteenth centuries but with apparently little effect on ship- building methods in the islands.17 Ships produced around the Indian Ocean in the Middle Ages were typ- ically lighter than the Chinese junks whose hulls were made of sawn planks nailed together. Hulls of ships from Malaysian building traditions were made up of planks either sewn together with vegetal stiches or with small bumps on the interior of the planks through which lashings were strung. The shell went up first and any internal skeleton was added later. Even on larger ships there could be outriggers for increased stability but for sea- going traders that was not the case. Vessels typically carried sails of some variety, starting with bi-pod or tri-pod masts. Elongated canted square sails became the preferred option by the early Middle Ages and the dominant one by 1300. The vessels could be sizeable with an eighth century example reported to have been about 60 metres long and capable of carrying 1,000 passengers.18 In the sixteenth century Europeans tended to call the trading vessels of the Bay of Bengal and the Indonesian Archipelago junks, but those ships were different from the Chinese ones for which travellers used the same name. The jong not only was apparently built without iron nails but also lacked an axial rudder, instead carrying two quarter rudders, one on each side. The jong was a product of a radical evolution in Malaysian ship design which began in the thirteenth and fourteenth centuries. Planks in the new form were held together by dowels fitted in holes in upper and lower planks, the dowels then fixed in place by wooden nails. The same

17 Farris, ‘Shipbuilding’, 269-78. 18 Pierre-Yves Manguin, ‘Trading ships of the South China Sea: Shipbuilding techniques and their role in the history of the development of Asian trade networks’, Journal of the Economic and Social History of the Orient 36 (1993) 254-64; Manguin, ‘The Southeast Asian Ship’, 269, 273; Pierre-Yves Manguin, ‘Southeast Asian Shipping in the Indian Ocean during the first millennium A.D.’, in: Himanshu Prabha Ray, Jean-François Salles (eds) Tradition and archaeology: Early maritime contacts in the Indian Ocean. Proceedings of the international seminar, techno-archaeological perspectives of seafaring in the Indian Ocean, 4th cent. B.C.- 15th cent. A.D., New Delhi, February 28-March 4, 1994 (New Delhi: Manohar Publishers and Distributors, 1996), 183-86; Needham, Civil engineering, 457-58. the technology and teaching of shipbuilding 169

Illustration 5.3. A sketch of a cog from the Venetian book on shipbuilding of the first half of the fifteenth century by Michael of Rhodes. The Book of Michael of Rhodes: A Fifteenth Century Maritime Manuscript, Volume 1, Pamela Long, David McGee and Alan Stahl (eds), Cambridge, MA: MIT Press, 2009, fol. 182b. I am grateful to Alan Stahl for supplying me with the image. type of fastening was used for the frames and bulkheads. The hulls, which had more of a V shape than the flat-bottomed Chinese junk, had multiple sheathing with as many as four layers of planking. There were multiple masts, as with the Chinese junk, to carry sails made of fibre matting but there was also a sail rigged on the forestay, the line running from the top of the foremast to the bowsprit. The ships were designed to take advantage of the consistent monsoon winds of the Indian Ocean and so put little emphasis on manoeuvrability. Builders smeared hulls with combinations of fish oil, lime and other ingredients to protect against fouling and attacks by shipworm but the coverings may have also helped to keep hulls water- tight. Built typically of teak or iron wood on north Java, Kalimantan or in southern Burma (Myanmar), jongs could be massive, reaching 1,000 tons, 170 richard w. unger though the average for sixteenth century sea-going traders going from India to the Red Sea and to Indonesia was in the range of 400 to 600 tons.19 The jongs that Portuguese sailors encountered in the South China Sea from the sixteenth century on seem to have been different from both the true Chinese junks and from the large sailing ships with sewn hulls of the western Indian Ocean. The big ships they found trading between the Chi- nese south coast and the Indonesian archipelago may in fact have been a hybrid, developed from the merger of Chinese and Malaysian practices. While junks built in northern China were clearly different from the vessels of the Indonesian archipelago a number of features of those Southeast Asian vessels turned up in ships built in southern China starting from the twelfth century if not before. Fastenings were of wood but could be held in place with iron nails. The ships could have either quarter rudders or a single axial rudder. The hulls were V-shaped and could have a keel. The hold rather than being open had bulkheads but they were not watertight and were often associated with heavy frames which also served to reinforce the whole construction. Merchants and traders in southern China seem to have used variants on the design, even having ships built for them in South- east Asia in the sixteenth and seventeenth centuries. That was part of a well-established pattern of commercial contact between the two regions, only disrupted but not stopped by bans on trade imposed by the imperial government in China. In the South China Sea at least it appears that a strict division between Chinese and Malay building traditions did not exist and sea-going ships were combinations, in varying degrees, of practices in the two different parts of Asia, something which further confused European observers.20

19 Irfan Habib, ‘The Technology and Economy of Mughal India’, Indian Economic and Social History Review 17 (1980), 14; Pierre-Yves Manguin, ‘Asian Shipbuilding Traditions in the Indian Ocean at the Dawn of European Expansion’, in: Om Prakash (ed.) The Trading World of the Indian Ocean, 1500-1800, (Calcutta: Centre for Studies in Civilisations, 2012), 597-629; Pierre-Yves Manguin, ‘New Ships for New Networks: Trends in Shipbuilding in the South China Sea in the 15th and 16th Centuries’, in: Geoff Wade and Sun Laichen (eds), Southeast Asia in the Fifteenth Century: The Ming Factor (Singapore: NUS Press, 2010), 333-58; Manguin, ‘The Southeast Asian Ship’, 268-72; A. Jan Qaisar, ‘Shipbuilding in the Mughal Empire’, Indian Economic and Social History Review 5 (1968), 150-54, 165-66; Sanjay Subrah- manyam, ‘A Note on Narsapur Peta: A “syncretic” shipbuilding centre in South India, 1570- 1700’, Journal of the Economic and Social History of the Orient 31 (1988), 305-06. I am grateful to Pierre-Yves Manguin for supplying me with copies of his publications in press as well as articles which proved difficult for me to obtain. 20 Manguin, ‘Asian Shipbuilding Traditions’, 22-24; Manguin, ‘New Ships for New Networks’, 341-49; Manguin, ‘Trading ships of the South China Sea’, 270-74. the technology and teaching of shipbuilding 171 Vessels built in the Malay tradition from the east coast of Africa to the South Pacific shared many characteristics and, small and large, they proved capable of impressive long distance sailing. Such ships carried settlers to Madagascar probably in the eighth and ninth centuries. They maintained trading connections between Arabia and India and between India and Southeast Asia through the High Middle Ages. Vessels from Gujurat im- pressed Portuguese sailors in the early sixteenth century because those ships were larger than their own. The hulls of those giants were probably sewn though it is possible that facing planks were formed with Z-shaped grooves so they would fit more tightly together, a technique mentioned in the seventeenth century and continued well into the nineteenth.21 Islam promoted travel, and vessels based on Malaysian design proved effective in carrying pilgrims for part of their journeys to Mecca. The dhows of the western Indian Ocean were a variant of the general type. Something of a departure from earlier designs used in western Indian, the dhow showed common features with its predecessors in the region as well as signs of possible influence from the Mediterranean. It emerged probably in the thirteenth or fourteenth century but by the fifteenth was in ordinary use for travel to and from Arabia and even further afield.22 Dhows had stitched hulls with no iron used at all in holding the planks in place. The anti-foul- ing compound was like that used in South Asia and because of the trades between India, Arabia and East Africa there was extensive cross-fertiliza- tion of shipbuilding techniques. There were variations on the dhow, a word that only appeared in the eighteenth century and was a generic European word for the sailing ships built along the shores of the Arabian Sea and Persian Gulf. Dhows were generally double-ended. Their sewn hull meant maximum size was less than that of junks, perhaps somewhere around 300 tons. Rig was one or two masts, each carrying a lateen sail.23 The dhow or its predecessors may have found use in the South China Sea as well but evidence of any sort is lacking on whether ships from Ara- bia and India served to maintain trade connections through the straits between Sumatra and the Malay peninsula.24 On the other hand there are indications from names used for different vessel types and also from certain

21 Manguin, ‘Asian Shipbuilding Traditions’, 12-13. 22 Ibid., 27-28. 23 Dionisius A. Agius, Classic ships of Islam: from Mesopotamia to the Indian Ocean (Leiden: Brill, 2008), 141, 147-155, 221-226; Dionisius A. Agius, In the wake of the dhow: the Arabian Gulf and Oman (1st ed. Reading, UK: Ithaca Press, 2002), 2-3. 34-35, 134-137. 24 Manguin, ‘New Ships for New Networks’, 334. 172 richard w. unger

Illustration 5.4. Fleet of Vasco da Gama, from the Livro de Lizuarte de Abreu, between 1558 and 1565, Pierpont Morgan Library, Manuscript M.525, fol. 18v-19. design features that Southeast Asian practice influenced shipbuilding tech- niques on both shores of the Bay of Bengal.25 The vessels that evolved in the western Indian Ocean might have shared roots with those built in the other geographical extreme of the Malaysian shipbuilding sphere, that is in the Pacific Ocean, but their appearance by the late Middle Ages was very different. As log rafts with outriggers or as double-hulled boats with varied rigs of sails made of matting those craft from far to the east were the vehicles for the settlement and irregular con- tact among the islands of the South Pacific. That was thanks in part to the complex navigational skills developed by Polynesian sailors and in part because of the durability of the vessels, sewn together in varying ways, which they used.26

25 Manguin, ‘Asian Shipbuilding Traditions’, 16-19; Manguin, ‘Southeast Asian Shipping’, 190-92. 26 Pierre-Yves Manguin, ‘Sewn-plank craft of S. E. Asia’, in: Sean McGrail and Eric Kentley (eds), Sewn plank boats: Archaeological and ethnographic papers based on those the technology and teaching of shipbuilding 173 In what was to Europeans a New World, boat-builders produced canoes of varying sizes. Dugouts or more generally vessels made from a single tree were in use in 1300 but so presumably were canoes made with wooden frames covered with skins or bark. Though capable of short sea passages the boats were more for inland and river transport. Paddles were the stan- dard source of propulsion, with little if any tradition of relying on wind power. The exception may have been rafts made up of multiple logs that sailed and drifted from the South American mainland to islands in the South Pacific but such voyages were at the very least exceptional.27 There were apparently no vessels capable of consistent trading in any volume across the high seas produced in the Americas in the years before 1500. The two shipbuilding traditions of Europe, products of the different sailing and tidal conditions in the Mediterranean and in the North, merged in the High and Late Middle Ages. In the South oared vessels also powered with a lateen sail, or possibly two, continued in use. They boasted some advances over classical predecessors. Large sailing ships with two steering oars and a lateen sail on each of one or two masts served as cargo carriers of more bulky goods. In the North by 1300 shipbuilders in North Sea and Baltic ports had adapted an earlier Celtic design to create the tubby cog which served as the vehicle for a breakthrough in bulk goods trading. Those shipwrights might have worked in shipyards in allotted stretches of river front in North German port towns but otherwise they and their Mediter- ranean counterparts operated enterprises on their own. The exception was the few shipbuilders who worked in government-operated yards, Arsenals, in Mediterranean ports. Those places of work were devoted to the produc- tion of galleys, typically destined for naval action.28 By 1300 ship design in general appears to have reached some measure of stability. The success of Malaysian and Chinese shipbuilding tended to slow further change or at least make change more a matter of modification and refinement than dramatic development of new types. A possible exception was the evolution of sea traders in the South Chi- na Sea. The articulation of established practice was the norm for shipbuild- ing and for most of the premodern period world-wide. Technological development was a process of many micro-inventions. ‘Craft innovation presented to a conference at Greenwich in November, 1984 (Oxford: B. A. R., 1985), 337; McGrail, Boats of the world, 314-45. 27 McGrail, Boats of the world, 397-98. 28 Richard W. Unger, ‘Admiralties and Warships of Europe and the Mediterranean, 1000-1500’, in: Robert William Love Jr. (ed.), Changing Interpretations and New Sources of Naval History (New York: Garland, 1980), 34-44. 174 richard w. unger was the outcome of small-scale and incremental practical experiment and of random variation’ and shipbuilding was no exception to the general rule.29 It was true in Europe as well as Asia with adjustments in design and construction going on at all times with each new small advance based on previous small improvements. There was a major exception to the general pattern. The pace of technical change became considerably more rapid on the Atlantic coast of Europe in the closing years of the Middle Ages. The transformation in European intellectual life which was self-con- sciously dubbed the Renaissance, coincided with a transformation in ship design between 1300 and 1650. European shipbuilding went through dra- matic advances with major breakthroughs on a number of fronts including related technologies such as navigation. The technological changes proved especially important since, thanks to the design changes themselves, the techniques were diffused to all part of Europe and then beyond to the rest of the world. In the thirteenth century sailors perfected ways of getting out of the Mediterranean into the Atlantic and directly to northern Europe. The mutual exposure to the designs and practices in other parts of the continent led to the merger of northern and southern shipbuilding tradi- tions. Shipwrights matched skeleton-first construction, long common in the South, with the square rig of the North but retaining one triangular lateen sail like those that powered Mediterranean ships. The lighter hull compared to earlier clinker-built northern ships and the composite rig with three masts created a versatile and manoeuvrable ship that could be scaled up or down and could carry fewer seamen per ton than many of its prede- cessors. Though the development of the full-rigged ship was dramatic and cre- ated a new vehicle with significantly greater range than any European pre- decessor, it may well be that there were even larger gains in efficiency in the Late Middle Ages from advances in the design of smaller vessels. Build- ers developed a simple rig which combined on a single-mast a sprit-rigged sail and a sail hanging down from the forestay. The new uncomplicated combination showed up on small boats used on rivers and estuaries. One man could handle the rig, though he presumably often had another man or boy along to give him a hand. Such small vessels then could match the ratios of tons served per man of even seagoing full-rigged ships but had much greater flexibility in providing transportation services. They served presumably to increase regional market integration.30

29 Epstein, ‘Craft guilds’, 70. 30 S.R. Epstein, ‘Regional Fairs, Institutional Innovation, and Economic Growth in Late Medieval Europe’, Economic History Review New Series 47 (1994), 459-82; Richard W. Unger, the technology and teaching of shipbuilding 175 Those smaller coasters and inland boats joined the new sea going full- rigged ships in giving Europeans in the years from 1300 to 1500 an array of vessels with greater capability than ever before. Shipcarpenters in an ongo- ing string of innovations created variants on the designs so that more spe- cialized types could fulfil ever more precise needs. Specialization in design suggested greater efficiency as well as technical advance. The superior sail- ing ships replaced galleys for many tasks. Shipwrights adapted caravels, originally built for fishermen, to make them into, first, vessels for exploring the coast of West Africa and then, second, for trading between Iberia and the Atlantic Islands. Fishing boats of various types received the same treat- ment, with designs adjusted for the types of fish to be caught and the meth- ods of catching them. Just as cargo ships became more specialized, so did warships. The new designs combined with advances in the technology of gunpowder weapons to create vessels that had only one function: visiting violence on others at sea. That process of differentiation was all but com- plete by 1650. European builders in the late seventeenth and through the eighteenth century fell into refining the designs that emerged during the transforma- tion of the Renaissance. Overall the appearance of stability reflected the results of exploring the potential in the new technology, finding the limits and adjusting to fit needs. The tendency to produce more differentiated vessels for more specific tasks continued and became even more apparent. One clear example was the development of a lightly or unarmed cargo ship of box-like cross-section with considerable tumble-home in the upper- works and a simple three-masted rig: the fluyt. Dutch shipbuilders pro- duced the relatively long bulk carrier by the end of the sixteenth century and then spent much of the seventeenth adapting the design for various uses in the trading network of northern Europe. By the eighteenth century the European cargo ship had evolved into the standard three-masted sail- ing packet, able to carry around 500 tons and well-suited for many trades around the world. There were continuing technical advances such as the introduction of the steering wheel, the use of increasingly divided sail plans to allow sailors to distribute their work in smaller units, an increase in the amount of canvas that ships could load on, and the emergence of larger and more serviceable two-masted ships that could supplant three-masters in a number of relatively short distance bulk trades.

‘Maritime Transport and the Integration of Low Countries Grain Markets in the Late Middle Ages’, in: Piet Van Cruyningen and Erik Thoen (eds), Town and Countryside from the late Middle Ages to the 19th Centuries: Supply and Demand of Food (Turnhout: Brepols Publishers, 2012), 101-122. 176 richard w. unger East and West: Cross-fertilization of Techniques

While Europeans continued to refine and expand the range of ship design Chinese and Malaysian shipbuilders appear to have been less able or will- ing to experiment. European voyages of exploration, made possible by the technical breakthroughs in ship design, exposed shipbuilders around the world to new options and methods. Europeans absorbed and exploited some of what they found in Asia. Their Asian counterparts borrowed as well, but the process of change in shipbuilding practice was a slow one, leaving the impression of stability in China, India and Southeast Asia. Chi- nese junks, Southeast Asian jongs and Arab dhows, after all, were flexible, could be built in varying sizes, were designed to deal with local conditions and worked effectively at fulfilling transport needs. The impression of sta- bility in Asian shipbuilding technology through the sixteenth century may be more apparent than real. The pace of change in Europe and the concen- tration of contemporary writers as well as later historians on the accom- plishments of European sailors could well serve to obscure continuing technical advances in Asian shipyards. Europeans were impressed with South Asian ships from their arrival by sea in the last years of the fifteenth century. One famous Portuguese commander early in the sixteenth cen- tury, for example, took on the jalba, a local type, for scouting work since it was superior to any skiff he brought from home.31 European traders recog- nized the advantages of jongs and junks from the outset as well. Portuguese, English and Dutch trading companies along with individual merchants were quick to use vessels of local design. Europeans fit into existing trading networks and so they had goods carried in locally built ships. The adoption of local designs by the interlopers may have even led to a decline in the use of large jongs by Southeast Asian traders.32 The Dutch as late as 1729 still carried on trade in tea with China in junks owned and operated by Chinese merchants. Even after establishing direct trade with Canton (Guangzhou) and even as late as the 1790s some of the ships the Dutch used in carrying goods back and forth to Batavia (Jakarta) were Chinese junks.33 In Makas- sar in the eighteenth century the Dutch East India Company (VOC) for

31 Agius, Classic ships of Islam, 316-20 32 Pierre-Yves Manguin, ‘The vanishing Jong: Insular Southeast Asian fleets in war and Trade (15th-17th Centuries)’, in: Anthony Reid (ed.), Southeast Asia in the Early Modern Era: Trade, Power, and Belief (Ithaca, NY: Cornell University Press, 1993), 197-201. 33 Liu Yong. The Dutch East India Company’s tea trade with China, 1757-1781 TANAP monographs on the history of the Asian-European interaction vol. 6. (Leiden: Brill, 2007), 3-5, 57-58. the technology and teaching of shipbuilding 177

Illustration 5.5. Dutch fluyts at anchor. Wenzel Hollar, print, 1647 Naues Mercatoriae Hollandicæ vulgo VLIETEN. trade between Sulawesi and Java used a variety of locally-built sailing ves- sels of indigenous design, the carpenters coming from a local ethnic group that specialized in shipbuilding. If anything the use of those regional types increased over time.34 In this the Dutch were following a practice, espe- cially with smaller types, which the Portuguese started from the earliest European penetration of Indian Ocean shipping, and the English followed in India well into the eighteenth century. The idea that European ships were so superior to Asian ones that they swept the seas immediately in the early sixteenth century is simply wrong.35 Though European ships did not enjoy a dramatic superiority that Asian builders needed to imitate, those shipwrights did borrow from designs of ships which came to their ports from the Far West. The Portuguese shipyard in Goa, set up soon after the town was taken, produced vessels of the latest European design and, in the process, trained local carpenters. By 1523 there were enough men who had learned the trade that there was no need to

34 G.J. Knaap and Heather Sutherland, Monsoon traders: Ships, skippers and commodi- ties in eighteenth-century Makassar, Verhandelingen van het Koninklijk Instituut voor Taal-, Land- en Volkenkunde; 224 (Leiden: KITLV Press, 2004), 46, 49-51. 35 G.V. Scammell, ‘European Shipowning in the Estado da India, 1500-1700’, Indica: Organ of St. Xaviers’s College Bombay Heras Institute of Indian History and Culture 26 (1989), 124-26. 178 richard w. unger

Illustration 5.6. A Kuwaiti baghla from the 1950s with a stern very similar to European sea-going sailing ships. Dionisius A. Agius, Classic ships of Islam: from Mesopotamia to the Indian Ocean. Leiden: Brill, 2008, 315. import more shipwrights from Europe.36 Enough Portuguese mercenaries and renegades were willing and able to pass on European practices so local buyers could get ships of the latest design or at least ships hard to distin- guish from those brought from Europe. By the late sixteenth century indig- enous shipyards in India were turning out vessels with features like those of European ships. At Aceh on Sumatra, for example, shipyards produced galleys modelled on Mediterranean techniques.37 Often Asian builders adapted or adjusted their own designs, giving the vessels that came from their yards borrowed attributes. The baghla, the largest and most ornate of ocean-going dhows, had a sternpost rudder and a hull that may have been influenced by European design, though it is also possible that the

36 Ibid., 119 37 Manguin, ‘The vanishing Jong’, 206; Subrahmanyam, ‘A Note on Narsapur Peta’, 308-09 the technology and teaching of shipbuilding 179 inspiration came from Chinese junks. With the Yemeni sanbūq there is little question that it got its square stern after European contact in the sixteenth century, another sign of Portuguese and, later, Dutch influence.38 In the Indonesian archipelago a type called a chialoup was a common car- rier among the islands and its roots were undoubtedly European.39 The expansion and prosperity of Indian shipping up to the mid seventeenth century may owe something to improvements in the design of locally-built ships that evolved in the prevailing atmosphere of cross-fertilization. The shift to imitating imported designs meant significant changes in practice and in components. Taking on European hull construction meant that planks were fastened with nails and in the western Indian Ocean there was a shortage of iron for nails. It may be that Indian shipwrights had started using nails before Europeans arrived by sea and instead borrowed the practice from China but what evidence that exists is inconclusive. In India where iron was available shipbuilding had an advantage and more so as, over time, locally-built ships became more like European ones.40 In the second half of the eighteenth century the thriving shipbuilding industry in southern Vietnam depended on skilled workers of varied back- grounds, including Europeans. The products of the yards were of both tra- ditional and European design. Building costs were low enough that Chinese merchants imported junks from yards along the Saigon River and local merchants bought Western style schooners from builders there. Thanks to migration from Vietnam Siam also developed an industry able to produce a varied range of ships. Those travelling carpenters took with them an abil- ity to build in Western ways. Builders in Vietnam were to some degree unique, since they showed more than any other Asian craftsmen, at least in the closing years of the eighteenth century, an ability to adopt and adapt European techniques. Oddly enough there were at the same time pockets where builders still sewed planks together on ships as large as 100-150 tons, a practice along with a few others which Malay shipwrights had long since abandoned. For various political reasons oddly some ports got cut off from the evolution in design in the rest of Southeast Asia.41 Another difference in Vietnam was that the much of the building, especially of large craft, was

38 Agius, In the wake of the dhow, 49-52, 180; Agius, Classic ships of Islam, 159. 39 Sutherland, Monsoon traders, 45-46, 52. 40 Agius, Classic ships of Islam, 156, 165-167; Manguin, ‘Asian Shipbuilding Traditions’, 13-15; Subrahmanyam, ‘A Note on Narsapur Peta’, 307. 41 Manguin, ‘Sewn-plank craft of S. E. Asia’, 321 180 richard w. unger in state-owned yards.42 Outside of the Portuguese in Goa and the Dutch in ports in north Java and imperial yards which came and went in China, al- most all shipbuilding in Asia, including innovative shipbuilding, took place in private enterprises and often with workers who enjoyed a variety of backgrounds. While shipwrights in South, Southeast and East Asia were far removed from the sites of European shipbuilding, their Ottoman counterparts in western Asia had direct access to Mediterranean practice and so could easily transfer their knowledge to the Red Sea. Penetration of European techniques along that route seems to have been less than through ships sailing directly from the European Atlantic front to the Far East and to ports in between. Ships in the Indian Ocean built in the Malay tradition could and did find specific tasks and continued through the eighteenth century to be the carriers of goods over the open ocean, though with diminishing importance from the mid seventeenth. Ships in the South China Sea built in the Chinese tradition continued to find use in many trades for indige- nous as well as for European shippers. Though local designs were subject to European influence and though there were modifications in the products of Asian shipyards, even to the point of building ships of western design, established building practices remained durable into the nineteenth cen- tury and even beyond. The drift in the Far East was possibly toward more specialization as European-style vessels took over some jobs formerly done by locally-built ships. In the New World local shipbuilding continued and Europeans ­adopted some designs for use on inland waterways. There, though, Euro- peans brought their own ships and shipbuilding methods which replaced most previous practices. The immigrants imported a nexus of maritime technology which swamped what had existed before, drastically extending the potential for transportation. Repair and construction of ships of Euro- pean design was almost immediately a part of overseas settlements in the Americas. The same was true, if to a lesser extent, in Asia. Trans-oceanic and intercontinental trade were dominated by European ships with the products of other shipbuilding traditions reduced to a lesser but durable role. While that trend was well underway by 1650 it would become the common pattern in the following century and a half.

42 Li Tana, ‘Ships and Shipbuilding in the Mekong Delta, c. 1750-1840’, in: Nola Cooke and Li Tana (eds), Water Frontier Commerce and the Chinese in the Lower Mekong Region, 1750-1880 (Lanham, MD: Rowman & Littlefield Publishers, Inc., 2004), 120-131. the technology and teaching of shipbuilding 181 The spread of European shipbuilding to the rest of the world continued and at an increasing pace. The secure colonial positions of certain Euro- pean states in India and Southeast Asia made possible the establishment of permanent shipyards. The Portuguese and after them the Dutch set up yards which produced ships of European design.43 Builders set up supply chains, shipping in raw materials like wood from distant sites to assure the quality of their products. For example, Spanish galleons were built in Manila at the end of the seventeenth century with wood brought there from Cambodia.44 In the New World shipbuilding became European. The English colonies which would become the United States of America moved from making small boats for local travel and inshore fishing to producing an increasing array of different designs, some original with long term im- plications for improvements in sailing ships. As the colonists mastered the problems of finance and marketing they increased their sales to Europe. By 1774, on the eve of the rebellion against the rule of George III, one in three English ships was built in America.45 Europeans expanded their share of trade in Asian waters and while Asian shipwrights still produced large numbers of ships such as junks, the growth in tonnage produced was more rapid for vessels of European design. An obvious sign of the technical success of European shipbuilders was the growth is the size of the merchant fleet. Specific jurisdictions such as Portugal in the fifteenth and sixteenth centuries, the Dutch Republic in the seventeenth, and England in the eighteenth century saw marked increases in the number of ships built and the number in use. The total size of the European merchant fleet rose more than fifteen-fold from the early six- teenth to the early nineteenth century. Even more impressive than the more or less 1.0% compound growth rate in tonnage was the approximate- ly 0.7% growth rate in tonnage per capita.46 Shipping was one of the fastest growing sectors in the European economy and by definition shipbuilding was also among the sectors leading all others. Though comparable data is

43 C.R. Boxer, The Portuguese seaborne empire, 1415-1825. (1st American ed. New York: A.A. Knopf, 1969), 209-11 44 Alastair Lamb, The Mandarin road to old Hué; narratives of Anglo-Vietnamese diplo- macy from the 17th century to the eve of the French conquest (Hamden, Conn.: Archon Books, 1970), 55; Tana, ‘Ships and Shipbuilding in the Mekong Delta’, 123. 45 Jacob M. Price, ‘A Note on the Value of Colonial Exports of Shipping’, in: Price, The Atlantic frontier of the thirteen American colonies and states: essays in eighteenth century commercial and social history (Aldershot: Ashgate Variorum, 1996), 704-06, 709. 46 Jan Luiten van Zanden, ‘Early modern economic growth: A survey of the European economy, 1500-1800, in: Maarten Prak (ed.), Early Modern Capitalism Economic and social change in Europe, 1400-1800 (London: Routledge, 2001), 81-82; Richard W. Unger, ‘The Ton- nage of Europe’s Merchant Fleets 1300-1800’, The American Neptune 52 (1992), 247-61. 182 richard w. unger lacking, there is no evidence for anything like that pace of expansion in shipping tonnage and so in shipbuilding in any other part of the world. The pace of growth and of technical change quickened even more in Europe after around 1820, thanks to the transfer of steam engines from land to shipboard. In the earliest versions steamboats were inefficient and could only serve effectively as river or tug boats since they needed so much fuel. As tugs they did make clearing ports much easier for sailing ships and so contributed to a marked rise in shipping efficiency. Steam engines became better over time at putting the energy from fuel to effective use and ship design changed in response. Through the nineteenth century ships came to be built with different materials and equipped with different power sources. Even by 1820, though, there were signs of a new technological re- gime which would in the future share few attributes with the practices and organisational structures of the previous five centuries.

Creation and Transfer of Knowledge

From 1300 to 1800 shipbuilders around the world and especially in Europe did create a new body of knowledge about how to make better ships and, as important, they effectively transferred that knowledge to their succes- sors. There were clear signs of success over time in teaching shipbuilding and sustaining the process of learning and teaching. While knowledge creation may have been obvious with Europeans, continuing knowledge transfer was as, or even more obvious, in China, South and Southeast Asia. In order to establish the ability of institutions to promote or deter the creation and diffusion of shipbuilding technology the institutions them- selves need to be identified, the character of required knowledge needs to be identified, and the people who put that knowledge to use need to be identified. Across the globe and over a number of centuries not surpris- ingly the institutions, the techniques and the organization of the shipbuild- ing industry were far from static. That of course makes the necessary tasks more difficult and unfortunately reduces the accuracy and the generality of any conclusions. Still, if not conclusions, at least some observations about institutions and the teaching of the technology of shipbuilding are possible. One generalization seems true. Shipbuilding worldwide, typically, took place in private yards owned and managed by independent entrepreneurs employing a number of skilled workmen. The size of firms varied and though they tended to get larger as ships got bigger and work became more the technology and teaching of shipbuilding 183 regular, still as late as the end of the eighteenth century, in the middle of long drawn out wars which sharply increased demand for ships in Britain, the leading producer of shipping tonnage in the world, shipbuilding enter- prises were still rather informal with anything but robust financing and management.47 There were state-owned yards, generated in Europe large- ly for the building of warships, but also in China and later in Vietnam for the building of a variety of vessels of differing sizes for commercial as well as state use.48 As warships became more differentiated from cargo ships in Europe, governments created naval shipyards that not only fitted out vessels and replenished supplies but also built warships. Those were estab- lishments with some permanence. Private yards on the other hand were often temporary, certainly up through the sixteenth century. Builders turned to any convenient piece of beach for the duration of the job and then moved on. Towns in northern Germany and the Netherlands in the Late Middle Ages set aside a part of their harbour fronts for shipbuilding which meant that the trade was carried on in the same place but the crafts- men and the businesses came and went.49 Enterprises did over time be- come more stable and shipbuilding firms tended to continue over longer periods, but they were often family affairs which would shut down if there was no likely successor to the owner. Shaping wood was the primary skill of all those who built ships in Europe and China, though that was less true in India and the South Pa- cific. It was the types of ships built which dictated the tasks done on the wharves and the degree of differentiation among those tasks and the work- men completing them. The evolution of ship design in the Mediterranean in the early Middle Ages created a distinction between those who shaped and fitted the wood and those who laid out the design of the ship. With bigger and more complex vessels, the shipbuilder came to be the designer who had an understanding of the entire project even before the first nail was driven. The same may have been true among builders of junks in Chi- na and also with builders in the Arabian Peninsula, though when the de- signer and supervisor separated from the men shaping the ship is not clear.

47 Helen Doe, ‘The Smugglers’ Shipbuilder: The Customers, Trades and Vessels of a Mevagissey Shipyard, 1799-1816’, The Mariner’s Mirror 92 (2006), 427-42. 48 Hans Lothar Scheuring, Die Drachenfluss-Werft von Nanking: Das Lung-chiang ch’uan-ch’uang chih, eine Ming-zeitliche Quelle zur Geschichte des chinesischen Schiffbaus Heildelberger Schriften zur Ostasienkunde vol. 9. (Frankfurt/Main: Haag + Herchen, 1987), 18-40; Tana, ‘Ships and Shipbuilding in the Mekong Delta’, 122. 49 E.g. Karl Friedrich Olechnowitz, Der Schiffbau der hansischen Spätzeit (Weimar: H. Böhlaus Nachfolger, 1960), 27-28, 112-13. 184 richard w. unger Certainly it was the case in southern Europe in the early Middle Ages and, as Mediterranean designs and practices merged with those of northern Europe, the distinction on wharves between designer, who was the direc- tor-manager, and carpenters spread throughout the continent.50 Because of the way work came to be organized, the training and knowledge required was very different for the different people working on shipbuilding wharves. Most needed to know essential and basic carpentry skills. Though ships might require some variants on what was done in other woodworking trades, as far as most men were concerned the skill they had to have was the ability to shape wood. Some knowledge of how ships behaved was obviously an asset, but even for the construction of small river craft build- ers needed only the standard tool kit and the knowledge of any carpenter. There were beside the shapers of wood a few skilled craftsmen who were able to move higher and become the designers of ships. By the seventeenth century they were self-conscious about their skill and could enjoy some recognition, especially if they worked for royal buyers.51 Designers would draw out the plans for the principal frames of the ship, first on the sand floor in a shed but by the early seventeenth century on paper, which then could serve as a guide for the men who worked with wood.52 Those design- ers, who were also owners of wharves, had to have all the regular skills of shipcarpenters, but with time and experience and training they moved beyond their counterparts. In Arabia builders worked by eye and experi- ence, while European builders became increasingly the slaves of drawn plans.53 It was the designers, the men who created those plans, probably more than the carpenters who were the source of new techniques in the seventeenth century and beyond. It was those designers, the foremen and owners of wharves, who passed on skills and understanding of the enter- prise to successors. By the eighteenth century in some prosperous ports the men who owned wharves even moved away from design work and left the day-to-day operation of the business, as well as decisions about imple- mentation of technology, to foremen who in turn became the source of training for new shipbuilders. There were guilds of shipcarpenters in Western Europe. To the New World and Asia Europeans did not export guilds. It appears that nothing

50 Richard W. Unger, The Art of medieval technology: images of Noah the shipbuilder (New Brunswick, NJ: Rutgers University Press, 1991). 51 Phineas Pett, The autobiography of Phineas Pett (London: Printed for the Navy Records Society, 1918). 52 Unger, The Art of medieval technology, figure 12. 53 Agius, In the wake of the dhow, 140-45. the technology and teaching of shipbuilding 185 like the European organizations existed elsewhere. Even within Europe not all towns with shipbuilding had shipcarpenters’ guilds. In the many port towns throughout Europe that did have shipcarpenters’ guilds the organi- zations tended to be more like trade associations with town governments assigning powers to them to regulate the industry.54 Guilds for the build- ers of ships were most common in North-western Europe, that is in Ger- many and the Low Countries. Efforts by absolutist monarchs in France and England to create national guild organizations in the late sixteenth and seventeenth centuries met with some limited success. French governments wanted to have industrial guilds to tax trades more effectively, but they also wanted guilds of shipcarpenters to assure supplies of necessary equipment and fittings for the navy. In England the Stuart kings failed to impose a universal guild for shipcarpenters because of local opposition and because of their general inability to impose a programme of centralization of au- thority. England then, like many parts of Europe, had few institutions for administering shipbuilding.55 The guilds that did exist mostly continued until the end of the eighteenth century, when liberal policy makers swept away as much regulation of the economy as they could. Though guilds of shipcarpenters were common they were by no means the dominant mode of organizing shipbuilding even in Europe between 1300 and 1800. The guilds that did exist trained new shipcarpenters. As with virtually all guilds, education of new tradesmen was one of the principal functions of the organisations. There were typically provisions for apprenticeships and, on completion, the new shipbuilder would, by the seventeenth cen- tury, get a letter stating he had completed formal training.56 Each aspiring craftsmen would work under the supervision of a master, learning the es- sentials of the trade. The guild regulated the length of the apprenticeship and the requirements for completion, generally just a period of instruction running to three years or more. The letter the guild produced at the end of the apprenticeship, like any educational certificate, became a source of validation and so of mobility. Men armed with evidence of successful train- ing could travel and expect to find work, carrying with them technology virtually and truly embodied. Guilds did not intend that their letters would promote tramping from one place to another but the structure of training

54 Frederic Chapin Lane, Venice, a maritime republic (Baltimore: Johns Hopkins Uni- versity Press, 1973), 165. 55 Richard W. Unger, Dutch Shipbuilding before 1800: Ships and guilds (Assen: Van Gorcum, 1978), 65-66. 56 Olechnowitz, Der Schiffbau, 68-81. 186 richard w. unger and the general openness to non-native skilled craftsmen common among guilds made the practice possible. Though those travelling journeymen might ultimately want to return home, as with many other crafts during their tramping years they were able to learn about new methods and to pass them on in their next job in some other place. Guilds in some cases did place restrictions on mobility by limiting the numbers of foreigners who could be hired but those limitations were few and often temporary.57 After a period of practising the trade either at home or elsewhere a ship- carpenter could rise to master status. That required completion of a task or set of tasks to the satisfaction of guild officers responsible for examining their brothers. More important than skill, at least by the eighteenth cen- tury, was access to capital since setting up business as an independent shipbuilder was expensive. Masters with their own wharves would be de- signers of the ships built but they had no specific training set down by the guilds. What set them apart was skill derived from earlier training and experience and the ability to finance and run their own business. There were typically no restrictions on the exchange of information within the guilds. Members were not required to divulge what they knew to their brothers but since they worked so closely together and often along- side each other, secrets, if they existed, were hard to keep. Members were not stopped from trying novel methods of building or novel designs. They were free to innovate in any way they chose. They had easy access to what was new in the industry because of their seeing what others were doing, but even more because ships visiting from other ports brought physical evidence of design changes elsewhere. In towns where repair rather than new construction was the principal form of work for shipcarpenters, there was a high degree of equality among the guild brothers. There can be little doubt that guildsmen did generate knowledge about how to build better ships through experience and experiment. Guilds did not restrict competi- tion and placed light requirements on entry through low initiation fees. The heavier requirements were the training regimens along established lines. The guilds promoted, preserved, and institutionalized a handicraft tradition, passing on knowledge from individual to individual while main- taining levels of freedom to allow for technical advance.

57 Epstein, ‘Craft guilds’, 73-76; Olechnowitz, Der Schiffbau, 71-72, 188-189; Reinhold Reith, ‘Circulation of Skilled Labour in Late Medieval and Early Modern Central Europe’, in: Epstein and Prak (eds), Guilds, innovation, and the European economy, 131-135, 141-142. the technology and teaching of shipbuilding 187 In Asia builders were able to pass on knowledge of how to design and build vessels of all sizes. They were also able to absorb and adapt new de- sign features, either indigenously developed or imported from other tradi- tions. They did not generate institutions similar to European ones for the development of their trade. Though there were informal methods of train- ing and knowledge transfer those did not prove as effective as European ones through much of the period in generating the necessary skills nor extent of innovation noticed especially in the far North-western reaches of Eurasia. European shipcarpenters’ guilds and guilds in general have re- ceived much more attention than the less institutionalized arrangements of Asia, undoubtedly because of the relative availability of surviving evi- dence. More careful examination of what went on among shipbuilders on the shores of the Indian Ocean and the China Seas will reveal data about how skills were generated and passed on but until that is done the undeni- able impression is that European guilds made a positive contribution to the development of ship design. Europeans also developed other ways for men to acquire knowledge about how to build a ship and again they seem to have moved beyond practices in Asia, especially in the seventeenth and eighteenth centuries.

Transferring Knowledge with the Written Word

Alongside the handicraft tradition was a written one with information transferred across space and time through works, both manuscript and later printed, dealing with shipbuilding. Learned works on shipbuilding appeared in China such as a history of Chinese shipbuilding by Tseng King- liang in 1044 and other writers in the Middle Ages commented on the de- sign and form of Chinese ships.58 In Europe the earliest works on the form of built ships appeared in the fifteenth century. In 1434 Michael of Rhodes writing in Venice produced what is probably the world’s first treatise on shipbuilding. Very possibly designed to train younger men on board ship and even more likely written to help with his ambitions to rise higher in the Venetian merchant service, the book devoted some 47 of its 440 pages to laying out precise dimensions for building different types of galleys and sailing ships. The list of measurements applied not only to the planks but also to the sails, cordage and anchors. Motivated perhaps by experience or interest in the mathematical basis for ship design, Michael of Rhodes pro-

58 Needham, Civil engineering, 460-65; Scheuring, Die Drachenfluss-Werft, 6. 188 richard w. unger vided extensive detail, of use as much for confirmation of building prac- tices as it was for guiding someone trying to build a ship.59 Other Venetians followed suit in the fifteenth and sixteenth centuries, offering descriptions of ships and categories to classify vessels.60 Li Chao-hsiang’s 1553 book on shipbuilding had the same air of a report. It was logical that the chief ad- ministrative officer of the largest shipyard in China, the Lung-chiang wharf in Nanjing, would concentrate on giving details of the classification of ships, controlling costs of construction and setting down instructions as well as policy recommendations for overseeing a shipyard.61 While books on carpentry and house building in particular, especially more popular works, might be infused with discussion of ritual and magic,62 Li Chao- hsiang’s work avoided any such issues. Michael of Rhodes was a sailor but even his book seemed to be more of a handbook for an administrator. The books were not so much shipbuilding treatises intended to show how to build ships, as reflections of how building was done. They bore a strong resemblance to records of administration that became increasingly com- mon in chanceries from the thirteenth century.63 If there was a difference in approach it may be that in Venice there was more interest in using ge- ometry to describe construction, as in the case of the mid fifteenth cen- tury work of Giorgio Timbotta.64 That may perhaps have been a sign of the general interest among Renaissance Europeans in classical authors, and in the case of shipbuilding the author offering inspiration would have been Euclid. The geometry was limited and practical, with no sense of some overarching theoretical concerns or indeed of any general understanding of the purpose, place or function of the ship. If the works on shipbuilding up to the middle of the sixteenth century transferred technology, it was

59 Michael of Rhodes, Michael of Rhodes, A Medieval Mariner and His Manuscript Dib- ner Institute, 2006 [accessed June 6, 2008]. Available from http://brunelleschi.imss.fi.it/ michaelofrhodes/; Pamela O. Long, David McGee, and Alan M. Stahl (eds), The Book of Michael of Rhodes, A Fifteenth-Century Maritime Miscellany 3 vols. (Cambridge MA: MIT Press, 2008). I am grateful to Alan Stahl for supplying me with the translation of the text on shipbuilding in Michael of Rhodes’ book before it appeared in print. 60 R.C. Anderson, ‘Italian Naval Architecture about 1445’, The Mariner’s Mirror 11 (1925), 135-63; Frederic C. Lane, ‘Venetian Naval Architecture about 1550’, The Mariner’s Mirror 20 (1934), 24-49. 61 Scheuring, Die Drachenfluss-Werft. 62 Klaas Ruitenbeek, Carpentry and building in late imperial China: A study of the fif- teenth-century carpenter’s manual Lu Ban jing Sinica Leidensia, vol. 23 (Leiden: Brill, 1993), 1-5, 25. 63 John Pryor, ‘The Galleys of Charles I of Anjou, King of Sicily: ca. 1269-84’, Studies in Medieval and Renaissance History New Series 14 (1993), 35-44. 64 Anderson, ‘Italian Naval Architecture’, 142-54 the technology and teaching of shipbuilding 189 how to administer an industrial operation, typically one owned by the state, rather than how to build ships. States promoted, though perhaps not ac- tively, the development of the written tradition but for administrative, not technical purposes. In the second half of the sixteenth century the number of works on shipbuilding increased as did the concentration in them on the building of ships. Fernando Oliveira’s Liuro da Fábrica das Naus of 1580 offered ­illustrations and instructions, if somewhat general and presented with a scholarly air, on the progress of building a large cargo ship. This was the beginning of a series of works produced for Iberian audiences. Diego de Palacios wrote the first printed book on shipbuilding, published in Havana in 1587.65 It was a vehicle for taking European and, more specifically, Span- ish shipbuilding practice to the New World where growing trade and the presence of raw materials created great interest in how to build a ship. Mathew Baker, the master shipwright of Queen Elizabeth I of England, produced ‘Fragments of English Shipwrightry’ around 1585. He was an ex- perienced shipbuilder who brought practical knowledge to his task, but had little success in presenting what he knew in a useful way. He did suggest graphically that since a mackerel swam easily through the water, then a ship’s hull should be like the body of a mackerel. How that goal might be achieved he left unclear. Baker’s book indicates the level of theory used and expressed by sixteenth century shipwrights. It also indicates that the books were not always meant to be understood and that they were not necessarily practical manuals. It is easy to overestimate his accomplish- ment.66 It would be some time before the written tradition would escape from the task of classification even though, buried in the early books, there might be some practical descriptions of what was an extremely complex process. Seventeenth century European works on shipbuilding continued along the same paths. The appearance of Joseph Furttenbach’s book on ship- building in 1629 indicated both a rising interest among scholars in the topic and the application of the term architecture to the design of ships. Though more learned than earlier works, the book was still committed to lengthy lists classifying types of ships.67 When a mayor of Amsterdam,

65 Vernon D. Tate, ‘The Instrvcion Nauthica of 1587’, The American Neptune 1 (1941), 191-95. 66 David McGee, ‘From craftsmanship to draftsmanship: Naval architecture and the three traditions of Early Modern design’, Technology and Culture 40 (1999), 222-24. 67 Josephum Furttenbach, Architectura Navalis Das is Von dem Schiff Gebäu; auff dem Meer und Seekusten zugerauchem ... (Ulm: Jonam Saurn, 1629). 190 richard w. unger Nicholas Witsen, in 1671 wanted to write about shipbuilding he went to an experienced shipbuilder to give him the necessary information. The second edition of 1690 of his book carried the title Architectura Navalis.68 The next large classic Dutch work on shipbuilding published in 1697 was, on the other hand, by a practising shipbuilder, Cornelis van Yk.69 By the close of the seventeenth century at least one shipbuilder was able to write about his trade and in a way that would appeal to a literate public. There was in the seventeenth century an effort to communicate the character of ship- building to buyers, perhaps because an increasing number of those buyers did not necessarily have an intimate knowledge of the sea. Builders began to produce half models and then full scale models to present to investors in ships but more frequently to the political appointees of admiralty boards responsible for procuring ships for navies. In that way as in others the state continued to influence what information was collected and how it was transferred. Navy boards certainly played a role in forcing builders to ar- ticulate practices on the wharf. The half models may have been helpful to the land-bound administrators but shipbuilders could also use them to take lines of the ship and guide construction. The advances in mathematics of the seventeenth century led in the eighteenth to the emergence of a scientific tradition in examining and writing about shipbuilding. Larrie Ferreiro in a recent book carefully de- scribed the process of the infiltration of science into what had been far from a scientific enterprise.70 Problems of determining the displacement of a ship before launch, the stability and the handling qualities had been topics of discussion among shipbuilders, but in the eighteenth century in a series of treatises men versed as much or more in science than in seafar- ing took on those questions. A mathematician as distinguished as Leonhard Euler might discuss the best proportions, theoretically, for a ship.71 It was

68 Nicolaas Witsen, Caspar Commelin, Broer Appelaer, and Jan Appelaer, Aeloude en hedendaegsche scheeps-bouw en bestier: : waer in wijtloopigh wert verhandelt de wijze van scheeps-timmeren, by Grieken en Romeynen, scheeps-oeffeningen, strijden, tucht, straffe, wet- ten, en gewoonten. Beneffens evenmatige grootheden van schepen onses tijts ... verschil van bouwen tusschen uitheemschen en onzen landaert, Indisch vaertuygh, galey-bouw, hedendaeg­ sche scheeps-plichten, verrijckt met een reex verklaerde zee-mans spreeck-woorden en bena- mingen (Amsterdam, 1671). 69 Cornelius van Yk, De nederlandsche scheeps-bouw-konst open gestelt: vertoonende naar wat regel, of evenredenheyd ... mitsgaders masten, zeylen, ankers (Delft: Andries Voorstad ... voor Jan ten Hoorn, 1697). 70 Larrie D. Ferreiro, Ships and science: The birth of naval architecture in the scientific revolution, 1600-1800 (Cambridge, MA: MIT Press, 2007). 71 Leonhard Euler, Scientia navalis, sev Tractatvs de constrvendis ac dirigendis navibvs (St. Petersburg: typis Academiae scientiarvm, 1749). the technology and teaching of shipbuilding 191 in France where a combination of committed naval officers, scientists, and a government at times devoted to naval expansion created by mid-century not only a literature on ship design, but also a school, small though it was, for training naval architects.72 While Pierre Bouguer launched the discus- sion in France, applying his knowledge of calculus to questions of sailing performance, it was Henri-Louis Duhamel du Monceau who made naval architecture a discipline, both through his teaching and through his text- book on Elémens de l’Architecture navale ... of 1752. Calculus became, along with many other mathematical and geometric skills, part of the tool kit of accomplished ship designers. Works in Spanish, in Swedish, and in English followed in the last years of the eighteenth century.73 Much of the work of bringing science to naval architecture was done within navies and with the support of governments. The invasion of science hardly reached commercial shipbuilding. Especially for smaller ships and in lesser ports handicraft traditions and traditional methods and training of new shipcarpenters continued to prevail well into the nineteenth cen- tury. The use of plans for ship construction was almost exclusively the re- serve of naval shipbuilding until the mid-nineteenth century.74 The influence of science and engineering on shipbuilders was apparently great- er in France than in England or the Netherlands, or indeed in China or the New World. The effect of the advances in thinking about shipbuilding is hard to measure since those new ideas only slowly reached beyond design- ing warships. Though the state, principally through the growing naval es- tablishments, had been and was increasingly a force for formulation of theories about shipbuilding, the creation of new methods and the trans- mission of that information remained in the hands of experienced practi- tioners. The use of science, of conscious research and of experiment in the building of ships was a product of the nineteenth century. The shift to a new basis for design served to remove or at the least weaken constraints on innovation.

Disincentives for Technical Advance: Guilds and Market Structure

Guilds of shipcarpenters in Europe were certainly involved in the creation and diffusion of technical knowledge about how to build ships. The men who built ships faced incentives and disincentives to developing new tech-

72 Ferreiro, Ships and science, 282-90. 73 Anderson, ‘Italian Naval Architecture’; Ferreiro, Ships and science, 269-78. 74 McGee, ‘From craftsmanship to draftsmanship’, 222. 192 richard w. unger nologies and training others. The place of guilds in the spectrum of support or deterrence of technical advance remains difficult to evaluate. There was technical advance, worldwide, from 1300 to 1800, and especially in Europe. There were guilds administering shipbuilding, at least in some parts of Europe. The obvious conclusion, or at least suggestion, is that guilds were a positive force. The danger of a post hoc ergo propter hoc argument looms, and even if guilds did promote technical advance it is possible that in their absence the pace would have been faster. Certainly improvements in build- ing methods took place outside guilds. The impetus for the regularization of knowledge, the development of a written tradition describing the pro- cess of building a ship and the application of new scientific learning to designing ships did not come from guilds. The impetus came from other sources and especially from navies and only developed after the middle of the seventeenth century. The character of the product, that is the ship, and the market for it were as important or more important than institutions in determining the pace of technical change. That was true as much in Asia as in Europe. Against the shipbuilder between 1300 and 1800 was arrayed a mass of disincentives to experiment, to try new designs or building methods and so create new techniques. The reputation for conservatism enjoyed by shipbuilders was a product of the high cost of failure. The disincentive to take risks was in- timately related to the size of the capital investment involved. Ships were the largest capital good, that is other than land, that was an option for in- vestors in pre modern Europe. Land had the advantage of being divisible while ships were single units. Investors in China as in Europe tried to mit- igate the indivisibility by part ownership but that did not solve the problem of the shipbuilder who, through error, would destroy the entire ship and with it all the sunk capital. In a recent paper Maryanne Kowaleski reported sale prices of ships, new and used, in fourteenth and fifteenth-century England. The figures varied widely because of the different kinds of ships involved and the conditions the vessels were in. She estimated that a 10-year old ship in the 1390s would have brought about 25 shillings per ton, the equivalent of slightly more than a tonne of wheat at prevailing prices.75 A 100-ton ship would cost as much as enough grain to feed about 275 ­people

75 Maryanne Kowaleski, ‘Investment and Ship-owning in Medieval England’, Paper presented at Measurable Advantages: Transportation, Trade, and Technology between Ancient Rome and the Modern West, University of British Columbia 2008, 2-3 April; Robert Allen and Richard W. Unger. Allen – Unger Database European Commodity Prices 1260-1914 2004 [accessed June 8, 2008]. Available from http://www.history.ubc.ca/faculty/unger/ ECPdb/about.html. the technology and teaching of shipbuilding 193 for a year. With large quantities of capital involved, extending the service- able lives of ships had obvious advantages for the economy. Success ap- pears to have been limited. Federigo Melis, working from results of research in the Datini Archive, set the average life expectancy of ships sailing in the western Mediterranean around 1400 at a bit under five years.76 Estimates for the life expectancy of Dutch ships in the seventeenth century, based on what Nicolaes Witsen said, put the number at ten years on average for cargo ships and twenty for fishing vessels.77 Almost contemporary Chinese regulations suggested eight years to be easily attainable with refits creating a good possibility of vessels reaching thirteen years of age before they might be scrapped. Survival rates depended on the quality of maintenance.78 Using Lloyd’s list it appears that ships built for ocean trading in the future United States by the 1770s survived on average between 13.3 and 16.8 years.79 The data is at best only suggestive though it is consistent with estimates of life expectancy offered to the new Secretary of the Treasury, Alexander Hamilton, in 1789 by customs collectors in some of the main United States ports.80 It does seem that some gains were made in the durability of the expensive capital good over the centuries in the western shipbuilding tra- dition, but it may be that Chinese junks survived longer on average than European packets in 1800. The tendency toward longer-lived ships, given the pressing need to decrease risk, was an incentive for builders to work on incremental changes in design and building methods. If ships were more durable and if total tonnage was the same or falling then opportunities for experiment were restricted. It is possible that Chinese shipbuilders from the sixteenth century faced that combination of restrictions. The rapid growth in the European merchant marine would have given builder more chances to make exactly the incremental changes which led to increas- ingly superior ships. Another disincentive to dramatic change, or for that matter to any change in shipbuilding techniques, was ways in which the process of design was conceived or understood. Architects responsible for massive building projects in the Central Middle Ages, faced with producing monastic estab- lishments and soaring cathedrals with roofs of stone that had to stay in

76 Federigo Melis and Luciana Frangioni, I trasporti e le comunicazioni nel Medioevo (Florence: Le Monnier, 1984), 32. 77 Unger, Dutch shipbuilding, 11. 78 Moll-Murata, ‘State and crafts’, 145-46. 79 Price, ‘A note on the value’, 717-19. 80 Marshall Smelser and William I. Davisson, ‘The Longevity of Colonial Ships’, The American Neptune 33 (1973), 17. 194 richard w. unger

Illustrations 5.7 & 5.8. Progress of ship construction in the past from Nicolaes Witsen, Aeloude en hedendaegsche scheeps-bouw en bestier of 1671 but copied from → the technology and teaching of shipbuilding 195

Fernando Oliveira's Liuro da Fábrica das Naus of 1580. Leiden University Library, 678 A 15, plates XVIII and XIX. 196 richard w. unger place, started from general concepts, in the first instance philosophical but more specifically geometrical.81 Shipbuilders do not seem to have ever had a philosophical understanding of the role of geometry but rather were interested in generating a workable result. Though early written works on shipbuilding from fifteenth century Venice suggest an interest in geometry, even there most writers were taken up with specific measures of planks and ropes, derived presumably from common practice. Without some theoretical underpinning designers were limited in their ability to imagine or try out novel designs. Certainly before the eighteenth century there was no solid basis for what to expect from a ship once built. Shakedown cruis- es were a necessity and adaptation and accommodation, the addition of extra wood here and there to improve stability, was accepted as normal.82 Another disincentive was the lack of competition among shipbuilders. In the Middle Ages in general, and in smaller centres even into the nine- teenth century, there was often only one builder available to those wanting to have a ship built. Over time that changed and in major ports the range of choice increased. In some towns wharves were concentrated in one place so buyers had easy access to competitors. In places like Venice and Genoa in the Renaissance and Holland in the seventeenth century, and then Eng- land in the eighteenth, there were sizeable numbers of builders available at least in or near major ports. In the last years of the seventeenth century new cargo ship construction in the Dutch Republic came to be concen- trated in the Zaanstreek, just to the northwest of Amsterdam. The number of builders producing big ships there was up to 60 by 1669.83 In England there was concentration on the lower Thames at about the same time, though shipbuilding spread more widely around the country as production expanded in the following century. The thicker market for ships and gener- ally easier travel as time went on, meant improved communication be- tween buyers and builders, generating closer relationships and easier consultation as well as exchange of knowledge.84 The presence of some competition and the ability of buyers to relay their needs effectively to

81 Nigel Hiscock, The wise master builder: Platonic geometry in plans of medieval abbeys and cathedrals (Aldershot: Ashgate, 2000), 273-282, 290-291. 82 E.g. Pett, The autobiography of Phineas Pett. 83 Diederik Aten, ‘Amsterdamse gilden en regenten contra de Zaanse nijverheid, 1600- 1800’, in: Clé Lesger and Leo Noordegraaf (eds), Ondernemers en bestuurders: Economie en politiek in de Noordelijke Nederlanden in de Late Middeleeuwen en Vroegmoderne Tijd (Amsterdam: NEHA, 1999), 75-76; Unger, Dutch shipbuilding, 7. 84 Richard W. Unger, ‘Selling Dutch Ships in the Sixteenth Century’, Maritime History 3 (1973), 125-46. the technology and teaching of shipbuilding 197 builders over time, to some degree did mitigate the disincentive to innovate from a lack of competition. In Asia thin markets may have had a similar effect of restricting innovation though in some Chinese ports, even without government sponsorship some level of competition may have existed. The small number and small size of most firms involved in building ships also acted as a break on technical advance. Joseph Schumpeter began the discussion of the relationship of firm size, and indeed market structure in general, to technical change by claiming that free competition was a dis- incentive to research and development.85 He thought that monopolists with greater resources and the guarantee of rewards in the form of inflated profits would be more likely to make the effort to find new technologies and implement them. He implied that in a world of free competition, since any gains from superior products or processes would be wiped out by imi- tators, monopoly was advantageous in the long run for an economy.86 The increase in spending on research and development in the 1970s led econo- mists to a closer examination of the relationship between market structure and research and development. The recent rise in patenting in the United States of America has led to even more extensive examination of issues related to the character of the market and technical change.87 Discussion is centred on spending on research and development which, so the heroic assumption goes, is a measure of technical change. Research generates intellectual property, which in turn has greater value the less widely it is known or the less widely it can be used to produce anything to be sold. Research and development expenditures will continue, so the theory goes, until the costs exceed the anticipated returns from the intellectual prop- erty. Since monopoly implies greater profits then, by implication, the larg- er the market share of the firm the more likely it will be able to undertake ambitious research and development projects. With shipbuilding between 1300 and 1800 there is some indication that monopoly promoted larger scale research and development. It was the needs of institutions like the Venetian Arsenal or the Lung-chiang shipyard in Nanking that led to the creation of written works on shipbuilding. It was the minister of the navy in France who promoted the application of science to ship design and the training of naval architects. The Portuguese royal

85 E.g. Joseph Alois Schumpeter, Capitalism, socialism, and democracy (3rd ed. New York: Harper, 1950), 104-06. 86 F.M. Scherer, Industry structure, strategy, and public policy (New York: HarperCollins College Publishers, 1996), 14-15. 87 Nancy T. Gallini, ‘The Economics of Patents: Lessons from Recent U.S. Patent Reform’, Journal of Economic Perspectives 16 (2002), 131-32. 198 richard w. unger shipyards in Lisbon in the sixteenth century developed the largest wooden ships ever built, the giants for the trade to India.88 The shipyards of the Dutch East India Company, the largest business enterprise in the world, in the eighteenth century produced purpose-built and large cargo ships for long distance trade, developing the designs themselves. The argument for the advantages of size and of monopoly in generating technical advance falls down, however, both historically and theoretically. Though the Dutch VOC did create the East Indiaman, company wharves were known to be almost as conservative as naval wharves in matters of design. Complaints about failure to keep up with technical changes in ship- building were common in the eighteenth century Dutch Republic, com- plaints directed at both the offices of the VOC as well as the navy. The model was the British Navy for some but it, like its counterparts elsewhere in Europe, insisted on precise classification and standardization of all as- pects of their vessels for the obvious and easily justified reasons of making them, their parts and their crews more interchangeable. Navies were slow about inventing. In the eighteenth century navies did solve problems of building every bigger warships, a sign of some technical advance, but it could be they could only move in one direction. Improvements in warships at least through the seventeenth century seem to have been borrowed from private yards where work was done to suit buyers. Admiralties showed common characteristics of large firms where long chains of command and slow decision making lead to the suppression or simply the loss of innova- tive ideas.89 Theoretically, monopoly will generate more research and de- velopment but over the long term a monopolist may not maintain the pace of technical improvement. There is every reason for a producer, in the absence of competition, to continue to reap inflated profits from any tech- nical advance and become lackadaisical about further improvement. Mo- nopoly is far from the optimal form of organization to generate technical change. The ideal structure depends on many factors to do with the char- acter of the product, the market for the product, the number of firms, and the ability of inventors to appropriate returns from innovation.90 It is not certain that small firm size, which typified shipbuilding between 1300 and

88 Boxer, The Portuguese seaborne empire, 57. 89 F.M. Scherer, Industrial market structure and economic performance (Chicago: Rand McNally, 1970), 354. 90 Partha Dasgupta and Joseph Stiglitz, ‘Uncertainty, Industrial Structure, and the Speed of R&D’, The Bell Journal of Economics 11 (1980), 25-27; Scherer, Industry structure, 424. the technology and teaching of shipbuilding 199 1800 in Europe and apparently in Asia as well, was a disincentive to techni- cal change or to the advancement and dissemination of learning.

Incentives for Technical Advance: Guilds, Profits and Training

There were certain incentives that fed the drive toward better, more effi- cient and more durable ships. The buyers of ships were often knowledge- able, with future captains being among the partners who put up the capital. Merchants, many with experience of travel by sea, were another common component of buyers’ syndicates. That was certainly true in Chi- na. Such men would have at least understood the constraints and problems facing shipbuilders, as well as understood the advantages of any innova- tion. The written contracts between buyers and sellers in Europe, confined to governments in the Middle Ages but by the late sixteenth century con- cluded by shipbuilders with non-governmental customers, set few ­constraints on the technology builders could select.91 The contracts might specify the essential dimensions and perhaps include a few comments about the function of the ship and any unusual elements. Otherwise the builder was free to exercise his judgement. Undoubtedly there was consul- tation throughout the construction process but the limits on shipbuilders’ experimentation within a wide range were not set by the contract and, by implication, not by the owners. The market structure could also serve as an incentive to technical change. A number of small firms could engage in a variety of experiments, though typically small ones, generating micro-inventions one after the other. They would not, as with a monopolist, be capable of taking on big changes with a high risk of failure, but they might well generate consider- able cumulative gains over time through incremental improvements. The easy access to information about such incremental changes could serve as an asset as well. Intellectual property protection in shipbuilding, as with other crafts thanks in part to the presence of guilds,92 was virtually non- existent. While something like patents date from the close of the Middle Ages in Europe, and though governments did award exclusive use of some experimental methods by the seventeenth century, for shipbuilding in practice no method or design could remain the exclusive property of any- one for long. The movement of ships, of sailors and of shipbuilders made

91 John Dotson, ‘Jal’s Nef X’, The Mariner’s Mirror 59 (1973), 161-70. 92 Epstein and Prak, ‘Introduction’, 14. 200 richard w. unger it impossible to keep secrets, even if governments had tried to enforce secrecy, which they did not. Within guilds, but also across guilds and across time and space, there were no restrictions on the use of the ideas and practices of others. Within guilds information would have flowed even more easily. The absence of similar institutions elsewhere in the world did not necessarily stop the transfer of knowledge nor inhibit imitation of superior practice. The presence of guilds, however, which facilitated move- ment of skilled craftsmen and offered a forum for exchange of information among practitioners seemingly added another dimension to knowledge exchange. There was an incentive to develop improvements because of the poten- tial to create short term quasi-rents. There could be considerable econom- ic rewards to having a superior method and, given the riskiness of innovation, imitators were probably slow to respond, increasing the dura- tion and value of the quasi-rents. There was no monopoly in shipbuilding, not even in the construction of warships, down through the eighteenth century so supernormal profits of monopolies in possession of a superior technology could never be accrued. The market for ships could probably best be described as one of monopolistic competition because of the char- acter of supply and demand. Under that regime, while there are no super- normal profits, prices are not driven down to a point where consumers reap all the benefits from cost savings. That regime leaves quasi-rents above what would be expected under free competition for innovators to share, at least for a while.93 In fact it may be possible that the gains to producers in total are greater under monopolistic competition than under monopoly. With more firms the pace of innovation should be faster, thanks to com- petition, than with one or a very small number of firms.94 Conditions of monopolistic competition with a restricted number of players seem to have favoured trying new methods and designs. A little bit of monopoly, in the case of shipbuidling created by market conditions, was a good thing. Com- petitors could imitate and exploit any innovation. They decreased the pio- neer’s profits but did not destroy the entire rent generated by the novel ways of building or designing ships. In addition, a reputation for a better product could attract business not only to an individual but to a port. Even if other builders, whether members of a guild or not, followed best practice they, and in fact all producers and residents in the town would have ­benefited.

93 Scherer, Industry structure, 11-12. 94 Scherer, Industrial market structure, 365-68. the technology and teaching of shipbuilding 201 The ease, or lack of it, of transferring information does not necessarily affect the pace of innovation in a simple and straightforward way.95 Ship- builders faced various incentives and disincentives for handing on what they knew. By teaching they could create, through the apprenticeship sys- tem in guilds for example, a workforce for their own wharves and for the industry in general. Since shipbuilding was highly volatile with dramatic swings in demand guilds could also maintain a continuing supply of trained workers, available when necessary no matter the fluctuations in the mar- ket.96 For the individual shipcarpenter training made him mobile but mo- bility could be a disincentive for the master shipwright to train him. If the newly minted shipwright left town, then the master reaped no benefit from the investment in the newly skilled worker. If the successful trainee stayed, then he could even become the master’s competitor. For inventors retain- ing intellectual property as long as possible maximizes gain. They have an immediate and obvious disincentive to share new methods for fear of ero- sion of any quasi-rents from new designs. For shipbuilders the disincentive was mitigated by the fact that many of the things they taught were ge- neric skills for working with wood. It was also mitigated by the fact that most innovations generated small rents. A well-trained and skilled indi- vidual, despite the best of tutelage, was not likely to be able to drive a competitor, least of all an established one, out of business. Gains to the mystery through having more practitioners who could attract business and could promote sales through better ships probably made training appren- tices attractive to master builders. That assured the passing on of acquired knowledge and potential for invention. While it might be extreme to claim that, ‘Craft-based invention and the multi-centred, competitive institu- tional setting in which it was embedded came close to resembling an ideal market structure for innovation’ at least with shipbuilding guilds had a positive impact.97

Conclusions

Between the Middle Ages and the Industrial Revolution a number of ­forces worked to promote technical change in shipbuilding but deterrents existed as well. There were significant improvements in ships built, in the

95 Gallini, ‘The Economics of Patents’, 137. 96 Epstein, ‘Craft guilds, apprenticeship’, 61-62. 97 Ibid., 77. 202 richard w. unger equipment used to build them and the ways that shipcarpenters went about building them. The biggest breakthrough came in Europe at the end of the Middle Ages. After that in the European, the Chinese and the Malay building traditions, designers were locked into a general approach. Within the established traditions shipwrights made incremental and often valu- able advances. The overarching understanding of a good ship did not change among any of the major shipbuilding traditions. Europeans may have been more willing than others to adopt design features from other traditions than their counterparts in the rest of the world, but the differ- ence was marginal. As late as 1800 there was no mistaking a junk for a dhow for a sailing packet. No single approach even as late as 1700 was overwhelm- ingly superior. By 1800, though, improving efficiency of European ships as measured by ratios of tons served per man, and the ever increasing imita- tion of European designs by Asian shipwrights, indicates the gap in qual- ity was widening and that by that date European ships were definitely superior.98 Conditions were such in the period from 1300 to 1800 that small chang- es within established paradigms were much more likely than wholesale transformation. The exception was when Europeans exported their tech- nology and understanding of ship design and construction holus bolus and created in India or the islands of what is now Indonesia or in the Americas, a new tradition separate from that which existed before. The different ap- proaches could be integrated in some places and in ways that made strict distinctions questionable. In the eighteenth century the Portuguese built East Indiamen for the Indian Ocean trade in Brazil,99 and foremen on Am- sterdam shipbuilding wharves were called javanen suggesting they came from Java. The pace of innovation in shipbuilding was determined only in the late eighteenth century by the pace of advance in scientific knowledge. Before that technical change was driven in large part by the anticipated rewards of introducing new designs and new building methods. How much build- ers would benefit depended on the size of the market for their products. Sales in turn depended on the pace of population growth or decline, the pace of change in trade volumes, investment decisions, and the actions of governments among many other things. Sales of ships also depended on

98 Paul A. Van Dyke, ‘Operational Efficiencies and the Decline of the Chinese Junk Trade in the Eighteenth and Nineteenth Centuries: The Connection’, in: Richard Unger (ed.) Shipping Efficiency and Economic Growth 1350-1850 (Leiden: Brill, 2011), 231, 236. 99 Boxer, The Portuguese seaborne empire, 211. the technology and teaching of shipbuilding 203 the pace of technical advances in a host of related technologies such as navigation, ropemaking, sailmaking, metallurgy, wood sawing and, more generally, in shipping. The market for ships was generally thin with few buyers and even fewer sellers. Numbers grew slowly over time and even in the eighteenth century only in certain specific places was accumulation considerable. The product of a shipbuilding yard was lumpy. While inves- tors might distribute risk through part ownership, the ship itself could not be divided up. That constrained owners in how they operated the ships they bought and made them even more risk averse, that is if the high price of the capital good alone was not enough to promote fear of novelty. The number of firms varied over time and from place to place and the level of competition undoubtedly did have an influence on the pace of innovation throughout the world. The exact influence is somewhat uncertain. The price elasticity of demand for ships was presumably low and that decreased the rewards from innovation. Firm size, the lumpiness of the product and the slow pace of change in basic knowledge about the performance of solids in liquids combined with other factors to make small innovations, carefully adopted, the sensible choice around the globe through the eigh- teenth century. The precise role of guilds in the long term evolution of shipbuilding technology remains unclear. Certainly, given the success of European ship- building and the presence of shipcarpenters’ guilds in a number of prom- inent European ports, it would appear that they did make a positive contribution. Shipcarpenters’ guilds were typically less strict in their regu- lations of brothers and the operation of their businesses than guilds in some other trades; cloth and beer making are obvious examples.100 Guilds- men producing for retail consumption drew much greater attention from town governments, and so shipbuilders often enjoyed a relatively greater degree of freedom even if they were in guilds. Those institutions did not stop technical advance since in places where guilds prevailed there were improvements in ships. Guild rules did not prevent innovation. There are many cases of guildsmen embracing new technology.101 By setting down strict rules on aspects of competition, the guilds may have forced shipcar- penters to try to innovate in order to differentiate themselves from their brothers and so gain an advantage. In that way guilds could act like cartels,

100 E.g. Richard W. Unger, Beer in the Middle Ages and the Renaissance (Philadelphia: University of Pennsylvania Press, 2004), 207-233. 101 Epstein and Prak, ‘Introduction’, 18-19. 204 richard w. unger as they did in a number of other ways.102 Guilds did create conditions for training new shipcarpenters and did maintain a flow of supply adequate to meet the long term human capital needs of European shipping. Their strength may have been in training for regular work, for repair and the simpler aspects of construction, rather than for new building. It should be no surprise that the Amsterdam shipcarpenters guild pressed the town government through the seventeenth and eighteenth centuries to guaran- tee them repair work on ships visiting the port, but left matters to do with new construction to builders outside of the town, builders beyond guild regulation.103 There was always a large non-guild sector in shipbuilding and that meant there was always competition and there were always outlets for guild-trained personnel. Though even in areas where no guilds existed there might be less formal organizations, as in the Zaanstreek outside Amsterdam,104 which offered a loose structure to the industry allowing for experiment and for relatively quick diffusion of information. Technology, both the prevailing methods and new ones, were taught and diffused through Europe and beyond in the sixteenth and more so in the seventeenth and more so in the eighteenth century. In sum, the results for shipbuilding were positive. The contribution of guilds may have been modest in some parts of Europe and some parts of the world, but they were still useful.105 The institutions for promoting the technology and teaching of shipbuilding had a net positive effect in the centuries before science separated Europe from the rest of the world and science separated the ship designer from the shipbuilders. Naval architects became independent pro- fessionals with their own training and professional organizations but it was not until 1860, though, that Britain got its Royal Institution of Naval Archi- tects. Even up to the mid nineteenth century in Europe, and later in the rest of the world, the handicraft tradition of training and innovation con- tinued to supply shipping with its needed tools.

102 Gunnar Mickwitz, Die Kartellfunktionen der Zünfte und ihre Bedeutung bei der Ent- stehung des Zunftwesens: Eine Studie in spätantiker und mittelalterlicher Wirtschaftsgeschichte (Amsterdam: Hakkert, 1968; orig. 1936); Unger, Dutch shipbuilding. 103 Aten, ‘Amsterdamse gilden’, 67-76. 104 Unger, Dutch shipbuilding, 84-85. 105 Karel Davids, ‘Guilds, guildsmen and technological innovation in early modern Europe: the case of the Dutch Republic’ unpublished paper International Institute of Social History, 2003 [accessed March 19, 2013]. Available from http://lowcountries.nl/papers/2003-2_ davids.pdf. moving machine-makers 205

CHAPTER SIX

MOVING MACHINE-MAKERS: CIRCULATION OF KNOWLEDGE ON MACHINE-BUILDING IN CHINA AND EUROPE BETWEEN C. 1400 AND THE EARLY NINETEENTH CENTURY

Karel Davids

Introduction

Machine-makers are among the most elusive craftsmen in history. Ma- chines were everywhere, but the people who built them are still barely known. Who were the makers of the mills, looms, pumps or windlasses that abounded even in pre-industrial times? Where did they come from? How did they acquire their skills? How did they manage to transmit their ac- cumulated knowledge to others? And more specifically: Were there any significant differences in this respect between machine-makers in different places and periods? None of these questions can be easily answered, be- cause machine-makers as a group in the pre-industrial era have scarcely been studied, except of course for the well-known body of Renaissance engineers.1 Despite this lack of comparative and in-depth investigations, Joseph Needham in volume 4 of Science and Civilisation in China, which exten- sively discusses mechanical engineering in China, not only claimed that ‘the building of mill-work and water-raising machinery of all kinds was spread throughout the length and breadth of the (Chinese) empire’ , but also maintained that China had artisans who ‘in skill and ingenuity (were) at least as eminent’ as the millwrights and engineers in Renaissance Europe and that the presence of such machine-makers was ‘therefore not enough’ to explain subsequent differences in the development in the science and technology.2

1 See e.g. William Barclay Parsons, Engineers and engineering in the Renaissance (Cambridge Mass., 1939) Bertrand Gille, Les ingénieurs de la Renaissance (Paris, 1964), Paolo Galuzzi, Renaissance engineers from Brunelleschi to Leonardo da Vinci (Florence, 1996). 2 Joseph Needham, Science and civilisation in China, vol. IV, Physics and physical technology Part 2 Mechanical engineering (Cambridge, 1965), 545-46 206 karel davids While the last conclusion will hardly be disputed by anyone, this still leaves open the question whether the observed equivalence in skill and ingenuity between individual artisans also implies similarity in patterns of circulation of technical knowledge. Does it mean that there was no relevant difference at all in the ways machine-makers in China and Europe acquired and transmitted their skills? This question, focused on a comparison of China and Europe between the late fourteenth century and the beginning of the Industrial Revolution, will be the subject of this essay. The first two sections will look into the patterns of circulation of knowledge in machine- making in China and Europe, and the similarities and differences that can be observed. ‘Circulation of knowledge’ on machine-making concerns the whole gamut of movements and exchanges of knowledge on the construc- tion of machines, both in a spatial sense (whether on an intercontinental or a local scale) and in the sense of acquisition and transmission through time in a region, a city, a workshop or from one craft to another.3 In the third section I will discuss the possible implications of these findings for understanding the evolution of the knowledge economy in China and Eu- rope before the nineteenth century and the early stages of the Industrial Revolution.

Circulation of Knowledge on Machine-making in China, Late Fourteenth – Early Nineteenth Centuries

Machines were numerous in China long before the Industrial Revolution. Following Lewis Mumford’s classification, Needham describes a wide range of what he calls ‘eotechnic machinery’—wooden or bamboo devices with a number of (often revolving or spinning) components to carry out certain productive functions, powered by animals, humans or water, as they were once in use in China. Among these machines were querns, animal-powered grinding mills, hand-driven chain pumps, scoop-wheels moved by humans or animals for lifting water onto fields, winches for mine-shafts, human- driven winnowing-fans, water-powered bellows, water-driven multiple spinning frames for hemp and ramie and vertical double-roller sugar-mills powered by animals.4 The types and uses of machinery were not static, as

3 Liliane Hilaire-Pérez and Cathérine Verna, ‘Les circulations techniques: hommes, produits, savoirs au Moyen Âge et à l’époque moderne (Orient, Occident)’, in: Michel Cotte (ed.), Circulations techniques en amont de l’innovation: hommes, objets et idées en mouvement (Belfort, 2004), 11-35, esp. 11-12. 4 Needham, Science and civilisation, vol. IV 2, passim, esp. table 56. moving machine-makers 207 studies on separate sectors of the economy show. The water-driven mul- tiple spinning frames, for example, seem to have gone out of use in the fourteenth century.5 During the Qing, water-power for moving the bellows in blast furnaces in the iron industry was often replaced by human labour.6 How did the makers of these and other machines acquire their skills? Machine-building may have been a household activity or sometimes a sideline of carpentry or the blacksmith’s trade. Dagmar Schäfer suggests that looms in the silk industry were either constructed or repaired in the households of weavers or ‘by loosely attached but specialized carpentries’.7 Special terms for artisan-engineers, such as jiang and kong existed at an early date.8 Insofar as the construction of machines indeed did require a certain level of specialization, the most obvious possibility is that machine- makers learned their craft on the shop floor, in the framework of a sort of apprenticeship relation. Apprenticeship was after all a common institution for the transmission of knowledge in crafts and trades.9 If the art of ma- chine making was indeed learned in the context of an apprenticeship rela- tion, this could mean that it was subject to regulation by a guild. The guilds that emerged in China from the late sixteenth century onwards mostly required the completion of an apprenticeship period (usually three years) as a condition for entry.10 If guild-regulated apprenticeship requirements relating to machine- building really existed, it is doubtful, however, whether they had much impact on the acquisition of skills by machine makers. The length of the apprenticeship term, as Peter Golas has pointed out, normally did not cor- respond with the actual time needed to learn a particular craft or trade. It was ‘primarily a custom-sanctioned initiation … before one could enter a trade’.11 Moreover, in Chinese guilds formal examinations at the end of

5 Needham, Science and civilisation, vol. V part 9, Textile technology, spinning and reeling (Cambridge, 1988) 236. 6 Donald B. Wagner, Iron and steel in ancient China (Leiden, 1993); idem, The tradi- tional Chinese iron industry and its modern fate (Richmond, 1997), 31, 45. 7 Dagmar Schäfer, ‘Silken strands: Making technology work in China’, in: Dagmar Schäfer (ed.), Cultures of knowledge. Technology in Chinese history (Leiden, 2012), 45-73, esp.66-67. 8 Needham, Science and civilisation, IV, 2, 9-10. 9 Peter J. Golas, ‘Early Ch’ing guilds’, in: G.William Skinner (ed.), The city in Late Impe- rial China (Stanford, 1977), 555-80, esp. 565-567. 10 Golas, ‘Early Ch’ing guilds’; Christine Moll-Murata, ‘Chinese guilds from the seven- teenth to the twentieth century: An overview’, in: Jan Lucassen, Tine de Moor and Jan Luiten van Zanden (eds), The return of the guilds, Supplement 16 International Review of Social History (2008), 213-47, esp. 223. 11 Golas, ‘Early Ch’ing guilds’, 566. 208 karel davids the apprenticeship period were not required. The level of expertise to be reached at the completion of the apprenticeship was thus left undefined. And last but not least, it is not an established fact that machine-building in Late Imperial China was a guild-regulated activity at all. Machine-mak- ers living in the countryside cannot have been organized in guilds, as these institutions were typically confined to cities.12 Even in cities, guilds were not a ubiquitous phenomenon. The number of guilds founded before the end of the eighteenth century was quite small. A significant expansion did not start until the middle of the nineteenth century and the phenomenon did not spread equally across the entire empire. While most cities in the Yangzi basin and the coastal area from Guangdong to Shandong saw a substantial increase in guilds, inland regions, frontier areas and a few plac- es along the Yangzi or near the coast remained more or less ‘guild-free zones’.13 Within cities, not every occupation was subsumed in a guild-like organization. Guilds of carpenters and joiners certainly existed in a number of places in the eighteenth and nineteenth centuries.14 Whether these guilds included machine-makers too, is still impossible to tell. The safest conclusion is perhaps, that machine-making in China, in so far as it became a specialized art, normally was learned by way of an informal, custom- regulated apprenticeship. Informal apprenticeship does not necessarily imply that knowledge is purely tacit and skills are restricted to a particular locality. Knowledge and skills acquired through informal apprenticeship can still to some extent be codified and be carried from one place to another through mobility of humans or otherwise. In the case of machine-makers in Late Imperial China, there are several conceivable ways in which this circulation of knowledge and skills can have occurred. Some of these were embedded in formal structures, others were not. I would argue, that circulation embed- ded in formal structures in reality fell short of what theoretically might have happened. Knowledge and skills about machine-making in these cir- cuits probably moved only to a limited extent and not in a very continuous, constant manner. The first way in which circulation in a formal context may have taken place has been suggested by Klaas Ruitenbeek in his study on carpentry

12 Klaas Ruitenbeek, Carpentry and building in Late Imperial China. A study of the fif- teenth-century carpenter’s manual Lu Ban jing (Leiden, 1993), 18. 13 Moll-Murata, ‘Chinese guilds’, 221-22, 243-44. 14 As documented by the database on Chinese guilds constructed by Christine Moll- Murata. I thank Dr. Moll-Murata for kindly permitting me to consult this database. moving machine-makers 209 and the building trade in Late Imperial China. Ruitenbeek draws attention to the potential relevance of the corvée system, which was (re)introduced in the early Ming. Under this system, various categories of producers were obliged to render regular labour service to the state. The system was based on an elaborate system of registration, combined with tight rules that re- stricted change of occupation and set narrow bounds to physical mobility. Farmers, soldiers, merchants or artisans were not allowed to switch to an- other profession and people could not travel long distances without official permission. According to an occupational census taken in the early Ming, the empire then numbered nearly 300,000 artisans. These artisans were divided into two groups: ‘resident artisans’, who lived in or near the capital and had to report for duty in the palace workshops ten days each month; and ‘shift artisans’, who lived and worked everywhere in the country and had to render service during a number of months in the capital once every few years (fixed at four years in 1445). Only at those occasions were they allowed to travel further than the normal limit of 100 li (58 kilometres) from their place of residence. 15 Shift artisans made up eighty per cent of all registered artisans. Of this category of artisans, one-half were active in the building trade, including, for instance, tens of thousands of carpenters and thousands of sawyers, masons and bricklayers.16 Ruitenbeek suggests that ‘for a carpenter living in a remote corner of the empire it must have been quite an experience to travel to the capital and meet with fellow craftsmen from other provinces. To him, it may have meant something comparable to the tour de France des compagnons or the Wanderjahre of the artisans in Europe’.17 If this were true, and if artisans in the building trade (notably carpenters) were also engaged in the con- struction of machines, this would mean that knowledge and skills on machine-making could indeed circulate constantly and widely across the entire empire. It is doubtful, however, whether the corvée system really has fulfilled this function of state-sponsored equivalent to the practice of tramping. Ruitenbeek himself admits that ‘it remains unclear … to what extent, and to what distance from the capital, the artisans of the empire were affected by the corvée system’.18 It is thus an intriguing hypothesis that for the time

15 H. Friese, Das Dienstleistungs-System der Ming-Zeit (1368-1644) (Hamburg, 1959), 117-30; Ruitenbeek, Carpentry, 16; Timothy Brook, The confusions of pleasure: Commerce and culture in Ming China (Berkeley, 1998), 19. 16 Ruitenbeek, Carpentry, 16-17. 17 Ibid., 18. 18 Ibid. 210 karel davids being remains unproved. Moreover, the system already began to lose its edge in the second half of the fifteenth century and it disappeared alto- gether in the mid-sixteenth century, as far as artisans were concerned. Since 1562, shift artisans could buy off their labour service with a money pay- ment.19 If the system of obligatory labour declined, the regular pattern of travelling to the capital must have lost its rationale as well. And this reduced the opportunity for the circulation of technical knowledge through human mobility. The guild system, which by and large arose after the corvée system had gone into decline,20 could have provided another formal channel for the circulation of knowledge and skills across the country. Not by a tramping system (which does not appear to have existed in Late Imperial China), but by virtue of the fact that guilds were often composed of people from the same geographic area. The frequent requirement that members of a guild have a common geographic origin has been highlighted as a striking dif- ference between Chinese and European guilds.21 Peter Golas mentions the example of the Leather Box guild in Beijing, founded in 1689, which as late as the 1940s still consisted exclusively of people originating from Shan- dong.22 The principle of common geographical origin could imply that knowledge and skills in particular occupations travelled indeed a very long way. It is not clear, however, to what extent this also applied to craftsmen engaged in machine-building. Besides, the principle of geographical ho- mogeneity of guilds was no more invariable than the system of obligatory labour. Common geographical origin tended to become less important as a basis for organization from the eighteenth century onwards.23 As a con- sequence, knowledge and skills moved perhaps across a shorter distance than before. The demise of the corvée system and the progressive decline of the principle of common geographical origins in guilds did not imply that long-distance mobility of craftsmen, including machine-makers, came to a halt. Circulation of knowledge through mobility of people could take place outside such formal structures. An example of this was the migration of Chinese sugar-mill workers to the environs of Dutch Batavia, Java, in the late seventeenth and early eighteenth centuries. About 1710, there were

19 Friese, Dienstleistungs-System, 128-29. 20 As Moll-Murata, ‘Chinese guilds’, 219 points out. 21 Golas, ‘Early Ch’ ing guilds’, 563; Moll-Murata, ‘Chinese guilds’, 220, 240. 22 Golas, ‘Early Ch’ ing guilds’, 564. 23 Ibid., 558, 564-65; William T. Rowe, Hankow. Commerce and society in a Chinese city, 1796-1889 (Stanford, 1984), 252-88. moving machine-makers 211 some 130 sugar plantations in operation in this area, mostly managed by Chinese entrepreneurs.24 The technology for processing sugar cane, includ- ing the use of animal-driven mills, was imported from China.25 The spread of these mills suggests that at least some machine-makers must have im- migrated from China as well. Like most Chinese living in Batavia26, they probably originated from the hinterland of Amoy. This migration move- ment was not organized by any formal institution either at the Chinese or at the Dutch end of the chain. However, how many skilled people, includ- ing machine-makers, moved around in this way in China or outlying areas of Chinese settlement between the fifteenth and nineteenth centuries, is as yet impossible to tell. Circulation of technical knowledge is not necessarily dependent on the mobility of people. Knowledge can also travel through space or time in codified form, for example embodied in technological literature. In his study on technological development in agriculture in China, Kent Deng attached special importance to the availability of agricultural books. Trea- tises on agricultural subjects (nongshu) were written in sizeable numbers throughout Chinese history, especially during the Song, Ming and Qing dynasties. Under these dynasties, total nongshu production reached an estimated 105, 127 and 199 titles, respectively. Nongshu were not only im- portant for the accumulation, standardisation and dissemination of knowl- edge on agriculture as such, Deng has argued, but they also ‘functioned as a means to bridge technological gaps’. Without these books, he claimed, agrarian recovery after repeated natural or man-made disasters would have been much more difficult (even though not all nongshu have been pre- served). ‘In this sense, Chinese technological literature, of which the nongshu formed an important part, contributed greatly to the survival of the Chinese agriculture-based society’, he concluded.27 Were books equally important in the domain of machine-building? This is not very likely. Written or printed books describing the construction of machines have been much more rare than treatises on topics relevant to agriculture. Machine-making was sometimes dealt with in general ­manuals

24 Leonard Blussé, Strange company: Chinese settlers, Mestizo women and the Dutch in VOC Batavia (Dordrecht, 1986), 90-91. 25 Margaret Leidelmeier, Van suikermolen tot grootbedrijf: Technische vernieuwing in de Java-suikerindustrie in de negentiende eeuw (Amsterdam, 1997), 18-21. 26 Blussé, Strange company, 147. 27 Gang Deng, Development versus stagnation. Technological continuity and agricultural progress in premodern China (Westport, Conn., 1993), xxii, 3-5, 176. 212 karel davids on carpentry or books on agriculture. Examples from the late Song, Yuan and early Ming are ‘Traditions of the Joiner’s Craft’ (Ziren yizhi) by Xue Jingshi, dated c.1260, the Nongshu by Wang Zhen, dated c.1304 and the Lu Jang Bin, compiled in the fifteenth century. The Ziren yizhi contained, among other things, a description of the construction of looms, including illustrations and lists of parts and measurements. Wang Zhen’s Nongshu gave descriptions and illustrations of water-powered bellows, winnowing- machines, irrigation equipment and other mechanical devices. The Lu Jang Bin provided explanations (plus lists of parts and measurements) of the making of a hand- or foot-driven irrigation wheels.28 The Ziren yizhi has not been preserved as a separate work but only as part of an encyclopedia composed at a later date. The other manuals remained in circulation for a long time. The Lu Jang Bin was still on sale in bookshops in China in the early twentieth century.29 In the last decades of the Ming dynasty new books dealing with ma- chines and machine-making appeared somewhat more frequently than before. The best known of this cluster of writings is the Tiangong kaiwu (Heavenly crafts revealing the use of things), compiled by Song Yingxing in the 1630s. The Tiangong kaiwu, which Needham calls ‘China’s greatest technological classic’,30 presented an overview of techniques in eighteen different fields, ranging from agricultural technology, textile making, salt making and sugar making, via ceramics, iron working, paper making and the building of boats and carts to weapon manufacture and the making of beverages. These descriptions were accompanied by dozens of illustrations of mills, wheels, looms, kilns and technical processes. Although the book did not discuss in detail the construction of machines themselves, it gave at least some idea of what they looked like.31 Machines were also depict- ed in some encyclopedias for daily use, such as the agricultural compen- dium Nongzhen Quanshu and Wang Zheng’s Explanations and diagrams of various machines, which presented his improvements of traditional agri-

28 Ruitenbeek, Carpentry, 31-32, 270-76; Deng, Development versus stagnation, 42; Needham, Science and Civilisation, IV 2, e.g. 153-154, 204, 371-73. 29 Ruitenbeek, Carpentry, 1, 32; Deng, Development versus stagnation, 38-39. 30 Needham, Science and civilisation, IV 2, 171. 31 Needham, Science and civilisation, IV 2.; Zen-Sun, E-tu and Shiou-Chuan Su (tr.), Sung Ying-Hsing: Chinese technology in the seventeenth century (Dover, 1966); Dagmar Schäfer, ‘The congruence of knowledge and action: The Tiangong kaiwu and its author Song Yingxing’, in: Christine Moll-Murata, Song Jianze, Hans Ulrich Vogel (eds), Chinese handicraft regulations of the Qing dynasty (Munich, 2005), 35-60; Benjamin A. Elman, On their own terms. Science in China, 1550-1900 (Cambridge Mass., 2005), 122. moving machine-makers 213 cultural equipment.32 A novel category were books derived from European examples. Jesuit missionaries and their first Chinese converts produced dozens of writings that built directly on technological literature composed in Europe. Sabbatino de Ursis and Xu Quangqi, for example, published Western techniques of hydraulics (1612), Johannes Schreck and Wang Zheng compiled Diagrams and explanations of the marvelous devices of the Far West (1627) and Adam Schall von Bell with Chinese assistants made a trans- lation of parts of Georg Agricola’s De re metallica (1638-1640), which has not been preserved.33 Many of the illustrations and descriptions contained in the first two works were copied from European books on machines and mechanics published between the fifteenth and early seventeenth centu- ries, notably those by Agricola, Besson, Ramelli, Stevin, Verantius, Zeising and Zonca.34 Except the translation of Agricola that appears to have been lost during the transition to the Qing,35 none of these works from the last decades of the Ming dropped completely out of circulation in China since. But their survival mainly took place in an indirect way. Parts of the works produced by the Jesuits and their Chinese collaborators (including many illustra- tions) were included in dynastic encyclopedias composed in the eighteenth century. They did not circulate as separate books. About three-quarters of the original text of the Tiangong kaiwu likewise was incorporated into an encyclopedia, the Gujin tushu jicheng (Complete collection of pictures and books of old and new times), published in 1728 in 64 ‘officially sanctioned’ copies.36 As a separate entity, Dagmar Schäfer concluded, the book prob- ably ‘was just preserved over a relatively short period of approximately forty years and then totally disappeared until the nineteenth century’. 37 It is therefore not very likely that many machine-makers in China ever saw, or consulted, a copy of the cluster of books on machines published during

32 Needham, Science and civilisation, IV 2, 170-71; Peter M. Engelfriet, Euclid in China. The genesis of the first Chinese translation of Euclid’s Elements books I-VI (Jihe yanben; Beijing 1627) and its reception up to 1723 (Leiden, 1998), 340-41; Elman, On their own terms, 110. 33 Pan Jixing, Hans-Ulrich Vogel and Elisabeth Theisen-Vogel, ‘Die Übersetzung und Verbreitung von Georgius Agricolas “De Re Metallica” im China der späten Ming-Zeit (1368- 1644)’, Journal of the Economic and Social History of the Orient, 32 (1989) 153-202, esp. 171, 191-92; Engelfriet, Euclid in China , 327-28; Elman, On their own terms, 110. 34 F. Jäger, ‘Das Buch von den wunderbaren Maschinen: Ein Kapitel aus der Geschichte der Abendländisch-Chinesischen Kulturbeziehungen’, Asia Major 1.1. (New Series) (1944), 78-96; Engelfriet, Euclid in China , 338-40; Needham, Science and Civilisation, IV 2, 211-18. 35 Vogel, ‘Übersetzung’, 190. 36 Schäfer, ‘Congruence’, 42. 37 Ibid., 44. 214 karel davids the Late Ming. In contradistinction to the domain of agriculture, a special- ist body of literature on machine-making did not develop before the nine- teenth century.38 In this field, circulation of technical knowledge through writings can not have been a very significant complement to circulation through mobility of people.

Circulation of Knowledge on Machine-making in Europe, Late Fourteenth – Early Nineteenth Centuries

The use of machines in Europe between the late fourteenth and early nine- teenth centuries appears to have been at least as varied as in China. All the types of machines and mechanical devices described in Science and civili- sation in China, volume 4, can also be found in late medieval and early modern Europe. This does not mean, of course, that Chinese and Euro- pean societies were ‘mechanized’ to the same degree or in the same way, nor that they produced goods of the same quality or at the same level of productivity. There were all kinds of differences, to be sure, depending in part on the relative use of particular raw materials, processes, mechanical principles or sources of energy. Benjamin Elman, for example, has re- marked that39 arguably, by 1600 Europe was ahead of Asia in producing basic machines such as clocks, screws, levers and pulleys that would be applied increasingly to the mechanization of agricultural and industrial production. In the sev- enteenth and eighteenth centuries, however, Europeans still sought the technological secrets for silk production, textile weaving, porcelain making, and large-scale tea production from the Chinese. Regarding inanimate sources of energy, shifts back from water-power to human power do not seem to have occurred in Europe, but there were changes in the relative frequency of water, wind and horses as sources of power40 and an extension of the application of wind-power to a much greater degree than in China. From the sixteenth century onwards, wind- power was harnessed not only for grinding corn or raising water (as in coastal areas of China)41 but also for all sorts of industrial processes, such as fulling, oil pressing, hemp crushing, timber sawing, hulling, cannon-

38 Cf. Schäfer, ‘Silken strands’, 73. 39 Elman, On their own terms, xxxi. 40 See for example John Langdon, Mills in the medieval economy 1300-1540 (Oxford, 2004), 34-40. 41 Needham, Science and civilisation, IV 2, 558 moving machine-makers 215 boring, grinding mortar or powder and cutting rags as raw material for paper. Northwestern France, Flanders and, especially, Holland were the centres of innovation of these new applications of wind-power. By the beginning of the nineteenth century, tens of thousands of windmills for all sorts of purposes were in operation across Europe.42 Yet, such variations do not alter the basic fact that China and Europe resembled each other in the extensive use of machines. As in China, machine-making in Europe was to some degree a special- ized activity. Experts in the construction of machines could be described with titles like artifex, ‘engineer’ (ingeniator, ingegnere, ingeniero, Werkmeister),43 ‘millwright’ or ‘wheelwright’ but could also be called by general names such as ‘carpenter’ or more specific ones such as ‘pump- maker’, ‘loom-maker’, or ‘clockmaker’. As in China, machine-making in Europe was probably normally learnt in the setting of a kind of apprentice- ship relation. Apprenticeship could be regulated by craft guilds. Guilds in Europe were in all probability of greater importance for the training of craftsmen—including machine-makers—than in China, if only because they emerged earlier and became much more widespread. Guild regulation of apprenticeship of machine-makers was even in Eu- rope by no means the general rule, however. Engineers or millwrights were after all not always or everywhere organized in guilds. In the Dutch Repub- lic, for example, loom-makers in Haarlem and millwrights in Rotterdam and Utrecht and belonged to the same guild as carpenters and pump-mak- ers in Haarlem were incorporated into the guild of St.Luke,44 but in the heartland of mill construction, the Zaanstreek, millwrights remained com- pletely outside the guild-system. The Zaanstreek was an almost entirely ‘guild-free zone’. Craft guilds were absent in many industrial towns in Eng- land, too. An alternative for regulation by guilds was regulation by the state. This was the prevailing system in England from the mid-sixteenth century

42 Terry S. Reynolds, Stronger than a hundred men: A history of the vertical water wheel (Baltimore, 1983), passim; Karel Davids, ‘Innovations in windmill technology in Europe, 1500-1800’, Atti XXXIV Settimana di Studi Economia ed energia, secc.XIII-XVIII ‘Economia e energia secc. XIII-XVIII’ (Prato, 2003), 271-91. 43 See for the various names e.g. Gerhard Dohrn-Van Rossum, ‘Migration—Innova- tion—Städtenetze. Ingenieure und technische Experten’, in: Matthieu Arnoux and Pierre Monnet (eds), Le technicien dans la cité en Europe Occidentale 1250-1650 (Rome, 2004), 291- 307. 44 C. Wiskerke, De afschaffing der gilden in Nederland (Amsterdam, 1938), 79 ; H.C. Hazewinkel, Geschiedenis van Rotterdam (Zaltbommel, 1974²) III, 846 ; Karel Davids, ‘Guilds, guildsmen and technological innovation in early modern Europe: the case of the Dutch Republic’ (Working paper 2003, www.lowcountries.nl/workingpapers.php). 216 karel davids onwards. The Statute of Artificers, introduced in 1563 and in force until 1814, laid down rules regarding apprenticeship for ‘every craft, mystery, or oc- cupation’ that applied to both the cities and the countryside of the English realm. According to this law, the term of apprenticeship in all sectors was seven years and the maximum number of apprentices per master was three. An apprentice was bound by indenture to a craftsman, who was obliged to teach him his trade. Once an apprentice had served out his con- tract, he was allowed to practice his craft anywhere in the country. These regulations were for a long time very effectively enforced.45 However, unregulated apprenticeship in machine-making was certain- ly not unusual. The Zaanstreek was a case in point. A millwright Pieter Johannes de Vries from Zaandam, for example, who in 1769 applied for a job as master millwright in the service of the Dutch East-India Company, established his credentials not by reference to a formal proof supervised by a guild, but by the presentation of a CV stating that ‘from childhood he had been taught by his uncle Jan Douwes de Vries the art of drawing and the making of all sorts of mills and gear of mills’ and that in the meantime ‘he had made such progress that for a few years past he had built with suc- cess a number of mills in the Netherlands as a master millwright himself (lastly in Flushing in 1768)’.46 A survey conducted in 1800 reveals that the villages along the Zaan then numbered ten millwright workshops, each of which, apart from the master millwright, employed 20 to 30 adult craftsmen and five boys, presumably apprentices.47 Within craft guilds, regulations on apprenticeship were seldom very specific as far as the contents of the training were concerned. By-laws rarely spelled out in detail what exactly apprentices were supposed to learn or how their competence should be measured at the end of their term. A masterpiece was sometimes required but the description of the proof was more often than not left somewhat vague. The guild of St. Luke in Haarlem, for example, introduced a separate test for pump makers in 1685.48 The by-law of the carpenters’ guild in 1711 described an exam for loom-makers

45 Margaret Gay Davies, The enforcement of English apprenticeship. A study in applied mercantilism 1563-1642 (Cambridge Mass., 1956); Jane Humphries, ‘English apprenticeship. A neglected factor in the First Industrial Revolution’, in: Paul A. David and Mark Thomas (eds), The economic future in historical perspective (Oxford, 2003), 73-102. 46 National Archives The Hague, Archives Verenigde Oostindische Compagnie (VOC) 277 27 April 1769. 47 A.M. Van der Woude (ed.), ‘De Goldberg-enquête in het Departement van Texel, 1801’, AAG Bijdragen 18 (1973), 95- 250, pp. 132, 136, 151, 159, 170, 179, 202, 218. 48 Regional Archive Kennemerland, Stadsarchief Haarlem 1581-1795 rood 60 G f.193v-200, Keuren en Ordonnantien 9 nr.16, 10 (1). moving machine-makers 217 but it did not give any specifications. The only requirement was that the item should be made in such as way that it could ‘pass the examination’.49 If skills acquisition through apprenticeship was specified, this was nor- mally done by means of an individual oral or written contract between a child’s family (or guardian) and a prospective master. Guilds normally did not intervene in the conditions, registration or supervision of these con- tracts. Written apprenticeship contracts can be found in large numbers in many countries in Europe from the Late Middle Ages onwards. These were also known in crafts related to machine-making.50 Sometimes skills acqui- sition and transmission was regulated by a collective contract. An example is an agreement concluded in 1678 for the duration of six years between the city government of Haarlem and six local makers of ribbon frames. The agreement in essence provided that the makers of ribbon frames would teach their art only to their own children and apprentices, that they would not emigrate from Haarlem or repair any frames outside the boundaries of the city (and would also hold back their pupils from doing so) and that they would duly inform the urban authorities about all frames (or parts thereof) sold to local customers, while the magistrates undertook to pay the six makers 900 guilders, to restrict new admissions to their craft and to keep a close watch on all extant ribbon frames to prevent any equipment from leaving the city. 51 Formal structures for apprenticeship in the form of guild or state regu- lations or written contracts thus appear to have been more common in Europe than in China, but this does not automatically imply that skill ac- quisition in machine-making went faster, more efficiently or led to a high- er level of technical accomplishment. More significant, perhaps, were differences in the actual extent of mobility of people and the degree to which knowledge on machine-making was embedded in codified forms. S.R. Epstein’s work on labour mobility and the circulation of technical knowledge can serve here as a heuristic model.

49 Regional Archive Kennemerland, Gildenarchieven Haarlem, nr. 315. 50 Bert de Munck, Steven L. Kaplan and Hugo Soly, ‘“Learning on the shop floor” in historical perspective’, in: idem (eds), Learning on the shop floor: Historical perspectives on apprenticeship (New York/London, 2007), 3-32; Karel Davids, ‘Apprenticeship and guild control in the Netherlands, c.1450-1800’, in: ibid., 65-84; and Davids, ‘Guilds, guildsmen’; Bert de Munck, Technologies of learning: Apprenticeship in Antwerp guilds from the 15th century to the end of the Ancien Régime (Turnhout, 2007), 41-49; Patrick Wallis, ‘Apprentice- ship and training in pre-moden England’, Journal of Economic History, 68 (2008), 832-61. 51 Regional Archive Kennemerland, Stadsarchief Haarlem, Loketkast 7-15-7- 1 contract 18 November 1678. 218 karel davids In several studies Epstein presented the following argument: (1) only migration by trained craftsmen was ‘likely to have had a significant impact’ on ‘technical diffusion’ in Europe in the Late Middle Ages and the Early Modern Period, given the fact that ‘mass schooling’ and ‘the harnessing of science to industry were unknown’ and ‘most technical knowledge was tacit and embodied in its practitioners.’ (2) Transitory migration in the form of journeymen tramping (Wanderschaft, compagnonnage) was more im- portant than permanent migration by master artisans. (3) Informal tramp- ing was numerically much more significant than tramping under the aegis of formally organized journeymen’s associations. (4) And finally, this regu- lar pattern of journeymen travelling was instrumental in the formation of regional ‘pools’ or ‘clusters’ with shared technical knowledge and skills, which did not extend across Europe as a whole, but largely coincided with ‘areas that were institutionally and culturally more homogeneous’, such as the German lands.52 Machine-makers in Europe partly conformed to this model. Both the similarities and variations tell us something about the differences between the circulation of technical knowledge in Europe and China. Among ma- chine-makers, too, spatial mobility was a frequent occurrence and it nor- mally assumed the form of temporary movements rather than permanent migration. These temporary movements mostly took place outside the context of formal organizations such as guilds or journeymen’s associa- tions, but they were often covered by formal arrangements like written contracts. In contrast with Epstein’s model, however, migration by master craftsmen was not less important than migration by journeymen, technical knowledge also circulated via other carriers than the machine-builders themselves, and the circulation of knowledge on machine-making was not really restricted to institutionally and culturally homogeneous areas. Knowledge on machine-making could easily cross regional, national and even continental boundaries. The high degree and broad range of spatial mobility of machine-makers as well as the extent to which technical knowl- edge could circulate in other ways than through the actual movement of people makes the pattern of circulation in Europe distinct from that in China. The spatial mobility of machine-makers showed some degree of varia- tion through time and by category of maker. The ingeniatores, or ‘engineers’,

52 S.R. Epstein, ‘Journeymen mobility and the circulation of technical knowledge, XVI- XVIIth centuries’, in: Liliane Hilaire-Pérez and Anne-Françoise Garçon (eds), Les chemins de la nouveauté. Innover, inventer au regard de l’histoire (Paris, 2003), 411-30 (and reprinted in: Arnoux and Monnet (eds), Le technicien dans la cité, 251-69). moving machine-makers 219 were almost from their very first appearance in eleventh-century sources a highly mobile group. They travelled long distances all over Europe to hire out their services to territorial princes, church authorities, urban govern- ments and, finally, states, for the construction of building equipment, siege engines, industrial mills, clocks, water-lifting devices and numerous other ingenious contraptions. From the fifteenth century onwards, engineers, especially in Italy and Germany, frequently composed ‘machine-books’— i.e. written or printed treatises, illustrated with drawings and other pictures—to advertise their skills, expertise and ingenuity.53 In sixteenth- century Spain fifteen per cent of all engineers were Italian by origin; an- other ten per cent were German, Flemish, French or English.54 Although it is true that the vast majority of engineers were Spanish, as García Tapia rightly stresses, the ‘pool’ of knowledge on machine-making evidently did not coincide with an institutionally or culturally homogeneous area. Many millwrights operated much closer to home, however. Millwrights and car- penters contracted for the construction of watermills or windmills in Eng- land between c. 1300 and 1540 mostly lived only a few miles, or at most fifty miles away.55 A similar picture emerges from accounts of mill construction in Holland between c. 1580 and 1800. Millwrights contracted for the building of wind- powered drainage mills often operated no further than a few dozen miles from their place of residence.56 But among this category of machine- makers, too, long-distance mobility was by no means exceptional. From the Late Middle Ages onwards, millwrights from Holland built drainage mills in Artois, Schleswig-Holstein and Polish Prussia. In the seventeenth and eighteenth centuries, they were hired by private entrepreneurs, trading companies governments or public institutions to construct drainage mills and wind- or water-powered industrial mills, not only in the Southern Neth-

53 Dohrn-van Rossum, ‘Migration-Innovation-Städtenetze’, passim ; Uta Lindgren, ‘Ordnungsprinzipien in technischen Handschriften und Drucken des Spätmittelalters und der frühen Neuzeit’, in: Arnoux and Monnet (eds), Le technicien dans la cité, 205-14 ; Pamela O. Long, Openness, secrecy, authorship: Technical arts and the culture of knowledge from Antiquity to the Renaissance (Baltimore, 2001), ch. 4; Dieter Lohrmann, ‘Das Maschinenbuch des Konrad Gruter für Erich VII., König von Dänemark (1424)’, Deutsches Archiv für Erforsc- hung des Mittelalters, 63 (2007), 71-92. 54 Nicolás García Tapia, Ingeniera y arquitectura en el Renacimiento español (Vallado- lid, 1990), 41. 55 Langdon, Mills in the medieval economy, 253-55. 56 Municipal Archive Rotterdam Ms. 3180, 3181; Siger Zeischka, Minerva in de polder. Waterstaat en techniek in het hoogheemraadschap van Rijnland (1500-1856) (Amsterdam, 2007), 248-50. 220 karel davids erlands, the Baltic area and northwest Germany, but also in the British Isles, France, Italy, Spain, the British, French and Dutch colonies in America, and in Dutch settlements at the Cape of Good Hope and in the Indonesian Archipelago.57 The rise of illustrated machine-books shows that knowledge on ma- chine-making in Europe since the Late Middle Ages also travelled in other ways than via the mobility of people. Even before the advent of print, these treatises on machines could circulate quite widely. Of Konrad Kyeser’s manuscript Bellifortis (1405), for example, at least 45 copies have been pre- served, almost all from the fifteenth century.58 Machine-books, normally in printed form, continued to appear throughout the early modern period. The books by Agricola, Besson, Ramelli, Stevin, Verantius, Zeising, Zonca and other authors, whose descriptions and illustrations were borrowed in Sabatino de Ursis and Xu Quangqi’s Western techniques of hydraulics and Johannes Schreck and Wang Zheng’s Diagrams and explanations of the marvelous devices of the Far West, formed part of of this very tradition.59. The tradition continued in the seventeenth and eighteenth centuries, with the publication of books on mills and other sorts of machines used in Holland. The first ‘mill-book’, called Architectura Mechanica Moole boek, composed by a Swedish millwright, Pieter Linpergh, was published in Am- sterdam in 1686. Linpergh’s book, reprinted in 1727, was in the 1730s sup- planted by two new mill-books authored by millwrights from Holland itself, which surpassed the work of their predecessor in both the quality and quantity of illustrations and descriptions: the Theatrum machinarum universale, of groot algemeen moolen-boek by Johannis van Zyl (with engrav- ings by Jan Schenk) and the Groot volkomen moolen-boek by Leendert van Natrus and Jacob Polly. In these works, every type of windmill was care- fully described and depicted, with measures and all. Engravings, printed on folio size pages, showed almost every relevant detail. These mill-books circulated both among literati and entrepreneurs at home and abroad and among millwrights themselves. Next to books and drawings, another medium in which knowledge could circulate in codified form was scaled- down models of machines.60 Models of machines, which were known in

57 Karel Davids, The rise and decline of Dutch technological leadership: Technology, economy and culture in the Netherlands, 1350-1800 (Leiden, 2008), ch. 5. 58 Lohrmann, ‘Das Maschinenbuch’, 74. 59 Machine drawings from the Late Middle Ages to c.1650 contained in these books are now accessible through the Database Machine Drawings, created in Berlin as part of a digital research library, The Archimedes Project, see http://dmd.mpiwg-berlin.mpg.de. 60 Davids, Rise and decline, ch. 7. moving machine-makers 221 Antiquity but fell into disuse since, reappeared in Europe in the Late Mid- dle Ages and became, from the sixteenth century onwards, an important means of representation in the communication between engineers and patrons, in instruction on machine-making, and in applications for privi- leges and patents. Individuals and institutions began to form collections of models, often accessible to the public as well.61 By 1800, both machine- makers and other interested persons in Europe thus could keep abreast of the state of the art in machine-making in a variety of ways. A perfect example of a first-rate consumer of knowledge at this time is Maarten Ne- derdijk, a carpenter and millwright from Rotterdam: at his death in 1809, he possessed a library of some 160 books, ranging from mill-books and treatises on architecture to manuals on physics, chemistry and mathemat- ics, as well as a collection of models, including a model steam engine. 62

Conclusion: Comparative Perspectives on the Circulation of Knowledge in China and Europe

The comparison of the circulation of knowledge on machine-making in China and Europe presented in this essay has revealed two significant dif- ferences in patterns. First of all, circulation of knowledge on machine- building through mobility of people, at least insofar as it was embedded in formal structures (such as the corvée system, guilds or formal arrangements on apprenticeship) developed in China to a lesser extent and happened at a less constant rate than in Europe. Secondly, circulation of knowledge on machine-making in a codified form (such as technological literature, draw- ings or models) occurred in China less frequently and was a less important complement to circulation through mobility of people than in Europe. What do these findings tell us about the evolution of the knowledge economy in Europe and China and the beginnings of the Industrial Revo- lution? One should beware of drawing very firm conclusions at this point, because so many variables are still unknown. The extent of mobility of machine-makers outside the context of formal regulations, for instance, is as yet hard to establish, although I suspect that this was lower among ma- chine-builders in China than in Europe. Nevertheless, a few suggestions can be made.

61 Marcus Popplow, Models of machines: A ‘missing link’ between early modern engineer- ing and mechanics? (Preprint nr.225 Max Planck-Institut für Wissenschaftsgeschichte, Berlin 2002). 62 Municipal Archive Rotterdam Ms. 834, probate inventory Maarten Nederdijk, July 1809. 222 karel davids If the above conclusions are correct and Kent Deng is right about the importance of agricultural books in China, then we can infer that from the fifteenth century onwards there were not only disparities in the develop- ment path of the ‘knowledge economy’ in China and (Western) Europe as a whole, as Jan Luiten van Zanden has claimed,63 but also between differ- ent sectors of the economy itself. Overall figures on book production, or on book production per capita, give us only a partial glimpse of the realities of knowledge circulation. Actual developments could vary by sector. Agri- cultural writings were produced in (Western) Europe, too (and they are still often consulted by historians as a source for agricultural history), but they are not considered a significant factor in the transmission of knowl- edge at the time, let alone in the continuity of agricultural development as such.64 In machine-making, by contrast, technological literature played a more important role in the circulation of knowledge in Europe than in China. This difference between sectors can only be partly explained by the attitude of literati in China, who were evidently more concerned with the writing and preservation of agricultural books than with the production and diffusion of literature on machine-making.65 There was, after all, a commercial printing industry in China since at least the Song dynasty (even though its status and relation with the state was somewhat different than in Europe), which ‘suddenly exploded’ in the 16th century.66 Commercial publishers ‘went well beyond’ the ‘Confucian canon and the standard his- tories as well as the The Ming Code and Ming Regulations’, according to Timothy Brook, ‘producing all manner of popular texts cheaply and in large volume to sell to a broader reading public than aspiring officials’.67 If there had been a large demand for books on machine-making, commercial pub- lishers might easily have produced them, too. Their relative scarcity may to some extent be explained by a lack of demand.

63 Jan Luiten van Zanden, The long road to the Industrial Revolution: The European economy in a global perspective, 1000-1800 (Leiden: Brill, 2009). See also Van Zanden’s con- tribution in this volume. 64 See for example Philip T. Hoffman, Growth in a traditional society: The French coun- tryside 1450-1815 (Princeton, 1996); Mark Overton, Agricultural revolution in England: The transformation of the agrarian economy 1500-1850 (Cambridge, 1996). 65 Cf. Deng, Development and stagnation, passim, on the role of literati. 66 Francesca Bray, Technology and society in Ming China (1368-1644) (Washington, 2000), esp. 11; cf. also Lucille Chia, Printing for profit: The commercial publishers of Jianyang, Fujian (11th-17th centuries) (Cambridge Mass., 2002), esp. 7-13 and Joseph P. McDermott, A social history of the Chinese book: Books and literati culture in Late Imperial China (Hong Kong, 2006) 70-71. 67 Brook, Confusions of pleasure, 131. moving machine-makers 223 A second conclusion concerns the flexibility of circulation. If knowledge on machine-making in Europe circulated more constantly and more wide- ly than in China and depended less on the mobility of people, then the ‘pool’ of available knowledge easily exceeded local or regional boundaries and innovations could travel without substantial frictions or barriers from one point of the Continent to another (and even beyond). This must have made the setting for technological change in machine-making in Europe between the late fourteenth and early nineteenth centuries more favour- able than in China. Guilds were a contributory factor to this process, but not an essential element. Finally, the distinction between mobility of people and circulation of knowledge embedded in codified forms (drawings, models, literature) may help to identify more clearly the connections between machine-making in the pre-industrial period and the beginnings of the Industrial Revolution. Robert Allen recently claimed that the watch-making industry, which ex- panded in England from the late seventeenth century onwards, was of crucial importance for the growth of the cotton industry in Lancashire. Watch-making provided the ‘high-quality, cheap gears … made of brass or iron’, which were the precision parts in Arkwright’s water frames, and watchmakers formed the skilled workforce that assembled the ‘clock work’ of the new machines.68 Yet Jennifer Tann and Gillian Cookson some years ago pointed to the remarkable fact that, contrary to conventional wisdom, traditional millwrights and clockmakers in England only played a minor role in textile machine-making after c. 1800. There are in fact very few ­examples of established machine-makers who made a successful transition to the building of machines for the rapidly expanding new branches of textile manufacture. The millwrights and engineers who did play an im- portant role in this stage were a new breed of experts who built ‘on the achievements of the more traditional kind of millwright’ but also had ‘a sound theoretical training’.69 According to Maxine Berg, millwrights and metalworkers like smiths and watchmakers provided a ‘primary source of skilled engineering labour’, but these were ‘not sufficient’. The manufacture of steam-engines from the 1790s, particularly, involved the formation of a

68 Robert C. Allen, The British Industrial Revolution in global perspective (Cambridge, 2009), 204-06. 69 Jennifer Tann, ‘The textile millwright in the Early Industrial Revolution’, Textile His- tory 5 (1974), 80-89, esp. 85 and 88; Gillian Cookson, ‘Millwrights, clockmakers and the origins of textile machine-making in Yorkshire’, Textile History 27 (1996), 43-57. 224 karel davids new workforce trained in engineering workshops created by, among others, James Watt.70 Although the construction and operation of new machines thus in the first stage of the Industrial Revolution may have been dependent on the mobility of machine-makers from existing crafts and trade who embodied the required skills, this was apparently no longer the case after c.1800. The explanation, I would submit, may reside in the fact that knowledge on machine-making at this time had become what Joel Mokyr has called71 ‘propositional knowledge’ , which could circulate in codified forms such as drawings, models and technological and scientific literature. This is ex- actly what Cookson’s observation of ‘sound theoretical training’ and Berg’s remarks on the importance of training under the supervision of a theo- retically skilled expert like James Watt suggest. Accumulated knowledge on machine-making during the Industrial Revolution in Britain evidently also could find its way into new industries because it was to some extent embedded in codified forms, which could be transmitted without the physical movements of experts. Differences in the ‘mechanics’ of the cir- culation of knowledge in China and Europe thus clearly mattered for the emergence of the ‘Great Divergence’. 72

70 Maxine Berg, The age of manufactures 1700-1820: Industry, innovation and work in Britain (London/New York, 1994²), 258-61. 71 Joel Mokyr, The gifts of Athena; Historical origins of the knowledge economy (Prince­ton, 2002), 4-21. 72 See now also Karel Davids, Religion, technology and the Great and Little Divergences: China and Europe compared, c. 700-1800 (Leiden, 2012). guilds and apprenticeship in china and europe 225

CHAPTER SEVEN

GUILDS AND APPRENTICESHIP IN CHINA AND EUROPE: THE JINGDEZHEN AND EUROPEAN CERAMICS INDUSTRIES

Christine Moll-Murata

Introduction

According to Epstein, the role of guilds in early modern Europe was positive since they supported the formation and circulation of human capital, and stimulated product innovation and the maintenance of quality standards. This revisionist position was contested, for instance, by Sheilagh Ogilvie, who claimed that, on the contrary, guilds were primarily rent-seeking in- stitutions and in the process adversely affected quality, skills, and innova- tion.1 In a time of academic controversy between those positions, it is useful to see to what extent these different views of the value (or harmful- ness) of European guilds are relevant to their counterparts in China. The present study proposes a look at ‘institutions of collective action’2 outside Europe and offers a perspective on whether Chinese corporations were useful and effective in the transmission of skills and in the circulation of skilled labour. In order to develop a comparative view, this chapter studies guilds in the porcelain industry of Jingdezhen in comparison to the related fayence industry in Delft, and alternative organisations for skill formation and circulation in the case of the first two European manufactures for hard-paste porcelain, in Meissen and Vienna. What can such intercultural comparisons show? Can they serve to bol- ster the arguments of those who are more convinced of the efficiency of the guilds, or those who are more sceptical about the qualities of this long-

1 S.R. Epstein, ‘Craft Guilds, Apprenticeship, and Technological Change in Preindus- trial Europe’, The Journal of Economic History 58 (1998), 684-713; Sheilagh C. Ogilvie, ‘Guilds, Efficiency, and Social Capital: Evidence from German Proto-industry’, Economic History Review 57 (2004), 286-333; S.R. Epstein, ‘Craft Guilds in the Pre-modern Economy: A Discus- sion’, Economic History Review 61 (2008), 155-74; Sheilagh C. Ogilvie, ‘Rehabilitating the Guilds: A Reply’, Economic History Review 61 (2008), 175-82. 2 For definitions, see ‘Institutions for Collective Action’, collaboratory website main- tained at Utrecht University, last accessed 2 December 2011, http://www.collective-action. info/. 226 christine moll-murata lived institution? Given the fact that even reading the same Central Euro- pean documents, Epstein and Ogilvie arrived at opposite conclusions,3 they would presumably have differed as well if they looked at the Chinese evi- dence. Certainly the way in which they have articulated the debate and sharpened the positions can help to focus the analysis of a guild record which appears less well-documented and more recent than its European counterparts.

Porcelain: Definitions and Early Exchange between China, the Middle East, and Europe

Definitions of ‘porcelain’ differ in China and the West. As Robert Finlay points out,4 Western categorisation distinguishes between the three differ- ent types of ceramics: ‘earthenware’, made from clay and fired between 600 and 1,000°C, yielding a porous result in brownish-red colours when fired; ‘stoneware’, baked at about 1,100-1,250°C, almost entirely nonporous and of light grey to black colour when fired; ‘porcelain’, made from porcelain stone and Kaolin (China clay), and fired at temperatures higher than 1,300°C. The latter comes from the kiln white and translucent.5 The Chinese traditional classification distinguished between tao, ‘earthenware’, and ci, which in- cludes both stoneware and porcelain. Porcelain ‘emerged’ in China, any- time between the second millennium BC and the thirteenth century AD, depending on the recognition of the point in time when ‘earthenware’ became ‘stoneware’ or ‘porcelain’.6 The Jingdezhen historians date the first genuine porcelain to 1004 AD, when the location was conferred the name of ‘market town [established during the reign of] Jingde [1004-1008]’. ­However, Kerr and her colleagues consider the Xing wares made in Hebei province during the early Tang dynasty in the seventh century AD to be ‘the world’s first true ’.7

3 Epstein, ‘Craft Guilds in the Pre-modern Economy’, 160, referring to Ogilvie, ‘Guilds, Efficiency, and Social Capital’, 296-97, for a diverging interpretation on a complaint that the Wildberg guild did not give approval to below standard quality cloth, and Ogilvie, ‘Rehabilitating the Guilds’, 177. 4 Robert Finlay, ‘The Pilgrim Art: The Culture of Porcelain in World History,’ Journal of World History 9 (1998), 144-46. 5 Rose Kerr and Nigel Wood, with Ts’ai Mei-fen and Zhang Fukang, Ceramic Technol- ogy. Vol. 5, Chemistry and Chemical Technology, pt. 12 of Joseph Needham (ed.), Science and Civilisation in China (Cambridge: Cambridge UP, 2004), 9, citing from Hamer & Hamer, The Potter’s Dictionary of Materials and Techniques 1975, 229. 6 Finlay, ‘The Pilgrim Art’, 145. 7 Kerr et al., Ceramic Technology, 151-52. guilds and apprenticeship in china and europe 227 Porcelain production has aptly been called a ‘pilgrim art’8, because be- tween the ninth and the eighteenth centuries, materials and decorations were exchanged and adapted in complex processes between China, the Middle East, and Europe. In this blue-and-white exchange, Middle Eastern potters strove to achieve the whiteness and the vitrified look of the glaze of Chinese porcelain, while Chinese potters were fascinated by the cobalt blue ornamentations used in Islamic ceramics. Persian and Egyptian arti- sans applied tin glaze to their wares to achieve the gloss of Chinese porce- lain, but instead of clay, used finely ground quartz powder, which turns out white when fired at temperatures lower than the 1300°C necessary for genuine porcelain.9 The Chinese conversely imported cobalt blue from Islamic traders and around 1300 started to decorate their products in blue and white.10 As Finlay explains, at first Chinese connoisseurs did not ap- preciate this new design and preferred the more subtle beauty of mono- chrome Song dynasty white and green porcelain and with carved or relief ornamentations that imitated, in form and colour, the venerated bronze vessels of Chinese antiquity. The blue and white wares initially remained confined to the Inner and Southeast Asian market.11 Only by the 1430s were they gradually accepted by the Chinese elite, whose taste had adapted to expressively painted porcelain, and thus to the ornamental style that was preferred in the Middle East.12

Jingdezhen: General Setting and Supply for the European Market

The production of fine porcelain in Jingdezhen, Jiangxi province, had started around 1004, but even before that date it had been a centre of ce- ramics. Since the eleventh century, the successive Song (960-1276) and Yuan (1276-1368) dynasties taxed porcelain production in money and in kind. During the Ming dynasty (1368-1644), the government established a state manufacture in Jingdezhen (the exact date is disputed, either 1369 or 1402). It stood under the direct or indirect supervision of court officials during

8 Finlay, ‘The Pilgrim Art’. 9 Regina Lee Blaszczyk, ‘Porcelain for everyone: The Chinaware aesthetic in the Early Modern era’, paper presented at the GEHN Conference ‘Global Histories of Economic Development: Cotton Textiles and Other Global Industries in the Early Modern Period’, Fondation Les Treilles, March 2006, http://www2.lse.ac.uk/economicHistory/Research/ GEHN/GEHNPDF/TREILLESBlaszczykPaper.pdf (last accessed 27 March 2013). 10 Finlay, ‘The Pilgrim Art’, 155. 11 Ibid., 156. 12 Ibid. 228 christine moll-murata the Ming and the Qing dynasties (1644-1911). However, at all times, private pottery and painting workshops and kilns existed as well and were fre- quently used by the state for their manpower and their firing facilities. Private potteries and kilns also operated during the years when the Impe- rial Manufacture was closed down. For centuries the Jingdezhen porcelain industry produced more pieces than all other porcelain centers in China together and until the eighteenth century it remained the world leader in both quality and volume. Chinese blue-and-white porcelain arguably was the most widely traded ceramics in world history and inspired imitations and adaptations by its East-Asian neighbours, the Middle East, Europe, and North-America. Europe experienced a kind of ‘porcelain craze’ that en- tranced the high and low nobility as well as the burgeoning middle classes. The greatest collectors were the French, Saxon, Prussian, Austrian, Italian, Portuguese, Spanish, English, Danish, and Swedish royal courts as well as all echelons of nobility in their lands. For some time in the late seventeenth and mid-eighteenth centuries, porcelain became the ultimate cultural pres- tige object, a high-priced import good sold at auctions in the great Euro- pean ports. From the beginnings in the early seventeenth century until the East-Asian porcelain trade gradually receded in the early 1800s, European East Asia Companies had shipped out huge quantities of porcelain. Esti- mates for the Swedish East Asia Company alone reckon that about 50 mil- lion pieces of Chinese porcelain were imported between 1732 and 1806.13 The Dutch East India Company vessel Geldermalsen that sank in 1752 on its way from Canton to Batavia, carried 140,000 pieces of Chinese porcelain, a load that has been assessed as typical for an East Asia Company ship in the mid-eighteenth century.14 Profiting from the newly established science of chemistry that devel- oped from the fundament of alchemy and was related to the desire to cre- ate gold, European nobles as well as private entrepreneurs sought not only to purchase, but also, if resources permitted, to produce imitations of Chi- nese porcelain in their own territories. On the one hand, this was accom- plished by way of experimentation until the necessary ingredients were found in the clays and minerals of the various European regions. Thus, the first specimens of true hard–paste white porcelain were presented to the

13 Jarl Vansvik, ‘ Collected in Sweden’, Website maintained by Jan-Erik Nilsson, Gothenburg, Sweden ‘China and Sweden: Treasured Memories’, http:// gotheborg.com/exhibition/vansvik.shtml, last accessed 2 December 2011. 14 Kerr et al., Ceramic Technology, 746. guilds and apprenticeship in china and europe 229 Saxon Elector August II by 1709,15 and the manufacture was opened in 1710. On the other hand, information on porcelain produc- tion leaked out directly from Jingdezhen through the French Jesuit François Xavier d’Entrecolles (1664-1741), who worked there as a missionary. He in- quired with Chinese proselytes, presumably merchants and potters, and although he cites no literature, it can be assumed that he also studied Chi- nese texts about procedures, components and proportions of the porcelain paste as well as the glazes. In two detailed letters on Jingdezhen and its porcelain, dated 1712 and 1722, he transmitted the results of his ‘fact-finding’ to his superior in France,16 who passed on the information, together with a sample of the Jingdezhen minerals, to the natural scientist Réaumur (1683-1757). On the basis of d’Entrecolles’ letters, Réaumur in 1727 published his analysis of porcelain production which was eventually also communi- cated to the Royal Society of London.17

Number of Kilns and Size of Workforce

The Jingdezhen Imperial Manufacture reportedly had been provided with 20 kilns in 1400 and 56 in 1430, and 23 specialised workshops for particular steps of production in the branch where the division of labour was quite advanced. In the early sixteenth century, 300 to 500 craftsmen, not only potters, but also masons, carpenters and blacksmiths, were permanently employed in the Imperial Manufacture under corvée obligations. Addi- tional labour was hired when demand was high.18 The corvée obligations encompassed relatively unskilled labour drafted from the surrounding counties in rotating work shifts, as well as skilled artisans who were as-

15 Ibid., 710. 16 François Xavier d’Entrecolles (S.J.), ‘Lettre du Père d’Entrecolles, Missionaire de la Compagnie de Jesus: Au Père Orry de la mesme Compagnie, Procureur des Missions de la Chine & des Indes’, in: Stephen Bushell, Description of Chinese Pottery and Porcelain: Being a Translation of the T’ao shuo (Oxford: Clarendon 1910; reprint New York: AMS Press 1973), 181-209, 210-22. 17 N.J.G. Pounds, ‘The Discovery of China Clay’, Economic History Review, New Series, 1 (1948), 22-23. 18 Fang Zhuofen, Hu Tiewen, Jian Rui, Fang Xing, ‘The Porcelain Industry of Jingdezhen’, in: Xu Dixin and Wu Chengming (eds), Chinese Capitalism, 1522-1840 (Basingstoke etc.: Macmillan, 2000), 311; Zhongguo taoci shi (History of Chinese porcelain), ed. by Feng Xian- ming, An Zhimin, An Jinhuai et al. (Beijing: Wenwu chubanshe, 1982), 366; Jiangxi tongzhi (Gazetteer of Jiangxi province), ed. by Liu Kunyi, comp. by Liu Duo, Zhao Zhiqian et al. 1881 (Reprint Taibei: Chengwen chubanshe, 1989), chap. 93, fol. 9a, Memorial by Xu Zhi. 230 christine moll-murata signed to work at the Manufacture either ten days every month, or on a work shift of three months per year over a period of three years. They worked in their own workshops during the rest of the time.19 Due to dimin- ishing resources and decreasing power to allocate unpaid labour, the pro- duction at the Imperial Manufacture was closed down in 1608. After the transition from the Ming to the Qing dynasty, the institution was re-opened in 1654, temporarily halted during the civil wars in the 1670s, and reno- vated in 1680. For about one hundred years, its most productive time with the highest level of artistic achievement, it stood under the direct control of court officials. Thereafter, local officials took over its direction. Between 1854 and 1864 it was destroyed in the Taiping rebellion, rebuilt in 1866 and productive until the final years of the Qing dynasty. In 1910 it was closed down and succeeded by the Jiangxi porcelain company Jiangxi ciye gongsi, which was financed partly by the government and partly by private entre- preneurs. In 1680, the Imperial Manufacture had 23 workshops and six kilns. The workforce was no longer subjected to corvée service, but was paid regular wages. As in the late Ming, the Manufacture after some time resorted to an outsourcing system for firing the wares produced at the Imperial workshops. The private sector in Jingdezhen consisted of larger and smaller family enterprises. A number of 300 private workshops and kilns was reported for the Yuan dynasty (1276-1368). Ming estimates ran to about 900. By the time of the Qing, d’Entrecolles wrote of about 3,000 ‘porcelain furnaces’ (four- neaux à porcelaine) in 1712,20 a number that seems rather high, especially in view of the more likely figure of 200-300 private kilns which Tang Ying, the Director of the Manufacture, mentioned for 1743.21 According to Tang Ying, 300 permanent artisans and administrators received wages and food from the Imperial Manufacture.22 The Imperial Maritime Customs Decen- nial Reports also pointed out the difficulty of giving precise figures; they quoted Morrison, who referred to 200-300 kilns and several hundred thou- sand workmen; Staunton, who wrote of 3,000 kilns (certainly following d’Entrecolles); and Huc, who mentioned 500 furnaces and over one million workmen, all before the Taiping rebellion (1853-1864). Thereafter, initially

19 Kerr et al., Ceramic Technology, 209. 20 d’Entrecolles (S.J.), ‘Lettre du Père d’Entrecolles’, 184. 21 Tang Ying, Tang Ying ji (Collected works of Tang Ying), ed. by Zhang Faying and Diao Yunzhan (Shenyang: Liao Shen shushe 1991), p. 959. Actually, Tang Ying refers to minyao er san bai qu 民窯二三百區 (two to three hundred locations of private kilns). 22 Tang Ying, ‘Taocheng jishi beiji’ (Stele inscription with a memorial on accomplished porcelain), 1735; for the date see Kerr et al., Ceramic Technology, 27; Tang Ying ji, 950. guilds and apprenticeship in china and europe 231 20 to 30 kilns resumed production. In 1869, 110 kilns (60 for fine and 50 for coarse wares) were in operation, and an estimated 150,000 workmen (500 for each fine ware kiln and 200 to 300 for the coarse ware kilns) found employment in the porcelain trade.23

Total Output Fang Zhuofen and colleagues point out that, based on consumption of firewood, the total output of Jingdezhen porcelain in the early Qing can be calculated at about 200,000 dan24 or 40 million pieces per year, but they assume that during the Qianlong reign the figure must have been about 300,000 dan or 60 million pieces a year. The maximum—but not undis- puted—estimated annual quota for imperial porcelain was ‘several hun- dred thousand pieces’.25 If this is taken to mean about 500,000 pieces, then Imperial porcelain amounted to about 0.8 percent of the private produc- tion during its most active period.26

Population Figures Population estimates based on taxed population remain vague, since mi- grants were not expected to register at their workplaces, but in their plac- es of origin. Recent research on the urban history of Jingdezhen assumes a figure of about 100,000 ‘hosts and guests’, implying permanent residents and migrants, for c. 1550.27 Of the total population, about 10,000 were hired in the porcelain sector. After 1680, population increased, and in the late Qing was estimated at about 200-250,000 permanent residents. However,

23 Decennial Reports on the Trade, Navigation, Industries etc. of the Ports open to Foreign Commerce in China and Corea […] , 1882-1891. I. Statistical Series, 6. Imperial Maritime Cus- toms (ed.), (Shanghai: Inspector General of Customs, 1896), 204-05. 24 dan 擔 is the unit for a kiln-load of 200 porcelain pieces. 25 Tang Ying ji, 144, in preface to the gazetteer of Fuliang, the district where the market town Jingdezhen was situated, dated 1740. However, this figure has been questioned Lu Jiaming, ‘Ming Qing shiqi Jingdezhen yuqichang ‘guanda minshao’ zhidu shitan, jian yu Pan Qun, Xu Dixin xiansheng shangque (Exploration of the system ‘official forming and private burning’ in the Ming and Qing Imperial porcelain manufacture of Jingdezhen: A debate with Pan Qun and Xu Dixin), Nanchang zhiye jishu shifan xueyuan xuebao (Journal of the Nanchang Normal Technical College) (1994/2), 32, who supposes that it spans several years, since Tang Ying in 1743 reported the quota of a maximum of 29,000 pieces per year which had to be despatched from the Imperial Manufactory to the capital. 26 Fang Zhuofen et al., ‘The Porcelain Industry of Jingdezhen’, 314. 27 Liang Miaotai, Ming Qing Jingdezhen chengshi jingji yanjiu (Study on the urban economy of Jingdezhen in the Ming and Qing periods) (Nanchang: Jiangxi renmin chuban- she, 1991), 16-19. 232 christine moll-murata

Table 7.1. Kilns and workforce of Jingdezhen. Year Kilns Workforce (persons) Source/Remarks Yuan (1276-1368) 300 Ming (1368-1644) 900? Calculated, probably too high 1712 3,000 D’Entrecolles, certainly exagger- ated 1743 200-300 Tang Ying 18th c. 200-300 several hundred Morisson, quoting from Fuliang thousand gazetteer (Decennial Reports) 18th c. 500 1,000,000 Huc (Decennial Reports= After 1864 20-30 Decennial Reports 1869 110 150,000 Decennial Reports 1907 80 Zhongguo cidu, p. 335

Table 7.2. Taxed and registered population of Fuliang district, of which Jingdezhen forms part. Year Households Taxed persons dingkou Before 1736 57,500 1782 55,896 250,290 1802 58,792 281,790 1821 59,606 288,220 1851 57,605 286,874 1869 57,584 286,894

Source: Jiangxi tongzhi (Gazetteer of Jiangxi province), chap. 47, fol. 28 a/b, p. 1042. in greater Jingdezhen an estimated several hundred thousand persons were engaged in the porcelain trade, including a large percentage of migrants.28

Professional Training in the Jingdezhen Porcelain Industry

A strict division of labour prevailed in the Jingdezhen ceramics industry. The basic steps of porcelain production, preparation of the raw materials (the clay used for the porcelain, the glazes, and the colours), the forming of the blanks on the wheel or with moulds, the production of protective clay containers (saggars) for firing the porcelain blanks, the firing, orna- mentation, and the packing were all separated. Such specialisations also

28 Ibid., 20. guilds and apprenticeship in china and europe 233 occur in other porcelain and ceramic production sites after these have at- tained a certain scope of production. Finer subdivisions of labour, for in- stance of genres of painting, for various phases of the forming process, or between work on or sculptured objects, are also frequently at- tested in porcelain production world-wide.29 However, in Jingdezhen, a further sub-division of the production process took place that seems to be quite unique. It distinguished the so-called ‘round’ or ‘open’ objects (yu- anqi) that were turned on the wheel or formed in moulds, and the ‘closed’ or ‘carved’ objects (zhuoqi) that could be round or polygonal, turned on the wheel and then carved when finished, or built up in segments and also formed with moulds. Potters would specialise in one of these sub-branch- es and were not supposed to try their hands in the other. In conventional reckoning, the Jingdezhen porcelain industry was said to have eight sectors and 36 trades, but actually the number of trades and subdivisions, most of which were controlled by particular guild branches, was even greater than that.30 As in the ceramics trades elsewhere, a certain hierarchy within the trade applied in Jingdezhen. It expressly ran along the lines of ‘fine production’ for well-off customers and ‘coarse production’ for everyday use. As a rule, the ornamental trade ranked higher than those crafts that demanded more physical strength, such as the preparation of raw materials, firing, and to a certain extent also the forming of porcelain. In principle, Chinese guilds or proto-guilds regulated or tried to regulate which groups of people could exercise what trade or procedure within the production process. They did so mainly according to place of origin. The same applies for the distributive trades, in which only merchants of distinct common-origin networks were supposed to trade with other specific regions. In the craft sector, apart from regulations about wage payment (terms of payment, and hiring and firing), production periods—porcelain produc- tion was seasonal work, and the period of rest between the twelfth and the second lunar month was strictly enforced—production ceilings, and sales

29 See, for instance, the long list of specific occupations in the Staffordshire potteries, ‘Pottery Jobs Index’, compiled by Steve Birks, Website The local history of Stoke-on-Trent, England, http://www.thepotteries.org/jobs/index.htm, last accessed 2 December 2011. 30 Sie Wun-Ci (Xie Yunqi), Chuantong yu zhidu chuangxin: Jingdezhen taoci chanye fazhan de bijiao yanjiu (Tradition and Institution Innovation: A Comparative Study of the Development of Jingdezhen Ceramics Industry). M.A. thesis, Guoli Zhongshan daxue (National Sun Yat-sen University, Taiwan), electronically released Feb. 2008, http://etd.lib. nsysu.edu.ts/ETD-db/ETD-search/view_etd?URN=etd-0213108-163726, accessed June 2008, 71fn192. 234 christine moll-murata conditions, guilds also made provisions about apprenticeship. So far, no written regulations before the twentieth century have been found. It is only from casual remarks in the literature on contemporary Jingdezhen porce- lain that we can trace such arrangements to the time when the guilds or common-origin artisan corporations began to become active. Chinese historians assume that this activity started in the late Ming, and that the Duchang guild house was the earliest example of common origin corporation.31 The fact that guilds were organised along lines of a common- origin rather than—as in Europe—on citizenship in the place where the guild was based, is a common Chinese phenomenon from the seventeenth through the twentieth centuries.32 Due to the prolonged civil wars during the dynastic transition from the Ming to the Qing, Jingdezhen was devas- tated in 1674,33 and the native workforce had all but disappeared in the process. After the situation had stabilised, new labour came in from the surrounding regions in Jiangxi, especially people from the Duchang district on the eastern shores of Lake Poyang, about one hundred kilometres to the west of Jingdezhen. From the rather recent guild steles34 that have been transmitted, it becomes clear that at first only a small group of Duchang migrants arrived in Jingdezhen; a twentieth-century stele mentions ‘four surnames’, and later ‘24 surnames’.35 In the interpretation of the Duchang guild representatives, the 24 surnames formed subgroups or ‘networks’ (bang) that organised the affairs of the guild on a rotational basis. This means that the restrictions of access to particular trades or skills not only concerned the region of origin, but were also related to family connections. Su Yongming estimates that there were as many as 400 subgroups and net-

31 Liang Miaotai, Ming Qing Jingdezhen, 231; Su Yongming, Hangbang yu Jingdezhen shehui bianqian—cong Mingmo dao Minchu (Guild networks and social change in Jingde- zhen—from the late Ming to the early Republic), Unpublished MA thesis, Nanchang University, 2005, 12-13. 32 Christine Moll-Murata, ‘Chinese Guilds, Seventeenth Through Twentieth Centuries: An Overview’, in: Jan Lucassen, Tine De Moor, Jan Luiten van Zanden (eds), ‘The Return of the Guilds’, International Review of Social History 53, Supplement (2008), 215. 33 Su Yongming, Hangbang yu Jingdezhen, 12. 34 Steles are publicly displayed stone tablets recording guild privileges and regulations for posterity. 35 Duchang ren yu Jingdezhen (Duchang people and Jingdezhen). Ed by Zhongguo renmin zhengzhi xieshang huiyi Duchang xian weiyuanhui wenshi ziliao yanjiu weiyuan- hui (Commission on historical and literary materials of the Duchang district commission of the Chinese People’s Consultative Conference), 1991. Duchang wenshi ziliao (Historical and literary materials on Duchang) Nr. 3, 23/24; interview with a retired guild representa- tive, 15, 18. guilds and apprenticeship in china and europe 235 works in Jingdezhen. 36 Above the level of the guild houses, most of them merchant guild houses (with a number of 24 given in the literature),37 three larger organisations—the so-called ‘three networks’ san bang: the Duch- ang, Huizhou, and mixed group—organised much of the routine affairs of the town.38 The advantage of keeping the skills within a native-place or- ganisation was to maintain the livelihood of the group to which one be- longed. The extreme specialisation in the Jingdezhen case also allowed the limitation of particular parts of the productive process within a given branch to people from particular regions. Sie Wun-ci stresses a fact which gave the firm a great advantage over Jingdezhen. In her opinion, the Chinese porcelain production and distribution channels remained rigidly divided.39 Unlike with Josiah Wedgwood and Sons, the Chinese merchants hardly had any say in production. They were obliged to buy exclusively from particular potters, remain in an approved branch and trade with specified marketing regions. The potters, on the other hand, were not supposed to turn to marketing.40 However, the potters, kiln work- ers and proprietors, and painters did not acquire their skills in their regions of origin. Rather, they were either sent to Jingdezhen workshops as children or adolescents, or they were born in town, but still identified themselves with their fathers’ or more remote ancestors’ native place. Rather detailed regulations are presented in a recent publication on porcelain production in Jingdezhen (2004) as part of the official city gazet- teer.41 These regulations were collected during a 1963 oral history project and recorded in a publication ‘for internal circulation’.42 Since it refers to the memories of elderly porcelain artisans, it gives the situation as of the

36 Su Yongming, Huang Zhifan, ‘Hangbang yu Qingdai Jingdezhen chengshi shehui’ (Trade networks and the urban society of Qing dynasty Jingdezhen), Nanchang daxue xuebao (Renwen shehui kexue ban) (Journal of Nanchang University, Humanities and Social Sciences Edition), 38 (2007), 82. 37 Liang Miaotai, Ming Qing Jingdezhen, 231. 38 Su Yongming, Huang Zhifan, ‘Hangbang yu Qingdai Jingdezhen chengshi shehui’, 84. 39 Sie Wun-ci, Chuantong yu zhidu chuangxin, 247. 40 Ibid., 246. 41 Zhongguo cidu: Jingdezhen shi Ziye zhi (China’s porcelain city: Monograph on porce- lain of Jingdezhen city. Vol. 2 of the Municipal Gazetteer), ed. by Jingdezhen shi difangzhi bangongshi (Compilation bureau of the Jingdezhen municipal gazetteer). 2 vols. (Beijing: Fangzhi chubanshe, 2004), 778-79. 42 Jingdezhen zhiciye lishi diaocha ziliao xuanji (Selected materials for the investigation of the history of the Jingdezhen porcelain industry), ed. by Jiangxi sheng lishi xuehui Jing- dezhen ciye lishi diacha zu. (N.p.: 1963). 236 christine moll-murata second half of the nineteenth century at the earliest. One of the rules refers to the years 1912 to 1916.43 The regulations—or memories of them—do not elaborate on the nature of the transferred skills, but give a framework of apprenticeship in Jingdezhen. Like the guilds, the apprenticeship system was strongly focused on regional cohesion. Many of the branches were—at least formally—reserved for people from particular home regions, most prominently from the Duchang district. The periods of apprenticeship varied considerably among the different craft branches. They extended between one and eight years, with an aver- age of 3.5 years. The eight-year term was intended for the very young ap- prentices of the ‘closed’ object turners; older apprentices of this trade had to serve for six years. Less demanding skills, such as odd jobs in the produc- tion of ‘open’ objects only required a one year apprenticeship. In some cases, no fixed terms are given, and it depended on the labour situation as to whether the fully trained apprentices could leave or were required to stay on. Only for the Jingdezhen branch of ‘round object’ potters is it stat- ed that a premium of 3,200 copper cash had to be paid by the apprentice to the master at the beginning of the training period. In the branch of the ‘closed objects’, apprentices would earn a slightly higher wage in their first year of learning (4 silver taels, corresponding to about 4,000 copper cash).44 The age of the apprentices depended on the branch they were engaged in. This ranged between twelve and twenty sui.45 In trades that demanded more physical strength, like mixing clay and transporting raw materials, saggars, and blanks, apprentices were taken on at a higher age. For the branch of ‘closed objects’ one apprenticeship contract would spec- ify: ‘Under the name of the master XY, an apprenticeship for a term of n years will be served. If the apprentice doesn’t listen to what he is being taught, secretly runs away or commits other mistakes, the master shall take no responsibility. This is set up as a contract’.46 This formulation, which may seem to be vague in European terms, implies that the master would

43 Zhongguo cidu, 778. 44 Ibid., 778-9. 45 Ibid., 612. Sui, ‘years of age’, traditionally are reckoned from the conception of the child, plus one year for every New Year passed. At birth, the child is one year old. 46 Zhongguo cidu, 779. For two more detailed contracts for tailor apprentices of the year 1896, see Jin Zhilin, ‘The Importance of Guilds for the Recruitment and Training of Skilled Workers’, in: Brian H.A. Ranson, S.R. Epstein et al. (eds), A Preliminary Study of Craft Guilds in China: Guild-Hall and Government: An Exploration of Power, Control and Resistance in Britain and China, Volume 1: A Preliminary Study of the Social Organisation of Guilds in China (Hongkong: Baptist University, David C. Lam Institute for East-West Studies 1997), 107-08. guilds and apprenticeship in china and europe 237 take no legal liability for any crime that the apprentice might commit, as parents would be expected to do. It also rejects possible demands for wage payment in case the apprentice absconded. In other words, it outlines that in the case of breach of contract on the part of the apprentice, the master would not be bound to the contract either. The apprentices were paid no or very low wages in some branches; in others they earned half of the masters’ wages in this third year. Entrance fees are mentioned for only one of their twenty different branches. In most cases, when the apprenticeship was finished, young people were still ex- pected to remain with their master for one year at submarket wages. Female apprentices are mentioned only for the spoon-makers; their wages were one yuan lower.47 The regulations did not state that the guilds or the mas- ters were obliged to conduct any formalised tests on the abilities of the apprentices after their training period was over. It merely stated (for the ‘carved object’ potters) that after the period of apprenticeship, the appren- tices had to give three banquets: one for the guild leaders, two for their own master; the saggar makers had to provide for ‘leaving the master wine’. If they agreed to produce an additional 200-300 loads of saggars without pay, the master would consent that they leave his workshop and drink the wine; if not, he would not appear for consuming wine and food, and the appren- tices had to stay on.48 A system that may be special for Jingdezhen is that shop owners as well as skilled artisans or ‘masters’ who had hired out to such proprietors as sub-contractors, could have their own apprentices. Such persons were ‘masters’ or ‘teachers’ in regard to their skills, not in the sense of indepen- dently operating a workshop of their own. The product of the apprentice’s labour belonged to the master, who provided for the apprentice’s board and lodging and in the last years of the apprenticeship gave them some remuneration. This system did not exist in all branches: for instance, hired

47 Male wages are not specified in this case. In the closed form trade, yearly wages of 4 silver yuan in the first year, 6 yuan in the second, and 8 silver yuan in the third are men- tioned. If these were also applied for the spoon makers, the girls’ wages would be 25%, 16%, and 12% lower than those of the boys’. Female wages were also lower in English and German porcelain production. See Maxine Berg, Luxury and Pleasure, 135-36 for Staffordshire between 1750 and 1765, and Helmut Gröger, ‘Die Arbeits- und Sozialverhältnisse der sta- atlichen Porzellanmanufaktur Meißen im 18. Jahrhundert’, Staatliche Archivverwaltung im Staatssekretariat für Innere Angelegenheiten (ed.), Forschungen aus mitteldeutschen Archiven. Zum 60. Geburtstag für H. Kretzschmar (Berlin: Rütten & Loening 1953), 175, who mentions unskilled female labour in Meissen since 1763, which at first was remunerated at a rate of one third of male unskilled labour. 48 Zhongguo cidu, 779. 238 christine moll-murata master carvers of porcelain blanks in the closed-form sector could have their own apprentices, but in all other sectors of closed-form pottery, the workshop owners personally employed the apprentices. To make things even more complicated, in open-form pottery, hired masters who filled saggars with small items of porcelain were allowed to have their own ap- prentices, but not those masters who filled the protective casings with larger porcelain blanks. The permission to train one’s own apprentice was usually confined in number (one apprentice at a time) and temporally (up to once every ten, in another version every twenty years). However, enforce- ment of these temporal limitations was no longer so strict by the twentieth century.49

Alternatives to Guild-organised Apprenticeship There were also craft branches which did not take on apprentices—or only as a last resort. These were for instance the kiln-builders from the Wei fam- ily, who preferred to keep their professional secrets within the extended family. They managed to do so for several generations, but finally the pro- duction secret leaked out to a member of the Yu family, probably a son-in- law.50 The Yu in turn would only apprentice sons and nephews of their own clan. Porcelain painting was also kept within the family if possible. Of this, the gazetteer bears ample evidence. Under the proud title ‘Hereditary Houses of the Porcelain Industry’,51 it contains group biographies of twen- ty-four families with at least three generations engaged in porcelain pro- duction; they had received recognition beyond the confines of the Jingdezhen porcelain circles. Apart from being a fascinating document for what it mentions and what it obviously conceals, it always specifies the training which the protagonists received. Clearly, knowledge tended to be kept within the family, as in the case of grandfather Wang Xiuchun 王秀 春 (1851-1904), father Wang Bu 王步 (1898-1968), the present family head Wang Enhuai 王恩怀 (b. 1935) and his brothers Wang Shenghuai 王声怀 (1929-1991) and Wang Xihuai 王希怀 (1932-1982), who were all blue-and- white underglaze painters.52 An apprenticeship like that of the master

49 Ibid., 612. According to Zhongguo cidu, 613, the apprenticeship system was abandoned in 1949, but in 1983, in order to enhance the transmission of old crafts, elderly masters were allowed to have their own, personal apprentices in the state factories where they were working. This could include their own children. Masters who trained such personal appren- tices earned a bonus. 50 Ibid., 611. 51 Ibid., 641-76, ‘Ciye shijia’. 52 Ibid., 648-50. guilds and apprenticeship in china and europe 239 sculptor Sun Hongyuan 孙洪元 (1883-1963) who did not originally come from a porcelain producing background, but was a farmer’s son, could start out very early, at age nine. In his case it took ten years, from age 9 to 19, to master the carving of closed-form vessels and reach accomplishment in design and glazing.53 Mrs. Bi Defang 毕德芳 (b. 1946) even began at age eight to carry on the family tradition of polychrome porcelain painting.54 Cases where the founder of the ‘hereditary house’ first learned another profession, like that of the painter Wang Yeting 汪野亭 (1885-1942) who started out as a village schoolteacher at age 19 and only began porcelain painting at age 27,55 or the porcelain sculptor Zeng Longsheng 曾龙升 (1901-1964) who first learned the family trade of wood carving before being sent to Jingdezhen as an apprentice in his uncle’s blank carving workshop, are the exception rather than the rule in this gallery of porcelain produc- ers.56 Such professional careers, although they can hardly be seen as rep- resentative for the average worker in the ceramics industry, strongly imply that family relations mattered a great deal, and conform with the policy of familial master-apprentice relations within state-run factories. Finally, there is at least one piece of evidence for an institution which fostered academic learning of ‘ceramics’—even if the scope and contents of its teachings remain vague. According to the Porcelain Gazetteer, Jing- yang Academy (Jingyang xueyuan 景仰學院), founded in 1776, was the first vocational school for porcelain training, and also the first academy at Jingdezhen at that time. As a market town Jingdezhen fell outside the state education system which provided for government operated schools for learning the Confucian scriptures from the lowest administrative level, the district, upwards through prefectures and provincial capitals to the impe- rial capital. Land and funds for the Academy were donated by private per- sons, so that it can hardly have been considered a state institution, but it definitely operated with government sanction. What is more, when it was enlarged in 1816, the two large guilds Taocheng (Perfect Porcelain) and Taoqing (Celebrated Porcelain) of the pine faggot and the brushwood kiln operators donated money. In fact, it is in another context also referred to as the guild house of these two corporations. The first teacher of the acad- emy was Zheng Tinggui 鄭廷桂, the editor of one of the few Qing mono- graphs on porcelain, Jingdezhen taolu (A record of Jingdezhen porcelain),

53 Ibid., 648-50. 54 Ibid., 659. 55 Ibid., 663. 56 Ibid., 672. 240 christine moll-murata which was written by Lan Pu 藍浦 (d. 1795). It is conceivable that this book was used as a textbook in the Academy. Useful as it is for an overview of the historical development and especially matters relating to the Imperial Manufacture and for the identification of historical porcelain, it can hard- ly be considered a textbook for producing porcelain. Set directly against a comparable contemporary work from Delft, De Plateelbakker by Gerrit Paape (1752-1803), the differences are striking. Paape, who had been ap- prenticed as a porcelain painter in Delft,57 after a short introduction in which he deplores the decline of the trade, provides detailed hands-on information on all steps of production from the preparation of the paste to the firing of the wares. For glazes and paints, he states the exact reci- pes—quite unlike Jingdezhen taolu. Thus, although the Porcelain Gazetteer presents the Jingyang Academy as a place for professional learning, we may assume that it served for the formation of merchants and entrepreneurs, for overseeing crafts rather than for directly instructing the craftspeople, who in most cases did not know how to read and write. We can imagine that it was intended for the merchants’ and elite porcelain artists’ offspring, who according to the ­Porcelain Gazetteer received a scholarship, and after graduation were ­expected to do practical studies in the Jingdezhen porcelain industry. In the last years of the Republic of China, a dozen of such historical guild house–academies existed in Jingdezhen, and served as official buildings such as printing shops, post offices, and schools.58

Delft: General Setting

In Europe, tin glazes and whitish wares were introduced by Islamic traders in Spain and in Italy since the thirteenth century. In Italy, it was adapted under the name of majolica, and produced in the city of Faenza, whence

57 Peter Altena, ‘Inleiding’, in Gerrit Paape, Mijne vrolijke wijsgeerte in mijne balling­schap [1792], web edition installed at Digitale bibliotheek van nederlandse letteren (DBNL), last accessed 2 December 2011. http://www.dbnl.org/tekst/paap004mijn01_01/paap- 004mijn01_01_0003.htm, 24. 58 They are identified in Jingdezhen taolu tushuo (Illustrated explanations on the Jing- dezhen porcelain kilns), by Lan Pu and Zheng Yangui, ed. by Lian Mian (Jinan: Shandong huabao chubanshe, 2004), 14, citing Fang Lili, ‘Chuantong yu bianqian—Jingdezhen xin jiu minyao ye kaocha’ (Tradition and change—an investigation of old and new private kilns in Jingdezhen), 264-65, and Jingdezhen wenshi ziliao (Literary and historical materials of Jingdezhen), No. 4, 190. The Hubei guild house has survived until today, see Cui Peng, ‘Jingdezhen huiguan yanjiu’ (A study of the Jingdezhen guild houses), Cangsang 2009/1, 141. guilds and apprenticeship in china and europe 241 the designation fayence. Majolica and fayence are used more or less syn- onymously to designate tin-glazed polychrome ceramics. This type of pot- tery was brought to North-western Europe by Italian migrants. In the sixteenth century, tin-glaze potteries were started in Antwerp, Haarlem, Delft, and Rotterdam.59 In 1600, two fayence potters were officially regis- tered in Delft.60 At about that time (1602), the Dutch East India Company VOC had ob- tained a monopoly on Dutch trade with Asia and started to import Chinese porcelain. Around 1620 the trade volume was about 100,000 pieces of por- celain per year, intended for a well-off group of consumers of perhaps 250,000 persons.61 The Delft fayence workshops reacted to the competition from China in various ways, producing objects which were not available from China, especially wall tiles, but also trying to imitate the Chinese product as closely as possible. By choosing a type of clay that turned out whitish when fired, and experimenting with a more transparent glaze, the products, which were fired at 1,000-1,100°C in a cylindrical kiln, became increasingly sophisticated and similar to the Chinese porcelain. The so- called Delftware or Dutch porcelain came to enjoy great popularity far beyond the borders of the Dutch Republic. Imports from China rose to 200,000 pieces in 1644, but thereafter de- clined rapidly. Due to wars and destruction at the end of the Ming dynasty and struggles for hegemony between the Manchu Qing and several con- tending factions, between 1654 and 1684, no Chinese porcelain could be imported.62 Japanese porcelain potters in Arita quickly imitated the Chi- nese motifs, and traded with the Dutch, but due to prices that were two or three times higher, these imports remained limited (1661: 11,500, 1663: 48,000, 1665: 65,000 items).63 For Delftware, the cataclysm of the Ming-Qing transition in the mid- seventeenth century proved a blessing, since the consumer demand for Chinese blue-and-white porcelain had already settled in, and the substitute

59 Blasczyk, ‘Porcelain for everyone’, 5-6. 60 Jan Daniel van Dam, Delffse Porceleyne: Dutch Delftware 1620-1850 (Amsterdam: Waanders, 2004), 11. For the changed consumer attitudes, demanding porcelain rather than the common wooden or metal tableware, see Blasczyk, ‘Porcelain for Everyone’, 2; Berg, Luxury and Pleasure in Eighteenth-century Britain (Oxford: Oxford UP, 2005), 130. 61 Van Dam, Delffse Porceleyne, 11. 62 Ibid., 28. 63 Ibid.; Christiaan Jörg, Porcelain and the Dutch China Trade (The Hague: Martinus Nijhoff, 1982), 93, warns that much more porcelain was imported unofficially from Japan than recorded by the VOC. 242 christine moll-murata product sold extremely well. In the years when the Japanese wares were the only porcelain available, the price of a porcelain plate or bowl was four to nine times higher than for an item of a majolica or fayence.64 Van Dam thus identifies 1647-1680 as a period of great increase of mass production in Delft, 1680-1700 as the twenty best years, 1700-1720 as still successful in a stagnating economy, 1720-1750 as a deep recession, but nev- ertheless containing technical and artistic peaks, and 1750-1853 as an un- expected revival and still productive period, which ended however in the decline of the industry and the closing down of the last shop in 1853. None- theless, only a few decades later antiquarian interest awoke, and since the 1880s copies of Delftware are being produced to this day by manufacturer De Porceleyne Fles. In 1913 De Porceleyne Fles was bestowed with the title ‘Royal Dutch Manufacture’, but continues to be a private enterprise. No part of the Delft production was ever financed by the state or the community.

Number of Enterprises and Size of Workforce The number of fayence workshops and kilns in Delft rose from two in 1600 to 21 in 1660 and 31 in 1700, but by 1815 all but five had been closed down.65 While exact data is available on the numbers of enterprises, estimates of the number of workers engaged in them are more speculative. A range of about 40-50 workers per pottery has been estimated,66 but this may have varied from place to place and according to the customer demand situation. If during the heyday of the trade between 1680 and 1720 an average of 40 workers were engaged in the trade, the branch directly employed about 1200 people altogether. The annual output of a fayence manufacture in Delft is estimated at 20-30,000 painted items.67 If the maximum annual imports were 200,000 items before the porcelain trade was interrupted by the fall of the Ming, that means that quite a number of Dutch workshops could hope to find a

64 Van Dam, Delffse Porceleyne, 29. 65 For a comparison of the smaller extent of the Delft earthenware potteries and their relation to the Delftware workshops, see Marie-Cornélie Roodenburg, De Delftse pottenbak- kersnering in de Gouden Eeuw (1575-1675): De produktie van rood pottengoed (Hilversum: Verloren 1993), 30. 66 Thera Wijsenbeek-Olthuis, Achter de gevels van Delft: Bezit en bestaan van rijk en arm in een periode van achteruitgang (1700-1800) (Hilversum: Verloren 1987), 71-72. Gerrit Paape, De Plateelbakker of Delftsch Aardewerkmaaker (Dordrecht: Blussé 1794), 2, reckons that the decrease by 20 of the number of fayence producers, caused about five hundred household members to lose their incomes, thus about 25 persons per pottery enterprise. 67 Van Dam, Delffse Porceleyne, 31. guilds and apprenticeship in china and europe 243 profitable sales market in this substitute industry. Therefore their number proliferated. Within Holland, Delft acquired a quasi-monopoly position for fine painted wares. As the production process became more complex, this led to the foundation of specialised supplier businesses, such as earth wash- ers and finishers, and glaze mills. The main reason for the decline of delftware according to Van Dam was the competition from Chinese producers after the situation at the Jingde- zhen production sites was consolidated in the late seventeenth century. Jingdezhen now provided not only blue-and-white, but also polychromes of the ‘’ type, which were applied after the first firing in over- glaze decoration. This new décor fashion had fascinated the Chinese em- perors from Kangxi (reg. 1662-1722) onward, when they had seen the European missionaries execute them in enamels on copper and in glass manufacturing at the palace,68 and it was introduced in Jingdezhen por- celain in the 1720s.69 During the period of Japanese substitution for Chinese porcelain, the Japanese had also delivered polychromes in the imari/Arita wares, which became quite popular in Europe. Finally, non-Dutch European competitors produced fayence as well as porcelain tableware. Although the market started to contract by the 1730s, there still was demand for high quality pieces, and some Delft manufactures adapted to the changed tastes in ornamentation by using underglaze blue and over- glaze enamel coloration.70 However, when after 1750 the prices of porcelain imported by the VOC fell, ‘the difference in price between Delftware and genuine porcelain became too marginal to secure the position of this imi- tation article on the international market’.71 The fact that business was poor was also due to the competition that arose from porcelain and stone- ware production on the Continent and in England.72 As Delft historians have observed, the decline of the city in the eighteenth century was gen- eral and extended across all industries; people were leaving the city, as the

68 Kerr et al., Ceramic Technology, 638-44. Kerr remarks that although the fashion and the impetus came from the palace and the missionaries, the technique may have been known in China earlier. This adds complexity to the issue of whether ‘famille rose’ enamel colours were a direct import from Europe or had been produced as an adaption of previ- ously known Roman and Central Asian techniques (Kerr et al., Ceramic Technology, 638-39). 69 Kerr et al., Ceramic Technology, 634. 70 Van Dam, Delffse Porceleyne, 136, 140. 71 Wijsenbeek-Olthuis, Achter de gevels van Delft, 339. 72 Jan de Vries, A.M. van der Woude, The first modern economy: Success, failure, and perseverance of the Dutch economy, 1500-1815 (Cambridge: Cambridge University Press, 1997), 309. 244 christine moll-murata

Table 7.3. Fayence potteries and the workforce of Delft. Year Registered potteries Workforce (persons) Source/Remarks 1581 1 Date of registration in the St. Lucas guild 1600 2 (3) Montias, p. 54, citing the Haardsteeden Register has 3. 1611 6 50 Workforce 9-10 persons per pottery 1660 21 1700 31 1,200 If 40 workers on average per pottery 1768 4,000 men, women, Altena, p. 24, citing Joseph and children Marshall ‘7,000 in better times’ 1815 5 1853 1 Closed down 18 Source: Roodenburg, 30.

Table 7.4. Population of Delft, Delfshaven and its jurisdiction, 1600-1795. Year Persons Remarks 1600 18,750 1622 22,769 1633 23,500 1680 24,000 Delft only (without Delfshaven) 1733 17,100 1749 13,910 Delft only (without Delfshaven) 1795 16,339 Source: Wijsenbeek-Olthuis, 27 population figures show. By the end of the eighteenth century, the once busy industrial town Delft had become a quiet place, popular with the af- fluent who appreciated its cleanliness and relatively sparse population.73 After the 1790s, it once again became a trading centre, but did not regain its previous industrial profile.

73 Ibid., 82-83. guilds and apprenticeship in china and europe 245 Guilds and Apprenticeship in Delft

The Delft fayence industry and its corporative structure was quite distinct from that in Jingdezhen. The scope of enterprises, as can be seen from the population figures and the number of enterprises and persons engaged in the trade, are of different dimensions. The industry in Delft started later and came to a halt earlier than Jingdezhen, even accounting for the fact that Delft, like Jingdezhen, started with the production of ceramics earlier than its more famous blue-white and subsequent polychrome phases. From the outset, larger ‘proto-capitalist’ enterprises had been established in both places; however, in the period of decline, the Delft development was one of merging in a situation of population decrease, while in Jingdezhen, spe- cialisation and labour fragmentation took place as the migrant population increased. Due to these factors, guild structures also differed between Jingdezhen and Delft. The most significant was that Delft had one large guild, named after the patron saint of the painters, St. Luke, not only for all Delft fayenc- ers, but also for artists and a range of other arts and crafts trades, such as painters, sculptors, art dealers, printers, booksellers, architects, stone carv- ers, glassmakers, furniture makers, embroiderers, and a variety of other professions.74 The accounts of this guild are extant from 1537,75 but the preserved register of its masters starts in 1613 and ends in 1714.76 The guild was dissolved in 1833.77 Its main function was to enforce the monopoly of production and trade that tried to exclude other cities and non-guilded

74 Kaldenbach, who analysed the master books of the guild, mentions ‘fine art painters, painting dealers, Delftware faience manufacturers, printers, engravers, sculptors in wood— also sculptors in stone or other fine substances such as ivory. Furthermore we find included within this guild: architects, tapestry weavers, embroiderers, glass makers, glass painters, glass sellers, engineers, surveyors, mapmakers, map coloring specialists, calligraphers, typeface makers, printers, book binders, and as the proverbial odd ones out, chair painters, and a furniture joiner. In Delft the goldsmiths were grouped in a separate guild’: Kees Kaldenbach, Fully searchable, complete text of the 1877 Obreen publication of the Guild Book of Delft Master Painters, Engravers, Sculptors, Potters etc. in the seventeenth century. Anno- tated internet version, 2002, last update October 2007. http://www.xs4all.nl/~kalden/ dart/d-a-ab-obreen-lucasgilde1.htm. Accessed March 2013. 75 Roodenburg, De Delftse pottenbakkersnering, 57. 76 Kaldenbach, Fully searchable, complete text of the 1877 Obreen publication. 77 The Saint Luke’s Guild of Delft, website essentialvermeer.com, maintained by Jona- than Janson, http:// www.essentialvermeer.com/saint_luke’s_guild_delft.html, last updated April 2008, accessed June 2008. For the several acts of abolition of guilds since 1798 and the Northern Dutch resistance against it until 1818, see Bert De Munck, Piet Lourens, and Jan Lucassen, ‘The Establishment and Distribution of Craft Guilds in the Low Countries’, in: Prak, Lis, Lucassen and Soly (eds), Craft Guilds in the Early Modern Low Countries, 63-64. 246 christine moll-murata competitors, and to control the training of the workforce, the quality and the prices of the products. Masters who wanted to produce and trade in Delft were required to become citizens of the town, and were registered after paying an entrance fee. Apprentices and journeymen, although not full members, were also registered and monitored by the guild. From indicators such as the prices of real estate property, taxes, and death donations, Michael Montias has divided guild members and depen- dents into four large groups. Fayencers could be found in all of these cat- egories. Most affluent and ‘the closest thing to capitalists in the guild’ were the owners of fayence works, who had ‘both large assets and liabilities arising from their heavy needs for capital’.78 Next came the less successful artists: fayencers not established on their own, housepainters, glassmakers, most sculptors and embroiderers, and printers. Montias places in the third rank the ‘artists and artisans not registered in the guild, but who by virtue of their education and training could aspire to master status.’ This includes apprentices and journeymen not established on their own. Finally, the least well-off were the apprentices and journeymen, mostly of the delftware industry, who ‘may be said to belong to the proletariat of Delft’. According to a household register of Delft, by 1600 a total of 88 house- holds of artists and artisans were registered in Delft, three of whom were fayencers. The number is probably understated, but it can give an impres- sion of the scope in comparison to Jingdezhen. The first Delft fayencer mentioned in that register, Hermann Pietersz., joined the guild of St. Luke in 1581.79 In 1611, six fayence potteries were active, with a workforce of about nine to ten people each, thus about fifty people were in the trade.80 The guild of St. Luke took in as members the masters who owned their shops and those who worked for others, sometimes for entrepreneurs who were not specialists in the trade. Since a delftware pottery enterprise, quite different from a Jingdezhen workshop, typically covered the entire produc- tion process, from forming to firing and decorating, much more capital was necessary to set up and maintain an enterprise of this kind. Montias has observed that in the initial phase, from the 1630s to the 1650s, few masters registered with the guild, and those who did operated their own potteries. In 1650, out of fourteen active master fayencers, thirteen owned a pottery of their own or with partners. However, in 1670, the relation was 48

78 John Michael Montias, Artists and artisans in Delft: A socio-economic study of the seventeenth century (Princeton, NJ: Princeton UP, 1982), 131. 79 Ibid., 58. 80 Ibid., 65. guilds and apprenticeship in china and europe 247 registered master fayencers against 28 owners or part-owners of potteries.81 Apprenticeship was regulated by the guild. In this sector, we find an arena where the guild cooperated with another municipal institution, the ‘Charity Chamber’ (Camer van Charitate). This welfare organisation ­arranged apprenticeships for orphans or children from poor families. From its ‘register of craft apprentices’ we learn of a boy who in 1612 was appren- ticed at age nine.82 Apprenticeship contracts reveal more details about the duration, which was four years, the fact that the apprentice was remu- nerated, but that board and lodging was not provided, and of the binding force of the contract. For instance, in 1607, the earliest established Delft fayencer Hermann Pietersz. took on Abram Gillisz. as an apprentice to learn the craft of fayence-making for four years. ‘In the first three years, he will learn to make dishes or plates, large or small. If Abram Gillisz. proves apt at this work, his employer will at his discretion teach him potting. During the four years, the apprentice will be paid 9 stuivers per day, board and lodging not provided. He must not quit his master’s employment for any reason, including marriage. In case his employer should die, he will remain in the service of his widow if she undertakes to continue the business’.83 A contract from 1611 shows that in comparison to Jingdezhen, produc- tion was far less divided. This contract stresses the indentured nature of the relationship, but also makes provisions in case the master could not keep the apprentice occupied. The apprentice in question was much older than the orphan boy of the first case: On March 11, 1611, Abraham Davidsz. took on the 19-year-old Jan Bartholomeusz. as an apprentice for four years. Jan was to learn ‘how to pot, load the ovens, prepare lead and tin glazes, and also help to put on colour and cover and tread the clay.’ His wage was 8 stuivers a day for the first two years and 10 stuivers a day the last two. He was not supposed to work for any other master ‘by day or night’, but if Abraham could not provide him with work, Jan would be allowed to earn a regular wage elsewhere. After expiration of the four years the apprentice was to stay in Abraham’s pottery and should be paid the same wages ‘as other masters give’.84 Montias plausibly argues that the fact that apprentices were paid and bound to their employer shows the scarceness of labour in this expansion

81 Ibid., 105. 82 Ibid., 66, 117. 83 Ibid., 67. 84 Ibid. 248 christine moll-murata phase of the industry.85 Considering the division of labour in Jingdezhen in the nineteenth and early twentieth century, the wide range of skills that the Dutch master was prepared to teach his apprentice also points to this situation. The guild regulations initially (1614) limited the number of new apprentices to be taken on at a time to two, and if apprentices were ac- cepted in one year, no new boys should be engaged in the next,86 but this was changed in 1620. Fast mobility, or in Epstein’s words, ‘opportunism’87 was discouraged by the provision (1658) that if external journeymen or servants were engaged, they should have been working for their previous master for at least six years. Finally, in an investigation of the literacy of masters and workers of the guild of St. Luke based on the members’ ability to sign their names, Montias found that between 1613 and 1675, among the fayence workers, almost two- thirds were illiterate. The relation of those who signed in a clear signature, clumsy signature, or only with a mark was 61 : 22 : 101.88 Those who could not read and write mostly belonged to the social categories three and four outlined above. Could ‘total institutions’ substitute for the guilds in Delft? The answer is negative, since the ‘total institution’ of the orphanage on the contrary collaborated with the guild. For the question of gender discrimination, we find that especially in the case of the Delft fayence industry, women who were represented in the guild as widows could be prominent entrepre- neurs, engaged in trade, and developing new techniques and design fash- ions.89 We do not know for certain how gender discrimination applied on the level of the craftspeople. Most probably women’s wages were lower than those of their male colleagues. For painters, Kaldenbach suggests that women produced paintings in the lower price range, although one of them

85 Ibid., 69. The situation was quite different with the painters. Ibid., 118, estimates an apprentice fee of 20 to 50 guilders per year for a boy living with his parents, and 50 to 110 guilders for those living with their masters. Other costly apprenticeships were embroider- ing, glassmaking, and printing. The Charity Chamber did not send their subsidised charges to such masters, but rather to fayencers, furniture makers and stone carvers (Ibid., 116). 86 Ibid., 95. 87 Epstein, ‘Craft Guilds, Apprenticeship’, 687. 88 Montias, Artists and Artisans in Delft, 116, table 5.1. 89 See Van Dam, Delffse Porceleyne, 62, for the technique of baking overglaze decora- tions in Het Moriaenshooft, which was launched by the proprietress Jannetje van Straeten. The widow Johanna van der Heul set the trend for imitation of Japanese Imari styles in her pottery De Grieksche A. In 1713 she made a contract with four fayencer journeymen that they should work for her exclusively as long as she was in charge of the company. (Ibid., 124-25). guilds and apprenticeship in china and europe 249 ‘got highly paid commissions’.90 Obreen’s list of the masters of the guild of St. Luke includes two widows and one daughter of owners of fayence pot- teries shops who had themselves registered as successor to their husbands’ or fathers’ heritage, but also two independent women’s names.91 Although these are clearly the exceptions, at least some women could register and be represented in the guild.

Starting from Scratch: The Meissen and Vienna Porcelain Manufactures

In his presentation for the 2006 GEHN conference on ‘The Return of the Guilds’, which is included in chapter two of this book, Epstein outlined a roadmap for approaches to guild studies and the issues he thought of as most important. In this presentation, he had also considered the alternative non-guild modes of transmitting skills and circulating labour—namely the family, rural proto-industry, ‘total institutions’ such as hospitals, asylums, and prisons, and centralised (manu)factories.92 Turning from the substitute product of delftware to the first two manu- factures of genuine hard-paste porcelain in Europe, we find them in a com- pletely different, in Epstein’s sense ‘alternative’ organisational setting than that of the guilded Jingdezhen and Delft ceramics industries. This is be- cause these were isolated state-financed enterprises without private com- petitors in their immediate environments. Moreover, they stood at the beginning of a scientific investigation of the chemical ingredients and properties of porcelain. At that time, the necessary raw materials and pro- cedures were experimentally determined in Dresden and Meissen, but scientific research and development was also carried out in the other cen- tres of European porcelain production. After the composition of the paste and the necessary temperatures for firing had been ascertained, the Meissen Manufacture was set up in 1710.

90 Kaldenbach, Fully searchable, complete text of the 1877 Obreen publication, ‘Note on Female Painters’. 91 Ibid., shows that two widows (of Meijnaert Garrebrantsz. and Gerrit Ecberssoon), and an unnamed master’s daughter in Delfshaven, but also two women without mention of deceased husband or father, Annetge van den Rijsheuvel (25.02.1667) and Barbara Rot- teveel (31.03.1671), the last one expressly designated as owner of her fayence pottery ‘De 3 Klocken’, registered in the St. Lucas Guild. On the women as shop managers, see Kees Kaldenbach, ‘How did they produce Delft blue faience or ‘Delft porcelain’ in the delftware potteries?’ http://www.xs4all.nl/~kalden/ Accessed June 2008. 92 S.R. Epstein, ‘Innovation’, Powerpoint presentation at the GEHN conference ‘The Return of the Guilds’, Oct. 2006, Utrecht University, available at the website of the Interna- tional Institute for Social History, http://iisg.nl/hpw/papers/guilds-epstein.pdf. 250 christine moll-murata

Table 7.5. Workforce at the Meissen Manufacture, 1765-1880. Year Employed Persons Persons Persons Number of per- persons employed in employed in employed in sons employed in production the artistic the technical technical depart- departments departments ments vs. 100 per- sons in artistic departments 1765 731 629 441 188 42.6 1775 611 503 355 148 41.7 1780 564 468 318 150 47.2 1790 524 429 272 157 57.7 1800 537 426 247 179 72.4 1805 521 412 247 165 66.8 1810 435 344 202 142 70.3 1814 402 307 167 140 83.8 1815 403 304 165 139 84.2 1820 418 317 193 124 64.2 1825 358 277 185 92 49.7 1833 358 256 188 68 36.2 1838 391 292 206 86 41.7 1840 378 278 195 83 42.6 1850 326 239 156 83 52.3 1855 336 260 194 66 34 1865 334 259 207 52 15.1 1870 396 320 262 58 22.1 1875 667 571 483 88 18.2 1879 607 520 434 86 19.1 1880 655 563 477 86 18

Source: Scholz, p. 74, table 3.

A few years later, the production secret leaked out and was transmitted to Vienna, in the competing Habsburg domain. The Military Councillor Claudius Innocenz du Pasquier who had organised this act of industrial espionage, in 1718 established a manufacture in Vienna. It was granted a sales and production monopoly, but initially received no further financial advantages from the Habsburg monarchy.93 In 1744 it was bought by the state and operated as an Imperial Manufacture until it was closed down in 1864.94

93 Jacob von Falke, Die K.K. Wiener Porzellanfabrik: Ihre Geschichte und die Sammlung ihrer Arbeiten im K.K. Österreichischen Museum (Vienna: Carl Gerold’s Sohn, 1887), 5-9. 94 Ibid., 44. guilds and apprenticeship in china and europe 251

Table 7.6. Workforce at the Vienna Manufacture. Year Workers Remarks 1744 20 1750 40 1761 140 1770 200 1780 320 1783 316 Folnesics, p. 55 1784 280 1799 500 35 kilns, six to seven firings per day 1800 branch factory at the clay production site Engelhardtzell near Passau: 60 work- ers, 7 kilns, one firing per day 1828 273 72 painters, 1 former (Modellierer), 7 Bossirer, 8 blue-and white painters, 41 turners 1829 252 (entire personnel, Folnesics, p. 138 incl. of administra- tion) 1830 151 50 painters, 6 blue-and-white painters 1864 Closed down

Source: Falke, 11, 22, 38; Folnesics and Braun, 55, 138.

How were skills trained and transmitted in this alternative model? With maximum workforces of five hundred (Vienna 1799), and 629 (Meissen 1765), the number of people to be trained was significant and came close to, or even surpassed that of a ceramics industry of a middle-sized, guilded European town. From Meissen, we know more about the artistic than the technical side of production. Art schools that taught design and sculpture provided for a six-year apprenticeship. They were directed by the leading artists of the manufacture, and later by professors of the Dresden Academy of Arts. After 1764, boys as well as girls were accepted. They received a scholarship and could increase their incomes by sales of their products.95 Apprentices’ performances were formally examined every half year. As a rule, they were permanently engaged after the apprenticeship. The direc- tors were convinced that this was better than to engage painters from out- side.96 In the same way, when new positions were open, preference was

95 Gröger, ‘Die Arbeits- und Sozialverhältnisse, 171, 179-83. 96 Traute Scholz, ‘Produktivkraftentwicklung, Arbeitskräftestruktur und betriebliche Lohnarbeitsverhältnisse in der Porzellanmanufaktur Meissen im 18. und 19. Jahrhundert’, 252 christine moll-murata given to children and relatives of the workers at the Manufacture.97 A schoolteacher was engaged for the elementary education of the workers’ children. Wages were high and social benefits such as old-age and invalid- ity insurances were provided. Last but not least, the image of working in a luxury production sector in the direct service of the ruler gave these work- ers an elite status that they were well aware of. In the Vienna Manufacture, the situation was comparable insofar as it also operated a school for painting and sculpture.98 Some of the appren- tices, as in Delft, came from the orphanages—in 1752 seven boys, two of whom were sent with the sponsorship of the Grand Duchy of Liechten- stein.99 However, when the Manufacture tried to obtain an apprentice fee of 18 Rheinische Gulden from the orphanages and the Grand Duchess of Liechtenstein, this proved to be so difficult that by 1777, state-sponsored apprentices were no longer taken in.100 The period of apprenticeship was six or seven years, and after finishing their term, the young painters did not receive full wages for another three years, or in an unusually well-docu- mented case in 1782, painters, turners, and formers could reach full masters’ wages fifteen years after they had started the apprenticeship.101 If the new employees proved unqualified after three years, they were dismissed. Due to the long apprenticeship term, the school preferred apprentice boys of about ten years old.102 Besides professional training, pupils were also taught elementary reading and writing skills and, after 1752, were placed under the custody of a priest who was to look after them outside their working hours.103 There was an active infusion of skills from outside. The imperial court initiated and fostered this introduction of artists and specialists for raw materials and their preparation. Since the early nineteenth century, the directors of the Vienna Manufacture had changed from artists and proto- industrialists into natural scientists.104 Economic efficiency rather than

Jahrbuch für Wirtschaftsgeschichte 1981/2 (Beiträge zur Geschichte der Sozialstruktur im 19. und 20. Jahrhundert), 101. 97 Ibid., 85-86. 98 Falke, Die K.K. Wiener Porzellanfabrik, 24. 99 Helene Bruder, Arbeitsverhältnisse in der K.K. Wiener Porzellanfabrik zur Zeit Maria Theresias. Unpublished PhDthesis, Universität Wien, 1932, 21. 100 Ibid., 22-23. 101 Ibid., 24. 102 Ibid., 24. 103 Ibid., 25. 104 Falke, Die K.K. Wiener Porzellanfabrik, 38, J. Folnesics and E.W. Braun, Geschichte der K.K. Wiener Porzellanmanufaktur (Vienna: Verlag der K.K. Hof- und Staatsdruckerei, 1907), 137. guilds and apprenticeship in china and europe 253 artistic excellence was now in demand. First experiments with steam en- gines were carried out, and the long kilns were replaced by more efficient round kilns. Taking into consideration that the Bohemian private ceramics industry was catching up and could produce more cheaply than the ­Vienna Manufacture, the directors expressed the hope that their institute could serve as a model institution for the private industry—comparable to the French state manufacture at Sèvres. In any case, they tried to at least break even and cause no extra expenses. However, they failed to achieve this and as a result the factory was closed down in 1864. In contrast with the situation in continental Europe, British porcelain manufacturers developed in an environment that was largely free from state sponsorship or control. The classical studies, especially those on the Staffordshire potteries by John Thomas, and on the Worcester manufacture by Lorna Weatherill, all describe how entrepreneurs acted independent- ly.105 Sie Wun-Ci analyses in great detail the entrepreneurial differences between the Staffordshire Potteries and Jingdezhen, and arrives at a quite negative assessment of the role of guilds in China. We will not contrast the British case with Jingdezhen, since the structural similarity (state control) in the continental manufactures lends itself better for comparison of guild production. As for distribution, which is highlighted in Sie’s thesis, the similarities are greater.

Conclusion

The Jingdezhen guilds and common-origin associations have had a poor image in the historiography in China and abroad. They were made partly responsible for the fact that the Jingdezhen industry did not develop and modernise in the twentieth century.106 Moreover, their precarious status as representatives of the interests of workers from certain regions, sanc- tioned by age-old usage and customary law rather than by official acknowl-

105 Ray Jones, Porcelain in Worcester 1751–1951: An illustrated social history (Hallow: Parkbarn, 1993); Lorna Weatherill, Pottery trade and Northern Staffordshire 1660-1760 (Man- chester: Manchester UP, 1971). 106 Jiang Siqing, Jingdezhen ciye shi (History of the Jingdezhen ceramics industry). Zhonghua shuju, 185; Kerr et al., Ceramic Technology, 771. For a generally skeptical assess- ment of the role of the Peking guilds in apprenticeship, see Zhang Li-hua, ‘Aspects of the Apprenticeship System of Craft and Trade Guilds in Beijing During the late Qing and early Republican Era’, in: Ranson, Epstein et al. (eds), Preliminary study of craft guilds in China, 52-61. 254 christine moll-murata edgement, made their existence difficult and forced them, especially since the 1930s, into a semi-legal position of secret societies, officially for the veneration of a common patron saint, but with criminal potential. These negative aspects may have shaped the overall assessment of their roles in earlier times. It is only recently that the customary regulations are per- ceived in a more positive, even nostalgic light.107 What makes the evaluation of their economic function difficult is that the textual evidence for their activities is scarcer than for European guilds. Liang Miaotai argues that the Chinese guilds were not as politically influ- ential as European guilds, and did not participate in any kind of municipal self-government. Although their control was (or could be) strictly enforced, their rules were not often formulated in writing, and therefore potentially more flexible.108 We should qualify this view to the extent that some Chi- nese cities, such as Beijing, Suzhou, Shanghai, and Hankou, have a much richer record of guild materials, and at least for Hankou, the importance of the guilds for informal city administration has been demonstrated.109 Liang Miaotai’s other argument for explaining the character of the Jingde- zhen guilds is that for centuries, the Jingdezhen porcelain industry was a market leader, and therefore the common-origin guilds were not worried about securing their share of the sales market. They rather concentrated on retaining their share of the labour market, since replacement of certain common-origin groups by others had occurred previously and led to violent conflict.110 Migrants had come to Jingdezhen since the fifteenth century, first from Raozhou and Leping, and also from the Wuyuan and Qimen districts in adjacent Anhui province, but later Duchang population pre- vailed. Conflicts had emerged in the mid-sixteenth century, resulting even in killings between Leping and local Fuliang (Jingdezhen) groups, and in the late sixteenth and early seventeenth centuries between Duchang and Fuliang people, who felt threatened by what they saw as a takeover by

107 ‘Jingdezhen de fengqing yu meili de zhuanshuo’ (Jingdezhen customs and beautiful legends), Website Sohu.com, http://auto.sohu.com/20050727/n240199132.shtml, installed July 2005, accessed March 2013; ‘Guojia ji fei wuzhi wenhua yichan daibiao zuo—Jingdezhen taoye xisu xiangmu jieshao (A representative case of immaterial cultural heritage on the national level: Introduction to the project ‘Customs in the Jingdezhen porcelain trade’), on the website Jiangxi sheng wenhuating (Department of Culture of Jiangxi Province), http:// www.jxwh.gov.cn/whzt/whzt4_4_6_10.asp, installed October 2007, accessed March 2013. 108 Liang Miaotai, Ming Qing Jingdezhen, 221. 109 William T. Rowe, Hankow: Commerce and Society in a Chinese City, 1796-1889 (Stan- ford: Stanford UP, 1984), 289-321. 110 Liang Miaotai, Ming Qing Jingdezhen, 221. guilds and apprenticeship in china and europe 255 strangers.111 Similar riots are also reported from the 1930s, at a time when the market for Jingdezhen goods had contracted.112 Thus, one important characteristic and function of the Chinese common-origin guilds for mi- grant workers was to protect the interests of their constituencies. Was that effective and beneficial for the economy? Or was this particularistic and deleterious to the advancement of Chinese industry? If these guilds functioned well, they had the positive effect of accom- modating the newcomers—who, this needs to be emphasised, came from relatively nearby regions. What has recently been argued in defense of the Jingdezhen guilds is that although they were still ‘feudal’ in structure, they at least eased the transmission of skills not only within the narrow confines of family and kin, but also to outsiders from the same region.113 Younger Chinese scholars such as Su Yongming and Huang Zhifan, while still insist- ing on the impeding character of the corporate organisation for modernisation,114 point to the positive effect of the guilds for peacekeep- ing, maintaining civilian order and the upkeep of the city infrastructure.115 In comparison with Delft, we find in China less clearly formulated ap- prenticeship contracts and also of a much more recent date, less formal recognition of the guilds by municipal governments, and in general less written documentation of their activities. However, the Jingdezhen indus- try and its guilds proved more enduring than the delftware substitute pro- duction. Compared with the non-guilded porcelain manufactures of Meissen and Vienna, there are more similarities: the Jingdezhen Imperial Manufacture also had an exemplary function, which means that first-class design and execution were strictly maintained. If anything, skill transmis- sion, especially from the state to the private sector, was easier and more effectively promoted by the government and the officials of the Imperial Manufacture. Finally, in this industry, which was first developed in China, the labour-intensive path was continued for a long time until innovations

111 Ibid., 223-224. Kerr et al., Ceramic Technology, 210 112 Hans-Wilm Schütte, ‘Perfektion als Hemmschuh? Anpassungsprobleme im chine- sischen Handwerk der späten Qing- und der Republikzeit’, in: Bernd Eberstein and Brunhild Staiger (eds), China: Wege in die Welt. Festschrift für Wolfgang Franke zum 80. Geburtstag (Hamburg: Institut für Asienkunde, 1992), 61–80, 73. 113 Liang Miaotai, Ming Qing Jingdezhen, 219. 114 Su and Huang, ‘Hangbang yu Qingdai Jingdezhen chengshi shehui’, 87. 115 Su and Huang, ‘Hangbang yu Qingdai Jingdezhen chengshi shehui’; see also Duchang ren zai Jingdezhen, 20-21, which specifies the public tasks of the Duchang guild house for maintaining a public ferry and a certain level of public hygiene, disposal of waste and corpses, poverty relief, and establishing schools. 256 christine moll-murata like coal-, oil-, and gas-fired kilns were finally introduced in the 1950s.116 In contrast, European manufacture of the new ‘synthetic’ material porcelain emphasised scientific methods, and since the nineteenth century, partly mechanised production and more economic and capital intensive methods of firing with fossil fuels.117 Contrasting the ‘guild-rehabilitationist’ and the ‘guild-critical’ posi- tions—in my view an absolute and static view of the entire guild-land­scape —is difficult to defend, and should be supplanted by a relative and dy- namic one. Between the critical perspective that blames the European guilds for ‘adversely affect[ing] quality, skill, innovation and economic policy’,118 and the more positive, ‘rehabilitationist’ view that insists on their economic and social efficiency,119 we find arguments supporting both propositions. In China, a more positive view of the corporations relates to the appraisal of their civic roles. Arguably, they also channelled and pro- vided security for the mobility of crafts people, and especially for mer- chants. However, the Chinese record does not yield information implying a particular concern of the guilds to ensure the transmission of skills to the next generation for the sake of the trade or the technology. Rather, the upkeep of the livelihood of the masters seemed to be the issue. This should be considered against the backdrop that the reports of craftspeople inter- viewed in the 1960s were influenced by the desire to criticise rather than neutrally appraise the contributions of the guilds. Taking up more directly the Epstein-Ogilivie controversy with regard to the area of skill formation, and acknowledging that it is difficult to transpose an European argument with all its different cultural underpinnings to China, the guilds in Jingde- zhen seemed less interested in fostering and evaluating the abilities of the young craftspeople, and reduced competition by restricting numbers of apprentices in every trade. This view is provisional insofar as the emergence of more textual evidence on guild activities, the unfolding of Chinese cap- italism, and the erosion of Marxist positions critical of guilds may all affect our current ideas about Chinese guilds in the future.

116 Kerr et al., Ceramic Technology, 368-69. 117 In Meissen, since 1865 human and animal labour for preparing the paste was entirely substituted with water power, and since 1872 with steam power. The first steam turbine for electrical power was built in 1904. From 1839 onward, the kilns were fired with coal. See Scholz, ‘Produktivkraftentwicklung’, 62-63. In English kilns for stoneware, coal was used already in the eighteenth century, but according to Kerr et al., Ceramic Technology, 768, for the first true hard-paste porcelain fired by William Cookworthy in the 1740s, wood firing was necessary since ‘coal stained the wares with sulphur’. 118 Ogilvie, ‘Rehabilitating the Guilds, 175. 119 Epstein, ‘Craft Guilds in the Pre-modern Economy, 171. guilds and apprenticeship in china and europe 257

Appendix

Figure 7.1. Jiangxi Province with Jingdezhen and Home Places of the Common-origin Guilds. Source: http://www.maps-of-china.com/jiangxi-s-ow.shtml. Note: Provincial borders should be ignored, as they were subject to change over time. 258 christine moll-murata labour relations, efficiency and the great divergence 259

CHAPTER EIGHT

LABOUR RELATIONS, EFFICIENCY AND THE GREAT DIVERGENCE: COMPARING PRE-INDUSTRIAL BRICKMAKING ACROSS EURASIA, 1500-2000

Gijs Kessler and Jan Lucassen

One of the main challenges in the emerging discipline of world history has been to find an explanation for the so-called Great Divergence—the wid- ening gap in technology, productivity and living standards between the West and its settler colonies, and the rest of the world, in particular Asia, after roughly 1800.1 Whereas attention initially focused on the proximate factors explaining the different historical trajectories of the ‘West’ and the ‘Rest’, such as technology, the rise of capitalism, and the intertwinement of Western military and commercial power, recent scholarship has wid- ened the scope of the debate to the question why these factors emerged in one part of the world, but not in the other.2 Arguments have been put forward in support of human capital forma- tion as a crucial variable. S.R. Epstein, in particular, focused on the social mechanisms regulating the intra- and intergenerational transfer of occu- pational skills, and has attributed an important role in this transmission to craft guilds, at least in Europe, an idea which has recently been expand- ed upon by scholars studying guilds and guild-like phenomena in Asia and other places.3

1 The term Great Divergence has been coined by Kenneth Pomeranz in his seminal work The great divergence: China, Europe, and the making of the modern world economy (Princeton, N.J: Princeton University Press, 2000). 2 Joel Mokyr, The Gifts of Athena: Historical Origins of the Knowledge Economy (Princ- eton: Princeton University Press, 2002); Karel Davids, The rise and decline of Dutch techno- logical leadership technology, economy and culture in the Netherlands, 1350-1800, 2 vols. (Leiden: Brill, 2008). 3 S.R. Epstein, ‘Craft Guilds, Apprenticeship and Technological Change in Pre-industrial Europe’, in: S.R. Epstein and Maarten Prak (eds), Guilds, Innovation, and the European Economy, 1400-1800 (Cambridge: CUP, 2008), 52-80 [reprinted from The Journal of Economic History 58 (1998) 684-713]; S.R. Epstein and Maarten Prak ‘Introduction: Guilds, Innovation, and the European Economy, 1400-1800’, in: Epstein and Prak (eds), Guilds, Innovation, and the European Economy, 1400-1800 (Cambridge: CUP, 2008), 1-24.; Maarten Prak, Catharina Lis, Jan Lucassen, and Hugo Soly (eds), Craft Guilds in the Early Modern Low Countries: Work, 260 gijs kessler and jan lucassen A more explicit argument on behalf of differential trajectories of human capital formation in East and West has recently been made by Gregory Clark.4 Reviewing existing explanations for the Great Divergence and find- ing them insufficient to fully account for the differential development of the West and the Rest, Clark offers a rival explanation, which focuses on the quality and the efficiency of the labour force as the single most impor- tant determinant of differences in industrial labour productivity. Outside the Western core and its settler colonies the introduction of industrial technologies failed to improve labour productivity because labour force characteristics differed from those required by the new production tech- nologies. The root cause of these differences, Clark argues, lie in the pre- modern and pre-industrial period, when work ethics and the social organisation of production in Europe came to exhibit a peculiar set of characteristics absent in the rest of the world. Crucial among these char- acteristics were high levels of labour discipline and well-developed habits of co-operation in production, i.e. the ability to effectively work together in a production setting. In this chapter we engage with Clark’s ideas in a comparative study of labour relations and the co-operation in production in pre-industrial brick- making in Western Europe, Russia and Northern India between 1500 and 2000. The principal question of the investigation is whether fundamental differences in labour force characteristics, work ethics and the co-operation in production can be observed between the cases under comparison, and, consequently, whether there is evidence, at this level of analysis, of differ- ent pre-industrial trajectories of human capital formation along the lines of Clark’s argument. Brickmaking is a particularly suitable sector of indus- try for such a comparative venture, firstly because it is indigenous in most places and secondly, because its production technique before mechaniza- tion (in Europe from 1900, in India starting hesitantly now) shows very little variation across the vast Eurasian land-mass (and indeed elsewhere). Differences in labour organisation and the co-operation in production can therefore be interpreted as indicative of local, non industry-specific cir- cumstances.

Power and Representation (Aldershot: Ashgate, 2006); Jan Lucassen, Tine De Moor, and Jan Luiten van Zanden, ‘The Return of the Guilds: Towards a Global History of the Guilds in Pre-Industrial Times’, International Review of Social History 53 Supplement (2008), 5-18. 4 Gregory Clark, A farewell to alms: A brief economic history of the world (Princeton, N.J.: Princeton UP, 2007). labour relations, efficiency and the great divergence 261 Clark’s case is built essentially on evidence for the industrial period. In his argument the differential productivity of Western-European and Asian industrial workforces at a given level of technology and capital endowment is caused by vastly different levels of efficiency, and, above all, of co-oper- ation in production to reduce error-rates between the consecutive stages in production processes. The distinction is that of labour discipline and the regular, repetitive and co-ordinated performance of certain tasks by European workers as against the perhaps equally hard but irregularly per- formed labour of workforces in other parts of the world, coming in bursts and bouts and with frequent intervals and disruptions.5 It is these differ- ences in what Clark calls ‘the quality of the labour force’, which explain why the introduction of industrial technology, capital and modern produc- tion methods outside the West-European core failed to reproduce the rise in productivity and levels of income which they had brought to the societ- ies of origin. Clark ascribes the emergence of these differences in the labour force characteristics to the pre-industrial period. For pre-industrial brickmaking, however, we find only minimal differences in the way the co-operation in production was organised in the three geographical areas under consider- ation. Although the organisation of production has varied both across time and space, the mechanisms of error-rate reduction adopted to ensure ef- ficiency within these different systems largely belong to the same reper- toire. Of course, brickmaking could have been exceptional in this respect and therefore further research on other sectors is necessary to produce more conclusive evidence, but so far our findings are not supportive of an interpretation that attributes the origins of the divergence in labour pro- ductivity between West and East to the pre-industrial period.

The Historical Dynamics of Co-operation in Production

Before we embark on our comparative venture, let us first address the main analytical issue at stake in some more detail. The focus of our investigation is the co-operation in production—the complementarity of efforts be-

5 Clark draws on examples from Western Europe and Asia. Most interestingly, though, lack of co-ordination of efforts between different stages in the production process, a highly irregular pace of work and ensuing high error-rates have also been singled out in the lit- erature as the main factors behind low industrial labour productivity in the Soviet Union, cf. Donald Filtzer, ‘Labor Discipline, the Use of Work Time, and the Decline of the Soviet System, 1928-1991’, International Labor and Working-Class History 50 (1996), 9. 262 gijs kessler and jan lucassen tween the various people engaged in a production process. Of crucial im- portance in determining the efficiency of this interaction in modern production is the reduction of ‘error-rates’. To quote Clark at length:6 Modern production technologies, developed in rich countries, are designed for labour forces that are disciplined, conscientious, and engaged. Products flow through many sets of hands, each one capable of destroying most of the value of the final output. Error rates by individual workers must be kept low to allow such processes to succeed. The introduction of such techniques in nineteenth century England was accompanied by greater attention to worker discipline. When workers in poorer countries lack these qualities of discipline and engagement, modern production systems are feasible only when little is demanded of each worker, to keep error rates as low as pos- sible. High error rates result in higher amounts of labour input per unit of output, and, hence, low overall labour productivity. The co-operation in production and error-rate reduction are of course issues confronting not just modern manufacturing, but any production process involving a division of labour. In modern production relations the collaboration of different specialists is achieved through direct management, training and supervision of the workforce by employers or their authorised representatives. Direct super- vision is essential in situations where time wages are paid, because only this way employers can be certain people actually work during the hours they are paid for. Also, the enforcement of labour discipline and supervi- sion are prerequisites for mechanisation, because it assures costly machin- ery will not be standing idle. Losses due to inefficiencies and errors are borne entirely by the employer. Historically, though, direct supervision is the exception rather than the rule. A form of indirect supervision which will feature prominently in this article is what David Frederick Schloss has called ‘co-operative sub­ contracting’.7 Co-operative (or collective) subcontractors usually combine more than one stage in the production process or the process in its en- tirety and are remunerated as a group per item of finished product. In such a set-up the organisation of the labour process is left to the collective and

6 Clark, Farewell to alms, 13. 7 Extensively discussed in Piet Lourens and Jan Lucassen, Arbeitswanderung und berufliche Spezialisierung: die lippischen Ziegler im 18. und 19. Jahrhundert, Studien zur his- torischen Migrationsforschung 6 (Osnabrück: Rasch, 1999), based a.o. on David Frederick Schloss, Methods of industrial remuneration (London: Williams and Norgate, 1898). See also Jan Lucassen, ‘Brickmakers in Western Europe (1700-1900) and Northern India (1800-2000): Some Comparisons’, in: Jan Lucassen (ed.), Global labour history: A state of the art, Interna- tional and Comparative Social History (Bern: Peter Lang, 2006), 513-71. labour relations, efficiency and the great divergence 263 the employer’s role is limited to the provision of the raw materials and production establishments at the input side and quality control on the finished product at the output side. In terms of Clark’s error-rate reduction co-operative subcontracting offers great benefits to the employer, because losses due to inefficient co-ordination and co-operation between workers are borne entirely by the collective in the form of a lower output and, hence, lower earnings. Workers’ own responsibility, on the other hand, to effec- tively organise their co-operation in production provides them with a great deal of autonomy and immunity from employers’ oppressive oversight. Mutual responsibility for the quality of the end-product, and therewith levels of remuneration, in other words, sets a premium on the minimisation of error-rates between the different members of the collective and the ef- fective application of available skills in the production process. The intra- and inter-generational transfer of occupational skills is accomplished through on-the-spot training, combined with differential redistribution of earnings according to skill. Often, subcontracting units are based on fam- ily groups, which further enhances mutual responsibility and efficient col- laboration, and facilitates training. Co-operative subcontracting has been extraordinarily widespread, but has eventually given way almost everywhere to direct hiring and modern production relations. The reasons for its disappearance have been insuf- ficiently elucidated, but since it tends to coincide with the advent of mech- anisation a causal link is strongly suggested. A crucial factor could have been the degree of control exercised by subcontractors over production processes, and their strong bargaining position in negotiating levels of remuneration which followed from this. Depending on the size of capital investments, and market requirements and demand, this could, at a certain point, have prompted employers-entrepreneurs to attempt to shift to direct, individual hiring, when this offered benefits in imposing levels of remuneration and enforcing adaptation to modern production techniques required by, for instance, mechanisation. At this point, however, minimi- sation of error rates becomes a matter of direct supervision and the main- tenance of labour discipline. These are the modern production relations which allow for the surge in productivity that, in Clark’s argument, powered the rise of the West but failed to be reproduced by the Rest. 264 gijs kessler and jan lucassen A Global, Comparative Perspective

Brick manufacturing is indigenous in many parts of the world, and basic production techniques exhibit a remarkable degree of uniformity across time and space. This study discusses the large-scale production of fired bricks—as opposed to sun-dried bricks—before the advent of mechanisa- tion, in particular of machine-moulding.8 For the various areas discussed here the time-scale of the investigation differs. Large-scale production of bricks, as distinct from incidental production for domestic use, is associ- ated either with large-scale building projects, often of fortifications or re- ligious buildings, or with a more or less constant demand from urban agglomerations, and this determines the starting point of our investigation. At the other end of the scale, the timeframe of the investigation is deter- mined by the mechanisation of brickmaking, which occurred from the late nineteenth century on in Western Europe, in the 1930s in Russia and is still underway in India. With some variation according to the area under study our investigation therefore covers approximately five hundred years, from 1500 to the present. Invented in the Neolithic Middle East, India and China, manual brick- making techniques have essentially changed very little since. Sufficient clay of good quality and enough fuel of any kind, as well as sufficiently long periods of warm, dry weather are the most important requirements. Thus, most brickworks are found in temperate or tropical zones, usually along river banks or other places where the right sort of clay can be found. Fuel at hand is convenient, but it can also be transported from elsewhere by water. Because of the cheapness of bricks in relation to their weight, and because they break easily, they will usually be transported to the construc- tion sites only across short distances, and preferably over water. That is why brickfields are, as a rule, concentrated near towns, where demand is stron- gest and most constant. Because of climatic constraints, pre-modern brick production was sea- sonal: in temperate zones during winter time ‘green’ (i.e. moulded but not yet baked) bricks could not be dried and frost would cause them to crack even before firing. While in this case temperature determines the season, in the tropics it is the humidity of the air that causes interruption. The monsoon rains make it impossible to properly dry the green bricks. Con- sequently, in Europe the moulding season lasted roughly six months, from

8 More extensively in: Lucassen, ‘Brickmakers in Western Europe .. and India’, 513-25. See also James W.P. Campbell, Brick: A world history (London: Thames & Hudson, 2003). labour relations, efficiency and the great divergence 265 April to September, in Russia three and a half months, from mid-May to late August, and in northern India four months, from October till February. Only the advent of indoor drying and storage in the mid-twentieth cen- tury brought an end to the seasonal character of brickproduction in Europe; in Northern India it still is a seasonal industry. Apart from these differences in the months when the production took place, historically all other conditions are strikingly similar. In nearly all cases selected here the industry depended not only on seasonal workers, but more specifically on seasonal migrant workers. Only where the first stages of the production process, i.e. the digging and tempering, required much work, year-round employment as a brickmaker was feasible and seasonal migration could be avoided. Much more frequently however, even in such cases local workers dug the clay in winter and seasonal migrant workers did the rest of the work during the dry and warm season. In all three cases the production was organised according to the same sequence of activities in working the clay (brickmakers did not have to bother about the fuel). As illustrated in Figure 8.1, it started with the dig- ging, tempering and mixing of the clay; next came the moulding of the bricks (or for that matter tiles and similar products) and the drying of the green bricks, flat on the ground and afterwards piled up on their sides in stacks, allowing maximum access to the wind. When sufficiently dry, the green bricks could be fired. Finally they had to be selected according to quality and piled up for transportation to the building sites.

Figure 8.1. Manual brickmaking—stages in the production process.

The greatest variation in the manual brickmaking process we encounter is in the firing technique. Two basic alternatives existed: the shifting clamp or the fixed kiln. In a clamp green bricks are mixed with fuel and the stack is enclosed by bricks which have already been fired. After firing the whole clamp is dismantled. The fixed kiln, a permanent structure, tended to be used primarily when bricks were made for a market with a steady high demand in a region with a plentiful supply of very good clay. The end of traditional manual brickmaking came with the introduction of the so-called Hoffmann kiln, patented in Prussia and Austria in 1858, but widely used only much later in the nineteenth century. This was a multi- 266 gijs kessler and jan lucassen chamber kiln which made superior use of heat and fuel, and allowed brick production to continue uninterruptedly, using part of the chambers for the final drying, others for firing, and yet others for the cooling down of the fired bricks. In order to draw the air through these different chambers, kilns of this type were equipped with characteristic tall chimneys. The Hoffmann kiln greatly enlarged production capacity, which enhanced the replace- ment of hand-moulding by mechanised moulding to keep up a steady sup- ply of green bricks to the ever-hungry kiln. This was by no means a clear-cut transition, however. Mechanical moulding completely replaced hand moulding long after the introduction of the Hoffmann-kiln. Although the number of stages in the brick production process is fairly limited, the quality of the work performed at each of them has important consequences for the subsequent one in the production chain. An added complication is that defects mostly become apparent only at the very end of the production chain, when the bricks are fired. Clay, for example, comes in different varieties and grades, often deposited in different layers atop of one another, and the harder, lumpy, or pebble-ridden varieties greatly add to the workload of the mixers and moulders upstream. If pebbles, stones, roots and sticks are not properly removed, or the clay is not thoroughly mixed, it can do severe damage: bricks made from non-homogeneous clay, or containing extraneous elements will crack and burst during the firing process. The most crucial stage is the moulding process, which requires consid- erable expertise. Not only should the moulder transform the lumps of clay into neat bricks, he should also take care to apply the necessary pressure to make them as compact as possible, because internal cavities and air bells will also cause the bricks to burst in the oven. A proper drying of the bricks is also essential, but requires less expertise than attendance and care, so as to protect the bricks from too much direct sun, from wind and from rain, and, towards the end of the season, from ground frost. The firing of the bricks finally, which should take account of the variety of clay used, the residual moisture in the bricks, the interplay of oven and fuel used, and a host of other factors, puts the work of preceding stages to the test. Good bricks should be fired long enough to fuse the clay particles in the bricks, but not too long, as this will deform them and lead to the formation of layers of glass, rendering the bricks hard, but brittle and therefore useless. Preventing errors from affecting brick quality and production loss is one thing. Equally decisive for productivity is the efficient co-operation be- tween the workers engaged in the various stages of the production process. labour relations, efficiency and the great divergence 267 Moulders should be assured of a steady supply of clay, water and sand, and should have sufficient empty space to put their wet bricks to dry, which is dependent on the work of those responsible for the drying process. Kiln- men are dependent on the supply of well-dried green bricks and fuel, as well as on the work of those unloading the kiln after firing. The more ef- fective the co-operation of these different categories of workers, the high- er the over-all productivity at the brick work.

Brickmaking in Western Europe from the Late Middle Ages until the Nineteenth Century

Brickmaking was ubiquitous in Early Modern and Modern Europe, except in mountainous regions where stone, and forest regions where wood was abundant.9 We concentrate on brick fabrication in the most important urbanised regions with a steady high demand (Northern Italy, the Nether- lands and Southwest England) as well as in prosperous rural areas in espe- cially Holland and Germany where specialised groups from the German principality of Lippe-Detmold and from Wallonia found employment. Before the advent of the Hoffmann kiln, European brick kilns and there- fore brick factories were relatively small. Leaving aside small potters’ kilns, dedicated fixed brick kilns in medieval England did not exceed a capacity of 50,000 bricks at a time, requiring the work of only one brickmaker and a few assistants.10 In the Dutch Golden Age, increased demand stretched the capacity of fixed brick kilns substantially (see below, table 8.3). The increased capacity of the Dutch fixed kilns coincided with attempts to improve other stages of the production process. In the first half of the seventeenth century, horse-drawn pug-mills were invented for improved clay-mixing.11

9 Cf. Campbell, Brick, 43 ff, who distinguish two distinct periods for Western Europe – the Roman (1st century BCE – 4th century CE) and a new start from the 11th century onwards. 10 Terence Paul Smith, The medieval brickmaking industry in England 1400-1450 (Oxford: B.A.R., 1985), 50-52. 11 Friedrich August Alexander Eversmann, Technologische Bemerkungen auf einer Reise durch Holland (Freyberg und Annaberg: Grazische Buchhandlung, 1792), 179-80; G.B. Jans- sen, Baksteenfabricage in Nederland: Van nijverheid tot industrie, 1850-1920, Gelderse Histo- rische Reeks, 17 (Zutphen: De Walburg Pers, 1987), 85-87; Piet Lourens and Jan Lucassen, Lipsker op de Groninger tichelwerken: Een geschiedenis van de Groningse steenindustrie met bijzondere nadruk op de Lipper trekarbeiders 1700-1900 (Groningen: Wolters-Noordhoff / Forsten, 1987), 7-8; Willy Bender and Hans H. Böger, ‘A short history of the extruder in 268 gijs kessler and jan lucassen The production capacity of clamps varied much more than that of kilns, from only a few tens of thousands to 200,000. As an average capacity we may take, very roughly, 100,000, equalling that of the large fixed kilns. Be- cause several clamps could be fired at the same time in one field during one season, such a production unit could run into millions.12 The largest production units were found at public works, like canal and fortress build- ing. There, one employer (often the state) could employ hundreds of brick- makers.13 However, in general, the manual European brick industry between the fifteenth and the mid-nineteenth century was characterised by small production units of between five and ten workers—at the maxi- mum a few dozen.

Division of Labour and Co-operation in Production

The division of tasks among the individuals within the workforce ran par- allel to the sequence of tasks described above. In a group of six workers, for example, one was the mixer (also called ‘temperer’), one the moulder and one the fireman, whereas two or three extra individuals were needed to assist the moulder, both with the transportation of the clay from the mixing pool or mill to the moulder’s table and to present it there in the form of a proper ball (the work of the so-called ‘up-striker’), as well as for the transportation from this table to the drying fields, including the turning and stacking (responsibilities of the ‘upganger’). Everybody had to assist the fireman when required, and assistance was also needed in loading and unloading the kiln. Thus, brickmakers were mostly working on their own, passing on the half-finished product to the next in line, but regularly all hands were needed as well, as in maintaining the fires day and night. In the trade some tasks were considered to require more experience than others. The transportation of the clay, the green and the ready bricks demanded the least skill, whereas for the presentation of clay balls to the moulder one already needed some more experience. Next in the hierarchy ceramics’, in: Frank Händle (ed.), Extrusion in Ceramics (Berlin: Springer, 2007), 100-101; Davids, The rise and decline of Dutch technological leadership, 121. 12 Lucassen, ‘Brickmakers in Western Europe .. and India’, 520; Joseph Arnold Foster (ed.) Contributions to a Study of Brickmaking in America, Part 3, Accounts of brickmaking in England published during the 17th and 18th centuries (Claremont: privately printed, 1965), 14. 13 Lourens and Lucassen, Arbeitswanderung und berufliche Spezialisierung, 109-11; Philippe Destable, Les chantiers du roi: La fortification du ‘précarré’ sous le règne de Louis XVI (unpublished PhD thesis in History, Université de Lille III, 2006), 158. labour relations, efficiency and the great divergence 269

Illustration 8.1. Overview of Cathrinesminde Teglvaerk, a Danish brick factory at Iller Strand bordering on the Flensburger Förde. Since the late 1840s brickmakers from Lippe were engaged in factories along this coast, characterized by horse-drawn clay pits and ovens under a roof without chimneys (note the smoke escaping through the tiled roof). Source: Copy of an oil painting in a private collection, circa 1860, Museet på Sønderborg Slot (Denmark). came the mixing of the clay with water and the elimination of stones, pebbles, roots and other impurities. Where this work was done in a horse- drawn mill, the worker also had to know how to handle the horse. Higher up in the hierarchy came the moulder, and the fireman. In some cases, the fireman was in charge of the team, like among the Lippe brick makers. We do not know why, but sometimes the firemen were separate from the oth- er brickmakers, who were organised around the moulder.14 The quality of the final product depended on the fireman’s ability to add the expensive fuel in such a way that a maximum number of bricks was baked not too

14 This was the rule among the seasonal brickmakers of Wallonia, for example, as well as in medieval Italy: Lucassen, ‘Brickmakers in Western Europe .. and India’, 534-39; Richard A. Goldthwaite, The building of Renaissance Florence: An economic and social history (Bal- timore, MD: John Hopkins University Press, 1980); Manuel Vaquero Piñeiro, ‘L’Università dei fornaciai e la produzione di laterizi a Roma tra la fine del ’500 e la metà del ’700’, Roma moderna e contemporanea IV, n. 2 (maggio-agosto 1996), 471-94; Manuel Vaquero Piñeiro, ‘La gabella dei calcarari: Note sulla produzione di calce e laterizi a Roma nel quattrocento’, in: Angela Lanconelli and Ivana Ait (eds), Maestranze e cantieri edili a Roma e nel Lazio: Lavoro, tecniche, materiali nel secoli XIII-XV (Manziana, Roma : Vecchiarelli, 2002), 137-54. 270 gijs kessler and jan lucassen hard nor too soft, as this would reduce their value by at least half. Accord- ing to the size of the enterprise some tasks might be combined or split, but the flowing of the product through six to ten pairs of hands was crucial, as was the occasional but regular co-operation of many or even all workers in one task. The labour process was organised differently in every region, but two basic systems can be distinguished, which will be analysed in detail in the following sections. The first system was one based on co-operative subcon- tracting and accounted for the great majority of cases. Gangs of brickmak- ers contracted all or part of the production process from the employers against a collective piece-rate per thousand bricks. Such a system existed in London, in Northern Italy after the sixteenth century, as well as in areas in the Netherlands and Germany that attracted migrant brickworkers from Wallonia and the German principality of Lippe-Detmold. Such gangs usu- ally enjoyed a great deal of autonomy in the organisation of their work, and because they were remunerated as a gang according to their finished production of the right size and quality, they had a direct interest in the efficient co-operation in production. Apart from providing the capital and the raw materials, the role of the employer in the production process amounted to little more than that of a paymaster; his control over the la- bour process was minimal. The second system is represented by the Dutch Rijnland (and to a cer- tain extent the Hollandse IJssel region), where since the second half of the seventeenth century a very different kind of labour relations prevailed. Here, something akin to modern production relations existed, with a sub- contractor hiring workers on an individual basis and directly supervising their collaboration in the production process. An essential element of the system of co-operative subcontracting prac- tised in the majority of regions was the redistribution of collective earnings within the gang. This was not simply an equal distribution, but instead reflected the contribution made to the collective effort, levels of skill and dexterity, and often age. Inexperienced young workers who had to learn the trade were put on time-wages, whereas experienced workers would receive a piece-rate. The redistribution of earnings within the collective was the cornerstone in the mechanism of error-rate reduction in the pro- duction process, as it made remuneration conditional, not so much on effort as such, but on effort which was functional within the existing divi- sion of labour. In this sense, it provided an extra incentive to match one’s efforts to the requirements of the collective and to acquire skills, particu- larly for the younger workers still learning the trade. Table 8.1 summarises labour relations, efficiency and the great divergence 271 the evidence on the redistribution of earnings within the gangs for the regional cases investigated in this study. The following paragraphs describe the formation, internal organisation and specific characteristics of each of the cases separately.

Table 8.1. Remuneration modes among brickmakers working in gangs, North-Western Europe c. 1600-1900.

England c. 1670- Lippe brickmakers Walloon brickmakers 1700 c. 1750-1900 c. 1700-1900 Remuneration for 5 Thaler per worker supervision by the or 0.35-0.65% of the gang leader total wage sum High share in total Moulders Firemen, moulders Moulders sum earned on and temperers piece rates Medium share in Earth maker / tem- Others 50% (1750- Moulders’ assistants: total sum earned on perer, carter, 1850) – 75% (1850- 50% of what mould- piece rates upganger, off-bearer 1900) of what ers earn : 80% of what firemen, moulders moulders earn and temperers earn Low share in total Up-striker, off- Youngsters (porters): sum earned on bearer (boys): 40% 25% of what mould- piece rates of what moulders ers earn earn Time wages Clerk New and less valid New workers workers: c. 15-30% of what the firemen, moulders and tem- perers earn Sources: Foster, Contributions to a study of brickmaking in America, Part 3, 6, 13-14, 19, 32 (remuneration for making green bricks, for firing the bricks see ibid. 17); Lucassen, ‘Brickmakers in Western Europe .. and India’; Lourens and Lucassen, Arbeitswanderung und berufliche Spezialisierung, 183-85

The Italian case is not included in this table for lack of sufficient data. Two overall characteristics of the system of remuneration should be mentioned, though. In the first place production in Italy moved towards an integrated system from the sixteenth century onwards, in which the fornaciai or kiln- men subcontracted all work from digging until burning. Within these in- tegrated groups the fornaciai were in the top position, responsible for paying taxes and membership fees to the guild, followed by the moulders in second position and the kiln-men and the stokers at the bottom.15

15 Lucassen, ‘Brickmakers in Western Europe .. and India’, 534-39; Vaquero Piñeiro, ‘L’Università dei fornaciai’, 476; Vaquero Piñeiro, ‘La gabella dei calcarari’; Manuel Vaquero 272 gijs kessler and jan lucassen Around London, brickmakers worked together in a gang of six called ‘stool’, which controlled the whole production process, and was paid as a group per 1,000 well-made bricks. This piece-rate system of remuneration was called ‘work by the Great’.16 At the end of the seventeenth century, Houghton concluded that in this way the members of a stool ‘put them- selves forward and work hardest of any sort of People: This very Brick- moulder has work’d till the white foam has arisen on his breast’.17 The workers divided the total wage sum among themselves according to their skills in three classes (cf. table 8.1). At the top was the moulder with six units; next came the temperer, the upganger (the man responsible for the drying of the green bricks) and sometimes the carter with four units, and then all others with three units. The same system of partition applied to the earnest-money or ‘Over-Money’ which the members of the stool re- ceived on St. Thomas (21st December) when they heard whether they would be welcome next season. In that case the moulder received three pounds 10 shillings, the earth maker three pounds one shilling, the off- bearer 40 shillings, the upganger and the carter 22.5 shillings each, and the up-striker 10 shillings.18 The Walloon brickmakers used a remarkably similar expression for their gangs which they called ‘banc’ and ‘table’.19 A banc consisted of two groups, one all-male for firing the clamps, and the other called table, composed of members of one or more families, for the entire moulding process. In par- ticular the skillful female Walloon moulders made a deep impression on contemporary observers.20 These workers built their own huts at the work site, shared a common kitchen and the chef the table calculated the ac- counts fortnightly. After receiving the full sum for the season the chef de table—as a rule the moulder—divided it into full-parts for the inner circle

Piñeiro, ‘Produrre laterizi a Roma nel XVIII secolo: le fornaci della Congregazione dell’ Oratorio’ (unpublished paper 2004); Goldthwaite, The building of Renaissance Florence, 211. 16 Foster, Contributions to a study of brickmaking in America, 10, 13-14, 17-18 (near London 1683 and 1693), and 54-60 (England 1780/1801). Similar systems appear to have existed in the first half of the nineteenth century in Nottingham and southern Lancashire, see Lucas- sen, ‘Brickmakers in Western Europe .. and India’, 528-33. 17 Foster, Contributions to a study of brickmaking in America, 16 (cf. also 14: they worked seven days a week, performing double standard time wage tasks, i.e. 14 per week). 18 Ibid., 14. 19 Lucassen, ‘Brickmakers in Western Europe .. and India’, 534-39. The earliest recorded date so far for this group is 1695, cf. J. Hollestelle, ‘De Nederlandse steenbakkerij in de zeventiende en achttiende eeuw’, Economisch- en Sociaal-Historisch Jaarboek 44 (1982), 19. 20 Eversmann, Technologische Bemerkungen, 149 labour relations, efficiency and the great divergence 273 of the gang (himself and the temperer), half-parts (for his assistant who prepared the clay balls), and fourth-parts (for the youngsters, except for the new workers who received time wages). The all-male group of firemen divided the season’s wage sum as follows: three or four men who piled the clamp at a piece wage which equalled about three fifths of that of the chief- fireman, and two boys who fetched the fuel on time wages. The Walloon migratory brickmakers enjoyed great independence while at work. The factory owners left the complete production process to them. Between the Walloon brickmaker gangs and the factory owners there was the figure of the maître-briquetier (literally, master brickmaker) who had made contracts with public bodies or private factory owners to deliver a specified number of bricks. These maître-briquetiers came in springtime to brickmakers’ markets to engage their groups, for instance at the busy Easter market in the town of Charleroi. They also had to provide straw mats for the gangs to build their huts, as well as one barrel of ‘small beer’ per 10,000 bricks. How much they were paid for these intermediary services by factory owners and gangs is not known. Only in very exceptional cases did this system show cracks. In 1810, when the builders of the Grand Canal du Nord, constructed under supervision of the French authorities between the Dutch towns of Weert and Venlo, failed to pay out the wages, this led to a strike and violent collective action by the brickmakers of at least seven clamps, all originating from villages northeast of Liège.21 Another conflict took place in 1894 when the Walloon brickmakers struck in protest to the restriction of female labour, the banning of child labour under twelve years, and the restriction of working hours for children aged twelve to eighteen years—changes stipulated by a Belgian law which came into force on 1 January 1894.22 The brickmaker gangs from the German principality of Lippe-Detmold, who worked primarily in the North-Eastern part of the Netherlands, North- ern Germany and Scandinavia, controlled the whole production process.23

21 Lucassen, ‘Brickmakers in Western Europe .. and India’, 534-39; Nel Mulders-­Thijssen, ‘Le Grand Canal du Nord van de Schelde tot de Maas’, in: C.M. Bos-van Bers, N. Mulders- Thijssen, and J. Vullings (eds), Le Grand Canal du Nord van de Schelde tot de Maas: De Noordervaart en haar waterwegen (Geysteren: Stichting Nieuwland, 2008), 56-62; Nel ­Mulders-Thijssen, ‘De onvoltooide Noordervaart in Maasbree’, in: Bos-van Bers, Mulders- Thijssen, Vullings (eds), Le Grand Canal du Nord, 174-76; Henk Willems, ‘Meijel, kana­ lenkruispunt in de Peel’, in: Bos-van Bers, Mulders-Thijssen, Vullings (eds), Le Grand Canal du Nord, 101-103. 22 Lucassen, ‘Brickmakers in Western Europe .. and India’, 537. 23 Lourens and Lucassen, Arbeitswanderung und berufliche Spezialisierung. 274 gijs kessler and jan lucassen

Illustration 8.2. Gang of Lippe brick makers in Sunday dress in front of a roofed oven without chimney, Oberndorf an der Oste (Northern Germany), 1903. Source: Photograph Landesarchiv NRW Abt. OWL, Detmold (Germany), D 75 Nr. 2670.

These all-male gangs of five to ten men who were responsible for the entire process from tempering to firing, were called Lipper Communes, and were formed by the firemen (Brandmeister) during winter time when all had returned home.24 Remarkably, Lippe brickmakers did not stick necessar- ily to the same gang over the years. Instead, they showed a high volatility and each year tried to be in a gang with a higher labour productivity and higher seasonal earnings.25 They formed a common household (the Lipper Commune proper), lived together in one or two rooms where two men shared one bed, and divided any earnings according to a fixed ratio. At the end of the season the fireman received from the owner of the kiln a sum

24 Lucassen, ‘Brickmakers in Western Europe .. and India’, 540-43, based mainly on Lourens and Lucassen, Arbeitswanderung und berufliche Spezialisierung. 25 Piet Lourens and Jan Lucassen, Labour mediation among seasonal workers, in par- ticular the Lippe brick makers, 1650-1900, version of 14 June 2011, available at: http://www. iisg.nl/migration/ziegler/documents/lourens-lucassen.pdf. This publication is based on individual level data documenting the life histories of over 800 mid- and late-nineteenth century Lippe brickmakers, made available online through the joint efforts of the Interna- tional Institute of Social History and the Landesarchiv Nordrhein-Westfalen at: http://www. iisg.nl/migration/ziegler/. labour relations, efficiency and the great divergence 275 of earnings which was derived from the number of bricks, well baked, prop- erly sorted and ready for sale, multiplied by the previously agreed rate per thousand. ‘Boarding money’ was deducted to cover fortnightly advances allocated for the purchase of food. The fireman himself was entitled to a fixed sum for his supervisory work. Finally, he divided the remainder according to fixed ratios: the fireman, the moulder and the temperer all usually received equal shares, but less experienced workers—again accord- ing to pre-arranged principles—got less, perhaps 80 percent of the full amount. Only inexperienced youngsters received time wages during one or two seasons.26 Of all cases known to us, the Lippe brickmakers must have enjoyed the maximum degree of self-rule.27 This was greatly enhanced by a system of mediation, similar to that of the Walloon maître-briquetiers, but set up much more formally by the government of the principality Lippe-Detmold from where the workers hailed. Since the last quarter of the seventeenth century the Lippe government sold a privilege to a Ziegelbote (literally brick messenger) to mediate between its brick workers and factory owners abroad. One family managed to hold on to this exclusive mediation job up to 1802, when the government divided the world outside the principality into two ‘brick districts’, then in 1842 into three, and in 1867 into four. The brick messenger received a fixed fee from the owners of the brickworks he had visited during autumn and early winter, and another from the firemen at home in Lippe whom he met later in winter and early spring. In ex- change, he made sure that the Lippe workers knew where to go in spring and what their uniform piece rate would be for the coming season. By this abolition of wage competition among Lippe brickmakers only their skill and the common efforts of the gang determined their earnings at the sea- son’s end. As soon as their employment was arranged, the firemen com- pleted their gangs. Although formally abolished in 1869 when Lippe was forced to abandon all remaining monopolies under the Gewerbefreiheit principles of (North-) German law, the messenger arrangement was so successful that it survived informally for another several decades. Besides, it was a flexible system open to innovation. Under pressure of the revolutionary events in 1848-49

26 Lucassen, ‘Brickmakers in Western Europe .. and India’, 541-42. 27 Ibid., mainly after Lourens and Lucassen, Arbeitswanderung und berufliche Spezia­ lisierung; P. Wessels, Ziegeleien an der Ems: Ein Beitrag zur Wirtschaftsgeschichte Ostfries- lands (Aurich: Ostfriesische Landschaftliche Verlags- und Vertriebsgesellschaft, 2004), 188-95. 276 gijs kessler and jan lucassen the new Lippe brick industry law of 1851 required the election, by all work- ers over 25, of messengers for each district, a complaints procedure, nom- inations of new firemen only by firemen already active in the same district, and a mediation court consisting of the messenger, three firemen, two ordinary brickmakers and a lawyer. Moreover, all messengers had to intro- duce an insurance system against the adversities of diseases and death. No wonder that open conflict was rare under these circumstances. In a typical case of disagreement between workers’ gangs and factory owners the fire- men asked for mediation by the messenger, who sometimes agreed with the point of view of the employer, but certainly not always. In 1772 the Lippe brickmakers united against military empressment in East Frisia. In 1794 and 1816 the messenger of the first district took up the cudgels for his men who had drawn up a ‘Resolution’ against the united employers in Gronin- gen and East Frisia. But the gangs themselves were also able to hold their own, as is illustrated by the successful court case between 1849 and 1852 of a Lippe gang in protest against a lock-out in the Dutch province of Gro­ ningen. Although widespread, co-operative subcontracting was not universal in the West-European brick industry. An alternative system of labour relations is represented by two areas in the maritime provinces of the Netherlands in the seventeenth and eighteenth centuries, then the centre of early-mod- ern brick production and technically perhaps the most developed of its times.28 First, along the Hollandse IJssel river between Gouda and Rotter- dam (henceforward IJssel) an individualised work organisation based on piece rates prevailed, and, secondly, along the Oude Rijn river and centred around Leiden and Woerden (hence Rijnland) an individualised system based on time rates.29 In both cases the workfolk were not migratory, but of local origin and appear to have lived in houses close to and owned by the factory where they worked. Although the oven capacity of the factories along the IJssel was large and even increased during the eighteenth cen-

28 We know too little about the Italian case, where the more successful firemen or fornaciai could subcontract all work at a kiln. In those cases they themselves were entitled to sell the product themselves and paid a small part of the workforce according to time rates (the stokers and their unskilled assistants), but most according to piece rates (the moulders and their families as well as the carters), see Vaquero Piñeiro, ‘La gabelle dei calcarari’, 154; Goldthwaite, The building of Renaissance Florence, 199, 201, 207. 29 Lucassen, ‘Brickmakers in Western Europe .. and India’, 526-29; besides the literature cited there, two newly discovered sources: Private collection, Account book Reeckeninge Anno 1662, 21 August 1662 until 2 July 1729, and Minutes Steenbacker Gilde Onder Rhijnlandt, 31 March 1753 until 7 August 1820, used here by kind permission of the owner. labour relations, efficiency and the great divergence 277 tury, labour productivity apparently did not. And in Rijnland and Woerden the system of individual time rates asked for extra managerial techniques and could not do without a strict work discipline, enforced through penal sanctions by a united front of employers. The large factories along the IJssel got their clay by dredging silt from the river.30 Around 1700 one factory owner boasted that a single oven, when fired three times, produced 1,800,000 bricks per season. In 1762 the biggest ovens were reported as twice that size, and in 1788 one oven could yield 977,778 first class bricks per firing. According to Eversmann the daily supervision of one factory with three ovens was the responsibility of one master and encompassed two groups of moulders and one of kiln-folk per oven. The centre of the moulding process was the moulding table with one moulder, one clay cutter, and for the transportation and handling of green bricks on the drying fields two women and two children, aged between eight and fifteen. All of them were paid piece rates per 1,000 bricks: the moulder two stuivers (stiver, i.e. 1/20th guilder), the clay cutter one and the children (and probably the women as well) half a stuiver. At the end of the nineteenth century payments were made individually or to the head of the household if more of its members were engaged at the same factory. Gabriel Jars le jeune reports that the children’s wages depended on their age, which implies a skill premium built into their piece rates. Average daily production per member of a moulders’ group varied from 1,000 via 1,166 to 1,133 bricks in 1788 and between 1,200 to 1,400 bricks in 1854. The kiln personnel consisted of 8 women who were engaged for three weeks per season in the loading of the kiln and six weeks in the unloading of the bricks. Such a group was enough for two ovens at one factory. Besides,

30 Mainly based on the following sources: a petition by the united employers to the Estates of Holland, s.d. but around 1703/1704 (Regional Archive Leiden, Stadsarchief Leiden 1574-1816, no. 3103), a description of brick factories in Moordrecht at the end of July 1766 by the French technician Gabriel Jars, le jeune (1732-1769) (Gabriel Jars (le jeune), ‘Die Kunst- wie in Holland Ziegelngestrichen, und mit Torfe gebrannt werden; zur Vollständigkeit der Zieglerkunst’, in: Daniel Gottfried Schreber (ed.), Schauplatz der Künste und Handwerke (Leipzig und Königsberg: Jacob Kanter, 1768), 131-1742), a description of brick factories at Gouderak in the late summer of 1783 by the Prussian technician Friedrich August Alexan- der Eversmann (Eversmann, Technologische Bemerkungen, 148-160; for a comment see C.M. Kooi, ‘Een open Hollandse steenoven in Gouderak in 1783 getekend en beschreven door een Duitse bezoeker’, Nieuwsbrief Stichting Historie Grofkeramiek 29 (January 2008), 16-20), an industrial survey of 1812 (Dutch National Archives, Biza 1796-1813, 1193: Nijverheidsen- quête 1812), and W.T. Gevers Deynoot, ‘Iets over de steenbakkerijen langs den Hollandschen IJssel’, Staatkundig en Staathuishoudkundig Jaarboekje, 2e serie 1 (1854), 392-93. See also D.W. Hoogendijk, C.C.J. Lans, J. Bakker, ‘De steenbakkerijen’, Historische Encyclopedie Krimpenerwaard 4 (1979), 49-70. 278 gijs kessler and jan lucassen there was an unknown number of firemen and men who, we think, fetched the peat and threw it onto the fire. As Eversmann put the total number of workers per factory with two ovens at 30, which is confirmed by other sources, we suppose that per factory there were nine firemen cum peat- men who took care of the two alternately but continuously burning ovens. Each oven was fired three times per season, with the last one allowed to be finished by the end of September. The women at the kiln, most dexterous in throwing and catching piles of eight bricks at a time, received time wages: 10 stuivers or half a guilder per day, whereas the firemen and peat- men were paid piece rates related to the amount of peat they fed into the ovens—a clear incentive not to waste the precious combustible. The fac- tory owners, or bazen, maintained an informal organisation which served to create a level playing field by setting certain rules. Each year, on Septem- ber 8, they would attentively inspect all factories to reach agreements on things like a fixed duration of the moulding season (exactly twenty weeks, from April 20 to September 10), the size of the moulds they would use, and, accordingly, the size of the bricks eventually produced. This self-imposed standardisation apparently served as a mechanism to undercut unfair com- petition by starting early or producing on the cheap. The Rijnland production area was situated a few dozen kilometres north of the river IJssel. Along the river Oude Rijn between the North Sea and Zwammerdam it comprised no fewer than sixty factories in the 1660s, pro- ducing several million bricks per year per unit. Comparing these factories with contemporaneous production levels of 100,000 in Italy and one million in England, it is no exaggeration to call Rijnland the centre of capitalist brick production in the Early Modern period. Its production was geared to the building industry of Amsterdam. This centre of commercial capitalism had a centralised organisation for the provisioning of masonry bricks, which struck centralised production and price deals with the Rijnland guild of brick factory owners, established in 1632.31 This might explain why labour relations in Rijnland were so exceptional, even when compared to the IJs- sel area. The first peculiarity was the engagement of individual workers (possibly also parents with children) by subcontractors (called aannemer, also ondermeester or meesterknecht) who agreed with the kiln owners on the basis of a remuneration per 1,000 well-baked bricks, but apparently subsequently paid weekly individual time wages to the workers.32 Accord-

31 Hollestelle, ‘De Nederlandse steenbakkerij’, 15-16. 32 The evidence on this crucial point so far is not abundant. However, contrary to all other regions in Western Europe before mechanisation, we have not found any evidence labour relations, efficiency and the great divergence 279 ingly, the differential between his piece wage and his workers’ time wage determined the subcontractor’s earnings, which provided an incentive to make workers produce as high a number of good quality bricks per time unit as possible. Because workers would not earn more from these efforts, the desired result could be achieved by close supervision only.33 The second peculiarity, closely linked to the first, is the strict supervision of the workforce. In 1640 a blacklist of all workers who had left a factory without consent or had been sacked, was introduced by the recently estab- lished Rijnland factory owners’ guild.34 Their names were put on a list and communicated to all guild members with the clear understanding that the workers thus blacklisted should under no circumstances be re-engaged at another factory during that same season. Thus, unemployment was the ultimate punishment for the disobedient. For the period 1754-1818 we have evidence that at least 156 individuals were blacklisted in this way.

Table 8.2. Workers blacklisted by the Leiden based Rijnland brickmakers guild, 1754-1818. Units black- Individuals listed male female unknown total man 65 65 65 woman 22 22 22 boy 1 1 1 girl 1 1 1 family units couple 4 4 4 8 couple with 8 10 (incl 2 12 (incl 4 8 or more 30 or more children boys) girls) children

of piece rates in the Rijnland brick factories. For the seventeenth century one author sug- gests time rates, see B.W. van der Kloot Meyburg, ‘Eenige gegevens over de Hollandsche steenindustrie in de zeventiende eeuw’, Economisch-Historisch Jaarboek 11 (1925), 85-88, 95. Only in the beginning of the twentieth century a mixture of piece rates (most of the earn- ings) and time rates is mentioned explicitly for roof tile makers, see B.W. van der Kloot- Meyburg, De economische ontwikkeling van een zuid-hollandsch dorp (Oudshoorn) tot in den aanvang der twintigste eeuw (The Hague: Martinus Nijhoff, 1920), 223. 33 Time wages for sedentary brickmakers were the rule in the nineteenth and twenti- eth centuries in the Netherlands, see G.B. Janssen, Twee eeuwen Gelderse bakstenen: De baksteenfabricage in Gelderland in de negentiende en twintigste eeuw (Aalten: Fagus, 2008). 34 B.W. van der Kloot Meyburg, ‘Een productiekartel in de Hollandsche steenindustrie in de zeventiende eeuw’, Economisch-Historisch Jaarboek 2 (1916), 215 (regulation 1640, article 6), 225 (regulation 1661, articles 27-32). 280 gijs kessler and jan lucassen

Table 8.2. Cont. Units black- Individuals listed male female unknown total father or 4 4 (incl 3 (incl 5 children 12 mother with 2 boys) 1 girl) children siblings 4 5 4 9 siblings 2 4 2 or more 6 or more (children) unknown relation: men 1 2 2 Total 112 91 (incl 4 50 (incl 9 15 or more 156 or boys) girls) children more indi- viduals Source: Private collection, Minutes Steenbacker Gilde Onder Rhijnlandt, 31March 1753 until 7 August 1820

This is on average three per year, apparently enough to put the fear of God into the work force. And what is more, this system apparently worked so well that it was emulated in the Woerden production area along the same Oude Rijn. There, blacklisting was introduced by the brick factory owners’ guild in 1687 and the system was still in use in 1879.35 More importantly, no incidences of protest have been documented—apart from a few incidental cases of dissatisfied workers who subsequently absconded and therefore were blacklisted (see table 8.2). Apparently, the brickworkers of the Oude Rijn were properly tamed. Nonetheless, two improvements in these harsh labour relations have to be mentioned. In the same 1687 regulation of the Woerden brick and tile- makers’ guild, the possibility of mediation between masters, subcontrac- tors and workers was introduced, a novelty which was copied by the tile­makers’ guild of Rijnland in 1728 and by the brickmakers’ guild of Rijn- land in 1743.36 To what extent this possibility of mediation has been used in practice remains unclear. In the minutes of the brickmakers’ guild of

35 Rob van der Laarse, Bevoogding en bevinding: Heren en kerkvolk in een Hollandse provinciestad, Woerden 1780-1930 (The Hague: Stichting Hollandse Historische Reeks, 1989), 39; ‘Eenige aantekeningen bij de Gildebrief van de pan- tichel- en steenbakkers’, Heemtijd- inghen Historische Vereniging Woerden en omgeving 12 (1972), 33-34 (articles V-VII); Van der Kloot Meyburg, De economische ontwikkeling, 86. 36 ‘Eenige aantekeningen’, 33-34 (articles VIII-IX); Regional Archive Leiden, Bibliotheek Leiden en omgeving 59897/03 plano (this Ordre en Reglement van ’t Pannebackers-gilde is not formally subdivided into articles); Ordonnantie en reglement 1743 (article XXIV). labour relations, efficiency and the great divergence 281 Rijnland between 1753 and 1820 we have found only three cases: in 1754 an ondermeester successfully lodged a complaint against his factory owner, while in 1764 and 1771 ordinary workers saw their protests upheld– at least partially.37 The second improvement was the foundation in 1764 of a mu- tual benefit society of the Woerden brick- and tilemakers’ journeymen. It still existed in 1844 and provided insurance to the male workers and their co-working spouses against the adverse consequences of illness, old age or other misfortune, as well as a decent burial where all members had to at- tend.38 This manifestation of self-rule notwithstanding, no collective sub- contracting ever appears to have emerged in the Woerden brick and tile industry.

Production and Productivity Thus, the brickmaking industry in Western Europe exhibited a great deal of organisational diversity. To what extent did this result in different levels of productivity? Table 8.3 presents the available evidence, expressed as the average production of bricks per gang worker per day (not including the firing). Unfortunately we have no comparable productivity figures for time- wage remuneration in the Rijnland area, but when we contrast individual piece remuneration at the IJssel with collective piece remuneration in the English, Walloon and Lippe cases, it suggests the superiority of the system of co-operative subcontracting based on piece rates. Given the fact that production technology and industry-specific skills were essentially similar, the crucial difference must have been the efficiency of the co-operation within the different systems of labour relations. The comparative advantages of co-operative subcontracting in this re- spect would appear to have been twofold. In the first place it provided an in-built system of error-rate reduction because of the common interest in maximising collective earnings, which was discussed above. Secondly, co- operative subcontracting offered mechanisms of intra- and intergenera- tional skill acquisition. This brings us back to the work of S.R. Epstein, but with a slightly different emphasis, because what appears to have been im- portant was not so much the transfer of production skills as such, but of knowledge and experience in the application of such skills as part of a

37 Private collection, Minutes Steenbacker Gilde Onder Rhijnlandt, 31 March 1753 until 7 August 1820. 38 Lucassen, ‘Brickmakers in Western Europe .. and India’, 529. 282 gijs kessler and jan lucassen

Table 8.3. Production capacities and labour productivity, Europe c. 1600-1900. Average production Average production Average production of bricks ( × 1,000) of bricks ( × 1,000) of bricks per gang per kiln per firing per factory per year worker per day (not including the firing) Italy (c. 1500-1700) 10-20 100 Italy (Rome 1700- 500-1,000 1800) Holland (IJssel, 200 ( 1703) – 1,800 (1703) – 1,000 / 1,166 / 1,333 1700-1854) 350/400 (1766) – 3,000 (1788) – (1788) – 1,200 / 370 (1854) 4,300 (1854) 1,400 (1854) England (Middle 50 Ages) England (around London c. 1680- 100-200 (clamps) 1,000 1,300-1,500 1700) Walloon brickmak- 800-1,000 1,300-1,650 ers (c. 1700-1900) Lippe brickmakers 100 500 1,650 (c. 1650- 1900)

Sources: Regional Archives Leiden, Stadsarchief Leiden 1574-1816, no. 3102 (Memorie c. 1703/1704); Lucassen, ‘Brickmakers in Western Europe .. and India’; Goldthwaite, The Building of Renaissance Florence, 178-179, 191, 201; works on the IJssel quoted in footnote 30; Foster,Contributions to a Study of Brickmaking in America, Part 3, 13, 19 (9,000 per day per stool of six men plus a clerk); R.S. Stirling, Report on two brickmaking machines sent out to India in 1851(Alipoor: Iron Bridge Yard, 1858), 2 gives a lower average of 1,000 per day (a gang of expert London brickmakers, consisting of a feeder, a moulder, an off-bearer and two men to wheel and hack the bricks ordinarily produces 5,000 bricks per day); Jules Vandereuse and Roger Pinon, ‘Contributions à l’étude de la briquetterie en Wallonie’, Annuaire 1958-1959 de la Commission Royale Belge de Folklore, Section Wallonne, 12 (1961), 334-37 which is not totally clear about the number of workers, but we conclude that includ- ing the firemen some 10 persons must have been involved; cf. also Hans Seeling, ‘Belgische Ziegel-Wallonen und Feldbrand am Niederrhein (1)’, Düsseldorfer Jahrbuch, 51 (1963), 243, 247; Smith, The Medieval Brickmaking Industry, 50-52; Lourens and Lucassen, Arbeitswanderung und berufliche Spezialisierung, 50-56 (a group of 6 Lippe brickmakers is able to fill five kilns of on average 100,000 bricks each in c. 30 weeks; Wessels, Ziegeleien an der Ems, 23 reaches a lower figure but this includes the more labour intensive roof tile production). collective effort. In other words, within co-operative subcontracting units workers did not only learn how to best make bricks, but also how to do this effectively together. These mechanisms of skill-acquisition were entirely informal and spanned many years. The fact that youngsters were paid time wages while piece rates were the rule is a clear indication of the conviction among brickmakers that their productivity was below par. The special status of labour relations, efficiency and the great divergence 283 the youngsters who still had to learn the trade is also underlined by the ragging of freshmen at Whitsun in their first season, as is documented for the Lippe brickmakers.39 The step-by-step increase of the shares in the common wage sum to which labourers with different skills were entitled, is the most direct indication of the premium put on skill. A small sample of 123 Lippe workers active in the eastern part of Groningen province in 1855 shows that for most boys who started their career at age 15 it took on average twenty years before they could become fireman and act as leader of a gang—that is, if they had survived, because mortality was high.40 The effectiveness of the gang as a production unit depended heavily on mutual relations between the members of a gang and their ability to func- tion smoothly as links in a chain. We should distinguish between family- based gangs, like parents working together with their children, and groups of non-related workers. Whereas family groups can rely on parental author- ity to enforce production discipline, such groups have the disadvantage that their members are condemned to each other. ‘For you another’ is no option, as in non-related groups, where the possibility of exclusion from the collective acts as a powerful factor undergirding group cohesion, par- ticularly for seasonal workers with a common origin working far away from home. During the pre-industrial period family-based groups appear to have been the exception, rather than the rule. Apart from the artisanal phase in which all family members might be engaged, like among the moulders in late medieval and early-modern Italy,41 the Walloon brickmakers consisted of mixed family groups, including boys of 12 years and over, and girls of 14 years and over.42 But brickmakers from Lippe-Detmold, on the contrary, worked in all-male gangs, including boys from 14 upwards.43

39 Lourens and Lucassen, Arbeitswanderung und berufliche Spezialisierung, 75-76, 79. On the individual level data documenting the life histories of over 800 Lippe brickmakers who died in 1840-1875, we have found a case of a young boy who did not survive the mal- treatment of his gang mates, possibly during a ragging ceremony (cf.Piet Lourens and Jan Lucassen, ‘Karrieren lippischer Ziegler: Das Beispiel Delfzijl 1855’, Lippische Mitteilungen aus Geschichte und Landeskunde 76 (2007), 63-80). For the vulnerable position of youngsters, see also Hans Seeling, ‘Belgische Ziegel-Wallonen und Feldbrand am Niederrhein (2)’, Düsseldorfer Jahrbuch, 54 (1972), 122. 40 Lourens and Lucassen, ‘Karrieren lippischer Ziegler’; Lourens and Lucassen, Ar­beits­ wanderung und berufliche Spezialisierung. 41 Lucassen, ‘Brickmakers in Western Europe .. and India’, 524-527 (mainly after ­Goldthwaite The Building of Renaissance Florence). See also Vaquero Piñeiro, ‘L’Università dei fornaciai’, and idem, ‘La gabella dei calcarari’. 42 Lucassen, ‘Brickmakers in Western Europe .. and India’, 535. 43 Lourens and Lucassen, Arbeitswanderung und berufliche Spezialisierung, 75, 80-81, 84; Lourens and Lucassen, ‘Karrieren lippischer Ziegler’, 70; Lourens and Lucassen, Labour 284 gijs kessler and jan lucassen Well-entrenched schemes for vocational training within sub-contract- ing units seem to suggest little or no contribution of the guilds towards the training and transfer of skills. Brickmakers’ guilds were employers’ organ- isations, trying to dominate an industry where substantial investments and large-scale production processes had become dominant, and, at that, very different organisations from the small-scale artisan workshops usually as- sociated with guilds that Epstein probably had in mind. Guilds were far from omnipresent in the industry, and tended to func- tion primarily as production-cartels. The oldest guilds of brickmakers are documented for Italy.44 But also in the Netherlands a few cases are known, like the already mentioned guild of Rijnland which united the brick fac- tory owners along the Oude Rijn river, between 1632 and 1818, and the tile producers’ guild which existed in the same region from 1728 onwards. It lived on—after a new start in 1818 under the name ‘Eendracht maakt macht’—at least until 1891. Further to the East another guild existed, that of the brick and tilemakers of Woerden and surroundings, active between 1687 and 1871/1879.45 These guilds are remarkable because they continued to exist long after the official abolition of the guilds in the Netherlands in 1798. Apart from checking the movements of the workforce, their main concern was to secure their supply-channels to the building industry in the main towns of the western parts of the Netherlands.46 Although the other adjacent important production regions, viz. the fac- tories along the Hollandse IJssel river between Haastrecht, Gouda and Rot- terdam, and those in the province of Utrecht along the river Vecht never had guilds, they nevertheless regularly forged coalitions in order to pro- mote their interests, sometimes together with, sometimes in competition with the guilds that were just mentioned. In Groningen province as well, producers occasionally united, for example in 1737/1738.47 In Southern Eng- land there was only one short-lived brickmakers’ guild, in Westminster mediation among seasonal workers, based on the life histories of over 800 Lippe brickmak- ers, show that only a few had started their career at age 13, while most were 14 or 15 years of age. 44 Lucassen, ‘Brickmakers in Western Europe .. and India’, 524-27, based on Goldthwaite, The Building of Renaissance Florence; see also Vaquero Piñeiro, ‘L’Università dei fornaciai’, and idem, ‘La gabella dei calcarari’, and idem, ‘Produrre laterizi a Roma’, and Campbell and Pryce, Brick, 100-101. 45 Lucassen, ‘Brickmakers in Western Europe .. and India’. 46 Private collection, Reekeninge Anno 1662 and Steenbakkersgilde Onder Rhijnlandt; Eversmann, Technologische Bemerkungen, 149-50, 156-57. 47 Private collection, Reekeninge Anno 1662 and Steenbakkersgilde Onder Rhijnlandt; Regional Archive Leiden, Stadsarchief Leiden 1574-1816. no. 3102. labour relations, efficiency and the great divergence 285 between 1636 and 1639.48 Nonetheless, also without a formal organisation London brick producers regularly organised combinations in order to raise prices, like in 1728-29, 1777, and 1784.49 Thus, at least in pre-industrial brickmaking, it was the gang or subcon- tracting unit, rather than the guilds, which functioned as the main vehicle for the transmission of skills, at least in those areas where the industry operated on the basis of this system of labour relations. Exactly how train- ing and the transmission of skills were organised in the areas where a dif- ferent system of labour relations dominated the industry, notably along the IJssel and in the Rijnland, is much less apparent. This is regrettable, because the labour relations in these production areas, based on direct supervision, could in a certain sense be seen as the forerunners of the modern production relations which would eventually come to dominate many industries. Its apparent advantages in terms of error-rate reduction, training and productivity notwithstanding, the system of co-operative sub- contracting in brickmaking in Western Europe came under pressure in the course of the nineteenth century, and was eventually replaced by forms of direct hiring and individual labour contracts. Mechanisation played a pivotal role in this process, both as a cause and a catalyst of change. In a powerful illustration of the value of the gang and its self-rule for the workers concerned, the transition from co-operative subcontracting to individual labour relations sometimes elicited bitter conflict. This was the case, for example, in Lancashire over the years 1855- 67.50 Probably already at the beginning of the century trade unions and

48 Lucassen, ‘Brickmakers in Western Europe .. and India’, 529; Joern Janssen, ‘On the Historical relevance of the Dispute about the Westminster Corporation of the Brickmakers 1636-1639’ (unpublished manuscript, London, July 1985, part of his ‘Fragments on the forma- tion of wage labour in building production’, ms. London, 11 September 1985); Joern Janssen, ‘The transformation of brickmaking in 17th century London’, Construction History Society Newsletter 71 (May 2005), 4, 9 (fn. 35, where an attempt is mentioned to establish an inde- pendent brickmakers’ guild for York in 1595). 49 Joseph Arnold Foster (ed.), Contributions to a Study of Brickmaking in America, Part 2, Statutes relating to brickmaking in England, 1729 to 1777. Together with excerpts from the journals of the House of Commons and the House of Lords relating to brickmaking in England (Claremont: privately printed, 1963), 1-34, 69-70; cf. also idem (ed.), Contributions to a study of brickmaking in America, Part 1, Statutes and proclamations relating to brickmaking in England, 1477 to 1728 (Claremont: privately printed, 1962), Foster, Contributions to a study of brickmaking in America Part 3; Joseph Arnold Foster (ed.), Contributions to a study of brickmaking in America, Part 4, Accounts of brickmaking in England published between 1800 and 1850 (Claremont: privately printed, 1969); The Gentleman’s Magazine, Vol. 54 (1784), 874. 50 Lucassen, ‘Brickmakers in Western Europe .. and India’, 530-33. Earlier on we only know of possible tensions because of an attempt in 1651 to fix maximum wage rates for 286 gijs kessler and jan lucassen burial societies of brickmakers had emerged in Manchester, Stockport and other industrial centres nearby. Labour conflicts concentrated on low piece rates for moulding and low advances, or a lack of sufficient work during the winter. Attempts of a combination of unions to enforce a closed-shop system and to prevent the employers from selling bricks outside the area covered by the union (obviously in an attempt to maintain the market price of bricks above a certain minimum) repeatedly invoked the ire of kiln own- ers. In such cases boycotts were initiated, often combined with violence towards the boss and his family, as well as other property. In 1862 a police constable was murdered, which was followed by the execution of one of the perpetrators. Apparently, in the aftermath of this conflict employers began to introduce moulding machinery which could be worked on a scheme of individual time wages instead of a cooperative subcontract with piece wages. This threat to the independent moulders and their gangs in turn led to machine breakage by union members. A similar transition took a much longer and gradual course in the case of the Lippe brickmakers. With the advent of mechanisation and the fading away of the seasonal character of brickmaking the fireman vanished as the primus inter pares in his gang; and during the 1890s the Zwischenmeister system gradually emerged, in which a fireman subcontracted with a fac- tory on a piece-rate basis, but employed the many men he needed on a time-wage basis. Existing organisations of Lippe brickmakers, which until then had united firemen and ordinary brickmakers, now fell apart, and by 1911 it had become clear that they could no longer be united in a single organisation.51 Not incidentally, the demise of the subcontracting system coincided with the emergence of more formal training in the brickmaking industry. During the nineteenth century the first proper manuals on brickmaking were published, aimed at producers and not at architects or builders.52 In

‘makers of brick and tile’ in Essex, see Janssen, ‘The transformation of brickmaking in 17th century London’, 6; John B. Smethurst and Peter Carter, Historical Directory of Trade Unions. Vol. 6, Including Unions in Building and Construction, Agriculture, Fishing, Chemicals, Wood and Woodworking, Transport, Engineering and Metal Working, Government, Civil and Public Service, Energy and Extraction in the United Kingdom and Ireland, Shipbuilding (Farnham: Ashgate, 2009), pp. 6-7 lists only 11 trade unions of brickmakers in England, first mentioned 1842-1874. 51 Lucassen, ‘Brickmakers in Western Europe .. and India’, 541; Lourens and Lucassen, Arbeitswanderung und berufliche Spezialisierung, 71-73. Apart from the strike movement in 1894, mentioned above, we have no information on the demise of the typically Walloon system. 52 For books treating the art of brickmaking see Campbell and Pryce, Brick, 99, 128-30, 170-80, 186-89, 206ff; for Germany see Seeling, ‘Belgische Ziegel-Wallonen und Feldbrand labour relations, efficiency and the great divergence 287 the wake of mechanisation professional schooling emerged with a winter course for brickmakers in the Lippe town of Lage, founded in 1891, and the more formal year-round Zieglerfortbildungsschule (brickmakers school) in the same place in 1906.53 In 1896 the Lippe town of Lemgo also founded a Zieglerschule with such a good reputation that it was visited by two Russian experts, the engineer Alexei Skachkov, director of the roof tile school of the Zemstvo of Novgorod, and Nicolai Chernov, director of the brick fac- tory attached to the same school.54 At the same time, mechanisation also appears to have involved a certain process of de-skilling, as witnessed by the increase of child and female labour. The widespread use of child labour in brickmaking, for which the industry gained a certain notoriety, is related mostly to these early phases of industrialisation, between c. 1850 and 1914. In South Staffordshire around 1870, for example, 75 percent of the labour force consisted of women and children; their husbands, fathers, brothers and sons worked at local collier- ies and ironworks.55 Another example are the seasonal workers, predomi- nantly women and children, from northern Italy who started to come to southern and central Germany at the end of the nineteenth century. In 1910-11 Gisela Lindner, the wife of sociologist Robert Michels published a detailed study about the abuses they had to endure.56

Russian Brickmaking between the Late Sixteenth and the Early Twentieth Centuries

The production of fired bricks presumably came to the Russian lands from Byzantium. The oldest known brick architecture, dating from the tenth and eleventh centuries, was built in so-called ‘plinfa’, a flat and broad type of brick originating from Byzantium, which subsequently spread over large

am Niederrhein (1)’, 225-58; cf. also Jars, ‘Die Kunst wie in Holland Ziegeln’ (1768), and Eversmann, Technologische Bemerkungen (1792). 53 Fritz Fleege-Althoff, Die lippischen Wanderarbeiter (Detmold: Meyer, 1928), 383-84; H. Linderkamp, ‘Auf Ziegelei’ an der Niederelbe. Zur saisonalen Wanderarbeit lippischer Ziegler im 19. und beginnenden 20. Jahrhundert (Stade: Stader geschichts- und Heimatverein, 1992), 67-68; Andrea Immenkamp and Ulrike Gilhaus, Museumsführer Ziegelei Lage: West- fälisches Industriemuseum. (Essen: Klartext, 2001), 65. 54 Staatsarchiv Detmold, L75 IV-15b, vol III (Cabinets-Acten betreffend ders. Zieg- lergewerbe 1890-1931), fol. 164-165. 55 Lucassen, ‘Brickmakers in Western Europe .. andIndia’, 531. 56 Lourensand Lucassen, Arbeitswanderung und berufliche Spezialisierung, 140-44. 288 gijs kessler and jan lucassen parts of Europe and Central Asia, from Georgia to Kiev and Bukhara.57 Although important, particularly in Kiev Rus’, brick remained, for a very long time, a construction material of secondary significance in the Russian lands. In the Northern parts of the country, which took the upper hand after the disintegration of Kiev Rus’, stone, and in particular wood were more widely used as construction materials.58 Abundantly available and relatively cheap, wood was the number one mass construction material until well into the nineteenth century in Mos- cow and St. Petersburg, and until well into the twentieth century in all other parts of the country. Brick construction was essentially limited to large-scale construction projects, and because of this the industry devel- oped in bursts and bouts, coinciding with periods of increased state and church construction activity. In the sixteenth and seventeenth centuries brick production was related above all to the construction of churches and monasteries, as well as of fortifications and city walls, necessitated by the territorial expansion of the Muscovy state.59 A new burst of activity fol- lowed at the start of the eighteenth century with the founding of St. Pe- tersburg, which was planned, after the Dutch example, as a city built in brick and stone. Brickmakers were mobilised from all over Russia and sent to St. Petersburg.60 In Moscow the industry went into almost complete decline until the late eighteenth century when Catherine II embarked on a large-scale urban reconstruction programme.61 By the early nineteenth century, however, the industry had largely disappeared again and was un- able to meet the huge demand for bricks required for rebuilding Moscow after the great fire of 1812. As a result the city was largely reconstructed in wood.62

57 A.V. Konorov, ‘K istorii kirpicha v Rossii v XI-XX veke’, Trudy instituta istorii estestven- noznaniia i tekhniki T. 7 (Moscow: Izdatel‘stvo akademii nauk SSSR, 1956), 182-85; P.A. Rappoport, Stroitel‘noe proizvodstvo Drevnei Rusi (X-XIII vv.) (St. Petersburg: Nauka, 1994), 5; Campbell, Brick, 144-47. 58 I.N. Cherniak, Ocherki po istorii kirpichnogo proizvodstva v Rossii (Moscow: Prom- stroiizdat, 1957), 10. 59 A.N. Speranskii, Ocherki po istorii Prikaza kamennykh del Moskovskogo gosudarstva (Moscow: RANION, 1930), 18-26, 44-48, 199-201. 60 S.P. Luppov, Istoriia stroitel‘stva Peterburga v pervoi chetverti XVIII veka (Moscow- Leningrad: Izdatel‘stvo akademii nauk SSSR, 1957), 99-105. 61 N.V. Voronov, ‘Stachka moskovskikh kirpichedel’chev letom 1779 g.’, Istoricheskie zapiski, 37 (Moscow, 1951), 290-91. 62 N.V. Voronov, ‘O rynke rabochei sily v Rossii v XVIII veke (po materialam kirpichnoi promyshlennosti)’, Voprosy istorii, 55 (1955), 98. labour relations, efficiency and the great divergence 289

Illustration 8.3. St. Petersburg 1904—handmoulder at work. Source: T.F. Sanotskii, Kirpichnoe proizvodstvo na r. Neve i ee pritokakh (St. Petersburg: A.E. Vineke, 1904), between pp. 44-45.

Only by the mid-nineteenth century a constant and expanding market for bricks emerged, sustained by both industrial and civil construction. This market was concentrated around the large industrial centres, mainly Mos- cow and St. Petersburg, but also Warsaw and, later, a couple of industrial centres in what is now Ukraine (Kharkov, the Donbass area).63 Because of the availability of secondary literature and published reports and other materials we focus on the two most important production areas, Moscow and St. Petersburg. The St. Petersburg brickworks were located on the banks of the Neva river and its tributaries, the Izhora, the Izhorka, the Slavyanka and the Tosna, as well as further upstream at Schlüsselburg, the gateway to lake Ladoga. The brickworks relied on barges for the delivery of fuel and

63 Cherniak, Ocherki po istorii kirpichnogo proizvodstva, 103-06. 290 gijs kessler and jan lucassen the transport of the fired bricks to the construction sites of the capital.64 In Moscow the brickworks were located on the southern outskirts of the town, along the highway to Kaluga, and in the North-East on the banks of the Khapilov pond. The lack of navigable rivers, with the exception of the Moskva river, which itself was apparently not suitable for brick production as there were no factories on its banks, meant that the Moscow brick in- dustry primarily relied on road and rail transport.65

Division of Labour and Co-operation in Production There is firm evidence for a division of labour among different categories of workers in Russian brick production only from the last quarter of the seventeenth century. Before that, brickmakers appear to have been essen- tially all-rounders, who moved from one construction site to the other to produce the bricks necessary for the erection of churches, monasteries, fortifications and other large edifices.66 Because of the predominance of construction in wood, they were few in numbers, and when the territorial expansion of the Muscovy state under Ivan the Terrible at the end of the sixteenth century required large-scale construction of brick fortifications, a centralised system of labour allocation was set up through which they were mobilised to ply their trade where state interests required this. The brickmakers enlisted through this Masonry Chancellery (Prikaz kamen- nykh del), set up in 1584, were annually despatched to construction sites to make bricks, which was an obligatory but remunerated labour duty, for which they were compensated through a number of tax privileges on the economic activities with which they supplemented their income from brickmaking, mostly trade.67 As a result, well into the seventeenth century brickmakers appear to have handled the entire chain of operations, from locating clay deposits to the construction of clamps and kilns to fire the bricks. Perhaps the actual excavation of the clay was carried out by unskilled local labourers or peas- ants. According to a decree issued by Boris Godunov in 1600 brickmakers

64 T.F. Sanotskii, Kirpichnoe proizvodstvo na: Neve i ee pritokakh (St. Petersburg: A.E. Vineke, 1904), 3-26, 61. 65 A.V. Pogozhev, Kirpichnogoncharnoe proizvodstvo Moskovskogo uezda: Opyt’ sani- tarno-promyshlennogo issledovaniia, Sbornik statisticheskikh svedenii po Moskovskoi guber- nii: Otdel sanitarnoi statistiki (Moscow: Izdanie Moskovskogo zemstva, 1881), T. 3, Vypusk 2, 47-184, contains short descriptions of most of the Moscow brickworks active in 1880. 66 Speranskii, Ocherki po istorii Prikaza kamennykh del, 13-15. 67 Ibid., 7-13. This 1930 in-depth archival study of the Masonry Chancellery is a monu- ment to the quality and vision of pre-Stalinist Soviet labour history. labour relations, efficiency and the great divergence 291

Illustration 8.4. St. Petersburg 1904—horse-drawn pug-mill and drying sheds. Of the eight to nine workers who served a pug-mill in operation, four can be seen at work in the pit on this photograph. Two of them are transporting the clay they have just dug out to the pug-mill on wheelbarrows, one is taking care of the mill and one is tending the horses. Of the other four men of the team, who supplied the moulders with clay, one can be seen at the top of the pit. Source: T.F. Sanotskii, Kirpichnoe proizvodstvo na r. Neve i ee pritokakh (St. Petersburg: A.E. Vineke, 1904), between pp. 44-45. were remunerated by piece-rates per 1,000 bricks.68 By the 1660s and 1670s, when construction activity increased and data become more abundantly available, a division of labour between kiln-men and moulders had emerged, in which the latter handled all stages in the production chain save for the actual firing of the bricks. The kiln-men were on time-wages, supplemented by a piece-rate for every fired kiln or clamp of bricks, mould- ers were on piece-rates per 1,000 bricks. We here find the first mechanism of error-rate reduction between stages in the production process. Moulders had to make between fifty and a hundred bricks more on top of every

68 Ibid., 94. 292 gijs kessler and jan lucassen thousand bricks they were paid for. This margin was to cover the percent- age of bricks which would be lost in the firing process.69 Also in the late seventeenth century fixed kilns and moulding sheds make their appearance at the Moscow factories of the Masonry Chancel- lery, and some of the enlisted brickmakers became settled kiln-men. Al- though nominally still hired workers in the service of the Chancellery, they effectively became small-scale entrepreneurs, who were not only engaged in the production of bricks, but also traded in construction materials and acted as subcontractors in the building trade.70 More or less simultane- ously, a change in hiring practices took place. Departing from the former exclusive reliance on enlisted urban craftsmen from the Masonry Chancel- lery for the moulding of the bricks, subcontracting of peasant brickmakers’ gangs, or so-called artely, emerged. Subcontracting by artely first makes its appearance in the mid-seventeenth century at ecclesiastical construction sites. Initially, these artely had no distinct division of labour—they per- formed all operations from digging the clay to firing of the bricks and dis- mantling of the clamps.71 But soon the kiln-men at the Chancellery brickworks in Moscow shifted to a different model of operations, contract- ing the production of a certain number of fired bricks from the Masonry Chancellery, and subcontracting peasant artely for the production of the green bricks.72 This became the dominant system of production until well into the nineteenth century, when kiln-men came to be recruited from the coun- tryside as well. The Masonry Chancellery was formally abolished at the start of the eighteenth century, and the mobilisation of large numbers of brickmakers from all over Russia notwithstanding, brick production for the construction of St. Petersburg in the early eighteenth century was also predominantly carried out through the subcontracting of migrant peasant labourers.73 Whether these peasant brickmakers knew a further division of labour remains unclear. Only by the late eighteenth century changes in production technique had resulted in the emergence of a new category of workers, at least in Moscow. This was due to the spread of a new moulding technique, which came from St. Petersburg.

69 Ibid., 82-87, 149-150. 70 Ibid., 150; Voronov, ‘O rynke rabochei sily v Rossii’, 93. 71 Voronov, ‘O rynke rabochei sily v Rossii’, 93. 72 Speranskii, Ocherki po istorii Prikaza kamennykh del, 194-196. 73 Cherniak, Ocherki po istorii kirpichnogo proizvodstva, 29-32.; Luppov, Istoriia stroitel’stva Peterburga, 99-105; Voronov, ‘O rynke rabochei sily v Rossii’, 94-95. labour relations, efficiency and the great divergence 293 Traditionally, Russian bricks had been moulded by foot, using a mould open on two sides (a bottomless box) placed on a low bench. In the St. Petersburg area a new moulding technique had been introduced during the early eighteenth century, in all likelihood imported from Holland, either by Peter the Great himself, or by a Dutchman called Van Ammers, con- tracted by Peter to set up the brick industry in St. Petersburg.74 This new technique relied on hand-moulding on a table in a mould closed on five sides (a topless box) and a more runny dough than was used in foot-mould- ing. Because of this last aspect it was probably more suited to the much wetter clay in the St. Petersburg area, but from there it spread to Moscow and other production areas. Its main advantage was a higher productivity in the moulding stage, but it allegedly complicated the drying-process.75 This gave rise, in the late eighteenth century, to the figure of the ‘dryer’ (sushnik), most likely performing the same task as the upganger in Western Europe. As soon as the moulded bricks—which first dried in rows on the floor or on stacked boards—were stiff enough he turned them on their edges. Subsequently, he carried them to the hacks where he rowed them up like a wall with some intervals between them to admit the wind.76 Due to this new division of labour, productivity allegedly went up by about one third during the 1770s.77 Moulders and dryers were migrant peasant labourers—serfs from seig­ niorial and ecclesiastical estates permitted to temporarily work elsewhere in exchange for a share of their earnings, the so-called quitrent or obrok, exacted by their legal owners. During the 1770s, for which data are available, annually between 600 and 700 such migrant peasant labourers worked in the Moscow brick industry. The season lasted from 10 May to 18 August. They were hired collectively in artely, and most of them returned every

74 Jean-Benoît Schérer, Histoire raisonnée du commerce de la Russie (Paris, 1788), 2 v, p. 45. Back in Holland, the members of the Van Ammers family who settled in Russia around 1700, belonged to a clan which delivered many a specialist brickmaker to the kilns in the Utrecht-Vianen region between the 1660s and the 1760s (preliminary results of our ongoing research on this topic; we thank Mr. Rob van Ammers for his genealogical information on this issue). 75 Cherniak, Ocherki po istorii kirpichnogo proizvodstva, 31, 41, 77, 131-31, 41, 77, 133; On the characteristics of the clay in the St. Petersburg area, cf. Sanotskii, Kirpichnoe proizvod- stvo na r. Neve, 41. 76 Cf. Foster, Contributions to a study of brickmaking in America part 3, 28-29 for late seventeenth-century England and Pogozhev, Kirpichnogoncharnoe proizvodstvo Moskovsk- ogo uezda, 20 for a description of the work of the dryer from late nineteenth century Mos- cow. 77 Voronov,‘O rynke rabochei sily v Rossii’, 96-97. 294 gijs kessler and jan lucassen year, often to the same factory. Most of these rural workers were moulders. Dryers accounted for only 12 to 15 percent. Kiln-men and overseers were hired from among the free population of the towns. The origin of the dig- gers is as yet unclear. Altogether the workforce engaged in brickmaking around Moscow amounted to some 800 people.78 The workforce per fac- tory varied greatly. Of the 25 registered factories in 1781 only 3 (12%) had fewer than 10 workers, 16 had between 10 and 50 workers (64%), and six factories had more than 50 workers (24%).79 Moulders were remunerated in piece-rates, apparently collectively per artel, and according to the same principle which had been in force in the late seventeenth century—for every 1,000 moulded bricks they had to make an extra 150 bricks to cover breakage in the course of the drying and firing process. Probably, dryers were remunerated according to the same principle, perhaps even collec- tively with the moulders, with the artel redistributing earnings afterwards. It is unclear how kiln-men were remunerated. Much more elaborate data are available for the late nineteenth and early twentieth centuries. A detailed report on the Moscow brick industry was published in 1881 by factory inspector A.V. Pogozhev, and a similar, if somewhat less detailed survey of production in the St. Petersburg area by inspector T.F. Sanotskii appeared in 1904.80 In addition there is a short description of the Moscow brickmaking industry in the 1870s in the journal of the Imperial Free Economic Society.81 The annual numbers of workers engaged in brick production had by then increased to 6,873 in Moscow (1880) and 13,104 in St. Petersburg (1903), which makes it a sizeable industry in both cases.82 These were exclusively migrant peasant workers. The divi- sion of labour by now involved separate diggers, moulders, dryers, transport workers and kiln-men. Sometimes the kiln-men were further subdivided into those placing the green bricks into the kiln (sadchiki), those doing the first, preliminary firing to extract the last moisture from the bricks (parsh- chiki), and those actually firing the bricks (vzvarshchiki). The loading of the kiln was also sometimes done by the dryers. The actual firing of the bricks was carried out exclusively by the two categories of kiln-men, collectively

78 Voronov, Stachka moskovski khkirpichedel’chev, 292-95. 79 Voronov, ‘O rynke rabochei sily v Rossii’, 97-98. 80 Pogozhev, Kirpichnogoncharnoe proizvodstvo Moskovskogo uezda; Sanotskii, Kirpi- chnoe proizvodstvo na r. Neve. 81 V.B., ‘Otkhozhii promysel Moskovskogo uezda (Polozhenie rabochego na kirpich- nykh zavodakh)’, Trudy imperatorskogo vol’nogo ekonomicheskogo obshchestva, T. 3 (St. Petersburg, 1878), 75-89. 82 Pogozhev, Kirpichnogoncharnoe proizvodstvo Moskovskogo uezda, 66. labour relations, efficiency and the great divergence 295 indicated as obzhigaly and to all evidence remunerated equally. Because of the more than twenty years between the two cases we will treat the Moscow and St. Petersburg production areas separately. In the Moscow brick industry around 1880 three main categories of workers can be distinguished: the diggers, the moulders and the kiln-men. Moulders came predominantly from the Kaluga province to the south, diggers from Kaluga and Riazan’, and kiln-men from Tula province. The main subdivision of the production process into stages ran parallel to the division between the three categories of workers. The diggers excavated the clay and deposited it at the moulding sheds, the moulders mixed and kneaded the clay, moulded the bricks and, depending on the moulding technique, took care of the drying and stacking, while the kiln-men han- dled the firing process. Hoffmann kilns had made their appearance at many Moscow brickworks, but still existed side by side with traditional fixed kilns. The predominant moulding techniques were the traditional foot- moulding, some table-moulding in a mould closed on five sides, some hand-moulding in a mould open on two sides, a few instances of fully mechanised moulding, but particularly the so-called ‘hand-machine moulding’ using cast-iron moulds set in tables with hand-operated presses and a foot-pedal lifting the brick out of its mould. This half-mechanised moulding technique had replaced the other techniques at most Moscow factories, but was itself in the process of being replaced by fully mechanised moulding using steam-powered machinery.83 Foot-moulding and hand-moulding in a mould open on two sides, the oldest two production techniques, required finishing after the initial drying to correct rough and jagged edges and to further compact the bricks by firmly hitting them several times on their broad sides with a flat wooden bat. This was the job of the so-called ‘corrector’ (pravil’shchik), also some- times referred to as the dryer (sushnik), which suggests that with these production techniques he also handled the other aspects of the drying process. What is puzzling, is that hand-moulding in a mould closed on five sides, the so-called ‘table’-moulding, the spread of which had allegedly initially led to the emergence of the dryer as a specialisation in the late eighteenth century (see above), did not apparently require the presence of this figure anymore in late nineteenth-century Moscow. Hand-machine moulding also did not require the presence of any ‘dryers’ or ‘correctors’.

83 This description is based on Pogozhev, Kirpichnogoncharnoe proizvodstvo Moskov­ skogo uezda, in particular pp. 11-31 and the descriptions of the 15 factories visited by the author on pp. 47-184. 296 gijs kessler and jan lucassen Indeed, the main advantage of hand-machine moulding with its mechan- ical press was reportedly that it did away with the necessity of this second round of compacting the bricks, resulting in lower labour costs per unit of production. This raises the question, though, of who handled the interme- diate stages of drying and stacking the bricks and taking them to the kiln for firing. Descriptions of the production process we have located so far, do not mention a separate category of workers for these tasks, so we will have to assume that moulders took care of these tasks, possibly assisted by the kiln-men in taking the bricks to the kiln.84 With the exception of the kiln-men, who were usually paid a monthly or a seasonal wage, the system of remuneration relied on collective piece- rates. Collective piece-rates ensure the efficiency of co-operation between the different workers within one stage in the production chain because it ties workers’ earnings to the result of their collective, rather than individ- ual effort. Moulders, the central link in the production chain, were without exception remunerated on piece-rates per 1,000 plus 150 bricks. Bricks were not counted individually, but stacked in standard patterns to allow a quick calculation of earnings, which took place in the middle and at the end of the season. Quality control was based on random checks—if these revealed defective bricks or intentionally irregular stacking, the moulders had to make ten extra bricks for every bad brick found. Discovery of defective bricks was perceived by employers as a major breach of trust—it would invoke him or his overseer to start smashing up the defective bricks until his ire subsided, destroying many a good brick as well in the process. Work- ers apparently were not in a position to protest against such behaviour.85 This underlines the importance of quality controls in systems of piece-rate remuneration and in the absence of hand-on supervision. The finishing of the bricks required by the traditional moulding techniques was also some- times remunerated on piece-rates, as well as in a few cases the firing of the bricks. As far as the diggers were concerned, an additional mechanism was in place to ensure their efficient collaboration with the moulders. Diggers were mostly paid a piece rate per 1,000 moulded bricks, which effectively meant that their remuneration was made conditional upon the efforts of the moulders. This arrangement clearly reflected the hierarchy of profes- sions at the brickworks. The productivity of the moulders depended on a

84 V.B., ‘Otkhozhii promysel Moskovskogo uezda’, 77-79; Pogozhev, Kirpichnogonchar- noe proizvodstvo Moskovskogo uezda, 22-23. 85 V.B., ‘Otkhozhii promysel Moskovskogo uezda’, 75. labour relations, efficiency and the great divergence 297 timely and steady supply of the right type of clay, and remunerating the diggers according to the output generated by the moulders served to mi- nimise supply disturbances and to synchronise efforts in the two adjacent stages of the production chain. This payment system of course involved a risk for the diggers in case the moulders would not work hard enough, but in practice this risk must have been negligible, both because the moulders were on piece-rates themselves, and because they were, as a rule, very ex- perienced and highly skilled workers who had been working in the indus- try for years. The reliance on piece-rates, accompanied at every stage of the produc- tion process by the requirement to produce an extra 150 bricks on top of every 1,000 to cover losses during the subsequent stage in the production process was a system of remuneration which protected the employer from the financial consequences of poor quality work and inefficient co-opera- tion, while relieving him from the task of direct supervision and manage- ment of the work process. For workers it meant a great deal of autonomy and the promise of a material reward directly proportional to the effort applied. It should be noted, however, that not all workers fully benefited from their side of the bargain. Most diggers, as well as kiln-men in some cases, were in fact hired on time-wages by subcontractors who themselves were paid piece-rates by the employer, a system which we have also seen in the Dutch Rijnland.86 We should be under no illusions about who was to benefit from such subcontracting arrangements—workers would with- out any doubt have been able to earn more had they been remunerated directly on piece-rates, but for whatever reason they could or would not enter into a direct labour relation with the factory owner, and worked through middlemen instead. The only category of workers that was invariably hired directly and re- munerated by the employer were the moulders. They were contracted in winter in the village by recruiters sent down from the factories, paid an advance, which they mostly used to pay their land taxes, and obliged them- selves to work for the duration of the season, from early May until the middle or the end of August. Just as in late eighteenth-century Moscow, workers were hired in artely, but now the unit of collective remuneration appears to have been in most cases not the artel’ but the pair of moulders working together in one moulding shed, or, in case of hand-moulding using a mould closed on five sides, the three-men ‘table’ working as one unit. The

86 Pogozhev, Kirpichnogoncharnoe proizvodstvo Moskovskogo uezda, 62, 95, 104, 109-62, 110, 115-116. 298 gijs kessler and jan lucassen collective contracts drawn-up at recruitment were signed on behalf of all couples or ‘tables’ separately, and the workers of one couple or ‘table’ were mutually responsible for the advance they had been paid in case one of them failed to turn up at the factory.87 Technically, this means the artel’ no longer was the contracting unit, and therefore no longer the collective mechanism of error-rate reduction it had been before. Instead, it now had primarily a consumption function, for collectively hiring a cook and buying food when at the factory, and perhaps sharing some of the costs connected to the annual journey to Moscow and back.88 What were the reasons behind this reduction of the role of the artel’? It is tempting to see it as an employer-driven process, intended to link remu- neration more closely to individual effort. In any case, this was the aspect of the system which was evidently contested. When a couple of moulders would fall idle because of insufficient supply with clay, sand or water by the diggers (a prime example of error-rates between stages in the produc- tion chain), they would start to loudly bang with a stick on an upturned bucket for their fellow artel’ members to hear, and these would also im- mediately stop working until they were all supplied with raw materials again, and in an equal position to make the best possible returns on their skills.89 Spreading the negative effects of inefficient co-operation between different groups of workers to envelope the entire production process at a factory, these work stoppages were greatly resented by employers and con- tracts could contain explicit clauses forbidding it.90 From the perspective of the workers the rationale behind this artel’- solidarity was probably not so much to equalise earnings as such, as rather to equalise income-earning opportunities. This strikingly recalls the un- derlying principle of the re-distributive peasant land commune which was still in existence at that time in Russia, and, it has been argued, aimed to provide peasants with equal access to land, the single most important re- source for generating income in the rural economy.91 Indeed, given the fact that off-farm earnings like the work at the brickyards served in the very first place to pay taxes, for which the peasant commune was collectively

87 Ibid., 131-132, 185-186. 88 V.B., ‘Otkhozhii promysel Moskovskogo uezda’, 81-83. 89 Ibid., 47-73, 83-90. 90 Pogozhev,Kirpichnogoncharnoe proizvodstvo Moskovskogo uezda, 131-32. 91 Cf. Peter Gatrell, The Tsarist economy 1850-1917 (London: B.T. Batsford, 1986), 12; Dorothy Atkinson, ‘Egalitarianism and the commune’, in: Roger Bartlett (ed.), Land com- mune and peasant community in Russia: Communal forms in Imperial and early Soviet Society (New York: St. Martin’s Press, 1990), 11. labour relations, efficiency and the great divergence 299 responsible, equalising opportunities on the shop-floor not only reflected the central tenets of the rural peasant economy, but also made plain math- ematical sense, considering the fact that the failure of fellow villagers to earn enough to cover their share of the taxes would mean a larger burden for those who were in a position to pay.92 Brick production in the St. Petersburg area took place along the Neva river and its tributaries. Factory inspector T.F. Sanotskii’s 1904 description of the industry shows a much greater degree of standardisation and uni- formity in the organisation of the production process across factories than in late nineteenth-century Moscow. The production chain revolved around horse-powered pug-mills placed at or near the clay-pits and served by eight to nine workers, four of whom dug up the clay and fed it to the mill, while four supplied the moulders with clay from the mill. Each mill provided clay to eight moulders. A brickwork could consist of several such pug-mill cen- tred production units.93 Moulding was still mostly done manually, relying on the technique of ‘table’-moulding with a mould closed on five sides, as it had originated in the St. Petersburg area at the start of the eigh­teenth century. At 13 out of 52 factories visited by Sanotskii moulding had been mechanised and relied on steam-powered moulding machines. The Hoffmann kiln was by now universal, but production was nonetheless sea- sonal. The moulding season lasted from mid-May to 29 August, September was used for preparing the clay for next years’ season, and firing took place from mid-March to 1 November.94 Leaving the factories relying on mechanised moulding aside, five or six categories of workers were involved in the production process. First of all, these were the diggers already mentioned above, who extracted the clay, fed it to the pug-mill and subsequently supplied it to the moulders. Second in line were the moulders. Unlike in late nineteenth-century Moscow, they no longer kneaded and mixed the clay. They merely moulded the bricks and left them out to dry in the open. Subsequently the bricks were gathered and stacked in racks to dry by ‘trimmers’ (obrezchitsy). Each pug-mill re- quired two such trimmers, i.e. one for each four moulders. Next came the dryers (sushniki) who transported the dried bricks to the kiln when ready for firing. They also often loaded the kilns—if not, this was done by the

92 On the collective tax responsibility of the members of the commune and its relation to labour migration, cf. Jeffrey Burds, Peasant dreams and market politics: Labor migration and the Russian village, 1861-1905 (Pittsburgh: University of Pittsburgh Press, 1998), 51-64. 93 Sanotskii, Kirpichnoe proizvodstvo na r. Neve, 70-71. 94 Ibid., 41-47, 51, 65. 300 gijs kessler and jan lucassen ‘loaders’ or sadchiki. The firing of the bricks was the work of the kiln-men or vzvarshchiki. The fired bricks were transported to St. Petersburg on barges, which were loaded by a special category of workers, or sometimes by the dryers. The geographical origin of workers was more diverse than in Moscow; about half of all workers were from Vitebsk province in modern-day Be- larus, around ten percent each from Novgorod province, the St. Petersburg province itself, and the Kaluga province, but from two different districts than those sending workers to Moscow. The remaining twenty percent of the workers came from a great variety of other places. Regional specialisa- tion did exist, but was not overly pronounced. Most workers belonged to the Russian Orthodox faith, fifteen percent were Roman Catholics, and 2.5 percent were Old Believers. Most of the workforce were men; women ac- counted for just under fourteen percent. The latter worked at the factories which relied on machine-moulding or as trimmers, and mostly were mar- ried to the men working at the factory.95 Mechanisms of error-rate reduction focused primarily on the efficient collaboration between diggers, moulders, and trimmers, i.e. the three cat- egories of workers whose cumulative efforts determined the number and quality of the dried bricks that went into the kiln. Moulders and trimmers were paid in piece-rates per 1,000 plus 210 bricks according to the principle which set apart a fixed percentage to cover eventual breakage. It appears they were not dependent on each other in terms of remuneration—mould- ers were paid for the bricks they had moulded and the trimmers for the bricks they had placed in the drying-racks. Diggers, on the other hand, were almost exclusively remunerated in piece-rates per 1,000 plus 210 moulded bricks. Only in some cases they were paid time-wages. Diggers’ and mould- ers’ efforts were thus systematically linked at the level of remuneration, which enhanced their close collaboration in the small production units centred around pug-mills characteristic for the St. Petersburg area at that time. At 210 the margin for breakage was substantially higher than the 150 which had been customary in the 1880s in Moscow. Kiln-men were exclu- sively paid time-wages, dryers were sometimes paid in piece-rates per 1,000 bricks placed in the kiln (apparently without the margin of 210 extra), but mostly also in time-wages.96 The collaboration between workers paid in time-wages and those on piece-rates was a potential source of disruption in the production process,

95 Ibid., 61, 66-68, 70-75. 96 Ibid., 86-87. labour relations, efficiency and the great divergence 301 if the latter depended on the input of the former. Workers paid in piece- rates were free to determine their own working hours. As a rule they would work not only very long, but also very irregular hours and this sometimes caused problems when the moulders, who were on piece-rates, would start work at three o’clock in the morning, when the diggers, who were paid time-wages at some factories, would not have started work yet, causing the moulders to run out of clay. Against this background it is perfectly under- standable why employers would pay diggers piece-rates linked to the out- put of the moulders—indeed, the question is why they did not always do this. One of the reasons might have been the substantial diversity in geo- graphic origin of the workers at the Neva brickworks, in particular the difference between Roman-Catholic workers from Vitebsk and Orthodox workers from other provinces. Because of the different church calendars Roman-Catholic and Orthodox workers would observe religious holidays on different days. Obviously, this would make it difficult to directly link the remuneration of diggers to that of the moulders. 97 Workers, in any case moulders and diggers, were hired in winter in their villages of origin, to all evidence directly by the employer’s representatives without recourse to intermediate subcontractors. Workers signed a con- tract, were paid an advance, and handed over their passports (internal passports, which allowed them to travel outside their areas of residence). Nearer the start of the season the employer would send them an extra advance to cover travel expenses. Trusted older workers, who had often been employed by the factory for years in a row, would simply send their passports to the employer and would receive an advance in return. Workers would come to the brickworks in artely, which again seem to have fulfilled primarily a consumption function. They also sometimes brought along younger workers who would learn the trade on the job from the older moulders, often their direct relatives. While in training these would be paid a time-wage, not by the employer, but by the moulders themselves. From the point of view of the employer the artel’ offered another benefit—in case workers hired in winter would not turn up at the start of the season, employers would hold the other members of the artel’ collectively respon- sible for the advances paid to these absentees. In 1903, however, such forms of collective responsibility were outlawed, and this undermined the posi- tion of the employer in the advance-system.98

97 Ibid., 97. 98 Ibid., 69, 86. 302 gijs kessler and jan lucassen Whereas at the Moscow brickworks the artel’ had still played a role in equalising income-earning opportunities, through a self-imposed artel’ solidarity, a different practice had evolved in the St. Petersburg area. At the start of the season, as workers were assigned to the different pug-mill cen- tred production units, workers would hold an auction to compensate for the differences in the productive potential among pug-mills and clay-de- posits. The best mills and pits were auctioned and the proceeds were used to compensate the workers who had been assigned to the worst. What remained of the money was collectively consumed. Different auctions were held for diggers and moulders, as at one particular site the mill might be advantageous to work with, but the pits inconvenient, and vice versa. In such cases the moulders also often paid a small compensation to the dig- gers. The problems with these compensatory arrangements was that they took place at the very start of the season, when none of the workers had any cash. Only at pay-day, at the end of the season, would the compensa- tion actually change hands, and often by that time arguments arose over the exact agreements, or the workers who had to pay had already left, or had been fired. In such cases the employer usually honoured the arrange- ments agreed upon, even if he was under no obligation to do so.99 Thus, we see a mechanism at work similar to that in Moscow, the pur- pose of which was to provide all workers with equal opportunities to get the most out of piece-rate remuneration. In Moscow this practice had fitted in with the collective rural tax-responsibility of the workers, but this tax- responsibility had been abandoned in 1903, the year before the publication of Sanotskii’s report on the St. Petersburg brick-industry.

Production and Productivity In the light of what we have seen in the preceding sections, let us now try and establish how these changes in the division of labour and organisation of the production process affected productivity. The tables 8.4a and 8.4b line up a selection of the statistical evidence available for the Moscow and St. Petersburg production areas over the period covered by this chapter. More or less comprehensive data are available for the late eighteenth, the late nineteenth, and the early twentieth centuries. For the late seven- teenth century we only have some incidental data on moulders’ output. The upper five rows (a-e) of table 8.4a characterise production, size and number of production units and the average annual size of the workforce.

99 Ibid., 98. labour relations, efficiency and the great divergence 303

Table 8.4a. Brick production in Moscow and St. Petersburg, 1770-1904. Moscow Moscow uezd St. Peters­burg St. Peters­burg 1770-1782 1878-1880 guber­niia guberniia 1879 1900-1904 a Annual production 20 140 143 294 (millions of bricks) b Number of factories 30 70 48 57 c Number of workers 838 7,432 7,713 12,673 d Average annual 667 2,000 2,980 5,106 production of bricks per factory (thousands) (a/b) e Average number of 28 106 161 221 workers per factory (c/b) f Average annual 23,866 18,837 18,544 23,132 production of bricks per worker (a/c) g Average production 477 377 371 463 of bricks per worker per day (assuming 50 working days in a season) (f/50) h In m3 clay consider- 1.3 0.9 0.9 1.1 ing predominant brick size

Production increased many times over the period under study, from 20 million bricks in late eighteenth-century Moscow to 294 million in early twentieth-century St. Petersburg. The number of factories also increased, but above all they became larger, producing more bricks on average per season, and employing a larger workforce, increasing from an average 28 workers in late eighteenth-century Moscow to 221 in early twentieth-cen- tury St. Petersburg. Total employment in the industry rose from a modest 838 workers in late eighteenth-century Moscow to over 12,500 workers in St. Petersburg at the turn of the century. The lower three rows (f-h) of the table gauges productivity. A first exer- cise is the simple division of total annual production by the number of workers employed, which gives us the average annual production per in- dividual (f). For comparisons across continents, though, average annual production per worker is not a very good indicator, because it varies di- 304 gijs kessler and jan lucassen

Table 8.4b. Moulder productivity in Moscow and St. Petersburg, 1670-1904. Moscow Moscow Moscow uezd St. Peters­burg 1670-80s 1770-1782 1878-1880 guberniia 1900-1904 Reported average daily 200-460100 500 750101 1500 production of a moulder In m3 clay considering 0.7-1.2 1.4 1.8 3.6 predominant brick size Reported average sea- 28-40,000102 70-90,000 sonal production of a moulder Reported average sea- 83-85 60-100103 90-130 sonal earning of a moulder (rubles) Sources: Speranskii, Ocherki po istorii Prikaza kamennykh del, 85-87; Voronov, ‘O rynke rabochei sily v Rossii, 97-98; Voronov, ‘Stachka moskovskikh kirpichedel’chev letom 1779 g.’, 291, 295; V.B. 1878, 75, 78; Pogozhev, Kirpichnogoncharnoe proizvodstvo Moskovskogo uezda, 20, 25-26, 47-73; Sanotskii, Kirpichnoe proizvodstvo na r. Neve, 64, 87. rectly with the length of the production season, which differs around the globe. 100 101 102 103 A much better instrument of comparison is the average production per day, as listed in table 8.3 for Western Europe. In Russia the moulding season lasted from early May to mid-late August. Because of the strenuous nature of the work many days of rest were observed, though, as well as numerous religious holidays. The number of actual working days per season was re- ported to have been 42 in Moscow in 1878 (or about five days a week), and

100 Recalculated from a production norm of 10,000 per season for enlisted brickmakers and reported weekly production at a private brickyard of 2350 per week assuming that, as reported for the eighteenth and nineteenth century, moulders on average worked not more than 5 days a week, and therefore the moulding season numbered 50 working days. This gives a daily production of 200 for the enlisted brickmakers and of 460 for the private. Cf. Speranskii, Ocherki po istorii Prikaza kamennykh del, 85-87. 101 Recalculated from a reported average daily production of 1500 per couple, or 2000 per three-men ‘table’. Cf. Pogozhev, Kirpichnogoncharnoe proizvodstvo Moskovskogo uezda, 20, 25-26. 102 Recalculated from a seasonal production per three-man ‘table’ at the Gusarev factory of 85-110,000 bricks. Cf. Ibid., 26. V.B., ‘Otkhozhii promysel Moskovskogo uezda’, 78 gives a slightly higher upper limit of 40,000 bricks per moulder per season. 103 Based on a sample description of 12 factories in Moscow and immediate surround- ings. Cf. Pogozhev, Kirpichnogoncharnoe proizvodstvo Moskovskogo uezda, 47-184. labour relations, efficiency and the great divergence 305 50-60 in early twentieth-century St. Petersburg.104 Because start and end of the moulding season are not reported to have differed between the two locations, however, we have assumed that on average in both locations a production season accounted for approximately 50 working-days. Although other workers, like the firemen, worked a longer season, the production of the moulders eventually determined the total seasonal production and therefore we have used the number of their working days to calculate aver- age labour productivity in physical terms for the entire workforce engaged in brickmaking. Dividing average production per worker per season by the number of work-days produces an average daily production per worker (g). Finally, for a proper comparison over time, we have to take into account changes in brick size, and to correct for this we have recalculated daily production per worker into cubic metres of processed clay (h).105 Trends for all three rows amount to the same: a decline in productivity between the late eighteenth century and the late nineteenth century, and an increase between the late nineteenth century and turn-of-the-century St. Petersburg. This at greatly different levels of overall production, though, which increased almost fifteen times over the same period. The decline in output per worker between late eighteenth and late nineteenth-century Moscow is puzzling and raises doubts as to the accuracy and comparabil- ity of the late eighteenth-century data, unless we assume it was caused by the enlargement of the scale of production and the hiring of less experi- enced workers, but pending further research we will leave this trend aside in our analysis. The rise in productivity between the late nineteenth and the early twentieth century, meanwhile, conforms to expectation, consid- ering the changes in the division of labour and production technology which have been described above—the separation of moulding, mixing, and drying into different specialisms, and the mechanisation of the mixing allowed for further increases in average output per worker. Crucial in achieving these productivity gains was the role of the mould- er. If we look at the separate data on moulder productivity in table 8.4b we can easily see that it was here that the main increases in productivity were realised. Unlike table 8.4a, which relates to fired bricks, the finished pro­duct of the factory as a whole, table 8.4b measures productivity in green bricks—

104 Cf. Ibid., 25-26; Sanotskii, Kirpichnoe proizvodstvo na r. Neve, 97. 105 In 1584 a first standard was introduced, stipulating a brick size of 312 × 134 × 89 mm (or 3721 cubic centimetres). It was replaced in the early eighteenth century by a new stan- dard of 280 × 140 × 70 (or 2744 cubic centimetres). In 1847 this was replaced by a standard of 267 × 133 × 67 mm (or 2379 cubic centimetres), in existence until 1927. Cf. Konorov, ‘K istorii kirpicha v Rossii v XI-XX veke’, 199, 202, 208. 306 gijs kessler and jan lucassen the final product of moulders’ efforts. They reveal a somewhat different trend. To start with, the average daily production of green bricks of a moulder as reported in contemporary descriptions shows a constant up- ward trend, in particular from the late eighteenth century on, from 500 to 750 in the late nineteenth century (regardless of the moulding technique used), and doubles over the last quarter of the nineteenth century to 1500 in early twentieth-century St. Petersburg. Recalculated into cubic metres of clay processed to correct for changes in brick size the increase in pro- ductivity is slightly less steep, but the trend remains the same. The second last row of the table lists the reported average seasonal production of a moulder at these levels of productivity to check for any inconsistencies in our calculations or readings of the sources, but again the picture is the same—that of a continuous increase in moulder productivity over the period under study. But what the data in the tables 8.4a and 8.4b also reveal is that between the late nineteenth and early twentieth century the increase in moulder productivity was much steeper than that of productivity for the workforce as a whole. Whereas the daily average output per worker increased only by a quarter, moulder output doubled. Increased moulder productivity due to specialisation must have made it imperative to hire more workers to supply them with clay, to tend to the drying of the bricks and their transportation to the kiln. As a result average output per worker also grew, but less steeply. Meanwhile, because such auxiliary workers were cheaper, total labour costs to the employer remained practically stable, as we can see in table 8.5. These sample data on the costs of production per 1,000 bricks for 1880 and 1904 reveal two changes—a dramatic reduction of expenses on fuel and a redistribution of labour costs between different phases in the produc- tion chain in sync with the changes in the division of labour. Leaving aside expenses on fuel, which decreased manifold due to the shift to Hoffmann kilns and coal instead of wood, total production costs were nearly identical at 6.15 roubles per thousand in 1880 and 6.09 roubles in 1904. Within this overall figure, however, labour costs were distributed quite differently. In 1904 St. Petersburg, a larger share went to the diggers and the workers operating the pug-mills, which reflects the larger share of auxiliary workers among the workforce. Moulders accounted for a lower share, but for a more specialised task, because they no longer mixed and kneaded the clay nor attended to the drying and stacking of the bricks, which was now done by separate groups of workers. Comparison to 1880 is complicated by the fact, though, that we do not know exactly which part labour relations, efficiency and the great divergence 307

Table 8.5. Production costs per 1000 bricks 1880 (Moscow) and 1904 (St. Petersburg) 1880 1904 roubles % roubles % Supply of clay, water and 1.75 12.4 2.13 27.7 sand to the moulders Moulding 2.15 15.2 1.55 20.2 Drying, stacking and 0.95 12.3 transport to the kiln Firing (labour) 1.25 8.8 0.35 4.6 Firing (fuel) 8.00 56.5 1.60 20.8 Repairs a.o. 1.00 7.1 1.11 14.4 Total 14.15 100.0 7.69 100.0 Source: Pogozhev, Kirpichnogoncharnoe proizvodstvo Moskovskogo uezda, 2; Sanotskii, Kirpichnoe proizvodstvo na r. Neve, 58. of this work used to be carried out by the moulders and which by the kiln- men, who were both involved in it in the late nineteenth century. This also makes it difficult to interpret the smaller share of the total labour costs which went to the firemen, because it could be that they as well had shed some tasks to the newly specialised workers who took care of the drying, stacking and transport to the kiln. Thus, changes in the division of labour and a more narrow specialisation of the crucial moulders, together with mechanisation of part of the produc- tion process, raised output per worker at equal costs to the employer. On top of this, changes in the firing technique and ovens and fuel used, slashed total production costs per 1,000 bricks by almost half. But there is some evidence that these improvements came at a price: whereas in late nine- teenth-century Moscow employers applied a percentage for breakage of 15 to moulders’ piece rates, this percentage had increased to 21 in early twen- tieth-century St. Petersburg (see above). This could point in the direction of increased error-rates due to a more segmented division of labour, but it could of course also simply have been related to changes in the production technique, for example the mixing, which resulted in green bricks with a greater risk of impurities regardless of the moulders’ skills.

Indian Brickmaking from c. 1800

Both the first moulded and the first fired bricks were made in Mesopotamia around 5000 BCE, but only from 3000 BCE did the innovation spread to the 308 gijs kessler and jan lucassen west and east, possibly reaching India around 2000 BCE.106 Whether the large-scale production of moulded and fired bricks remained in existence ever after is still an open question. A continued production can be dem- onstrated, however, from the advent of Islam in the Indian Subcontinent until today.107 In the North the Delhi sultans, the Moghuls and their follow- ers, like the rulers of Awadh, as well as the disciples and followers of Hindu revivalism in Bengal (mainly in the eighteenth and nineteenth centuries) built extensively in brick.108 Unfortunately, though, substantial records about brickmakers so far are available only from the nineteenth century onwards. India, like Europe, traditionally has two quite different types of brick- making. At the village level, as well as for maintenance, bricks are produced by family craftsmen, but for the larger building sites, seasonal workers are attracted, usually from elsewhere. Large groups of local sedentary brick- makers, like we encountered in Early Modern Europe in Holland, around London, in Lancashire, and everywhere in Western Europe in the mecha- nised factories from the second half of the nineteenth century onwards, are not the rule in India. Small-scale brickmaking is part of the job of the kumhar or potters’ caste.109 There is, however, no evidence that those workers engaged in large-scale brickmaking tend to belong to the kumhars’ caste. The largest production units originally were found near public works, like temples or mosques, palaces, fortresses, water tanks or city walls, as well as later on canals, railways, barracks, jails and other essential buildings of the colonial state and its successors.

106 Campbell , Brick, 22, 30-37. 107 T.N. Mishra, Ancient Indian bricks and brick remains (New Delhi: Harman, 1997); Emani Siva Nagi Reddy, Evolution of building technology: Early and Medieval in Andhradesa, 2 vols., (Delhi: Bharatiya Kala Prakashan, 1998); Munni Pareek, Early Indian residential architecture (New Delhi: Harman, 2002); on the continuity in pottery fabrication see Baid- yanath Saraswati, Pottery-making cultures and Indian civilization. (New Delhi: Abhinav Publications, 1978), 102-13. 108 Rosie Llewellyn-Jones, A fatal friendship: The Nawabs, the British, and the city of Lucknow (Delhi: Oxford University Press, 1985), 214-25; George Michell (ed.), Brick temples of Bengal: From the archives of David McCutchion (Princeton: Princeton University Press, 1983); Ahsan Jan Qaisar, Building construction in Mughal India: The evidence from painting (Delhi: OUP, 1988). 109 Daniel Houston Buchanan, The development of capitalistic enterprise in India (New York: MacMillan, 1934), 84-85; Saraswati, Pottery-Making Cultures, 55-56; Carol Kramer, Pottery in Rajasthan: Ethnoarcheology in two Indian cities (Washington and London: Smith- sonian Institution Press, 1997), 46-47. labour relations, efficiency and the great divergence 309 The construction of the Ganges Canal in northern India in the mid- nineteenth century provides a good example. It involved the burning of half a billion bricks for bridges, aqueducts, locks, and so on. The gangs possibly amounted to over 3,000 men, most of whom seem to have come from a vast area with a radius of up to 500 kilometres.110 Modern brick factories, still largely dependent on traditional, manual techniques, have on average 200 labourers each, most of whom are engaged in the moulding. At the start of the current millennium this industry is estimated to employ some 6 million people in India alone. Around large cities, like New Delhi or Calcutta, factories are placed very near to each other, resulting in a con- centration of thousands of brickmakers in a small area.

Division of Labour and Co-operation in Production Like in Western and Eastern Europe, brick production in India is almost universally based on collective subcontracting, but contrary to European practice family-based groups are the rule within this system. Although the brick production for the Ganges Canal appears to have been carried out exclusively by men, from the 1860s onwards boys are mentioned and soon also women and complete families. As in Europe, collective subcontracting is the principal system for error-rate reduction within the production pro- cess through collective remuneration and the differential redistribution of earnings within the group according to levels of skill, experience, and dex- terity. In India, collective subcontracting divides the production chain into three parts, with different units for the excavation, tempering and mixing of the clay, the moulding, the loading and unloading of the dried bricks, and for the actual firing. At the Ganges Canal the moulding gangs, working on piece rates, consisted of eight men: three beldars for excavating and carrying the earth to the moulding terraces, one for breaking the clods and for the rough tempering, three for mixing and carrying the clay, ashes and water to the moulder, and lastly the moulder himself. Moulders received 6 to 7 rupees a month, beldars 4 rupees, and boys 2 to 3 (piece rates con- verted into time rates).111 The firing in clamps was subcontracted by sepa- rate gangs, as is documented at least since the 1840s. They were nearly twice as numerous as the moulders’ gangs.112 Labour productivity of the mould-

110 Lucassen, ‘Brickmakers in Western Europe .. and India’, 547. 111 Ibid., 548. We assume that these are in fact piece rates, converted for convenience’s sake into average monthly wages. 112 Ibid. 310 gijs kessler and jan lucassen ers around 1850 may be estimated at 800 bricks per day per man.113 This is more or less on a par with contemporary Russia (cf. table 8.4b), but com- pares less favourably to Western Europe (cf. table 8.3). If correct, this could possibly offer some background to British engineers’ constant complaints about the poor craftsmanship of the Indian labourer, even if they praised the quality of the bricks.114 Regarding the contemporary situation we have much more elaborate information on composition and social background of brick workers. The fact that manual brickmaking and collective subcontracting are not at all a thing of the past in India offers a unique opportunity to study such a system of labour relations in practice, in particular through observation and interviewing at brickworks in northern India.115 In the contemporary industry as well the production process is divided into three stages, con- tracted by different groups of workers, mostly family-based. Moulders work as small family groups consisting of parents and children old enough to assist them, sometimes other family members as well. Chil- dren can start working from 6-7 years of age and from age 10 they work at full speed. Moulders’ families organise the different tasks from tempering to drying completely among themselves. In fact, digging is no separate job as these families work inside the clay pit which is about two metres deep.116 Squatting under the walls of the pit, the clay is available to them at arm’s length, they scrape it from the wall with a hoe, mix it with water, make a ball and the moulding can start. The green brick is turned on the flattened ground and the moulder moves backward, leaving neatly arranged rows of

113 R.S. Stirling, Report on two brick making machines sent out to India in 1851 (Alipoor: Iron Bridge Yard, 1858), 10 (10,000 bricks per diem by 12 hand-brick moulders); Jan Lucassen,‘The Brickmakers’ Strikes on the Ganges Canal in 1848-1849’, in: Rana P. Behal and Marcel van der Linden (eds), India’s labouring poor: Historical studies, c.1600-c.2000 (New Delhi: Cambridge University Press, 2007; first published in International Review of Social History 51, 2006), 58. 114 Lucassen, ‘The Brickmakers’ Strikes’, 82-83. 115 Most information has been collected by Jan Lucassen during interviews in the years 1997-2001 on some twenty different factories all over northern India, but most near New Delhi. See esp. interview on Factory no. 4 (Interview 06-02-1997). All reports will be made available through the IISH (collection Brick Works India). On modern women and child labour at Indian brick kilns there is a vast literature, e.g. Mitha, Yameema et al. Solid foun- dations, solid contributions: Women in the brick kiln industry. Patterns of Female Employment in Mining and Construction Industries (Lahore: ASR, 1989). See also Georges Kristoffel Lieten, Ravi Srivastava, and Sukhadeo Thorat, Small hands in South Asia: Child labour in perspective. (New Delhi: Manohar Publishers & Distributors, 2004). 116 An exception are the alluvial clay pits south of Calcutta which are filled with silt, and subsequently dug out by specialized digger-gangs. labour relations, efficiency and the great divergence 311

Illustration 8.5. At the Akra brick factories south of Kolkata, established in the mid- nineteenth century and still owned by the West Bengal Government, a gang of 35 men is engaged in transporting the river sand to the moulding pit, digging the clay, mixing sand and clay in an iron pug-mill and delivering it to the moulders’ gangs. On this photograph, taken on February 21st, 1997, five gang members at the end of this chain are visible, plus a little boy keeping the oxen-drawn mill beam down. The gang, hailing from Murshidabad (250 kilometers north of Akra) receives a task wage. One of them, the primus inter pares called sardar, collects the total wage sum from the subcontractor who has rented the factory from the Government. On average the gang members earn 1,300 rupees per month. To the right of the pit four women are waiting to carry the mixed clay to the moulders, as a rule their husbands. Source: IISH, collection Jan Lucassen on Indian brickmaking, factory 10, photo 12-17. green bricks in front of him. Members of this group also take care of the hedging. Thus, in India, the most crucial production stages from clay to brick are all performed by the moulders’ gangs, with the issue of error-rate reduction between the stages solved by the inherent internal logic of col- lectively remunerated subcontracting units we have identified above. Separate gangs of loaders then take over the green bricks and deliver them to the kilns. They come in groups from different rural parts of north- ern India and have to travel across long distances. At one factory it is unusual to find more than two groups with different geographical origins. This is caused by the fact that factory owners work with only one or two intermediaries or mukadams for recruiting the workforce. Loaders and 312 gijs kessler and jan lucassen unloaders organise themselves in gangs of several young married couples, who as a rule seem to be related to each other. In Northern India most of them come from Rajasthan. Contrary to the moulders’ and loaders’ gangs, who are paid group-wise according to piece rates, the firemen or jelai are paid time wages, some- times supplemented with an extra small sum per thousand first-class out- put in excess of the normal 70 percent of the total kiln content.117 Like the others they are also seasonal migrants, but most of the firemen in Northern India seem to hail from a relatively small area in Uttar Pradesh, in particu- lar from KundaTahsil in the district of Pratabgarh in eastern UP. Per kiln a gang of four firemen, all adult males, is working under a head or mistri who may serve several factories at once. Two firemen are available for maintain- ing the fire on each of the watches, which commence at 6 and 12 o’clock, day and night. In South India firemen also receive time wages. The biggest difference is that here not just men, but families consisting of husband and wife are hired. The women break the coal while the men take care of the actual firing.118 At the brickfields the firemen are a special case: Kunda firemen always stick together and never participate in collective action of moulders or loaders. In fact they are not the respected top of the brick field pyramid, like most firemen in Europe, particularly among the Lipper. Although they can boast of a long tradition as specialists, possibly even going back to the last quarter of the nineteenth century when William Bull first experiment- ed with his newly invented trench kiln, their earnings are moderate and they are considered by the other labourers as a sort of labour underclass.119 How to explain this remarkable fact? We will first look at their middlemen and then at the firemen themselves. The remuneration of the firemen’s mistris sometimes consists of time wages,120 sometimes not. In one factory it is explained very clearly: the

117 Nagendra Nath Mitra, Materials of Construction as Used in India: their Nature, Pro- duction and Use (Calcutta ; Simla : Thacker, Spink, 1924), 164, 170. 118 International Institute of Social History (IISH), Coll. Brickworks India, Interviews Factory 13; Lucassen, ‘Brickmakers in Western Europe .. and India’, 555-556. 119 Bull could offer to send specialists to brick manufacturers who wanted to experi- ment with his new kiln. It is not sure where these people came from, but it must have been from eastern UP as Bull himself lived there then. Attempts to trace the history of the Kunda firemen to this episode have failed as the memory of the interviewees only allows us to suppose that this tradition already existed in the 1930s. The owner of factory 4 stated that his father started his Bull’s trench kiln in 1927 already with specialized firemen from UP (interview 06-02-1997). 120 IISH, Coll. Brickworks India, Interviews Factory 16: same wages for supervising mistri (not himself part of the gang) as firemen. labour relations, efficiency and the great divergence 313 jelai mistri receives 2 percent of the wages of ‘his’ men—in this case ten men—plus 2,000 rupees per kiln (he serves three of them). This would result in 40 rupees per month per person or 400 rupees per kiln, which makes 1,200 for three. On top of this come 2,000 rupees per kiln, or 6,000 for three. As we suppose this is not per month, but per season (so 1,000 per month) the total monthly income of a jelai mistri would amount to 2,200 rupees—not too much in comparison to the 2,000 rupees of a normal je- lai.121 There may be an in-built element of exploitation in their labour rela- tions in cases where the mistri receives piece wages (5 to 7 rupees supervision money per 1,000 bricks),122 whereas the firemen themselves receive only time wages—a situation in which it is in the interest of the mistri rather than of the firemen to work as quickly as possible. On the other hand, a proper firing of the bricks takes time and there is no point in rushing it if one wishes to avoid underburnt bricks. If then the Kunda firemen are not simply the victims of exploitation by their mistris, we have to look for another explanation for their subordinate status. A more detailed investigation in some of the villages where the firemen come from has revealed more about their class and caste back- ground.123 It turns out that they invariably belong to the poorest inhabit- ants and are classified as ‘scheduled casts (SC)’ or as ‘Other backward Casts (OBC)’. Some own, but most merely lease a few acres of land, taken care of by their wives and children, who occupy themselves with cottage industry, in particular basket and rope making. It is tempting to see a reflection of their subdued position back home in their weak position at the brick kilns. A central figure in the system of labour relations and the organisation of production in the Indian brick industry is the intermediary, often called ‘jobber’, mukadam, jamadar, thekedar, or mistri. This person is informed by a factory owner as to how many moulding families are required, and from him he also receives the advances in the wet season for distribution

121 IISH, Coll. Brickworks India, Interviews Factory 20. 122 IISH, Coll. Brickworks India, Interviews Factory 4. 123 The following information has been gathered during a field trip to Kunda from 7 to 12 February 2001, by Jan Lucassen together with Dr. Dhirendra Dangwal, whom we would like to thank deeply here. By train they travelled from New Delhi via Lucknow to Kunda. From Kunda they visited by two-wheeled horse drawn tonga the following villages: Mauli (in 1981 with 3,071 inhabitants), Barichitija (in 2001 with 100 families), Benti Uparhar (in 2001 with at least 3,400 inhabitants, in which they visited the purwa’s Benti-Khas, Man Singh Ka Purwa with 10-15 families, and Pure Aman Singh Ka Purwa), Maharajpur-Khaz (in 1981 with 108 families, in 2001 with 200), and Faredupur near Kunda (in 2001 with 100 families). In total 16 informants were interviewed (all reports in IISH, Coll. Brick Works India). See also Lucassen, ‘Brickmakers in Western Europe .. and India’, 555-56. 314 gijs kessler and jan lucassen among the moulders. At the same time the mukadam has to make sure that the workers will be ready at the start of the dry season to depart for the same brickfield owner from whom they had received their advances. As a rule, one factory has some four mukadams; one, two or even more for the moulders, one or two for the loaders and unloaders, and one for the fire- men. The mukadam holds a very ambiguous and difficult position: on the one hand he needs to enjoy a certain degree of trust from both parties, while on the other hand he has to make his money out of the deal, must be able to ‘flex his muscles’ in the face of any workers unwilling to repay advances or to obey him generally, and he must keep the factory owner at a distance when the latter is dissatisfied with the workers. Evidence—mainly based on qualitative materials collected so far among the moulders or loaders— reveals the whole gamut of characters, from primus inter pares to sadistic exploiter, and it seems to be extremely hard to find a meaningful average type of mukadam. During the wet season the intermediaries travel by train or bus to the factory owners (until the moment of the interviews travelling by car was not reported) in order to get a contract. They really seem to compete with each other. Villages near Kutch Behar in Bengal annually see four to six jamadars arriving to recruit people, which gives them ‘plenty of choice’ to choose which intermediary to work with.124 The factory owners have to trust the mukadams, but sometimes they check them and visit them at home to test their credit-worthiness and background. A good mukadam has the experience to select good workers and has the guts to tell them the truth if necessary. This at least, is the ideal intermediary in the eyes of the factory owners and their overseers (munshis).125 The mukadam has lived or still lives in the region where the brickmakers come from. The main problem for the employer is the enforcement of an indirect contract, ultimately made with a number of mukadams represent- ing hundreds of workers in different regions hundreds of kilometres away, people whom the employer does not know personally. He tries to overcome that problem by providing advances to the workers through the mukadam. It depends on the precise form this contract takes to what extent we can speak of a mukadam who is subcontracting the work from the employer, or of a group which is subcontracting from the employer with the help of a mukadam. For that purpose the way the work and the remuneration of the mukadam is organised informs us best.

124 IISH, Coll. Brickworks India, Interviews Factory 20. 125 IISH, Coll. Brickworks India, Interviews Factory 16. labour relations, efficiency and the great divergence 315

Illustration 8.6. The moulder on this photograph is the husband of the woman in the orange dress on Illustration 8.5. Both, possibly together with their daughter (left) form one of the 75 moulders’ groups who receive piece wages per unit. A moulders’ couple, migrant labourers from other parts of Bengal or Orissa, may earn, depending on its pro- ductivity, between 70 and 80 rupees per day. Living costs of adults are estimated at some 20 rupees per day. Besides clay mixers and moulders, the factory also employs 150 Biharis from Ranchi (400 kms) and also Patna (600 kms) and Chaibasa to carry dried bricks to the ovens. The carriers are exclusively women and children (both carry the bricks on their head) and boys using yokes, and they are paid individual piece rates. Source: IISH, collection Jan Lucassen on Indian brickmaking, factory 10, photo 13-OA.

The largest numbers are involved in the recruitment of moulders. One mukadam recruits dozens or even hundreds of people, as a rule married men representing their families. The advances a moulder receives from the employer amount to thousands of rupees per family, which means that such a unit has to work for several months in order to repay this debt. This can be deduced from the piece wages received by the workers, invariably expressed in so many rupees per 1000 bricks moulded and dried (respec- tively loaded, or unloaded). When employers are competing on the labour market it is tempting for both mukadams and workers to cash their ad- vances, but subsequently to choose another employer somewhere else.126

126 According to one factory owner (Interviews Factory 2) this happened to his neigh- bour (at Factory 3) in 1996 (Interview 03-02-1997). 316 gijs kessler and jan lucassen Legally the position of the employers seems to be weak as the handing out of advances is strictly speaking illegal. That is why fingerprints and similar signs of labourers on a contract are useless. Nevertheless, the employers have developed various remedies to prevent abuse of advances. One would be the enforcement of existing rules in their own organisations, which of course forbid such illicit competition between members. Cooperation between owners of brick factories started rather late—in the 1940s, possibly because the government and large railway com- panies played such a dominant role in India’s public works. Their impor- tance increased in reaction to moulders’ collective actions since the 1970s. In 1957 the first federation was set up and in 1975 the All India Brick and Tile Manufacturers Federation. In Pakistan we find a similar develop- ment.127 The only alternatives seem to be to recur to private terror, or to bear the loss in view of the huge profits that can be made in this business. It looks as if the last two options are most popular. At a certain point advances can become substantial loans which cannot be paid back any more by working part of the brickmakers season, or even a whole season. As Breman has shown for South-Gujarat, this is one of the origins of bonded labour.128 Given the importance of advances in hiring practices in the industry, this raises the question to what extent Indian brickmakers should be characterised as bonded labourers. Although we do not have direct and detailed evidence on indebtedness and bondage among the workforce, random interviews undertaken by Lucassen at a number of factories reveal rather high turnover rates, which would certainly not be a sign of bondage, because it strongly suggests workers are largely free to change employment. Loaders or lariwallas invariably worked only one, two or four years in one and the same gang.129At the next factory individual moulders say that they change on average every three years their jamadar (in this case a big man who caters for 200 to 250 gangs, consisting of husband, wife and chil- dren from age six, in total 1,000 individuals).130At another most labourers stay only one, two or three years, although their munshis prefer to work with men they know. Loaders there are more loyal. They stay with the same

127 Lucassen, ‘Brickmakers in Western Europe .. and India’, 560-61. 128 Jan Breman, Wage hunters and gatherers: Search for work in the urban and rural economy of South Gujarat (Delhi: Oxford University Press, 1994); idem, Footloose labour: Working in India’s informal economy (Cambridge: Cambridge University Press, 1996). 129 IISH, Coll. Brickworks India, Interviews Factory 14. 130 IISH, Coll. Brickworks India, Interviews Factory 15. labour relations, efficiency and the great divergence 317 boss for five to ten years. Labourers, and in particular moulders, are said to be able to afford a quick turnover because they have paid back their ad- vance money at half-season.131 A factory near Chennai shows a high turn- over rate for moulders because the workers decide to go to the brick works only when the monsoon does not bring enough rain. In a good harvest year they stay at home and the mistris have a lot of trouble finding enough workers.132 In effect, the relation between employers, middlemen and workers appears to be a generally loose one. Moulders’ families often not only hail from the same villages, but can also be related. At one factory the owner explained that his moulders consisted in fact of subgroups of four to seven related families.133 Their inclination to stick with the same factory owner or thekedar, especially if he serves hundreds of families at several factories at once, is minimal and this particular owner even could state: ‘I never see the same faces’. This distance between owner and thekedar on the one hand and workers on the other is also reflected in a comment by the same own- er, stating that it is not necessary that the thekedar is always present for supervision, as long as he is there in case of trouble. To conclude, although cases of bonded labour among brickmakers have been documented, it seems misleading to attach this characteristic to the group as a whole. One could even tentatively argue that in comparison to Western and Eastern Europe Indian brickmakers are more footloose and less prone to stick to one and the same employer. This conclusion has important implications for the labour-productivity of the Indian moulders: their primary work incentive is not so much coer- cion as rather compensation in the form of collective piece-wages. The position of the intermediaries may be more ambiguous than in the Euro- pean cases, but they certainly cannot be seen as simple instruments of coercion. Indeed, instances of collective action and worker organisation are numerous, confirming once more that labour in this industry should be characterised as free, rather than bonded. Surely, if we realise that in Indian parlance brickmaking still today belongs to the ‘unorganised sector’ (a fate it shares with most other sectors)134 one does not immediately ex- pect forms of collective action and workers’ organisation. The reality, how-

131 IISH, Coll. Brickworks India, Interviews Factory 16. 132 IISH, Coll. Brickworks India, Interviews Factory 13. 133 IISH, Coll. Brickworks India, Interviews Factory 19. 134 Sabyasachi Bhattacharya and Jan Lucassen (eds), Workers in the informal sector: Studies in labour history, 1800-2000 (Delhi: Macmillan India, 2005). 318 gijs kessler and jan lucassen ever, is and has been much more diverse and colourful. Even in the artisanal sphere we find collective actions.135 At the large works collective action had a long history. Most famous is the strike of the brickmakers on the Ganges Canal in 1848-49.136 Thousands of moulders protested in several ways against the threat of reductions of their piece rates—and with success. It started with the refusal to start work at the beginning of the dry season, followed by strikes, machine breaking and finally arson. British engineers appear to have been used to such forms of collective action by navvies and brickmakers, in any case until the begin- ning of the twentieth century.137 We do not know much about the con- tinuation of this tradition in the next fifty years but from the 1970s collective actions of brickmakers are documented for Pakistan, the Indian cities of Ahmedabad and New Delhi, as well as West Bengal.138 Formal and informal trade unions, strikes for piece rate hikes or the prevention of their lowering are numerous and continuing to this very day. Success depends to a large extent on the question whether the mu­kadam is more of a primus inter pares, as seems to be the case in Northern India, or himself an exploiter of the people he recruits. The latter is certainly the case in Pakistan, where intermediaries play as bad a role as their counter- parts (called ‘capo’) in the Friuli (Italian) gangs around the turn of the century in Germany.139An example of the opposite is provided by the role of mistris among firemen or jelai groups, in which one acts as the head of the group.140 One informant, some 45-50 year old and an experienced jelai for nearly thirty years, of which sixteen as a mistri, mentions the existence of some form of organisation. In his district—the Dhosa-border of Delhi— forty to fifty mistris meet occasionally to elect their pradhan (or president; the same word is used for village chiefs or for brick kiln owners). In the Dhosa-border area four pradhans are active. All mistris pay 50 rupees per

135 Cf. Tirthankar Roy, ‘The guild in modern South Asia’, in: Lucassen, De Moor, Van Zanden (eds), The Return of the Guilds, 95-120 on similar cases and on the question to what degree India has known guilds. 136 Extensively in Lucassen, ‘The brickmakers’ strikes’; see also Lucassen, ‘Brickmakers in Western Europe .. and India’, 547-52. 137 Lucassen, ‘Brickmakers in Western Europe .. and India’, 560; Lucassen,‘The brick- makers’ strikes’ , 50-51. 138 Lucassen, ‘Brickmakers in Western Europe .. and India’, 560-67. 139 Ibid. 140 One mistri (Interviews Factory 15) emphatically tells that mistris always try to put jelai from different villages together on one kiln as jelai from one village at one factory or batta together are supposed to become lazy very quickly. labour relations, efficiency and the great divergence 319 month to him during the brick season.141 This pradhan has no fixed tenure and his effectiveness to fix good wage rates142 and to defend the general interests of the mistris and their jelai depends on how much he is sup- ported by the mistris.

Conclusion

In this chapter we have been looking for evidence of different pre-indus- trial trajectories of human capital formation in the organisation of the labour process in manual brickmaking across the Eurasian landmass, with a special eye to mechanisms for the reduction of error-rates and to the co- operation in production along the lines of a set of ideas recently put for- ward by Gregory Clark. Linking the differential development of labour productivity in the West and East after 1800 to differences in the ability to effectively work together in industrial production environments requiring discipline and co-ordination of efforts, Clark argued that the origins of these different labour characteristics lay deep within the pre-industrial period. Having compared pre-modern labour force characteristics and the co- operation in production in the brick industry in Western Europe, Russia and Northern India, we find, however, little evidence of fundamentally different trajectories of human capital formation along such lines. Al- though the organisation of production has differed both across time and space, the mechanisms of error-rate reduction adopted to ensure efficien- cy within these different systems belong to one and the same repertoire. The quintessential mechanism of organising the collaboration within the labour process in all three geographical areas was the combination of col- lective subcontracting on piece-rates and differential redistribution of earnings within the group according to levels of proficiency and skill. Train- ing and the transfer of skills were organised within the collective, in some cases on the basis of family training. Vocational training of this kind, there- fore, not only taught workers how to apply their skills, but also how to apply them effectively within a collaborative environment. In such a system workers are collectively responsible for minimising error-rates, since this is what determines their earnings as a group, and therefore, eventually, as

141 This account (Interviews Factory 8) is confirmed by other informants who say, however, that this fee amounts to only 10 to 20 rupees. 142 It sometimes happens that factory-owners unofficially give a small tip (100-200 rupees) to the jelai as a group. 320 gijs kessler and jan lucassen

Figure 8.2. Collective subcontracting units in brickmaking in Western Europe, Russia and India, 1670-2000 (*Moscow 1880 = hand-machine moulding, the most widely used mould- ing technique at the time). individuals. Employers are relieved of the necessity to closely supervise the workforce and can limit themselves to the imposition of strict quality con- trol on the final output. Production was organised in this way among the Lipper and Walloon brickmakers in North-Western Continental Europe, around London, in Russia up to the late nineteenth century, and in Northern India up to the present day. The variation among these areas of production was mainly in the numbers of different tasks carried out by one and the same collective (cf. Figure 8.2). In Western Europe brickmakers’ gangs tended to contract the whole production process from digging the clay to firing the bricks, except in medieval Italy and among the Walloon brickmakers, where firemen were a separate group from the moulders. In late nineteenth-century Moscow the production process was mostly broken up into three clusters of tasks carried out by separately remunerated gangs of workers, responsible for, respectively, the digging of the clay, the tempering, mixing and moulding, and the firing of the bricks. In early twentieth-century St. Petersburg moulders had come to concentrate solely on their core task, whereas the labour relations, efficiency and the great divergence 321 rest of the production process was divided among three or four other groups of workers. In India, finally, the production process is broken up into three parts, divided between moulders, loaders and firemen. Two alternatives existed for this system of collective subcontracting. The first one was individual piece-rate remuneration, prevalent along the river IJssel in the late seventeenth century and, apparently, on the rise in late nineteenth and early twentieth-century Russia. While it equally shields the employer from the negative effects of bad work by individual workers, it does not provide in itself a mechanism for an efficient error-rate reduction between the stages in the production process. Unfortunately, we do not know how this problem was solved along the IJssel, but in Russia it gave rise to the practice of linking the remuneration of the diggers to the output of the moulders, thus ensuring the latter would be sufficiently and timely supplied with clay. Essentially, this was a modification of the system of collective subcontracting: although in name diggers were remunerated individually, in practice they were remunerated according to the collective effort performed in co-operation with the moulders, only the redistribution of earnings between the two categories was not left up to the collective itself. Moulders in this system were also apparently remunerated collec- tively, but per couple or three-man table rather than by gang. A completely different, second alternative we find in the Dutch Rijnland, where workers were remunerated individually, on time wages, and closely supervised and disciplined. Although workers were not hired directly, but through an intermediary, the system was nonetheless strongly reminiscent of modern production relations, and it is most likely no coincidence that this particular production area was the centre of early-modern capitalist brickmaking in Europe. How to interpret the findings of this chapter? Several observations can be made. The fact that no substantial differences between regions were found in the organisation of the labour process in pre-industrial brickmak- ing does not appear to provide immediate support for Clark’s hypothesis. But of course brickmaking could well be exceptional. Research on other industries and sectors of production is essential to further advance the argument. Ideally, such research should also aim to take in the shift to modern production relations in these industries. Equal pre-industrial la- bour force characteristics might have produced different results in combi- nation with the introduction of modern production relations in some cases, and similar results in others. Mapping this variety might enable us to isolate the importance of longer-term and short-term factors in explain- ing different outcomes in different parts of the world. 322 gijs kessler and jan lucassen This draws our attention to the actual circumstances surrounding the introduction and implementation of new forms of production organisa- tion, error-rate reduction and workforce supervision in areas outside the North-West European cradle of the Industrial Revolution. If it were to come out that industrial production relations were grafted onto one and the same stem, the fact that different trees grew from them would have to be ex- plained by the circumstances under which the grafting and further growth took place. Surely, these circumstances differed. Whereas in the West in- dustrial production relations developed from within, in most other parts of the world they were introduced from outside, often within a colonial or neo-colonial context. Exactly because they involve the switch from indirect to direct supervision this context might have severely undermined both the legitimacy of employers, managers and foremen, as well as their abil- ity to overcome initial resistance to these new production relations. explaining the global distribution of book production 323

CHAPTER NINE

EXPLAINING THE GLOBAL DISTRIBUTION OF BOOK PRODUCTION BEFORE 1800

Jan Luiten van Zanden

Introduction

What does the global map of book production look like as it emerged in the centuries before 1800? In different parts of the world, different tech- niques were used to copy the written word. The first and most ancient technique was copying by hand, practiced virtually everywhere where the written word was used on a certain scale. This was a relatively simple pro- cess—one only needed paper (or another material to write on, such as parchment or papyrus), ink and a goose feather, and, obviously, the ability to read and write. Compared with the other techniques that will be dis- cussed, it was relatively labour-intensive, and it took quite a lot of time to copy a manuscript. In Medieval monasteries we find that the average pro- ductivity may have been close to one manuscript per scribe per year, but this may have been so low because of the high standards maintained in the process, and the many illuminations added.1 The most advanced technique discussed here was moveable type printing, as developed by Gutenberg and his associates in the 1440s and 1450s, and before him, by Chinese and Korean printers who experimented with it from the tenth century onwards. This technique dominated the reproduction of the written word in large parts of Europe (with the exception of the south-east, which was part of the Ottoman Empire) and its offshoots (the Americas). It was, as we will see, a technology using a lot of capital, including human capital, with very high levels of labour productivity. In Japan, Korea and China an ‘intermedi- ate’ technology, woodblock printing, was dominant, but occasional ex- periments with moveable type printing occurred and hand-copying was probably more important than in large parts of Europe. Outside these re- gions (Europe, Eastern Asia, and the Americas), apart from the occasional

1 Eltjo Buringh, Medieval Manuscript Production in the Latin West (Leiden: Brill, 2010). 324 jan luiten van zanden western style printer which was part of the colonial world in Bombay, Goa or Batavia, the written word was solely being hand-copied. In other regions with relatively high levels of urbanization, such as the Arab World and India, woodblock printing nor moveable type printing really took off before the nineteenth century, although in the Ottoman Empire the ban on print- ing (by Muslims) was abolished in 1727. The question therefore is: why did different parts of the world, even before the ‘Great Divergence’ of the nineteenth century, chose such differ- ent strategies for reproducing the written word? The traditional view of this is that movable type printing was a ‘superior’ technology, which was only spreading slowly to other parts of the globe because of conservatism or inefficient institutions. Eventually, in the nineteenth century, it would succeed in conquering the rest of the world, demonstrating its superiority. In this approach, woodblock printing is a sign of ‘intermediate modernity’, whereas hand copying represents backwardness. Such a picture is highly simplified, however, and does not take into account a number of other factors—such as relative prices and the size of the market for books—that, this chapter argues, help to explain global patterns. A closer study of dif- fusion patterns could help clarify under which circumstances which tech- nologies were optimal—and what the long-term consequences of the different choices of techniques were. I begin by studying the two most famous cases of innovation: the spread of woodblock printing in China, and the rise of moveable type printing in Western Europe, and briefly dis- cuss the failure of the diffusion of these techniques in the Ottoman Empire.

Three Diffusion Processes

Woodblock printing was already an established technology in textiles manufacture—for the printing on cotton cloth in particular—before the development of its application to the reproduction of texts in China and Korea around 650 AD. A first breakthrough of the new technology occurred during the Sung (960-1279), when it began to be used commercially on a relatively large scale. It is unclear to what extent it really already replaced hand-copying during this period. A rapidly growing literature is suggesting that it was only during the (late) Ming (1368-1644) that printed books re- ally replaced manuscripts, and that in the intervening centuries the book industry was more or less stagnant, producing only a limited numbers of books. McDermott has reviewed the evidence, and concludes indeed that explaining the global distribution of book production 325 the major changes only occurred during the sixteenth century, a conclusion shared by most authors in the field now.2 McDermott also gives estimates of the numbers of imprints still in existence (in the National Library of China and the National Central library of Taipee), a series which is quite likely only the ‘tip of the iceberg’, but nonetheless allows us to study the long-term dynamics of the industry between the Sung and the Ming (Figure 9.1). If we assume that this series is representative of book production in China in this period, it is clear that between the late Sung and the early Ming it was more or less stagnant: the available estimates show a produc- tion level of about 10-25 imprints per decade, which is rather low. McDer- mott concludes on the basis of this evidence: ‘What is remarkable about the case of China is the extremely long time—eight centuries—that passed before imprints clearly dominated manuscripts in the collection and book markets of the most highly developed commercial and cultural region of the empire’.3 The acceleration of printing industry only occurred during the late fifteenth and sixteenth centuries, when output rose by a factor 10-15, but even during (and after) the boom of the printing industry in the late Ming, hand-copied manuscripts remained quite important. Chia’s es- timates suggest higher production levels of output in this period, of per- haps as much as 300 titles per decade during the 1522-1644 period, when woodblock printing really became the dominant new technology.4 To some extent the delayed and slow spread of the innovation—McDermott sug- gests a ‘delay’ of seven centuries, between the first breakthrough in the tenth century and ‘final’ diffusion in the sixteenth century—can be related to political changes. Traditionally, the Yuan/Mongols have been blamed for the collapse of the industry, but the early Ming was even worse—the ideo- logical move towards orthodox Confucianism after 1368 seems to have been bad for market demand as well. As will be discussed below, economic fac- tors perhaps offer a better explanation for the coexistence of these two technologies. The cost advantages of the new technology were simply not large enough to create a sudden and rapid diffusion. Instead, also after the

2 J. McDermott, ‘The Ascendancy of the Imprint in China’, in: C.J. Brokaw and Kai-wing Chow (eds), Printing and Book Culture in Late Imperial China (Berkeley, CA: Univ. of Cali- fornia Press, 2005), 55-107; cf. Cynthia J. Brokaw, ‘On the History of the Book in China’, in: Brokaw and Chow (eds), Printing and Book Culture in Late Imperial China, 3-55, and Lucille Chia, ‘Of Three Mountains Street: The Commercial Publishers of Ming Nanjing’, in: Brokaw and Chow (eds), Printing and Book Culture in Late Imperial China, 107-52. 3 McDermott, ‘Ascendancy’, 78. 4 Chia, ‘Three Mountains Street’, 128. 326 jan luiten van zanden sixteenth century, both technologies continued to occupy certain niches in the market for reproduced texts. The diffusion curve of moveable type printing in Europe was quite dif- ferent, as Figure 9.1 also shows. After its invention in the 1450s, the technol- ogy spread like wildfire over all of Western Europe. Within two generations, before 1500, there were hundreds of printing presses in almost all Euro- pean countries, including Poland and the Ottoman Empire where Sep- hardic Jews established the first press in 1493—the ban on printing applied to Muslims only. As a result, manuscript production already began to de- cline sharply during the 1470s.5 The data on European book production are better than those for China, and therefore Figure 9.1 may overestimate the gap between them, but the European estimates only cover books still available in libraries, like the China data. The point of the comparison is the different steepness of the two diffusion curves (or actually, production curves): the printing revolution in Europe was sudden and dramatic, sug- gesting huge economic advantages of the new technology. There was, therefore, a genuine revolution of the printing press; before the 1450s there had been many experiments with woodblocks, but mainly for printing (small) pictures. It is no coincidence that many of the experi- ments that preceded the invention of moveable type also took place in Germany: ‘some of these (wood)cuts were intended to be pasted onto walls, or onto pieces of furniture or other household objects. Others found their way into manuscripts, either as illustrations having some immediate rel- evance to the text or as additional matter’.6 But using woodblocks or other forms of xylography for the reproduction of books, was something of the post 1450 period, once Gutenberg’s invention had opened up all kinds of new avenues for combining techniques. It remained a technology used only rarely (except for illustrations in books that otherwise were printed using moveable type). The third diffusion process that we will try to explain concerns the Ot- toman Empire, where a first printing press was put to work in 1493 by a Jewish immigrant, who had just been expelled from Spain. The technology remained rather unsuccessful; a few other printing presses were intro- duced, by Jews and other religious minorities such as Armenians (1567) and Greeks (1627), but they were constrained to printing books in Hebrew

5 Eltjo Buringh, Jan Luiten van Zanden, ‘Charting the “Rise of the West”: Manuscripts and Printed Books in Europe, A Long-Term Perspective from the Sixth through Eighteenth Centuries’, Journal of Economic History 69 (2009), 409-45 for the underlying data. 6 L. Febvre, and H.J. Martin, The coming of the Book (London: Verso, 1976), 61. explaining the global distribution of book production 327

Figure 9.1. Diffusion curve of printing in China and Europe (total imprints per decade). Sources: McDermott, ‘Ascendancy’, 58; Buringh and Van Zanden, ‘Charting the ‘Rise of the West”’. or other minority languages, because printing in Arabic was forbidden.7 Other attempts to develop this market came from Italian printers, most famously the Medici Oriental Press, which started to publish books in Arabic script in 1590; but these too were largely unsuccessful, as previous attempts dating from 1514 and 1537-38 had been.8 Raimondi, who was fi- nanced by a member of the Medici family, hoped to sell his books in the East, rather than the West. Ironically, while the works failed to sell in the Ottoman Empire, they did significantly stimulate the study of the Middle East in Europe. In 1727 the ban on printing by Muslims was lifted, but Ibra- him Muteferrika, who convinced Sultan Ahmed III to end the prohibition, was only allowed to print non-religious texts. The output of the industry remained very limited, and the market consisted mainly of a small and relatively isolated community of students and scholars in Istanbul.9 Mute- ferrika published 17 titles, his successors another 7, but there clearly was

7 G. Nassi (ed.), Jewish Journalism and Printing Houses in the Ottoman Empire and Mod- ern Turkey (Istanbul: The Isis Press, 2001). 8 Robert Jones, ‘The Medici Oriental Press (Rome 1584-1614) and the impact of its Arabic Publications on northern Europe’, in: G.A. Russell (ed.) The ‘Arabick’ interest of the natural philosophers in seventeenth century England (Leiden: Brill, 1994), 88-108. 9 Maurits van den Boogert, ‘The Sultan’s answer to the Medici press?’, in: A. Hamilton et. al. (eds) The Republic of Letters and the Levant (Leiden: Brill, 2005), 265-93. 328 jan luiten van zanden no ‘take off’ like in fifteenth-century Western Europe or Ming China. Was it the religious ban that limited the demand for books, or was it the absence of a middle class able to read and write and interested in buying books, that explains the lack of success of the new technology? What is perhaps even more striking about the experience of the Arab world, is that woodblock printing also did not take off. For reasons dis- cussed below, there were certain problems with applying the technology of moveable type printing to Arabic, but these did not impact on the alter- native, the (wood)block. Moreover, woodblock printing was a well-known technique in the manufacture of high-quality textiles; the first example already dates from tenth-century Egypt.10 For the reproduction of the written word, the technique never really took off in the Arabic world. This is striking, in particular because there appears to have been a relatively large market for books during the eighth-twelfth centuries, when the pro- duction of manuscripts was booming thanks to, amongst others, the in- creased interest in the sciences. There are many stories of very large libraries being put together by state dignitaries or other wealthy men, which point at the availability of large numbers of books in the period.11 After the twelfth/thirteenth centuries this ‘explosion of books’ was over, however; although later Islamic societies (the Ottomans, the Mughals) also showed some interest in books, they probably did not have the same ap- petite for manuscripts as the Abbasids and the Fatimids. Finally, a brief note on Japan. In spite of its close commercial, religious and political contacts with China (and Korea, another center of woodblock printing), printing did not really take off there before the end of the six- teenth century. In that period a number of experiments with moveable type printing were carried out, using imported printing presses from Korea and the west (acquired via Jesuits), using wooden type-pieces rather than metal. The first Tokugawa emperor was also involved in these experiments. It ‘was soon decided that the running script style of Japanese writings would be better reproduced using woodblocks’, however, and after about 1600 woodblock printing became the dominant technology.12 Japan in the seventeenth and in particular the late eighteenth century developed— much like China—a mass market for (cheap) books, but the preferred

10 Jonathan Bloom, Paper Before Print. The History and Impact of Paper in the Islamic World (New Haven: Yale University Press, 2001), 218. 11 The best survey in Ibid., 110 ff. 12 Peter F. Kornicki, The book in Japan: A cultural history from the beginnings to the nineteenth century (Leiden: Brill, 1998). explaining the global distribution of book production 329 technology remained—as in China—the woodblock, although alternative techniques were known and used occasionally. The available estimates of the number of books in eighteenth century Japan suggest that the market was, in per capita terms, quite a bit larger than in China in the same peri- od.13 Why these remarkable differences? Why did not the whole world use the same technologies for copying text? I will discuss four factors behind the global dispersion of these technologies: the size of the market, relative prices, and the institutional environment (for example, government poli- cies such as the ban on printing); but I would like to begin with the most obvious problem of the various scriptures used in these regions, and the quality standards that were related to them.

Quality

A major problem when introducing moveable type printing was the change in the quality of the reproduced text. We already mentioned that in Japan, the loss of the ‘running style’ was seen as a disadvantage of moveable type printing. Arabic writing is, according to Bloom, ‘unique in always being cursive; letters within a word have to be joined wherever possible … A disconnected way to write Arabic, comparable to the individual charac- ters in Hebrew, Greek, and Latin, never evolved. The Arabic script therefore presents typographical problems quite unlike those presented by other alphabets, or even by Chinese, with its thousands of discrete characters. Arabic type requires an extremely high level of skill in punch cutting to imitate calligraphic norms, and the compositor must also know which form of a particular letter to use when. A complete font of Arabic type, including the vowel marks required for Koranic and other vocalized texts, can easily run to more than six hundred sorts, or individual characters, plus huge quantities of leads and quadrats to be placed between vowel marks and lines’.14 The first experiments with movable type printing in Arabic, by European printers interested in the Arabic market, were therefore often not successful because of these quality problems—the prints were consid- ered to be ‘clumsy, inelegant’ or having ‘an odd typeface … quite unaccept- able by Muslim calligraphic norms’.15

13 Kornicki, The book in Japan, 192-206; the comparison is presented in Buringh and Van Zanden, ‘Charting the “Rise of the West”’, 437-38. 14 Bloom, Paper before print , 217-18. 15 Ibid., 219-220. 330 jan luiten van zanden Of course, these problems could be overcome, as more elegant, later examples, demonstrate. But printers needed time to experiment and to invest in better typeface. There were however two main differences with the ‘take off’ of printing in Western Europe: the first full-scale experiment, Gutenberg’s Bible, was, thanks to Gutenberg being a perfectionist, in terms of quality standards really outstanding, because he had solved many prob- lems at the same time (or in fact, during a period of experimenting which lasted almost twenty years) and was therefore able to attain a level of per- fection that was really quite different from the first experiments with, for example, Arabic print. One of his main concerns was that the printed ver- sion of the Bible that he concentrated most of his efforts on, would look very similar to the manuscript versions of the Bible that dominated the market. Although it was, of course, quite different from a ‘regular’ manu- script, the similarities were and are quite striking. In short, he had a very good sense for marketing, as a result of which his product was almost im- mediately accepted. Secondly, and perhaps even more importantly, the market had become used to experimentation and to low-budget books. Hand-copying had become a real industry, aimed at cutting costs (for poor students, for example) and at creating a mass market for its products (via the pecia-system).16 Various kinds of woodcuts were also circulating, pos- sibly also preparing the market for the even greater revolution of moveable type printing. Had Gutenberg tried to introduce the same technology in, for example, the tenth century, before the commercialization of and mass production by hand-copying, he might have met with similar resistance as printers in the Arabic world met. Perhaps part of the eagerness of Europe to accept the new technology can be explained in this way, as well as the greater problems Arabic printers had in placing their products on the mar- ket.

Font Size

A related ‘cultural’ problem concerns the size of the font required for print- ing in the various languages. The problem with the Arabic script could be solved—as was shown subsequently—with the development of a much larger font, in which characters could be linked in different ways. Chinese (and Japanese) would-be printers wishing to make use of movable type

16 R.H. Rouse and M.A. Rouse, Manuscripts and Their Makers: Commercial Book Produc- ers in Medieval Paris 1200-1500 (London: Harvey Miller, 2000). explaining the global distribution of book production 331 were at a huge disadvantage here. One needed many more—perhaps as much as a few thousand—characters to print a standard text in Chinese than in a language using Latin characters. Imperial projects carried out in both countries, sometimes required 25,000 moveable types or more— which also led to huge problems in storing them and making them acces- sible for the compositor.17 Handling such a huge font—selecting the right characters quickly—must have been difficult and time-consuming. One argument is that moveable type printing was probably more capi- tal intensive, because one needs a much larger capital invested in font. Therefore, for Chinese printers it was relatively inefficient to use moveable type, as the costs involved were much larger than those for European print- ers using Latin script. A Chinese font consisted of many thousands differ- ent characters, much more than the hundred or so which are required for printing in Latin characters (Arabic would be in between, with, as we saw, about 700 characters). In practice, however, these differences were prob- ably not that large. A typical font scheme used by a European printer con- sisted of 100,000 to perhaps as much as 200,000 types, as we learn for example from the famous Antwerpen printer Plantin. Two examples of font schemes he bought in the 1560s (when his accounting system is par- ticularly detailed) are a roman Garamond font of 101,857 types, for which he paid 66 guilders, including the costs of 261 pounds of lead, and another even bigger font scheme of in total 200,000 types (8,104 a’s, 3,023 b’s etc.), worth 85 guilders. These were relatively large sums (the daily wage of an average compositor or carpenter was only 0,30-0,40 guilders), but they were dwarfed by the value of the collection of punches and matrices of the company, worth 1,200 guilders in the 1560s.18 By comparison, a printing press in the same years was bought for 50 to 60 guilders; Plantin operated seven presses at the time, their total investment was therefore about 400 guilders, or one third of the capital invested in the matrices and punches.19 To return to the point of the different font sizes: the disadvantage of Chinese printers in this area was relatively small, as European printers too needed at least many tens of thousands types for their printing operations. The point is that it is not the capital invested in the characters as such that mattered much; the matrices were the most expensive (as the example of

17 Christine Moll-Murata, State and Crafts in the Qing Dynasty, 1644-1911. Habilitation thesis, submitted to Tübingen University, Faculty for Cultural Sciences 2008, 241. 18 Leon Voet, The Golden Compasses. A history and evaluation of the printing and pub- lishing activities of the Officina Platiniana, vol. II (Amsterdam: Van Gendt & Co, 1972), 57-81. 19 Ibid., 133 332 jan luiten van zanden Plantin shows). These matrices, from which the actual font was being cast, were arguably the most important invention by Gutenberg, to guarantee the standard quality of the characters, which would in turn guarantee the consistent quality of the print itself. In the company contract that Plantin concluded in the 1560s, it was the only item specified separately. Chinese printers circumvented this big investment by having carvers make wooden or metal (mostly bronze) characters, instead of casting them from matrices. Tsien concluded that ‘to exploit fully the advantages of movable-type print- ing, casting seems the logical method, because it is much harder to cut characters in bronze than in wood, and since each font includes tens or hundreds of thousands of characters, to cut them individually by hand would be against the very economic principle which dictated the use of movable type’.20 He also maintained that much of the evidence suggested, however, that engraving of moveable type had been the standard practice in China. Chinese printers, when using moveable type, often preferred wooden font, which, given the cheap supply of carvers (about which more below), was relatively easy to acquire. The big disadvantage of this was the much lesser durability of the material—after printing a few thousand copies, they had to be renewed again, whereas metal could be used much longer. Met- al movable type had been invented in Korea, perhaps already in the elev- enth century, and probably reinvented in China during the Ming, but it did not become the standard technology. Moreover, when metal was used in East Asia, instead of casting the type, softer materials were used to carve it, resulting in similar problems; of course, the lead that was used most frequently in Europe, was also rather soft and would ultimately need to be replaced as well.

Relative Prices

We have established now that there were consistent differences in the choice of technique between East Asia and Europe (and the Arabic World), which cannot easily be explained with reference to the cultural factors alone. The argument is that economic factors—market size and relative prices—have played a dominant role. Moveable type printing was the most capital intensive way of reproducing texts, in particular when the font was

20 Tsien Tsuen-Hsiun, Paper and Printing, in Joseph Needham (ed.), Science and Civili- sation in China, vol. 5, Cambridge U.P., 1985), 218-19. explaining the global distribution of book production 333 not carved but cast, which required a large investment in very expensive matrices. One of the key differences was related to the ‘pressing’ stage of the production process. A key capital good in European printing was the printing press. Usually this press has been seen as the most important in- vention made by Gutenberg, who probably had the wine press as a source of inspiration. It is significant that Chinese printers did not use such a press at all, although the idea is rather simple, and similar technologies for press- ing for example oilseed—for the manufacturing of oil and bean cake— were well known. Chinese printers did not use a capital good, but manual labour, although might have been aware that a lot of labour could be saved by using such as press.21 As we saw already, Chinese and Japanese printers usually carved their texts in woodblocks, or, in the few cases that moveable type was used, carved their font, instead of using expensive matrices to cast font. Again, this can be interpreted as a choice for a more labour intensive mode of production, whereas Europe developed a capital-intensive technique. Usu- ally, economic historians tend to think that saving labour by using more advanced capital goods is under all circumstances rational, because it leads to lower costs and higher profits. We cannot compare the two technologies under European conditions (and factor prices), as only one technology dominated in that part of the world. But in China the two technologies did coexist, which gave rise to a few contemporary studies comparing them. In the late eighteenth century Jin Jian, who as head of the official printing office had a lot of experience with large printing jobs, compared the costs of printing with woodblock or with (wooden) movable type, and conclud- ed that the latter was more expensive. The material costs of printing a copy of a chapter of Shiji (Memories of the Grand Historian) using wooden mov- able type were 2,339 taels per volume compared with 1,450 taels when woodblocks would be used. The first required the carving out of a font of 235,000 types of the most common 6,700 characters; for woodblock printing ‘only’ 2,675 blocks had to be manufactured.22 Even for such a big project, woodblock printing seems to have been cheaper, but we have to realize that print runs of these imperial projects were extremely low— maybe as low as one single copy—and that movable type printing profited from economies of scale.

21 Kai-wing Chow, Publishing, Culture, and Power in Early Modern China (Stanford: Stanford U.P., 2004), 70-71. 22 Moll-Murata, State and Crafts, 244. 334 jan luiten van zanden A few decades later, in the 1830s, another interesting comparison was made by Medhurst, a Christian missionary who planned to print the Bible in Chinese. He arrived at another conclusion: producing 2,000 copies with woodblocks would costs 1,900 pounds sterling, and it would take three years to finish the project, during which nine blockcutters and five printers and binders had to be employed. Using metal moveable type would require seven workers during one year, at a total cost of 1,515 pounds. Moreover, whereas the woodblocks would wear out after a run of 2,000 copies, with letterpress it would be possible to make up to fifty runs.23 As this com- parison demonstrates, the labour productivity of the two technologies differed enormously: to make the same 2,000 volumes of the Bible, one needed either 42 men-year when using the woodblock technology, and 7 men-year when using metal movable type. Hand copying 2000 volumes of the Bible, it may be assumed, would have taken perhaps as much as 1-2,000 men-year. The economic advantages of moveable type printing in the Chi- nese context, as estimated by Medhurst, may have been a rather recent development, however. European printing technology developed rapidly in the late eighteenth and early nineteenth century, which must have led to even greater increases in labour (and perhaps also capital) productivity. Why did European printers systematically choose the more capital in- tensive and labour saving technologies, whereas Chinese and Japanese printers, who often must have been aware of the different technologies available, opted for more labour intensive and less capital intensive tech- niques? I propose that the difference is probably related to differences in relative prices, in particular of capital and labour. Capital was much cheap- er in Europe, and access to capital markets was relatively easy. Plantin, for example, borrowed at 6.25%, and used 5% for internal accounting pur- poses, and such very low interest rates were quite usual in Europe at the time. Interest rates in China were much higher, from 12 to perhaps 24%, as was already observed by Adam Smith, although they too may have been declining in the long run.24 By contrast, labour was relatively expensive in Europe, in particular during the fifteenth century, when the transition to- wards the capital intensive moveable type printing occurred. In the cen- tury and a half after the Black Death, real wages first increased to—by international standards—very high levels and they remained that high in

23 C.A. Reed, Gutenberg in Shanghai. Chinese Print Capitalism 1876-1937 (Honolulu: University of Hawai’i Press, 2004), 31-32 24 Jan Luiten van Zanden, The Long Road to the Industrial Revolution. The European Economy in a Global Perspective 1000-1800 (Leiden: Brill, 2009), 23-24. explaining the global distribution of book production 335 large parts of Europe until the price revolution of the sixteenth century. In England and Holland, they never fell back to the pre 1348 level again, in other parts of Europe there was a strong decline of real wages during the early modern period. During the late Medieval period, between 1350 and 1450, interest rates declined strongly whereas real wages went up a lot. This scissor movement led to a (modest) increase in labour saving technology, of which the printing press is probably the best example.25 Gutenberg’s invention should therefore be understood against the background of these changes in relative factor prices, which gave a strong incentive to develop labour saving technologies. Another factor might be—but this is more speculative—that metals such as copper and in particular lead were cheaper in Western Europe, whereas wood was perhaps less scarce in China. There is an intriguing table in one of Bob Allen’s papers, on prices in England and China at about 1700, which makes the point that in particular lead was much more expensive in the latter country than in England.26 It is striking that lead, which be- came the standard metal for typecasting in Western Europe, was probably never used in China: only bronze and tin are mentioned by Tsien, who of- fers the most detailed discussion of this issue.27 Chow also mentions only one example of an experiment with lead.28

Size of the Market

Matrices were a crucial part of the investment a European printer had to make, but even larger investments went into the books itself. The total invested capital of the Plantin firm in the 1560s was about 10,000 guilders, of which no more than 20-25 was invested in fixed assets such as matrices, font, presses etc., and the rest (about 70%) almost exclusively in books that had been produced.29 To complicate things a bit, Plantin, like other pub-

25 Ibid., 178-179. 26 Taken from Robert C. Allen, ‘Mr. Lockyer Meets the Index Number Problem: the Standard of Living in Canton and London in 1704’, Unpublished paper Conference ‘Towards a Global History of Prices and Wages’, Utrecht 2004; ratios of Chinese to English Prices (in silver): tea .04; rice .05; unskilled labour .27; cotton cloth .30; ironwork .32; copperwork .35; nails .69; beef .70; broad cloth .75; brass work .80; paper .83; milk 1.13; lead 1.20; mutton 1.68; bread 1.76; charcoal 5.28. Note how expensive paper and lead were in relation to for example unskilled labour and iron, not to mention rice. 27 Tsien, Paper and Printing, 194-222 28 Chow, Publishing, culture and power, 68. 29 Florence Edler, ‘Cost Accounting in the sixteenth century: the books of account of Christopher Plantin, Antwerp, Printer and Publisher’, The Accounting Review 12 (1937), 226-37. 336 jan luiten van zanden lishers, also traded in the books of other printers/publishers, which also covered part of the ‘remaining’ capital. In order to optimally use the rela- tively large capital invested in the fixed assets, and because each book represented a relatively large investment in itself, large print runs were optimal—and also possible given the durability of the characters. Average print runs therefore increased to more than 1,000 in early modern Europe.30 Print runs with woodblocks were dependent on the quality of the wood and the work of the carver, were often worn out after a few hundred, per- haps two thousand, copies had been made. In theory, very large print runs of (tens of) thousands of books were feasible, and would strongly affect the average price of the book if it was a commercial success. But such print runs also required a large investment in stocks of books, and of other ma- terials such as paper and ink. Two factors seem to be involved here: access to (cheap) capital, which was superior in Europe, and the size of the mar- ket. If the market is big enough, and books sell readily, the capital invested in large print runs will still be limited. If China really was such a huge market for books as has been suggested by recent literature, the question remains why the disadvantages of high interest rates were not offset by the advantages of being able to produce large print runs with the moveable type technology. That Chinese publishers struggled with the problem of how to finance their stocks of books, is evident from examples Chow gives of publishers who first printed part of the book, in order to finance the rest of the project by selling the first part.31 Woodblock printing had a certain advantage here as well: one could easily print a small run first, and try to sell the copies produced, and store the woodblocks for a possible second run. European printers could do this as well, of course, but probably usu- ally needed the characters (and the frames) for the next project, and where more induced to finish the project in one print run. This implies that move- able type printing was in fact even more capital intensive than is suggested by the technology only, because it required large print runs, whereas small- er print runs were already feasible with woodblock printing What does the fact that movable type printing did not really ‘take off’ in China before 1800 (or even before the second half of the nineteenth century), and that even the growth curve of woodblock printing as sug- gested by Figure 9.1 was much less dynamic than that in Western Europe, tell us about the size of the market in China? Much of the recent literature

30 Buringh and Van Zanden, ‘Charting the “Rise of the West”’, 415. 31 Chow, Publishing, culture, and power, 64. explaining the global distribution of book production 337 on Chinese printing industry makes the point that it was, from the late Ming onwards, quite dynamic and catering for a rapidly developing mass market. In principle, given the fact that the population of China was as large as that of Europe as a whole, but unified by one common language, the potential size of the market there was much bigger than in Europe. The gains that could be realized through economies of scale were therefore very substantial indeed. On the basis of sheer population numbers, one would therefore expect that a technology such as moveable type printing that was dependent on large markets would easily dominate the Chinese market, which was huge (again in terms of population) compared to the much smaller markets of the various European countries. This was not the case, however. The question therefore remains, why not? Factor costs are part of the story, but could a really large market not compensate for this? It may not have been a coincidence that this pattern of global tech- nologies coincided with huge differences in the number of books produced per capita. In a related paper it was estimated that China produced on average about 3 new titles per million inhabitants between 1750 and 1800, Japan 7, Russia 6, Poland 23, Germany 126, the USA 142, Sweden 219, the UK 198, and the Netherlands 538.32 Because print runs in Europe were prob- ably higher than in the countries with woodblock printing, the actual dif- ferences in terms of numbers of copies were probably even larger than this.33 In Western Europe, moveable type printing seems to coincide with the existence of a mass market, but does this imply that the market for books in China was much more limited as these estimates suggest? There are a number of explanations for this that can be considered: low levels of literacy, low real incomes, and high book prices. Although some estimates have been produced about the level of literacy of eighteenth-century Chi- na, we do not know a lot about this. There clearly was a class of literati, and literacy as such played a large role in society—most importantly, of course, in relation to the examination system. If the level of literacy was as high as has been suggested by Rawski—that is, 40% for men—the low level of book production is a puzzle, but the Rawski estimate may simply be too high.34 We know a little bit more about real incomes in China, compared with Western Europe; there is some evidence that real wages even in the most advanced parts of China were substantially lower than in the North-

32 Buringh and Van Zanden, ‘Charting the “Rise of the West”’, 437. 33 Chow, Publishing, culture, and power, 248. 34 E.S. Rawski, Education and Popular Literacy in Ch’ing China (Ann Arbor Mi.: Univer- sity of Michigan Press, 1979), 119. 338 jan luiten van zanden Western Europe, where we find the largest markets for books.35 Also, in the 1820s (when we have the first more or less reliable estimates), real incomes per head of the population in the most developed part of China, the Yang- zi Delta, was probably somewhat more than about half the level of that of the Netherlands, which seems to confirm the income gap estimated on the basis of real wages.36 Finally, there is the possibility that the printing industry was relatively inefficient, resulting in high prices for books, which may have restricted demand. In Western Europe, real book prices fell dramatically from the 1450s onwards, by more than 90% during the whole 1450-1800 period. In the Netherlands, the price of the cheapest Bible, to give a particularly spec- tacular example, fell from about the annual wage of a craftsmen in the 1460s to his daily wage in the eighteenth century.37 But the impressionis- tic evidence that is available suggests (as was already remarked by Ricci) that book prices in China were not extremely high—in fact, in nominal terms they may have been similar to European prices.38 The most plausible explanation for the smaller size of the market for books is probably the combination of a level of literacy that was lower than in the most advanced parts of Europe, and the generally lower real incomes that were being earned. Perhaps in the Yangzi delta the market for books on a per capita basis was much larger, but we do not have data on this. An alternative solution of the paradox is that economies of scale of moveable type printing were, given the factor costs in China, much more limited than in Europe. The costs estimates presented by Reed do however show that in the early nineteenth century large savings could be realized by switching to moveable type printing.

Conclusion

We have tentatively tried to explain the global distribution of technologies for reproducing the written word in the period before 1800, and in particu- lar the reasons behind the fact that in the Arabic World only hand-copying

35 R.C. Allen, J.P. Bassino, J.P., D. Ma, Chr. Moll-Murata, and J.L. van Zanden, ‘Wages, Prices, and Livings Standards in China, Japan, and Europe, 1738-1925’, Economic History Review 64 (2011), 8-38. 36 Li Bozhong, and Jan Luiten van Zanden, ‘Before the Great Divergence? Comparing the Yangzi Delta and the Netherlands at the beginning of the nineteenth century’. Journal of Economic History 72, 4 (2012) 956-989. 37 van Zanden, The Long Road, 182-83. 38 Chow, Publishing, culture, and power, 39-48. explaining the global distribution of book production 339 occurred, in China (and Japan) woodblock printing dominated the indus- try, whereas in Western Europe moveable type printing became all-impor- tant after 1454. Quality issues and the character of the script may have played a role, but these factors appear not to have been the only or even the main causes of the fundamental differences in regional spread of tech- nologies. For the Arabic world, for example, quality issues may have hin- dered the acceptance of moveable type printing, but these same issues made Arabic an ideal candidate for woodblock printing, which however also was not used as a technology there. The fact that in China the font size used in moveable type printing had to be much larger than in Western Europe, also does not seem to have been crucial; European printers, for example, also used very large fonts (with 100,000 to sometimes 200,000 characters)—their main investment went into the matrices to cast these characters, but Chinese printers who used moveable type did not develop this technique, and continued to have the characters carved by specialists, circumventing the problem. What seems to have been crucial in explaining the differences between China (and Japan) on the one hand, and Western Europe on the other hand, are differences in factor costs. When Gutenberg developed his version of moveable type printing, real wages were at a historical peak (after the Black Death of 1348 and its many echo-epidemics), whereas interest rates had fallen to—by historical standards—very low levels (5-6%). Parts of Western Europe continued to have relatively high real wages, and low interest rates, throughout the Early Modern period. Movable type printing was the ideal ‘high-tech’, capital intensive and labour saving technology for such an en- vironment; but it was only economically feasible when large print runs could be made, and therefore the success of this technology was also de- pendent on the size of the market for books in Western Europe. Factor costs in China and Japan favoured a less capital-intensive technology, and the size of the market was perhaps also not large enough to make movable type printing a feasible option before the nineteenth century. These differences in production technology therefore reflect wider disparities in economic development, relevant for understanding the ‘Great Divergence’. The Arab World is a different story, however. Government policy—a ban on printing in Arabic (until 1727)—is perhaps part of the explanation for the absence of a book industry there. However, the attempts by European printers to produce for the Arabic and Ottoman markets were also unsuc- cessful. Another part of the story is probably that, after the boom in man- uscript production in the Arab World between 800 and 1200, demand seems 340 jan luiten van zanden to have declined to levels that remained very low throughout our period. When the ban was lifted, in the early eighteenth century, this also did not lead to a strong increase in book production; on the contrary, levels re- mained extremely low. What is, therefore, behind these global patterns of the book industry is a global distribution of human capital—of the skills to read and write. It points at large concentrations of these skills in the two extremes of Eurasia, in China and Japan, and in Western Europe (increas- ingly concentrated in the North Sea region). In between, in the Arab World, levels of literacy were much lower. The way the written word was copied tells us a lot about the world economy at the start of the Industrial Revolu- tion. bibliography of the published works of s.r. epstein 341

Bibliography of the books and articles published by S.R. Epstein

1986 Alle origini della fattoria toscana. L’ospedale di Santa Maria della Scala di Siena e le sue terre, c.1250-c.1450 (Florence: Salimbeni) 1988 ‘Inventari di beni di grance dell’ospedale di Santa Maria della Scala di Siena (1356, 1385)’, Archeologia medievale, 15: 535-49 1989 ‘The textile industry and the foreign cloth trade in late medieval Sicily (1300-1500): a “colonial relationship”?’, Journal of medieval history, 15: 141-83 1990 ‘Cities, regions and the late medieval crisis: Sicily and Tuscany compared’, Past and Present, 130: 3-50 1992 An island for itself: Economic development and social transformation in late medieval Sicily Past and Present Publications Series (Cambridge: Cambridge University Press); revised edition translated as Potere e mercati in Sicilia: Secoli XIII-XVI (Turin: Einaudi, 1996) 1993 (Edited, with T. Szabó, D. Ciampoli, and M. Ginatempo), Lo Statuto dei Viarî di Siena: Viabilità e legislazione di uno stato cittadino del Duecento Siena: Accademia senese degli intronati [a volume of transcribed documents] ‘Town and country in late medieval Italy: economic and institutional aspects’, Economic history review, 2nd ser. 46: 453-77 ‘Manifatture tessili e strutture politico-istituzionali nella Lombardia tardo-medievale. Ipotesi di ricerca’, Studi di storia medioevale e diplomatica, 14: 55-89 ‘Marc Bloch: the identity of a historian’, Journal of medieval history, 19: 273-83 ‘Istituzioni politiche, economia regionale, commercio internazionale: il caso della Sicilia tardo-medievale’, in: Strutture del potere economico e politico nelle città dell’Europa medievale e moderna (Naples), 43-64 1994 ‘Tuscans and their farms: Moral hazard and risk sharing in late medieval Tuscany’, Rivista di storia economica, 12: 111-37 ‘Regional fairs, institutional innovation and economic growth in late medieval Europe’, Economic history review, 2nd ser., 47: 459-82 ‘Storia economica e storia istituzionale dello stato’, in: P. Schiera ed. Origini dello Stato: Processi di formazione statale in Italia fra medioevo ed età moderna (Bologna), 97-111 1995 ‘A reply’, in ‘Debat: Comerç internacional i desenvolupament regional en el Mediterrani: La Sicília medieval de S.R.Epstein’, Revista d’història medieval 5: 133-80 ‘Dualismo economico, pluralismo istituzionale nell’Italia rinascimentale’, Revista d’història medieval 6: 63-77 ‘Freedom and growth: The European miracle?’, in: E.V. Barker (ed.) LSE on freedom (London), 165-81; reprinted New Brunswick, N.J. 1997; Hamburg, 1997; translated as ‘Libertad y crecimiento: el milagro europeo?’, Anuario IEHS (Buenos Aires), 14 (1999) ‘Conflitti redistributivi, fisco e strutture sociali (1392-1516)’, in: F. Benigno and C. Torrisi (eds), Elites e potere in Sicilia: Dal medioevo ad oggi (Rome), 31-45 342 bibliography of the published works of s.r. epstein

1996 ‘Stato territoriale ed economia regionale nella Toscana del Quattrocento’, in: R. Fubini (ed.), La Toscana al tempo di Lorenzo il Magnifico: Politica economia cultura arte (Pisa), vol. III, 869-90 ‘Taxation and social representation in Italian territorial states’, in: M. Boone and W. Prevenier (eds), Finances publiques et finances privées au bas moyen âge (Leuven- Apeldoorn), 101-15 ‘Governo centrale e comunità del demanio nella Sicilia tardo-medievale: le fonti capi- tolari’, in: Atti del XVI Congresso di Storia della Corona d’Aragona, Sassari-Alghero 19-24 maggio 1990: La Corona d’Aragona, Sassari-Alghero 19-24 maggio 1990: La Corona d’Aragona in Italia (secc.XIII-XVIII). Presenza ed espansione della Corona d’Aragona in Italia (secc.XIII-XV) (Sassari), vol. III, 383-415 1997 ‘Eileen Power e la storia economica europea’, Storica 8: 201-7 ‘Power, resistance and authorities: Craft guilds and technological change in pre-indus- trial Europe’, in: B. Ranson (ed.), Guild-Hall and government: An exploration of power, control and resistance in Britain and China, vol. II. Power, resistance and authorities: Aspects of guild organisation in England (Hong Kong), 46-69 1998 (Edited, with Heinz-Gerhard Haupt, Carlo Poni and Hugo Soly), Guilds, economy and society, Madrid: Universidad de Sevilla) ‘Nuevas aproximaciones a la historia urbana de Italia: el Renacimiento temprano’, Història (Madrid), 417-38 ‘Craft guilds, apprenticeship and technological change in pre-modern Europe’, The journal of economic history 53 (1998), 684-713; reprinted in S.R. Epstein, Maarten Prak (eds), Guilds, innovation, and the European economy, 1400-1800 (Cambridge: Cambridge University Press), 52-80 ‘Italy’, in T. Scott (ed.), The peasantries of Europe from the fourteenth to the eighteenth century (London), 75-110 2000 Freedom and growth: Markets and states in Europe, 1300-1750 (London: Routledge) ‘Late medieval and early modern towns as focal points of market power: An interview with Stephan Epstein’, Itinerario 15, 2: 87-104 ‘Market structures’, in: W. Connell and A. Zorzi (eds), Florentine Tuscany: Structures and practices of power (Cambridge), 90-121 ‘Constitutions, liberties, and growth in pre-modern Europe’, in: M. Casson (ed.), Cultural factors and economic growth (Munich-New York), 152-81 ‘The rise and decline of Italian city-states’, in: M.H. Hansen (ed.), A comparative study of thirty city-state cultures (Copenhagen), 277-94 ‘The late medieval crisis as an “integration crisis”’, in: M. Prak (ed.) Early modern capital- ism: Economic and social change in Europe, 1400-1800 (London), 25-50 2001 (Edited), Town and country in Europe, 1300-1800 (Cambridge: Cambridge University Press 2001). ‘Town and country in Europe, 1300-1800’, in: S.R. Epstein (ed.), Town and country in Europe between the fourteenth and the eighteenth centuries (Cambridge), 1-29 ‘Labour and labour markets between town and countryside. Part II: Middle Ages-16th century. Introduction’, in B. Blondé, E. Vanhaute and M. Galand (eds), Labour and labour markets between town and countryside (Middle Ages – 19th century) (Turnhout), 76-81 ‘Fairs, towns, and states in Renaissance Europe’, in: S. Cavaciocchi (ed.), Fiere e mercati nella integrazione delle economie europee secc.XIII-XVIII (Florence), 71-90 bibliography of the published works of s.r. epstein 343

2002 ‘History matters: lessons from the marketplace’, LSE Magazine, Summer 2002 ‘Strutture di mercato’, in: A. Zorzi and W. Connell (eds), Lo Stato territoriale fiorentino (secoli XIV-XV): Ricerche, linguaggi, confronti (Pisa), 93-134 2004 (with M.M. Bullard, B.G. Kohl, and S.M. Stuard), ‘Where history and theory interact: Frederic C. Lane on the emergence of capitalism’, Speculum 79: 88-119 ‘Property rights to technical knowledge in premodern Europe, 1300-1800’, American Economic Review 94, 2: 382-87 ‘Labour mobility, journeyman organisations and markets in skilled labour Europe, 14th- 18th centuries’, in: L. Hilaire-Pérez and A. Garçon (eds), Les chemins de la nouveauté: Innover, inventer au regard de l’histoire, collection CTHS Histoire, vol. 9, (Paris), 411-30; and in: M. Arnoux and P. Monnet (eds), Le technicien dans la cité en Europe occidentale 1250-1650 (Rome), 251-69 ‘Apprenticeship’, in: J. Mokyr (ed.), The Oxford encyclopedia of economic history (Oxford), vol. 1: 146-49 ‘Craft guilds’, in: J. Mokyr (ed.), The Oxford encyclopedia of economic history (Oxford), vol. 2: 35-39 ‘Journeymen’, in: J. Mokyr (ed.), The Oxford encyclopedia of economic history (Oxford), vol. 3: 211-12 ‘Naples’, in: J. Mokyr (ed.), The Oxford encyclopedia of economic history (Oxford), vol. 4: 39-30 2006 ‘The rise of the West’, in: J. Hall and R. Schroeder (eds), An anatomy of power: The social theory of Michael Mann (Cambridge), 233-62 2007 ‘L’economia italiana nel quadro europeo’, in: F. Franceschi, R.A. Goldthwaite, R.C. Mueller (eds), Il Rinascimento Italiano e l’Europa, vol. 4: Commercio e cultura mercantile (Treviso), 3-47 2008 (Edited, with M. Prak), Guilds, innovation, and the European economy, 1400-1800 (Cambridge: Cambridge University Press) (with M. Prak), ‘Introduction: Guilds, innovation and the European economy, 1400-1800’, in: S.R. Epstein and M. Prak (eds), Guilds, innovation, and the European economy, 1400- 1800 (Cambridge), 1-24 ‘Craft guilds in the pre-modern economy: A discussion’, Economic History Review 61: 155-74 2013 ‘Transferring technical knowledge and innovating in Europe, c.1200 – c.1800’, in this volume 344 bibliography of the published works of s.r. epstein index 345

INDEX

Abbasids, 328 Aurangzeb, 72 Abdülmennan, Sinan Ibn, 136 Austria, 228, 265 Academy, 239, 240 Averlino, Antonio, 48 Aceh, 178 Afghanistan, 141 Bacon, Francis, 55 Africa, 141, 165, 166, 171, 175 Baghla, 178 Agra, 82 Baker, Mathew, 45, 189 Agricola, Georgius, 47, 213, 220 Baltic, 173, 220 Ahmed III, 327 Barovier, Angelo, 50 Ahmedabad, 80, 318 Batavia, 176, 210, 211, 324 Akra, 311 Battuta, Abu Abdullah Mohammed Ibn, Alberti, Leon Battista, 47, 48 165 Alchemy, 51, 56 Bay of Bengal (see also: Bengal), 168, 172 Allen, Robert, 8, 13, 223, 335 Beauvais, 40, 147 Americas, 82, 180, 202, 220, 323 Beijing, 102, 254 Amiens, 40 Belarus, 300 Ammers, Dirk van, 293 Belgium, 273 Amoy, 211 Bengal (see also: Bay of Bengal), 72, 81, 308, Amritsar, 82, 83 311, 314, 315, 318 Amsterdam, 11, 36, 97, 189, 196, 202, 204, 220, Bengalis, 88 278 Berg, Maxine, 223, 224 Anhui, 118, 119, 254 Bernier, François, 72, 73 Antwerp, 36, 97, 139, 241 Besson, Jaques, 213, 220 Application, 54 Bhayya, 87, 88 Apprenticeship, 29, 30, 31, 32, 59, 60, 61, 65, Bihar, 80, 91 69, 70, 71, 74, 77, 79, 81, 82, 83, 84, 85, 91, Black Death, 33, 59, 161, 334, 339 92, 105, 109, 110, 153, 154, 156, 163, 185, 187, Blacksmith, 81, 91, 207, 229 201, 204, 207, 208, 215, 216, 217, 221, 236, Bloom, Jonathan, 329 237, 238, 239, 246, 247, 248, 251, 252, 255 Bohemia, 26, 253 Arabia, 171, 183, 184, 324, 327, 328, 329, 330, Boilermakers, 115 331, 338, 339, 340 Bologna, 154 Architecture, 7, 36, 47, 53, 221 Bombay, 79, 80, 81, 85, 86, 87, 88, 89, 90, 91, Arita porcelain, 122 324 Arita, 241, 243 Book industry, 16, 50, 60, 211, 213, 222, 323- Arkwright, Richard, 223 340 Armenia, 326 Bouguer, Pierre, 191 Art schools, 251 Bowrey, Thomas, 72, 73 Artois, 219 Boyle, Robert, 8 Asia Minor, 132, 136 Brabant, 138 Asia, 6, 26, 94, 96, 97, 98, 109, 124, 125, 126, Bradstock, Timothy, 114 134, 162, 164, 165, 168, 170, 171, 172, 174, 176, Brazil, 202 177, 178, 179, 180, 181, 182, 184, 187, 192, 197, Breman, Jan, 316 199, 202, 227, 228, 259, 261, 288, 323, 332 Brickmaking, 2, 153, 209, 259-322 Atlantic, 174, 180 Bristol, 32 Augsburg, 39, 60 British Isles, 3, 220 August II, 229 Brook, Timothy, 222 346 index

Brunelleschi, Filippo, 40, 41, 47 241, 253, 254, 256, 264, 323, 325, 326, 327, Brussels, 139 328, 329, 330, 331, 332, 333, 334, 335, 336, Building industry (see also: cathedral build­ 337, 338, 339, 339, 340 ing), 2, 10, 11, 13, 19, 26, 39, 43, 45, 49, 53, Chow, Kai-Wing, 335, 336 59, 155, 157, 159, 209, 219, 264, 289 Civil War (English), 46 Bukhara, 148, 288 Clark, Gregory, 18, 89, 260, 261, 262, 263, 319, Bull, William, 312 321 Burma (Myanmar), 88, 169 Clockmakers, 35, 52, 59, 60, 73, 126, 215, 219 Bursa, 132 Cloth production, 26, 35 Byzantium, 136, 147, 148, 155, 287 Clusters, 12 Coblers, 73 Caen, 139 Codification, 64, 66 Cairo, 132 Colbert, Jean-Baptiste, 46 Calcutta, 85, 88, 89, 90, 91, 309 Collective invention, 64 Calendering, 103 Collective knowledge, 54 Cambodia, 132, 142, 181 Confucian, 239, 325 Cambridge, 145 Cookson, Gillian, 223, 224 Canterbury, 138 Coromandel, 88 Canton, 116, 176 Coster, Salomon, 37 Cape of Good Hope, 220 Cotton, 4, 79, 81, 85, 114, 223, 324 Caravels, 175 Cotton mills, 91 Cardano, Girolamo, 56 Cotton weaving, 60 Carpentry, 73, 80, 115, 116, 123, 207, 208, 209, Craft associations (see also: guild), 65 212, 215, 216, 219, 221, 229 Carpet weaving, 82, 83 Dam, Jan Daniel van, 242, 243 Caste, 80, 87, 88, 156, 308, 313 Davids, Karel, 12, 62 Cathedral building (see also: Building in- Deccan, 79 dustry), 40, 42, 43, 44, 46, 49, 52, 61, 65, Delft, 14, 15, 16, 225, 240, 241, 242, 243, 244, 288 245, 246, 247, 249, 252, 255 Catherine II, 288 Delhi, 308 Ceded, 88 Della Valle, Pietro, 73 Central Europe, 43 Deng, Kent, 222 Ceramics, 212, 225-256 Denmark, 228, 269 Ceredi, Giuseppe, 50, 51, 54 Descartes, René, 39, 55 Chaibasa, 315 Dhow, 171, 176, 178, 202 Chandavarkar, Rajnarayan, 86, 87, 88 Dondi, Giovanni de, 39 Changsha, 104 Douw, Simon, 37 Chaozhou, 103 Drainage, 50 Charleroi, 273 Drawing, 39, 41, 43, 46, 48, 55, 58, 64, 216, Chartres, 40 220, 223, 224 Chemistry, 64 Dresden, 249, 251 Chennai, 317 Duchang, 234, 235, 236, 254 Chernov, Nicolai, 287 Duhamel du Monceau, Henri-Louis, 191 Chia, Lucille, 325 Dutch East India Company (VOC), 176, 198, Children, 77 216, 228, 241, 243 China, 2, 3, 6, 8, 14, 15, 16, 17, 19, 20, 21, 26, Dutch painting, 57 59, 86, 93-127, 132, 137, 142, 150, 156, 157, Dutch Republic, 26, 61, 62 159, 164, 165, 166, 167, 168, 169, 170, 173, Dutch Rijnland, 270 176, 179, 180, 182, 183, 187, 188, 191, 192, 193, Dyers of cloth, 73 197, 199, 202, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215, 217, 218, 221, 222, 223, East Asia Companies, 228 224, 225, 226, 227, 228, 233, 234, 235, 240, East China Seas, 165 index 347

Edict of Nantes, 61 Finlay, Robert, 226, 227 Edison, Thomas, 1 Flanders, 215, 219 Edo, 122 Florence, 37, 40, 41, 47, 123 Egypt, 141, 156, 227, 328 Flushing, 216 Elman, Benjamin, 214 Flute (fluyt), 10, 175, 177 Elvin, Mark, 3 France, 37, 43, 48, 49, 52, 53, 59, 61, 185, 191, Embroidery industry, 81 197, 215, 220, 229 Energy production, 62 Free riders, 58 England (see also: Great Britain), 8, 20, 29, Free riding, 65 32, 45, 46, 47, 48, 49, 52, 53, 57, 61, 63, 70, French, 38, 40, 46, 51, 145, 156, 219, 220, 228, 82, 84, 109, 119, 135, 138, 139, 176, 177, 181, 229, 253, 273 185, 189, 191, 192, 196, 215, 216, 219, 223, Frisia, 276 228, 243, 262, 267, 271, 278, 281, 282, 284, Friuli, 318 335 Fromanteel, Ahasuerus I, 37 Entrecolles, François Xavier de, 229, 230 Fryer, John, 73 Epstein, S.R., 93, 94, 95, 96, 107, 108, 112, 114, Fujian, 103 120, 121, 123, 124, 125, 126, 157, 161, 163, 217, Fuliang, 254 218, 225, 226, 248, 249, 256, 259, 281, 284, Furniture industry, 80 341, 342, 343 Furttenbach, Joseph, 189 Ethnicity, 69, 92, 177 Euler, Leonhard, 190 Galilei, Galileo, 30, 55 Europe, 2, 3, 4, 5, 6, 7, 8, 9, 14, 15, 17, 18, 19, Galleons, 181 20, 21, 25, 26, 27, 31, 33, 35, 36, 42, 43, 56, Galleys, 175 57, 59, 60, 61, 65, 66, 70, 71, 72, 73, 74, 76, Ganges Canal, 309, 318 81, 91, 92, 93, 94, 95, 96, 97, 101, 105, 107, Ganges-Indus, 75 109, 110, 111, 112, 114, 120, 121, 122, 123, 124, Gaoyang, 120 125, 126, 131, 136, 138, 141, 153, 155, 156, 159, Gender, 248, 272, 273, 287 161, 162, 163, 164, 167, 168, 170, 173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, Geneva, 61 187, 188, 189, 192, 193, 199, 202, 203, 204, Genoa, 60, 154, 196 206, 213, 214, 215, 217, 218, 219, 220, 221, Geometrical, 43, 46, 55, 191, 196 222, 223, 224, 225, 226, 227, 228, 236, 240, Geometry, 143, 148, 159, 188 241, 249, 253, 256, 259, 260, 261, 264, 265, George III, 181 267, 268, 271, 276, 281, 282, 285, 288, 293, Georgia, 288 308, 309, 310, 312, 317, 319, 320, 322, 326, Germany (see also: Holy Roman Empire), 327, 328, 329, 330, 331, 333, 335, 336, 337, 26, 43, 48, 49, 53, 59, 60, 61, 62, 63, 135, 338, 339, 340 138, 140, 142, 144, 145, 154, 173, 183, 185, 218, Eversmann, Friedrich August Alexander, 219, 220, 267, 270, 273, 275, 287, 318, 337 277, 278 Ghent, 97, 139 Experiment, 49, 50, 51, 55, 199, 207, 228 Gil de Hontañon, Rodrigo, 44 Eyferth, Jacob, 118 Giorgio Martini, Francesco di, 47, 48 Glassmaking, 36, 51, 52, 56, 60, 62, 66, 251 Faenza, 240 Gmünd, 145 Falcon, Jean Baptiste, 38 Goa, 177, 180, 324 Family, 69, 71, 73, 74, 75, 77, 78, 80, 81, 82, 83, Godunov, Boris, 290 92, 98, 111, 113, 114, 118, 124, 125, 152, 153, Goldsmith, 52, 62, 72, 106 155, 157, 159, 216, 230, 234, 238, 239, 249, Gouda, 276, 284 255, 263, 283, 308, 309, 310, 312, 313, 315, Granada, 132 317, 319 Great Britain, 26, 59, 62, 76, 183, 204, 224, Fang, Zhuofen, 231 253, 310, 318 Fatimids, 328 Great Divergence debate, 4, 6, 8, 19, 21, 132, Ferreiro, Larrie, 190 224, 259, 260, 324, 339 348 index

Greece, 326 Indenture, 69, 70, 71, 216 Groningen, 276, 283, 284 India, 2, 4, 6, 14, 17, 18, 26, 69-92, 97, 142, 149, Guangdong, 102, 103, 208 156, 165, 167, 168, 170, 171, 172, 176, 177, 178, Guangzhou, 102, 176 179, 180, 181, 187, 198, 202, 260, 264, 265, Guild, 6, 9, 13, 15, 16, 18, 29, 31, 32, 33, 38, 39, 307-322, 319, 320, 321, 324 41, 42, 54, 55, 56, 57, 58, 59, 60, 61, 66, 69, Indonesia, 168, 170, 179, 202, 220 70, 73, 75, 92, 93, 94, 95, 98, 99, 100, 102, Industrial Revolution, 1, 3, 5, 17, 52, 96, 162, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 163, 201, 206, 221, 224, 322, 340 114, 120, 121, 123, 152, 153, 154, 155, 156, 159, Ingolstadt, 140 161, 162, 163, 164, 184, 185, 186, 187, 191, 192, Innovation, 52, 54, 56, 57, 58, 59, 62, 66, 67, 200, 201, 203, 204, 207, 208, 210, 215, 216, 76, 108, 114, 120, 125, 140, 162, 163, 175, 180, 217, 218, 221, 233, 234, 235, 237, 238, 239, 186, 187, 197, 198, 200, 202, 203, 225, 255, 240, 245, 246, 247, 248, 253, 254, 255, 256, 275, 307, 324 259, 271, 278, 279, 280, 284, 285 Instrument making, 2, 26, 36, 52, 54, 62, 126 Gujarat, 80, 88, 142, 149, 171, 316 Inventors, 49, 53 Gutenberg, Johannes, 323, 326, 330, 332, Iran, 141, 148, 156 333, 335, 339 Iraq, 156 Iron-making, 33, 113, 287 Haarlem, 215, 216, 217, 241 Irrigation, 50 Haastrecht, 284 Isfahan, 156 Habsburg Netherlands, 61 Islam, 132, 141, 148, 156, 171, 227, 308 Hamilton, Alexander, 73, 193 Istanbul, 132, 137, 327 Hangzhou, 102, 104, 137 Italy, 11, 26, 42, 47, 49, 51, 53, 57, 59, 60, 61, 63, Hankou, 254 154, 161, 219, 220, 228, 240, 241, 267, 270, Harris, John, 52 271, 278, 282, 283, 284, 287, 318, 320, 327 Havana, 189 Ivan the Terrible, 290 Hebei, 226 Henan, 118 Henry VI, 145 Jacquard (loom), 38, 77 Hideyoshi, Toyotomi, 168 Jakarta, 176 Hindu, 308 Japan, 6, 84, 86, 97, 113, 115, 116, 119, 121, 122, Hindustanis, 88 123, 124, 126, 167, 168, 241, 243, 323, 329, Hoffman Kiln, 265, 266, 267, 295, 299, 306 330, 333, 337, 339, 340 Hoffman, Phil, 21 Jars, Gabriel, 277 Holland, 196, 215, 219, 220, 243, 270, 282, 293, Java, 169, 177, 180, 202, 210 308, 335 Jesuit, 213, 328 Holy Roman Empire, 155 Jews, 61, 112, 326 Honnecourt, Villard de, 39, 47, 151 Jiajing, 118, 119, 126 Hoorn, 10 Jiangnan, 125 Horology, 36, 37 Jiangsu, 117 Hoysala Kingdom, 156 Jiangxi, 113, 227, 230, 234 Huang, Zhifan, 255 Jin, Jian, 333 Hubei, 103 Jingdezhen, 15, 16, 17, 226, 227, 228, 229, 230, Huguenots, 61 231, 232, 233, 234, 235, 237, 239, 243, 245, Huizhou, 235 246, 247, 248, 249, 253, 254, 255, 256 Humanist, 50 Johns Hopkins University, 141 Hunan, 103 Jones, Inigo, 53 Hungary, 138 Joshi, Chitra, 88 Huygens, Christiaan, 37 Journeymen, 59, 60, 61, 65, 154, 186, 218, 246 Junks, 165, 168, 169, 176, 179, 181, 183, 193, 202 Iberia, 175, 189 IJssel river, 276, 277, 278, 285, 321 Kaifeng, 137 index 349

Kalimantan, 169 Lombardy, 60, 123, 145 Kaluga (province), 290, 295, 300 London, 10, 14, 32, 36, 37, 41, 59, 77, 97, 154, Kanpur, 87, 88, 90 159, 229, 270, 272, 285, 308, 320 Karkhana, 75, 76 Loom-makers, 123, 215 Kashmir, 82 Low Countries, 26, 37, 53, 139, 153, 185 Keldermans family, 139 Lower Yangzi, 3, 20, 117 Kerala, 149 Lucknow, 81 Kerr, Rose, 226 Luther, Martin, 61 Khapilov pond, 290 Lyon, 9, 37, 38, 39, 57 Kharkov, 289 Kiev, 288 Maastricht, 139 King’s College, 145 Machinemaking, 17, 21, 35, 47, 48, 49, 54, 63, Klein, Ursula, 7 76, 205-224 Kolkata, 311 Macleod, Christine, 8 Konkan, 87 Madagascar, 171 Korea, 97, 323, 324, 328, 332 Maddison, Angus, 19 Kowaleski, Maryanne, 192 Madras, 88 Kunda Tahsil, 312 Makassar, 176 Kuran, Timur, 20 Malaysia, 168, 170, 171, 172, 173, 176, 179, 180, Kutch Behar, 314 202 Kyeser, Konrad, 220 Manchester, 286 Kyoto, 122 Manila, 181 Manuals, 17, 22, 52, 53, 111, 123, 152, 188, 189, Labour, 57, 58, 65, 70, 259-322, 333, 334, 335 211, 212, 219, 220 Lage, 287 Manucci, Niccalao, 73 Lake Ladoga, 289 Marathas, 87 Lake Poyang, 234 Marlborough, 139 Lambert, Richard, 89 Marshall, John, 37 Lan, Pu, 240 Masayuki, Tanimoto, 98 Lancashire, 87, 223, 285, 308 Masons, 30, 33, 38, 115, 154, 156, 209, 229 Lechler, Lorenz, 43, 142, 143 Master, 61, 65, 71, 74, 75, 79, 80, 83, 86, 92, 110, Ledderose, Lothar, 157 156, 186, 201, 216, 218, 236, 237, 238, 246, Leiden, 276 247, 248 Leipzig, 60 McDermott, Joseph, 324, 325 Lemgo, 287 Mecca, 171 Leping, 254 Medhurst, Walter Henry, 334 Li, Chao-hsiang, 188 Medici family, 327 Li, Jie, 149 Medicine, 111 Liang, Miaotai, 254 Mediterranean, 45, 46, 61, 62, 171, 173, 174, Libavius, Andreas, 56 180, 183, 184, 193 Liechtenstein, 252 Meissen, 229, 249, 250, 251, 255 Liège, 273 Melis, Federigo, 193 Lincoln, 154 Mesopotamia, 48, 307 Lindner, Gisela, 287 Metal industry, 52, 60, 61, 62, 64, 66, 76, 91, Linen, 62 115, 203, 223 Linpergh, Pieter, 220 Michels, Robert, 287 Lippe, 267, 269, 270, 271, 273, 274, 275, 276, Middle East, 2, 6, 16, 20, 97, 134, 227, 228, 281, 282, 283, 286, 287, 312, 320 264, 327 Lisbon, 198 Midlands, 13 Literacy, 110 Mignot, Jean, 43, 145 Literature, 224 Migration, 16, 35, 59, 60, 61, 67, 76, 80, 82, Liu, Yen, 166 92, 115, 116, 179, 180, 209, 210, 211, 218, 232, Lloyd (insurance), 193 234, 255, 294 350 index

Milan, 42, 43, 49, 60, 144 Osaka, 116, 122 Millwright, 216, 219, 223 Ottoman Empire, 136, 180, 323, 324, 326, 327, Mines, 26, 35, 52, 59 328 Mobility, 55, 62, 64, 65, 66, 138, 185, 186, 217, Oude Rijn river, 276, 280, 284 219, 220, 221, 223, 224 Ovington, John, 72 Models, 41, 46, 48, 49, 50, 51, 54, 64, 190, 220, Oxford, 97, 159 221, 224 Modules, 144, 148, 149, 150, 157, 158 Paape, Gerrit, 240 Mokyr, Joel, 1, 2, 6, 28, 133, 224 Pacific Ocean, 172 Moll-Murata, Christine, 109 Padmasali (weavers), 79, 80 Monghyr, 80 Padua, 50, 154 Mongols, 141, 325 Paid, 272 Montias, Michael, 246, 247, 248 Painting, 73, 115, 233, 235, 238, 239, 240, 248 Moscow, 288, 289, 290, 292, 293, 294, 295, Pakistan, 316, 318 297, 298, 299, 300, 302, 303, 304, 305, 307, Palacios, Diego de, 189 320 Palissy, Bernard, 51 Moskva river, 290 Palladio, Andrea, 53 Mumford, Lewis, 206 Palma, 40 Murano, 35, 36 Paper making, 11, 60, 63, 114, 117, 118, 119, 126, Murshidabad, 311 212, 215 Muslims (see also: Islam), 324, 326, 327, 329 Paracelsus, 56 Muteferrika, Ibrahim, 327 Pardeshi, 87, 88 Paris, 36, 37, 55, 97, 131, 135, 154, 155 Parler family, 152 Nanjing, 166, 188, 197 Parler, Heinrich, 145, 152 Natrus, Leendert van, 220 Parler, Peter, 152 Neckar, 142 Parthasarathi, Prasannan, 4, 14, 20 Nederdijk, Maarten, 221 Pasquier, Claudius Innocenz du, 250 Needham, Joseph, 205, 206, 212 Patent, 11, 37, 38, 42, 52, 53, 54, 57, 65, 66, 199 Netherlands (see also: Dutch Republic), 11, Patna, 315 14, 58, 59, 63, 176, 177, 179, 180, 181, 183, 190, Peddapuram, 79 191, 193, 241, 242, 248, 267, 270, 273, 276, Pérez, Liliane, 9 288, 284, 337, 338 Pergola, Paolo da, 50 Neva river, 289, 299 Persia, 82, 171, 227 Neva, 301 Perspective, 47 New Delhi (see also: Delhi), 309, 318 Peter the Great, 293 New World (see also: America), 162, 173, 181, Pharmacy, 111 184, 189, 191 Philip II, 41 Ningbo, 115, 116 Piacenza, 154 Nördlingen, 60 Plantin, Christophe, 331, 332, 334, 335 Normandy, 139 Pogozhev, 294 North Sea, 10, 173, 278, 340 Poland, 61, 138, 326, 337 Norwich, 32, 139 Polish Prussia, 219 Novgorod (province), 287, 300 Polly, Jacob, 220 Nuremberg, 39, 49, 60, 61, 66 Polo, Marco, 165 Nuvolari, Alessandro, 9 Polynesian, 172 Pomeranz, Kenneth, 3, 4, 19, 20, 133 O’Brien, Patrick, 6 Porcelain industry, 2, 14, 16, 17, 19, 21, 59, 106, Ogilvie, Sheilagh, 225, 226, 256 122, 225-256 Oliveira, Fernando, 189 Porter, Michael, 12 Orissa, 315 Portugal, 45, 46, 170, 176, 177, 178, 179, 180, Oriyas, 88 181, 197, 202, 228 index 351

Pottery, 51, 62, 246 Saxony, 228, 229 Prague, 138, 153 Scamozzi, Vincenzo, 53 Prak, Maarten, 125, 163 Scandinavia, 61, 273 Pratabgarh, 312 Schäfer, Dagmar, 8, 207, 213 Printing, 64, 104, 110, 115, 222 Schall von Bell, Adam, 213 Privas, 39 Schenk, Jan, 220 Proto-industry, 96, 106, 107, 122, 125, 249 Schleswig-Holstein, 219 Prussia, 228, 265 Schlüsselburg, 289 Punjab, 81, 88 Scholastics, 56 Schools, 124 Qimen, 254 Schooners, 179 Qiu, Pengsheng, 113, 114 Schreck, Johannes, 213, 220 Schumpeter, Joseph, 197 Quality control, 296 Schwarz, Leonard, 77 Queen Elizabeth I, 189 Scientific Revolution, 133 Scotland, 73 Raimondi, Marcantonio, 327 Scrafton, Luke, 73 Rajasthan, 312 Secrets, 55, 56, 58, 61, 65, 113, 114, 118, 119, 122, Ramelli, Agostino, 213, 220 186, 200, 238, 250 Ranchi, 315 Sèvres, 253 Raozhou, 254 Shandong, 208, 210 Rawski, Evelyn, 337 Shanghai, 115, 116, 117, 254 Real wages, 20 Shanxi, 103, 118, 137 Réaumur, 229 Shaoxing, 102, 103, 115 Red Sea, 170, 180 Shipbuilding, 2, 4, 10, 13, 17, 19, 21, 26, 33, 35, Regensburg, 39, 138, 155 45, 46, 48, 52, 59, 65, 102, 161-204, 212 Rent-seeking, 108, 110 Sholapur, 79 Rheims, 46 Siam, 179 Rhineland, 59 Sichuan, 104, 118, 119 Rhodes, Michael of, 187 Sie, Wun-ci, 235, 253 Rialto Bridge, 41 Silicon Valley, 13 Ricci, Matteo, 338 Silk industry, 9, 38, 39, 52, 57, 62, 76, 78, 80, Rijnland, 276, 277, 278, 279n32, 281, 284, 285, 106, 114, 207 297, 321 Silver wire, 80 Roberts, Lissa, 7 Silver-smiths, 73 Roman Empire, 165 Skachkov, Alexei, 287 Roriczer family, 153 Smelter, 81 Smith, Adam, 163, 334 Roriczer, Matthäus, 43, 144 Smith, Pamela, 7 Rosenthal, Jean-Laurent, 4, 94, 107 Soissons, 146 Rotterdam, 37, 215, 221, 241, 276, 284 Song, Yingxing, 8, 212 Royal Institution of Naval Architects, 204 South China Sea, 165, 170, 171, 173, 180 Royal Society, 37, 55 South Pacific, 171, 172, 173, 183 Ruitenbeek, Klaas, 208, 209 Spain, 44, 49, 52, 53, 61, 132, 181, 189, 191, 219, Russia, 18, 260, 264, 265, 287-307, 310, 319, 220, 228, 240, 326 320, 321, 337 Srinagar, 82 St. Petersburg, 288, 289, 291, 292, 293, 294, Saigon River, 179 295, 299, 300, 302, 303, 304, 305, 306, 307, Salviati (Filippo), 30 320 San Bartolomeo, Paolino da, 73, 74 Staffordshire, 253, 287 Sanbūq, 179 Statute of Artificers, 29, 70, 216 Sanotskii, 294, 299, 302 Steam-engine, 3, 4, 7, 158, 164, 182, 221, 223, Savona, 154 253, 295 352 index

Steelmaking, 63, 91 Urbanisation, 20, 63, 66, 67, 92 Stevin, Simon, 213, 220 Ursis, Sabbatino de, 213, 220 Stockport, 286 Utrecht, 215, 284 Stonemasons, 153 Uttar Pradesh, 312 Strasbourg, 153 Uzbekistan, 141, 148 Su, Yongming, 234, 255 Sub-contracting, 57, 263, 271, 278, 281, 284, Valeriani, Simona, 7 285, 286, 292, 297, 309, 310, 311, 314, 319, Vecht river, 284 320 Venetian glass industry, 9, 35, 42, 50, 56, 66 Sugar making, 210, 212 Venice, 36, 41, 45, 46, 50, 51, 53, 54, 56, 60, Suger (abbot), 151 154, 187, 188, 196, 197 Sugihara, Kaoru, 97, 124 Venlo, 273 Sulawesi, 177 Verantius, Faustus, 213, 220 Sumatra, 171, 178 Verona, 154 Surat, 80, 88 Vienna, 225, 250, 251, 252, 253, 255 Suzhou, 102, 103, 104, 113, 254 Vietnam, 179, 183 Sweden, 61, 63, 191, 228, 337 Vigevano, Guido da, 47 Villebressieu, Etienne de, 39 Taipee, 325 Vinci, Leonardo, da, 47, 55 Taiping rebellion, 230 Vitebsk province, 300, 301 Taiwan, 98, 103 Vries, Jan, de, 12 Tamils, 88 Vries, Pieter Johannes de, 216 Tang, Ying, 230 Tann, Jennifer, 223 Wages, 32, 43, 75, 83, 84, 86, 92, 99, 233, 236, Tapia, García, 219 237, 247, 248, 252, 262, 272, 273, 275, 277, Taylors, 73 278, 279, 281, 282, 283, 286, 290, 291, 296, Technological leadership, 26, 59, 62, 63, 66, 297, 300, 301, 311, 312, 313, 317, 338 67, 192 Telengana, 79 Wallonia, 267, 271, 272, 273, 275, 281, 282, Tests, 237 283, 320 Textile industry, 80, 101, 120, 212, 324, 328 Wang, Zheng, 212, 213, 220 Thames, 196 Warsaw, 289 The Hague, 37 Watchmaking, 54, 62, 223 Thomas, John, 253 Watt, James, 224 Timbotta, Giorgio, 188 Weapon manufacture, 21, 126, 212 Tool-making, 123 Weatherill, Lorna, 253 Trade unions, 285, 286 Weaving, 73, 76, 78, 79, 83, 207 Training, 163, 187, 204 Wedgwood, Josiah, 235 Travancore, 73 Weert, 273 Troyes, 147 Westminster, 140, 284 Tseng, Kingliang, 187 Windmills, 13, 215 Tsien, Tsuen-Hsiun, 332, 335 Wire-drawing, 49, 61 Tula (province), 295 Witsen, Nicolaes, 193 Turkey, 82, 156 Woerden, 276, 277, 280, 281, 284 Turkmenistan, 141 Wolcott, Susan, 89 Women, 77, 81, 86, 112, 118, 249, 251, 277, 278, Ukraine, 289 300, 309 United Kingdom (see also: Great Britain), Wong, Roy Bin, 4, 94, 107 337 Wool, 62 United States of America (see also: Ameri- Wren, Christopher, 11, 41, 159 ca), 181, 193, 197, 337 Wu, Renshu, 99 Uppada, 79 Wuxi, 115, 117 index 353

Wuyuan, 254 Zaan river, 11, 13, 216 Zaanstreek, 196, 204, 215 Xu, Quangqi, 213, 220 Zanden, Jan Luiten van, 97, 222 Zeising, Heinrich, 213, 220 Yangzi Delta, 115, 116, 124, 166, 208, 338 Zhe, Zong (emperor), 149 Yellow Sea, 123 Zheng, He, 166 Yemeni, 179 Zhili, 112 Yeola, 81 Zonca, Vittorio, 213, 220 Yk, Cornelis van, 190 Zwammerdam, 278 York, 153 Zyl, Johannes van, 220 354 index